Practitioners Guide to Incorporating Greenhouse Gas Emissions into the Collaborative Decision-Making Process (2012)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

87 INTRODUCTION This appendix provides detailed reference material related to the consideration of greenhouse gases (GHGs) in transportation planning and project development. Topics covered include • Federal and state requirements and guidance for GHG consideration in planning; • Surface transportation contribution to GHG emissions; • Contextual factors infl uencing GHG emissions; • Cost-effectiveness of transportation GHG reduction strategies; • GHG analysis tools; • Using trend analysis to project future vehicle miles traveled (VMT); • Converting highway vehicle VMT into emissions; • GHG emissions from transit vehicles; • GHG emissions from nonroad vehicles; • Emissions from construction, maintenance, and operations; • Vehicle and fuel life-cycle emissions; • Indirect effects and induced demand; and • Using the MOVES model to estimate GHG emissions. A RESOURCE MATERIAL

88 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS FEDERAL AND STATE REQUIREMENTS AND GUIDANCE FOR GHG CONSIDERATION IN TRANSPORTATION PLANNING Federal Guidance on GHGs in Statewide and Metropolitan Planning As of January 2011, there is no federal guidance on considering GHG emissions in the statewide and metropolitan transportation planning processes. The most recent sur- face transportation–authorizing legislation, the Safe, Accountable, Flexible, Efficient Transportation Equity Act: A Legacy for Users (SAFETEA-LU) of 2005, does not spe- cifically provide for any such consideration. However, states or metropolitan planning organizations (MPOs) that wish to consider them would not find any barriers in the legislation. GHGs could even fit under the general rubric of SAFETEA-LU’s fifth plan- ning factor, which includes among its goals to protect and enhance the environment and to promote energy conservation. Federal Guidance on GHGs in the National Environmental Policy Act In response to requests by federal agencies and a formal petition under the Admin- istrative Procedure Act, the Council on Environmental Quality (CEQ) released draft guidance in February 2010 on when and how federal agencies must consider GHG emissions and climate change in their proposed actions (Sutley 2010). Final guidance will not be released until sometime after the public comment period. The draft guid- ance explains how federal agencies should analyze the environmental impacts of GHG emissions and climate change when they describe the environmental impacts of a pro- posed action under the National Environmental Policy Act (NEPA). It provides practi- cal tools for agency reporting, including a presumptive threshold of 25,000 metric tons of carbon dioxide equivalent (CO2e) emissions from the proposed action to trigger a quantitative analysis, and instructs agencies how to assess the effects of climate change on the proposed action and their design. The draft guidance does not apply to land and resource management actions, nor does it propose to regulate GHGs. This guidance provides answers to some of the basic questions regarding how GHG analyses might be incorporated into the existing NEPA structure. Users of this Guide are encouraged to identify the latest guidance from CEQ and the U.S. Environmental Protection Agency (EPA) as it relates to considering GHG emissions in the NEPA process. Although NEPA analyses are specifically focused on impacts, the CEQ guidance recognizes that the global nature of climate change makes it impractical to literally assess the impacts of a given project on the climate. Instead, the level of GHG emis- sions is identified as a reasonable proxy for assessing potential climate change impacts. The CEQ guidance identifies a reference point of 25,000 metric tons of direct emissions per year as a useful indicator of significance. Although 25,000 tons is not a hard-line threshold, above that level, agencies should plan to provide an analysis of GHG emissions in their environmental documents. This level is based on Clean Air Act (CAA) requirements that stationary sources emitting 25,000 tons or more of CO2e emissions annually must report their emissions to the EPA; the idea is that this level provides comprehensive coverage of emissions while limiting reporting requirements to a reasonable number (U.S. Environmental Protection Agency 2009).

89 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS The CEQ guidance references several methodologies for quantifying both emis- sions and carbon sequestration. With regard to mitigation, the proposed guidance says, “CEQ proposes that the agency should also consider mitigation measures and reasonable alternatives to reduce action-related GHG emissions,” and further, “agen- cies should evaluate GHG emissions associated with energy use and mitigation oppor- tunities and use this as a point of comparison between reasonable alternatives.” Several other provisions are of interest to transportation practitioners. CEQ notes that some agencies may choose to examine GHG emissions in aggregate, citing trans- portation programs as lending themselves to this programmatic approach. In that case, subsequent NEPA analyses for actions implementing that program at the project level would tier from the programmatic NEPA analysis, summarizing relevant issues already dealt with at the programmatic level. Finally, the CEQ guidance identifies the effects of climate change on the proposed project as an impact that should be examined in NEPA. Specifically, climate change effects should be considered in the analysis of projects that are designed for long- term use and located in areas vulnerable to specific climate drivers (such as sea-level rise) within the facility’s lifetime. The project’s effect on the vulnerability of affected ecosystems should also be considered in the context of projected climate changes and resulting implications for that ecosystem’s ability to adapt. Rather than recommend- ing that agencies develop climate projections for each action, the guidance merely rec- ommends summarizing the relevant scientific literature on projected climate changes, particularly the synthesis and assessment products of the U.S. Global Change Research Program. (Although not discussed in the guidance, in the future the development of the National Oceanic and Atmospheric Administration’s National Climate Service may provide a unified federal source of climate projection data.) For adaptation, CEQ notes particularly that monitoring programs can be helpful not just to ensure that deci- sions are carried out as provided in the Record of Decision, but also because adaptive planning requires constant learning to reduce uncertainties. For example, adaptation is an iterative process in which monitoring is needed to assess how well the adaptations are working in the context of how the climate is actually changing (as compared with projections). Implications of EPA Authority to Regulate GHGs Under the Clean Air Act The CAA uses two main strategies for meeting clean air goals: emissions standards and national ambient air quality standards. EPA’s authority to regulate GHGs under the CAA is easily translated into GHG emissions standards for vehicles, and EPA has already shown its willingness to act by mandating new fuel economy and GHG emis- sions standards (U.S. Environmental Protection Agency 2010b). These new standards, and any that may follow, will contribute to reducing transportation GHG emissions significantly. State and local planners, however, are much more involved in the other side of the CAA: the national ambient air quality standards and the resulting regulatory structure for transportation conformity. The current structure of these standards under the CAA is not ideal to follow for GHG regulations. The national ambient air quality standards regulations are designed for pollutants in which local and regional concentrations

90 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS are important; that is, where emissions are released is as important as how much. Thus, emissions control measures are governed by local concentrations of different pollutants. Although GHGs could in theory be regulated through a conformity-style approach, such an approach does not address the global nature of GHG concentra- tions, particularly for project-level analysis. Any given project, or an entire region’s worth of projects, would have essentially no effect on ambient CO2 concentrations. However, elements of the conformity approach, such as developing regional GHG emissions budgets for transportation, could be used as a way to translate national GHG targets into state or local policies to reduce emissions. One method might be to require states and MPOs to report transportation GHG emissions as part of the transportation conformity process or separately (as a requirement for all areas, even those in attainment and maintenance status for all other pollutants), even if no budgets or restrictions are set. (By analogy, EPA has instituted mandatory GHG reporting for large stationary sources under the CAA, even though emissions standards have not yet been set.) At some point, it is likely that state and local planners will be required to report GHG emissions from transportation sources, regardless of the strategy EPA or Congress pursues. State Practice on Considering GHGs in Transportation Planning Some states have adopted requirements or provided guidance on considering GHG emissions in state and regional transportation planning. California Senate Bill (SB) 375 requires MPOs to develop a sustainable commu- nities strategy that lays out a plan to meet the region’s transportation, housing, eco- nomic, and environmental needs in a way that enables the area to meet the statewide GHG emissions reduction targets set by the California Air Resources Board (CARB) under Assembly Bill 32. CARB has worked with regional planning agencies through- out the state to set acceptable region-specific GHG targets from the transportation sector. To create incentives for compliance, funding for new transportation projects is linked to projects fitting into the sustainable communities strategy, and strategies such as transit-oriented development are given a streamlined state environmental review process or exempted from review altogether. California’s SB 375 GHG reductions illustrate the potential for achieving GHG reductions through land use and transportation planning. The final targets established by CARB for SB 375 GHG reductions statewide, across all the MPOs in California, amount to 3 million metric tons in 2020 (out of projected statewide GHG emissions of 596 million metric tons in 2020), or one-half of 1% of projected GHG emissions in California. The Massachusetts Department of Transportation’s GreenDOT policy, adopted in June 2010, sets reducing GHG emissions as one of three mutually reinforcing goals and establishes policies to achieve those goals. The policy requires that statewide and regional transportation planning documents, including MPO long-range transpor- tation plans, integrate the GreenDOT goals and that statewide and regional trans- portation improvement programs include GHG emissions reduction as a projection selection factor (Massachusetts Department of Transportation 2010).

91 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS The New York State Energy Plan requires MPOs to conduct a GHG analysis of their transportation plans, although it does not require them to meet any reduction targets (Volpe National Transportation Center 2009). Washington State has set per capita VMT reduction targets of 18% in 2020, 30% in 2035, and 50% in 2050. All of these targets are compared with a 2020 base- line. These benchmarks are set statewide, with no directives on regional target set- ting. However, Executive Order 09-05 (Washington’s Leadership on Climate Change) directs the Secretary of the Department of Transportation to work collaboratively with other state agencies, local and regional governments, and others organizations to esti- mate current and future statewide levels of VMT, evaluate potential changes to VMT benchmarks to address low- or no-emission vehicles, develop additional strategies to reduce emissions from the transportation sector, and cooperatively develop and adopt regional transportation plans that will reduce GHG emissions and achieve the statu- tory benchmarks to reduce annual VMT per capita. As of late 2010, a working group is in the process of developing reports on these issues. State Practice on Considering GHGs in State Environmental Review Even before CEQ released its federal NEPA guidance, some states had begun grappling with the issue of GHGs in their own state environmental reviews. Three notable examples are California, Massachusetts, and Washington. These three states pursue largely similar approaches but also show important differences in how thresholds of significance are defined, how life-cycle emissions are treated, which GHGs are to be considered, and recommended protocols for quantifying emissions. California Effective March 18, 2010, California adopted revisions to its California Environmental Quality Act (CEQA) regulations that introduce GHGs into the CEQA process (State of California 2010). The CEQA guidance does not establish criteria for setting thresholds of significance, and CARB has been asked to recommend a method for doing so. It does note that one consideration would be the extent to which a given project would help or hinder attainment of the state’s goals in reducing GHG emissions. In the con- text of SB 375 and other efforts throughout the state to reduce emissions, this may be understood as the extent to which a project is consistent with local or regional blue- print plans, sustainable community strategies, climate action plans, or other policies to reduce emissions. Similar to CEQ’s proposed NEPA GHG regulations, the CEQA rules allow agencies to assess GHG impacts at a programmatic level, such as in a gen- eral plan; a long-range development plan; or some other GHG reduction plan. Later project-specific environmental documents may tier from and incorporate by reference the existing programmatic review. CEQA also provides a streamlined process for some types of projects that are presumed to have beneficial GHG reduction impacts; as a result, certain residential, mixed-use, and transit projects do not need to perform an assessment of GHG emissions from light-duty vehicles (LDVs). Carbon offsets and carbon sequestration are both identified as potential mitiga- tion measures. The policy does not specify the methodology for doing these analyses, but it does reference models that could be used, including CARB’s mobile source

92 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS emission factor model, EMFAC. California defines GHGs to include the six Kyoto Protocol gases: CO2, methane (CH4), nitrous oxide (N2O), hydrofluorocarbons, per- fluorocarbons, and sulfur hexafluoride (SF6). It should be noted that there are some efforts in California to simplify or cre- ate exemptions to CEQA because of the level of difficulty involved in completing a CEQA review; for instance, the governor proposed allowing up to 100 exemptions from CEQA each year (Shigley 2010). Massachusetts In 2007, Massachusetts released a policy on GHGs for all new Massachusetts Envi- ronmental Policy Act (MEPA) reviews (Massachusetts Executive Office 2007). This policy calls for analysis of CO2 emissions only and does not address other GHGs. The transportation emissions calculation protocol is relatively straightforward compared with calculations performed for criteria air pollutants. For transportation emissions, it requires project sponsors to calculate net new VMT resulting from the project, and multiply that VMT by the MOBILE6.2 CO2 emissions factors using either individual factors for each vehicle type or using the Massachusetts fleetwide emission factor pro- vided in the guidance. (MOBILE6.2 CO2 emissions factors, unlike emissions factors for other pollutants in the model, do not differentiate emissions by speed or congestion conditions.) To calculate VMT reduction from travel demand management strategies, MEPA recommends the use of the EPA COMMUTER and CUTR Work Trip Reduc- tion models. The MEPA policy requires calculation of transportation emissions not just from transportation projects, but also from any development that has VMT impacts. For instance, an industrial facility doing a MEPA review would need to account for the VMT generated by trucks bringing supplies and by workers commuting in their cars. The policy allows for the use of carbon offsets as a mitigation tool, but it gives priority to on-site mitigation, as well as suggesting that local or regional offsets be given prior- ity. It allows project sponsors who propose exceptional measures to reduce GHGs to opt out of doing the GHG analysis. Washington The Washington State Department of Transportation’s (WSDOT) Guidance for Project-Level Greenhouse Gas and Climate Change Evaluations addresses GHG con- sideration in transportation project environmental review (Washington State Depart- ment of Transportation 2010). The guidance is particularly noteworthy in that it separates emissions into three categories: operational, construction, and embodied or life cycle. The guidance also provides boilerplate language to use in GHG discussions. The policy does not discuss the role of carbon sequestration and offsets. Rather than defining a threshold of significance based on the tonnage emitted, WSDOT defines the level of analysis required by the type of environmental documen- tation being prepared; for example, no analysis is done for categorical exclusions; qualitative analysis is done for environmental assessments; and quantitative analysis is done for environmental impact statements (see Table A.1). The analysis varies by

93 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS type of emission. For quantitative analyses, WSDOT recommends the use of EPA’s MOVES model for operations emissions and the Energy Discipline Report for con- struction emissions (CH2M HILL 2009). It only recommends qualitative discussions of embodied and life-cycle emissions in an environmental impact statement, and none in an environmental assessment. TABLE A.1. OVERVIEW OF ANALYSIS BY DOCUMENTATION TYPE, WASHINGTON STATE Type of Emission Categorical Exclusion Documented Categorical Exclusion/Checklist/ Environmental Assessment Environmental Impact Statement Operational No Qualitative Quantitative Construction No Qualitative Quantitative Embodied or life cycle No No Qualitative Source: Washington State Department of Transportation (2010). SURFACE TRANSPORTATION CONTRIBUTION TO GHG EMISSIONS Emissions by Sector The Inventory of Greenhouse Gas Emissions and Sinks (U.S. Environmental Protec- tion Agency 2010b) provides historic data on GHG emissions from transportation and other sectors. Direct transportation emissions from on-road sources accounted for approximately 23% of total U.S. GHG emissions in 2008. When considering all trans- portation sources (including aircraft, marine, rail, and pipeline), this figure increases to about 29%. As shown in Figure A.1, industry is the only economic sector with higher GHG emissions; however, recent trends show transportation and industry emissions converging to represent an almost equal share of U.S. GHG emissions, with transpor- tation emissions soon to be (or already) surpassing industrial emissions. The industrial sector also includes transportation-related emissions, including those associated with vehicle manufacture, fuels production, and production of cement and other materials for transportation facilities (see below). The growth in transportation GHG emissions between 1990 and 2008 was caused by an increase in VMT (especially for medium- and heavy-duty trucks) and stagna- tion of fuel efficiency across the U.S. vehicle fleet. Person miles traveled by LDVs increased 36% from 1990 to 2008, ton-miles carried by medium- and heavy-duty trucks increased 55% from 1990 to 2007, and passenger miles traveled by aircraft increased 63% from 1990 to 2008 (Bureau of Transportation Statistics 2009). The increases in aircraft passenger miles were offset by improvements in aircraft efficiency, operating efficiency, and higher load factors over this time period; aircraft emissions were roughly the same in 2006 and 2007 as in 1990, and slightly lower in 2008 (U.S. Environmental Protection Agency 2010b, Table 2.15). Although average fuel economy for the LDV fleet over this period increased slightly because of the retirement of older vehicles, average fuel economy among new vehicles sold actually declined between 1990 and 2004. This decline reflected the increasing

94 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS market share of light-duty trucks, which grew from about one-fifth of new vehicle sales in the 1970s to slightly more than half of the market by 2004. The trends of increasing VMT and declining fuel efficiency have reversed them- selves, at least temporarily, in recent years. Average new vehicle fuel economy improved in 2008 and 2009 as the market share of passenger cars increased. Growth in pas- senger VMT slowed from an annual rate of 2.6% from 1990 to 2004 to an average annual rate of 0.7% from 2004 to 2007, and in 2008 decreased for the first time since 1980 (due primarily to higher gasoline prices and the economic turndown) (Bureau of Transportation Statistics 2009, Table 1.32). There appears also to be a long-term structural lowering of VMT growth rates, as historic sources of VMT increases may well be plateauing because of factors such as the entry of women in the workforce, population growth, and LDV and licensed driver relationships coming close to satura- tion levels. The U.S. Department of Energy’s Annual Energy Outlook (AEO) provides forecasts of CO2 emissions by sector through 2035; these forecasts are referred to as the AEO reference case (Energy Information Administration 2010). The AEO reports only CO2 emissions, but the historic data from the EPA inventory include all GHG emissions. Since CO2 makes up more than 95% of all inventoried transportation GHGs, the data from the two sources can be considered roughly comparable for this sector. The Figure A.1. Historic trends in GHG emissions by sector.

95 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS difference is greater in the industrial sector, which is why the AEO forecasts show the transportation sector having higher CO2 emissions than the industrial sector in both the present and future years. The reference case projection considers the effects of LDV fuel economy standards through model year 2016 and the Renewable Fuel Standard 2 adopted in 2010, but not the effects of any post-2016 fuel economy requirements or proposed efficiency requirements for heavy-duty vehicles. Under the AEO reference case, transportation is forecast to be the economic sector with the largest contribution to total GHG emissions from the present until at least 2030, as shown in Figure A.2. The AEO forecasts transportation energy usage and GHG emissions based on projections of activity and fuel efficiency for each mode. The 2011 early release AEO reference case projects that for LDVs between 2007 and 2035, fuel economy gains are almost entirely offset by increases in VMT. LDVs include passenger cars, motorcycles, and light trucks with less than an 8,500 lb gross vehi- cle weight rating, most of which are used primarily for personal travel. Light trucks include almost all four-tire, two-axle vehicles such as SUVs, minivans, and pickup trucks. The AEO LDV forecasts consider underlying factors that drive these trends, such as how income per capita, population forecasts, and fuel costs affect the growth of personal travel and VMT. Forecasts for other modes consider different factors, such Figure A.2. Forecasted CO2 emissions by sector.

96 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS as how increases in industrial output increase heavy-duty vehicle (truck) activity, as well as rail, marine, and air transport activity. The 2011 early release AEO forecasts reflect an average annual increase in VMT of about 1.3% over the next decade (1.1% for LDVs and 2.6% for trucks) and 2.0% between 2020 and 2030 (1.5% for trucks), yielding an average annual 1.5% increase between 2006 and 2030. Although this is a reduction from previous forecasts that predicted a 1.8% increase, it still may be high considering recent economic and system usage trends. Emissions by Mode Figures A.3 and A.4 provide a detailed inventory of transportation-related GHG emis- sions sources for both historic and forecast scenarios. LDVs make up the largest por- tion of GHG emissions, followed by heavy-duty vehicles and aircraft. This is true for both the historic and forecasted inventories. When considering the breakdown of transportation GHG emissions by transportation mode in 2008, passenger modes made up about 71%, with freight modes constituting the remaining 29%. Figure A.3. Inventory of transportation-related GHG emissions by mode.

97 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS As discussed above, it is likely that the AEO forecasts overstate future GHG emis- sions, at least for LDVs. If VMT growth slows below 1.5% annually and vehicle effi- ciency standards continue to be increased beyond requirements that currently extend through model year 2016, emissions from LDVs will decrease in the future. These emissions may decrease more if proposals to further increase light-duty fuel efficiency standards, as well as to adopt heavy-duty emissions standards for the first time, are implemented. Figures A.5 and A.6 show contributions to GHG emissions by both passenger and freight modes. As shown in Figure A.5, the vast majority of passenger transpor- tation GHG emissions come from LDVs, which accounted for 87% of the passenger transportation GHG contribution and 62% of total GHG transportation emissions in 2008. Domestic air travel made up most of the remaining emissions (9% of passenger transportation emissions and 7% of total emissions). Travel by bus, motorcycle, rail, and ship accounted for the very small balance of passenger transportation and total emissions. Figure A.6 shows that about three-quarters of freight-related GHG emissions (21% of all transportation GHG emissions) come from trucks. Freight rail accounted for 9% of freight-related GHG emissions and 2.6% of total transportation GHG emis- sions, with GHG emissions from air, marine, and pipeline operations making up less than 2% each of total transportation GHG emissions. Figure A.4. Future inventory of transportation-related GHG emissions by mode.

98 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Figure A.6. Contribution to GHG emissions, freight modes. Figure A.5. Contribution to GHG emissions, passenger modes.

99 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Perhaps of greatest interest in freight-related GHG emissions is that emissions from heavy-duty trucks have increased rapidly since 1990, growing at three times the rate of LDV emissions. This increase is the product of decreasing fuel efficiency (per ton-mile carried) and increasing demand for freight movement by trucks. From 1990 to 2007, CO2 emissions per ton-mile carried increased almost 12%, and ton-miles car- ried increased 55%. The changes were driven by an expansion of freight trucking after economic deregulation of the trucking industry in the 1980s; widespread adoption of just-in-time manufacturing and retailing practices by business shippers and receivers; increasing highway congestion; and structural changes in the economy that produced higher-value, lower-weight, and more time-sensitive shipments that were best served by trucking. In October 2010 the federal government proposed heavy-duty fuel effi- ciency standards for the first time, which may begin to reverse this trend if imple- mented (U.S. Environmental Protection Agency and National Highway Traffic Safety Administration 2010b). CONTEXTUAL FACTORS INFLUENCING TRANSPORTATION GHG EMISSIONS Overview of Contextual Factors The AEO reference case presented above is just one potential scenario for transporta- tion GHG emissions. GHG emissions may be affected by a wide range of factors, some under varying degrees of influence by transportation agencies (e.g., speed, congestion, construction and maintenance practices, infrastructure investment, and pricing), and some over which they have little or no influence (e.g., population growth and vehicle and fuel technologies). As shown in Figure A.7, GHG emissions from passenger and freight travel are affected by five primary factors: total travel activity, the fuel efficiency of vehicles, the operational efficiency of drivers and the system (e.g., congestion, speed and aggressive driving), the carbon content of fuels, and energy use associated with construction and maintenance. Figure A.7. Different components of transportation-related GHG emissions.

100 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS TABLE A.2. CONTEXT FACTORS THAT COULD INFLUENCE GHG EMISSIONS AND SURFACE TRANSPORTATION ENERGY USE Factor Category Factor Influence Transportation costs and pricing • Congestion pricing • Parking pricing • User fees (e.g., gas taxes, VMT fees, and excise taxes) • Cost of fuel • Vehicle insurance and registration fees A, E, S, F Population and economic activity • Overall population growth, nationally and by region • Aging population • Increasing immigration • Continuing internal (to the U.S.) migration • Changing levels of affluence • Economic growth or stagnation • Service versus industrial economy • Magnitude and patterns of consumption • Tourism and recreational activity patterns • Patterns and variations in values, priorities, and political beliefs of the population • International trade and travel • Fiscal conditions for state DOTs, transit operators, and local transportation agencies A, E, S Table A.2 presents an overview of key contextual factors that could influence GHG emissions and surface transportation energy use. The table also identifies which of the components of transportation GHG emissions (identified above) each factor will likely affect. Additional discussion is provided in the following sections on several important factors that are most directly relevant to GHG planning and analysis. These factors include • Transportation costs and pricing (fuel cost, public-sector user fees, parking pricing, vehicle insurance pricing, congestion pricing, and vehicle registration fees); • Population and economic activity; • Passenger and truck VMT; • Vehicle technology and fuel efficiency; • Carbon intensity of transportation fuels; • Operational efficiency by drivers and system managers; • Construction, maintenance, and agency operations; and • Future scenarios for energy use, supply, and costs, including potential economy- wide federal policy initiatives directed at GHG emissions reductions (e.g., cap-and- trade carbon tax). (continued on next page)

101 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Factor Category Factor Influence Land use and urban form • Urban and rural land use patterns • Developing megaregions • Continuing and emerging challenges in rural and nonmetropolitan areas • Quality of schools as it affects locational choices • Crime and security as they affect locational choices • Comparative cost of housing and other services in different land use settings • Comparative fiscal and economic conditions in different local jurisdictions and statewide A Operational efficiency of drivers and system managers • Congestion • Intelligent transportation systems • Eco-driving and other driving behaviors • Speed (speed limits, speed enforcement, design speeds, flow management, traffic signal timing and synchronization, and use of roundabouts) • Freight routing, border-crossing procedures for freight, urban freight consolidation centers, urban goods movement policies, and other freight logistics S, A Passenger and truck VMT • Magnitude and type of costs and pricing for transportation use (e.g., cost of fuel, cost of vehicles, and user fees) • Passenger VMT per capita • Freight and logistics patterns and overall freight demand • Extent of use of telecommuting and alternative work schedules • Potential shifts to pay-as-you-drive insurance • Parking supply management and pricing A Policies and regulations • Emerging national approaches (e.g., cap-and-trade, taxation, and conformity) • Statewide and metropolitan surface transportation planning legislation and regulations • National Environmental Policy Act (NEPA) A, E, S, F, C Vehicle technology and fuel efficiency • Fuel economy: CAFE and California Pavley standards and consumer purchase decisions • Emerging alternative propulsion systems (e.g., hybrid and electric) and characteristics E, F Carbon intensity of transportation fuels • Corn ethanol • Cellulosic fuels • Algae-based fuels • Electricity as a vehicle power source (including differential of carbon intensity of electric power sources over time and across regions and states) • Low-carbon fuel standards and policies F Future scenarios for energy use, supply, and cost • Price of energy (especially petroleum) • Conservation incentives and education A, E, F TABLE A.2. CONTEXT FACTORS THAT COULD INFLUENCE GHG EMISSIONS AND SURFACE TRANSPORTATION ENERGY USE (CONTINUED) (continued on next page)

102 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Population and Economic Growth Forecasts The U.S. Bureau of the Census releases national population forecasts every 4 years using the cohort-component method, which is based on assumptions about future births, deaths, and net international migration. A 2008 Census release projects that the U.S. population will increase from 310 million people in 2010 to 374 million people in 2030—a growth of about 20%, or 0.93% per year. Out of this increase of 64 million people, 29 million (46%) are expected to be immigrants (U.S. Census Bureau 2008, Table 1). This is important to travel trends because immigrants are usually already working age and need to travel to work, unlike people born in the United States, who will not reach working age until much later in life. The percentage of the population aged 65 and older will also increase, with people 65 and older making up 19% of the population in 2030 compared with 13% in 2010 (U.S. Census Bureau 2008, Table 2). This increase in an older population will potentially reduce the demand for personal travel and especially work-related travel. Economic growth also affects transportation demand, because a growing economy will involve the production of more goods and services, many of which need to be trans- ported. The Congressional Budget Office, which produces 10-year economic forecasts, projects that gross domestic product will grow by about 3.5% annually between 2010 and 2015 (in real terms), and 2.3% annually between 2016 and 2019 (Congressional Budget Office 2009). A recent report for the U.S. Chamber of Commerce notes that international trade has continued to grow faster than the U.S. economy, increasing the volume of freight moving through international gateways, as well as along domestic Factor Category Factor Influence Construction, maintenance, and agency operations • Extent of new construction and type of construction (tunnels versus at-grade) • Energy intensity and carbon intensity of construction equipment and practices • Energy intensity of materials used in construction and maintenance (including extent of use of recycled materials) • Roadway lighting • Vegetation management along right-of-way (including vegetation choices and mowing practices) • Snow-plowing practices • Vehicles and fuels used in agency fleets • Paving frequency, pavement type, paving practices • Work zone management (as it affects traffic tie-ups and idling) • Energy efficiency of agency buildings and facilities • Asset management practices that affect energy and carbon generation • Increasing requirements for energy-efficient construction S, C Note: A = influences travel activity; E = influences vehicle fuel efficiency; S = influences system and driver efficiency; F = influences carbon content of fuels; C = influences GHGs from construction, maintenance, and agency operations; CAFE = corporate average fuel economy. TABLE A.2. CONTEXT FACTORS THAT COULD INFLUENCE GHG EMISSIONS AND SURFACE TRANSPORTATION ENERGY USE (CONTINUED)

103 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS trade corridors (Cambridge Systematics et al. 2008). All of these economic forecasts assume recovery from the economic downturn that began in 2008. Passenger and Truck VMT Forecasts Numerous groups and organizations have developed passenger and truck VMT fore- casts. The VMT growth rate assumption used in the AEO reference case works out to be an average of 1.5% per year between now and 2030, which is lower than the previous rate of 1.8%, but higher than the U.S. Census projection of 0.93% annual population growth. The 2009 AASHTO Bottom Line report (American Association of State Highway and Transportation Officials 2009) used a growth rate of 1.4% in VMT per year. However, some experts have come to view even this rate as too high. They suggest that factors such as rising fuel prices, lower economic growth, saturation of the workforce, plateauing of women’s entry into the workforce, an aging popula- tion, and a lower rate of transportation investment will further reduce VMT growth rates in the future. Since 2000, the annual VMT growth rate has been only 1.4%, with an absolute decline occurring in 2008. The early release of the 2011 Annual Energy Outlook projects an annual average growth in truck VMT of 1.9% between 2011 and 2020, moderating to 1.4% through 2035. The long-term growth rate is in line with the AASHTO Bottom Line report, which forecasts truck VMT growth at the same 1.4% annual rate as LDV VMT. The AASHTO forecast is based on the observation that freight VMT has recently been growing at about the same rate as passenger VMT. For example, between 1995 and 2006, passenger car and other two-axle, four-tire vehicle traffic grew by 24.4%, while combination truck traffic grew by 23.6%, and all truck traffic grew by 25.2%. In con- trast to light-duty VMT, which is primarily affected by socioeconomic, demographic, and land use factors, truck VMT is closely related to overall economic activity and the structure of how industries produce and ship goods. At first glance this seems to contradict the earlier observation that GHG emissions have increased more rapidly from trucks than from cars since 1990. This can be explained by two factors: first, the greatest increase in freight volumes occurred in the early part of this period (1990 to 1995); and second, the productivity of freight movement (ton-miles per VMT) has continued to decrease. Vehicle Technology and Fuel Efficiency Forecasts Significant increases in fuel economy standards for LDVs, coupled with higher prices and investments in alternative fuels infrastructure, are likely to have a dramatic im- pact on the development and sales of alternative fuel and advanced technology LDVs. The AEO reference case includes a sharp increase in sales of unconventional vehicle technologies, such as flex-fuel, hybrid, and diesel vehicles. For example, AEO projects hybrid vehicle sales of all varieties increase from 2% of new LDV sales in 2007 to 40% in 2030; diesel vehicles account for 16% of new LDV sales, and flex-fuel vehicles for 13% in the 2030 projections. Dramatic shifts away from spark- and compression -ignited engines are not anticipated in the next 20 years, however, because it is not anticipated that battery-powered electric or fuel cell vehicles will be able to replace the petroleum-based fleet in this time period.

104 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS In addition to the shift to unconventional vehicle technologies, the AEO reference case shows a shift in the LDV sales mix between cars and light trucks. Driven by rising fuel prices and the cost of corporate average fuel economy (CAFE) compliance, the market share of new light trucks is expected to decline. In 2007, light-duty truck sales accounted for approximately 50% of new LDV sales. In 2030, their share is expected to be down to 36%, mostly as a result of a shift in LDV sales from SUVs to midsize and large cars. The 2007 Energy Independence and Security Act (EISA) required a change in the federal fuel economy standards for the first time in 20 years. In May 2010, EPA and the National Highway Traffic Safety Administration (NHTSA) adopted a set of new light-duty fuel economy standards through 2016 consistent with the GHG emissions standards adopted by California (U.S. Environmental Protection Agency and National Highway Traffic Safety Administration 2010c). In October 2010, the agencies announced their intent to propose more stringent light-duty fuel efficiency standards for the 2017 through 2025 model years, with potential for a fuel economy standard as high as 62 mi/gal or as low as 47 mi/gal for model year 2025 (U.S. Environmental Protection Agency and National Highway Traffic Safety Administration 2010a). One of the uncertainties in future year motor vehicle technology and fuel efficiency forecasts is whether U.S. LDV sales will return to historic levels after the economic recession is over. Recent annual LDV sales have been near 16 million units, while the 2030 AEO 2009 forecast is for sales near 20 million units per year. Some analysts believe that the most recent historic sales are, for various reasons, artificially high, and that near-term vehicle sales will be closer to 12 million than 16 million. If this occurs, the penetration of new technologies and more fuel-efficient vehicles will be slower than expected, and baseline GHG emissions will be above expected values. This would make it more difficult for organizations to meet GHG emissions reduction targets. However, in most households with more than one vehicle, newer, fuel-efficient vehicles are likely to be used more intensively than older, less efficient vehicles, so the effect on VMT by more efficient vehicles is likely to be greater than the market penetration of new vehicles alone would suggest. Unlike LDVs, heavy-duty vehicles are not currently subject to fuel efficiency stan- dards. However, the 2007 EISA required that EPA evaluate fuel efficiency standards for trucks. In October 2010 EPA and NHTSA announced proposed GHG and fuel effi- ciency standards for heavy-duty trucks. The proposed standards would reduce energy consumption and GHG emissions by 7% to 20% for combination tractors, heavy-duty pickups and vans, and vocational vehicles by model year 2019 compared with a 2010 baseline. The reduction compared with the AEO reference case would be somewhat lower because this projection already assumes modest increases in fuel efficiency over this time period. The proposed standards are less aggressive than light-duty standards (as measured by the percentage improvement in fuel efficiency, as for LDVs), largely because market forces have already fostered more aggressive development and adop- tion of fuel economy improvements for U.S. trucks compared with LDVs.

105 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Trends in System Operations and Operational Efficiency of Drivers As gas tax revenues fall and the Highway Trust Fund realizes severe shortfalls, state and local agencies are facing significant budget constraints that affect their ability to operate the transportation system. This fiscal stress, along with constrained right- of-way, community impacts, and environmental concerns, limits expansions of the transportation system and maintenance and operational investments in the existing system. Many agencies, in particular state DOTs, have begun to use intelligent trans- portation systems (ITS) and other management and operations strategies to mitigate declines in reliability and increases in travel time as transportation demand outpaces infrastructure investment. This trend is likely to continue in the future. Given that the United States consumed an additional 2.9 billion gallons of fuel in 2005 because of congestion, a substantial increase from 0.5 billion gallons in 1982 (Texas Transporta- tion Institute 2007), the success of such strategies in reducing growth in delays and traffic congestion could help reduce GHG emissions as fuel is used more efficiently. Conversely, if VMT continues to increase without corresponding infrastructure or operational improvements, then congestion, delay, and associated emissions will con- tinue to increase. The application of dynamic technology, specifically ITS, is becoming a relatively common strategy for improving the operational efficiency of the transportation sys- tem. Examples include ramp meters that control the volume of drivers entering a highway, electronic signage that informs drivers of upcoming travel conditions, and traffic signalization that can encourage steady vehicular flow along a specific corri- dor ( Lockwood 2008). ITS technology also allows for traffic management centers to respond promptly to roadway incidents, thereby lessening delay and potentially reduc- ing GHG emissions. Lane management, a strategy that expands on the traffic management center and ITS concept, allows the transportation agency to actively manage travel lanes in real time for optimal flow conditions. Managed lanes, also known as high-occupancy toll lanes, allow carpools to ride for free, but charge other vehicles a toll that varies by time of the day and current traffic conditions. A high-occupancy toll lane increases high- way efficiency by allowing additional vehicles to use an underutilized high-occupancy vehicle lane. The U.S. DOT’s Urban Partnership Program provided funds for selected metropolitan areas to demonstrate different aspects of managed lanes operation. It is expected that the experiences of these metropolitan areas with the managed lane con- cept will provide the impetus for other metropolitan areas to adopt similar strategies. Over the long run, however, GHG emissions reductions due to fuel savings from management and operational strategies are likely to be partially offset by induced demand, or the increase in travel that results from improved travel conditions. The magnitude of the induced demand effect is a subject of considerable uncertainty and is likely to vary according to the type of strategy. For instance, strategies that reduce travel time or improve reliability (such as most ITS strategies) would be most likely to result in some amount of additional travel, thus reducing the magnitude of congestion and GHG benefits over the long run. In contrast, operations strategies that modestly increase travel time (such as enforcement of reduced speed limits) or raise monetary

106 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS travel costs (such as congestion pricing) are not believed to result in additional travel demand, and could actually have a slight suppressing effect. In calculating induced demand for GHG analysis purposes, it is critically impor- tant to exclude induced demand that is based on mere shifts in VMT from one facility to another or from one time of day to another, because these are not net increases in GHG but merely a shift in time or location. Moreover, induced demand associated with operational improvements may be lower than induced demand associated with adding highway lane capacity. Hymel et al. (2009) estimated the elasticity of statewide induced demand as being 0.037 in the short run and 0.186 in the long run; that is, their estimates are smaller than most previous estimates of induced demand. Only about 40% of this effect is associated with reducing congestion in urban areas; 60% of the induced demand effect related to shortening distances for road trips. In the future, as vehicle technologies (such as hybrids or electric vehicles) that are more efficient in low-speed operation become more widely adopted, the GHG emis- sions reduction effects of operational strategies will decline. Even without considering these effects, the efficiency benefits of congestion reduction will decline over time in proportion to increases in CAFE standards, as well as the adoption of less carbon- intensive fuels, as baseline GHG emissions decrease. This effect is by no means unique to operational strategies. It is equally true that the GHG emissions reduction effects of most other transportation strategies, including land use, transit, and other VMT- reducing strategies, will decline commensurate with success in decarbonizing vehicles and fuels. Driver behavior is another factor in improving the operational efficiency of the system. Eco-driving (defined as avoiding rapid accelerations and braking, avoiding speeding, proper gear shifting, and using cruise control) and enhanced maintenance of vehicles are estimated to have a 1% to 5% potential in reducing GHG emissions (U.S. Department of Transportation 2010). The benefits of such efforts, however, seem to vary by how long the behavior lasts. Two U.S. studies compared fuel consumption under standard versus more aggressive driving cycles. An EPA study found that aggres- sive driving can reduce gas mileage by 33% at highway speeds; a CARB study found an increase in fuel consumption of 5% to 14% accompanied more aggressive driving (International Energy Agency 2005). In Europe, where eco-driving campaigns have been more widespread and implemented for a longer time than in the United States, short-term savings have been found to be higher than long-term savings. For example, one estimate found reductions in fuel consumption of 15% to 25% for drivers in the first year; this reduction dropped to an average of 6.3% in subsequent years. A Dutch study showed a 10% overall long-term reduction from an eco-driving program (Lucke and Hennig 2007). Properly keeping an engine tuned saved 4% in fuel, proper tire inflation led to a 3% reduction in fuel consumption, and using the correct motor oil resulted in a 2% improvement (International Energy Agency 2005). The important issue for such initiatives in the United States is whether concerted efforts to change driver behavior will have any possibility of doing so in any significant way. It is likely that eco-driving campaigns and targeted marketing will occur much

107 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS more frequently in the future and that these efforts will have a positive benefit on GHG emissions reduction. However, the overall impact on GHG emissions will vary greatly by how much the efforts really affect driving behavior in the longer term. Future Scenarios for Energy Use, Supply, and Costs Because the vast majority of transportation energy in the United States comes from petroleum, importing oil will remain a political necessity for decades into the future. This requires ceding a certain level of political influence and control to oil-exporting nations. Many of these oil-producing nations are among the most politically unstable in the world, which results in unavoidable uncertainty with regard to oil supply. Further- more, although overall worldwide supplies of petroleum are nowhere near exhaustion, it is likely that the ability to expand oil supply capacity is nearing its peak and that in the near future, it will become prohibitively difficult to expand oil production beyond current levels. When this occurs, energy will need to come from other, nonpetroleum sources, most of which are likely to reduce life-cycle GHG emissions. During the transi- tion period, there will be pressure to extract petroleum from previously uneconomical sources, such as tar sands. Such production methods are more energy-intensive, and their use may result in increased life-cycle GHG emissions per unit of fuel produced. Several technologies are available or in development that could potentially reduce gasoline consumption and GHG emissions in the transportation sector. Many of these options, such as hydrogen fuel cells, would require a dramatic infrastructure invest- ment before the technology could be implemented on a large scale. Biofuels and elec- trification require far more modest infrastructure investments, and therefore are more likely to be implemented in the foreseeable future. Biofuels require feedstocks that can be produced with very little energy input in order to reduce overall carbon emis- sions. However, concerns have been raised that the demand for biofuel feedstocks may reduce agricultural land for other purposes while increasing pressure to convert nonagricultural lands (such as forests) to agricultural production, which could cause sequestered carbon to be released. This land use concern is not true of all biofuels and alternative fuels, such as cellulosic and algae-based fuels. Plug-in electric vehicles require electricity production from low-carbon sources such as wind, solar, nuclear, and biomass to significantly decrease emissions. The U.S. invests billions of dollars every year to promote energy efficiency, expand energy supply, develop energy technologies, and reduce energy costs. More than $16 billion was spent on energy subsidies in 2007 (Energy Information Administra- tion 2007). The 2007 Renewable Fuels Standard (RFS), signed into law as part of EISA, mandates that 36 billion gallons of biofuels will be used in the United States in the year 2022. In March 2010, EPA updated the RFS to encourage the production of low-GHG biofuels (U.S. Environmental Protection Agency 2010d). These changes include a higher standard in the short term to reflect existing production surpluses. In addition, the standards for advanced biofuels and biomass-based diesel have been modified to be stronger and more flexible. The RFS will result in a dramatic increase in the amount of ethanol being sold in the country over the next 15 years, and could potentially reduce overall gasoline consumption.

108 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS The impact of any of these alternative fuels on transportation GHG emissions will range from modest to quite significant, depending on the fuel and how it is produced. Figure A.8 shows relative GHG emissions, including full fuel-cycle emissions, for a variety of transportation fuels; the estimations shown are based on the Department of Energy’s Greenhouse Gases, Regulated Emissions, and Energy use in Transporta- tion (GREET) model, Version 1.8b. Compared with gasoline, emissions reductions range from about 16% for an 85% corn ethanol blend (E85) to 57% to 84% for ethanol from various cellulosic feedstocks. A 20% blend of soy-based biodiesel pro- vides roughly an 18% reduction, and natural gas results in a reduction in the range of 16% to 30%. (Note that the model does not reflect the latest research on bio- fuel impacts reported for the 2010 RFS2 rulemaking [U.S. Environmental Protection Agency 2010e].) Electricity shows roughly a 33% reduction with current technology and electricity generation mix. Benefits of hydrogen vary greatly depending on the production method. The net impact of any of these fuels on total GHG emissions will depend not only on the per vehicle benefit but also on the rate of market penetration, which will depend on a host of very uncertain factors, such as technology advance- ment, fuel supply, policy choices that may encourage or discourage specific fuels, and the relative prices of different fuels. Figure A.8. Relative GHG emissions from different fuels, GREET model.

109 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Potential Federal GHG Reduction Policy Initiatives A variety of policy actions have been proposed at the federal level to reduce GHG emissions from all sectors, including transportation. The federal climate change policy landscape is likely to evolve significantly over the next few years depending on what actions are taken on transportation reauthorization, as well as energy and/or climate change legislation and regulation. Federal policy actions affecting transportation GHGs can be grouped into five categories: • Implement cap-and-trade or carbon price strategies to establish an economywide carbon price; • Set vehicle and fuel standards to reduce carbon emissions per unit of travel; • Provide vehicle and fuel market incentives to accelerate adoption of more efficient and less carbon-intensive technology; • Fund expanded research and development of advanced vehicle and fuel technol- ogy and climate change research to realize long-term technology improvements and enhanced decision support; and • Revise transportation programs and funding to include a focus on GHG measure- ment and reduction. Some of these categories are likely to have only indirect effects on transportation planning. The magnitude of effects on transportation planning will vary based on the detailed formulation of any of these categories. A cap-and-trade system, and poten- tially other market incentives, will increase the price of gasoline, and the greater the increase in gasoline price, the greater the effect on transportation demand, revenues, and planning. Most cap-and-trade proposals are expected to have modest effects on VMT, at least in the next two decades. Most of these proposals would result in a gasoline price increase of 10 to 20 cents per gallon in the first few years of imple- mentation, increasing to up to 40 to 60 cents per gallon by 2030. Analysis by the Energy Information Administration of cap-and-trade legislation found reductions in transportation GHG emissions on the order of 5% below a reference case in 2030. This reduction results in part from a decrease in LDV and truck VMT of about 2.5% to 3% up to 2030, as well as small improvements in LDV (1.2% to 1.3%) and truck (0.5% to 0.6%) efficiency. LDV efficiency improvements are small because most of the lowest-cost efficiency improvements will already have been implemented as a result of the recently enacted CAFE standards. A significant part of the reduction is due to reduced volumes of fuel shipments, particularly coal, because of the shift away from coal-fired power plants, and fuel oil through pipelines. To the extent that pricing mechanisms, vehicle and fuel standards, and research and development are effective at reducing gasoline consumption, they will also reduce revenues for transportation infrastructure—at least until the motor fuel tax is replaced or supplemented with other revenue sources. However, it is possible that revenue

110 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS obtained through pricing may be reinvested in GHG reduction programs, including transportation programs. The last category of federal policy actions focuses specifically on the programs implemented by transportation agencies. It is possible that reauthorization of the sur- face transportation bill will include requirements and/or incentives to address GHG emissions and climate change issues in transportation planning. Such additions to the bill could take the form of any of the following: • Technical assistance on GHG data and analysis procedures and planning methods; • Regulations requiring consideration of GHGs in planning via inventory develop- ment, plan assessment, and development of mitigation measures; setting GHG emis- sions reduction targets at a state or metropolitan level; or requiring specific planning activities (such as integrated transportation and land use planning); and/or • Funding incentives, either by establishing performance criteria for GHG emissions reductions and distributing funding on the basis of these criteria, or setting aside fund- ing for implementation of specific GHG emissions reduction measures. EFFECTIVENESS AND COST-EFFECTIVENESS OF TRANSPORTATION GHG EMISSIONS REDUCTION STRATEGIES Both effectiveness (potential magnitude of GHG emissions reductions) and cost-ef- fectiveness (cost per unit of reduction) are important considerations when selecting a set of strategies through the transportation decision-making process. The focus is on strategies that can be directly influenced by transportation agencies, but information on other strategies (such as vehicle efficiency and fuel standards) is also presented for comparison. Considerations affecting the feasibility of each strategy are also ad- dressed. Users of the information in this section should recognize the considerable uncertainty present in the cost-effectiveness estimates provided. Strategy Assessment The strategies considered for reducing GHG emissions are found in nine major categories: 1. Transportation system planning and design; 2. Construction and maintenance practices; 3. Transportation system management and operations; 4. Vehicle and fuel policies; 5. Transportation planning and funding; 6. Land use codes, regulations, and other policies; 7. Taxation and pricing; 8. Travel demand management; and 9. Other public education.

111 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS A listing of strategies, and a definition or description of each, is provided in Table A.3. Inclusion of the type of strategy or project in this table does not guarantee that it will reduce GHG emissions; the GHG impacts of any given strategy or project must be evaluated based on local conditions and data. TABLE A.3. POTENTIAL GHG EMISSIONS REDUCTION STRATEGIES Strategy Definition or Description Transportation System Planning and Design Bottleneck relief Increased capacity at bottlenecks (specific points on the transportation network where demand exceeds capacity), such as interchanges, intersections, and lane drops. High-occupancy vehicle/high- occupancy toll (HOV/HOT) lanes HOV: Highway lanes reserved for the use of vehicles carrying a minimum of two or three persons. HOT: Lanes that single-occupant vehicles are permitted to use at a price, which is set to ensure that lane capacity is not exceeded. Toll lanes or roads Highway facilities for which a price is charged, whether fixed or variable, for their use. Truck-only toll lanes Priced lanes for the exclusive use of trucks. Fixed-guideway transit expansion Urban transit systems including bus rapid transit, light rail, heavy rail, and commuter rail operating on exclusive right-of-way. Intercity rail and high-speed rail Rail operating over long distances between major cities. Bicycle facilities and accommodation Bicycle lanes, paths, parking, racks on buses, and other infrastructure improvements for bicyclists. Pedestrian facilities and accommodation New or improved sidewalks, pedestrian crossings, and shared-use facilities; measures such as traffic calming to enhance the pedestrian environment. Rail system improvements Track upgrades, clearance improvements, railyard capacity expansion, or other improvements to increase the speed and/or reduce the cost of moving goods by rail. Marine system improvements Improvements to ports or waterways, such as dredging or lock upgrades, to increase the speed and/or reduce the cost of moving goods by boat or ship. Intermodal facility and access improvements Capacity, operational, or access enhancements at truck–rail, truck–marine, or rail– marine intermodal and transload facilities. Transportation System Management and Operations Traffic signal timing and synchronization Technologies and practices to reduce congestion and smooth traffic flow through improved signal timing and/or coordination of multiple signals. Incident management Technologies and practices to reduce response time to incidents, clear incidents more quickly, and alert travelers. Traveler information systems Provision of up-to-date information to travelers and truckers on traffic conditions, incidents, and expected delays; the availability of public transportation and other travel alternatives; weather conditions; road construction; and special events. Advanced traffic management systems Other systems, such as speed harmonization, integrated arterial and freeway control, and applying surveillance and control to improve traffic flow. Access management Strategies to reduce congestion and improve safety on arterial roadways by controlling the location, design, spacing, and operation of access to adjacent land uses. Congestion pricing Roadway pricing that varies with the actual or expected level of congestion on the facility, with the goal of keeping traffic levels below the capacity of the system. (continued on next page)

112 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Strategy Definition or Description Speed management Reduced speed limits on high-speed facilities, including the Interstate system, other limited-access highways, and high-speed rural major arterials, to no more than 55 or 60 mph; and/or greater enforcement of existing speed limits. Truck and bus idle reduction Education, laws, and/or incentives to introduce technology (such as electrical hook- ups at truck stops or on-board auxiliary power supplies) to reduce long-duration idling of heavy vehicles. Transit fare measures Transit fare subsidies or discounts to encourage transit use, targeted at the general population (e.g., free fare zones) or subpopulations such as workers. Transit frequency, Level of Service, and coverage Expanded frequency, geographic coverage, or temporal coverage of urban bus or rail transit. Transit priority measures Measures such as signal preemption, queue bypass lanes, and shoulder running to speed transit services relative to driving. Land Use and Smart Growth Integrated transportation and land use planning Regional or corridor activities to coordinate transportation and land use plans and projects to improve travel efficiency. Funding incentives and technical assistance to local governments Money or staff dedicated to helping local governments update plans, zoning, and other documents and practices consistent with smart growth principles. Parking management and pricing Providing disincentives to driving by pricing or limiting the amount of parking, or using pricing to encourage park-once trips. Designated growth areas, growth boundaries, and urban service boundaries Policy or regulatory designations to encourage growth in compact and/or central areas as an alternative to sprawl. Transit-oriented development, infill, and other location- targeting incentives Planning activities and fiscal and regulatory incentives to focus development in areas that can be efficiently served by transit, nonmotorized travel, and shorter automobile trips. Freight villages and consolidation facilities Freight facilities that are clustered together to reduce truck trip lengths and improve intermodal access; or locations where deliveries (retail, office, or residential) can be consolidated for subsequent delivery into the urban area in an appropriate vehicle with a high level of load utilization. Travel Demand Management and Public Education Employer-based commute programs Requirements for employers to reduce single-occupancy vehicle trips by their employees; or outreach, assistance, and incentive programs to encourage them to do so. Ridesharing and vanpooling programs Programs such as ride-matching databases, vanpooling programs, and other supportive actions to increase vehicle occupancies for work trips. Telework and compressed work week Working from a location other than the regular workplace using modern telecommunications and computer technology or working a regularly scheduled number of hours in a shortened span of time. Nonwork transportation demand management programs School pool, social marketing, individualized marketing, and other outreach and incentive-based programs aimed at reducing nonwork personal travel. Eco-driving Education programs directed at increasing vehicle fuel efficiency by affecting both driver behavior and vehicle maintenance. TABLE A.3. POTENTIAL GHG EMISSIONS REDUCTION STRATEGIES (CONTINUED) (continued on next page)

113 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Strategy Definition or Description Vehicle and Fuel Policies Alternative fuel and high- efficiency transit vehicle purchase Purchase of transit vehicles such as hybrid electric, natural gas, or electric buses to reduce energy use or use fuels with reduced carbon content. Alternative fuel and electric vehicle infrastructure Direct provision of alternative fueling infrastructure; or subsidies, incentives, or technical assistance to encourage other entities to provide such infrastructure. Government fleet purchases Purchase of high-efficiency and/or low-carbon fuel vehicles for government fleets. Construction, Maintenance, and Operations Practices Low-energy and GHG pavement and materials Use of less energy-intensive construction materials by state and local highway departments and other transportation agencies, such as recycled material in cement, and asphalt that is prepared at a lower temperature. Construction and maintenance equipment and operations Use of more efficient transportation agency or contractor equipment, and practices (such as idle reduction) to improve the efficiency of equipment utilization. Alternative energy sources or carbon offsets Use of transportation agency property for renewable energy generation or carbon sequestration. Right-of-way management Practices such as reduced mowing to reduce energy consumption and GHG emissions. Building and equipment energy efficiency improvements Improvements to transportation agency facilities and equipment, such as energy efficiency retrofits, to reduce energy use. TABLE A.3. POTENTIAL GHG EMISSIONS REDUCTION STRATEGIES (CONTINUED) Information Sources Information on effectiveness and cost-effectiveness is drawn from existing literature, with a focus on recent reports that summarize estimates across multiple strategies. The feasibility assessment is also based on information from the literature, as well as the judgment of the project team. The information provided must be interpreted with caution. The literature on transportation GHG emissions reduction strategies is for the most part fairly new and focuses on summary estimates at a national level. There is considerable uncer- tainty surrounding the estimates for many strategies, and both effectiveness and cost- effectiveness may vary significantly depending on local factors. The feasibility of a given strategy may also vary from location to location, and may change in the future depending on technological evolution, market trends, and changing political and soci- etal viewpoints. Furthermore, climate and transportation analysts have yet to devise a common framework for analyzing costs. For instance, there is no common agreement as to whether reduced transit fares or increased road prices should be represented as a change in public cost, a cost to consumers, or as a transfer payment from nonusers to users.

114 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Metrics and Methodological Issues Effectiveness is typically measured in terms of metric tons (tonnes) of carbon dioxide equivalent (CO2e) emissions reduced per year or cumulatively over a number of years. For comparison at different geographic scales, however, effectiveness must be measured as a percentage reduction of emissions, either in the total transportation sector or in a particu- lar transportation subsector (e.g., on-road vehicles). Use of different comparison bases in the literature creates challenges for the development of consistent effectiveness estimates. Cost-effectiveness is typically measured in terms of dollars per tonne of CO2e reduced and can be compared more consistently across studies. To evaluate a string of future year benefits, costs are typically discounted to current year dollars using a standard discount rate. Future GHG emissions are usually not discounted, although practices vary on this topic. It is generally agreed that the benefit of reducing a tonne of GHG emissions is roughly the same whether that reduction occurs now or 10 years in the future. The most important metric is cumulative GHG emissions reductions start- ing in the present and continuing through some analysis horizon (e.g., 2030 or 2050). Another important consideration related to cost-effectiveness is the specific costs included in the estimate. Some estimates of cost-effectiveness include public-sector implementation costs only. Others include benefits to travelers, such as vehicle operat- ing cost savings. Tolls and taxes (or rebates) are generally considered a transfer between one entity and another, and therefore not a net social cost, although they affect the distribution of costs. A particularly challenging issue is the incorporation of nonmon- etary costs, such as time savings or environmental externalities (e.g., air pollution and impacts on public health). For some strategies, these costs can be quite significant, but they are usually not monetized for the purpose of developing GHG cost-effectiveness estimates. Net included costs in Table A.4 refer to all the monetized costs included in the cost-effectiveness estimates; usually these are vehicle operating costs in addition to direct implementation costs. They do not include the monetary value of travel time savings nor crash reduction benefits. Finally, the estimates typically include only oper- ating emissions benefits, and not construction or other life-cycle emissions. Readers should be aware that the use of net included cost-effectiveness measures is controversial, with the primary argument against their use being that they ignore other positive benefits associated with such strategies and thus bias the results against highway improvement projects. Caution should be exercised when using cost-effectiveness indices alone. For example, a cost-effectiveness index could show that one strategy is better than another based on the relationship between benefits and costs, but that the overall reduction in GHG emissions might be greater from the strategy that has the lower cost-effectiveness index. This highlights the concept that cost-effectiveness evaluation must be done in the context of the overall goals of the policy or planning study. Other Considerations GHG reductions are just one of the benefits and impacts that must be considered when evaluating any transportation action. Many strategies also have important cobenefits (positive impacts) or negative impacts. For example, congestion reduction strategies reduce traveler delay and improve mobility in addition to reducing fuel consumption

115 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS and emissions. Provision of alternative modes (transit, walking, bicycling) can in- crease accessibility, especially for populations with limited car access. By increasing the cost of travel, pricing may have negative impacts unless these impacts are mitigated through revenue redistribution or enhancement of travel alternatives. Some strategies, especially pricing, may have equity impacts by disproportionately affecting a particu- lar subset of the population (e.g., low-income travelers). Table A.4 shows a typical effectiveness assessment. This table, taken from a report to Congress (U.S. Department of Transportation 2010), shows a typical range of GHG emis- sions reductions reported in the literature, as well as a subjective assessment of the cost per tonne and the net cost per tonne. In addition, the extent to which a particular strategy has associated cobenefits or disadvantages (i.e., whether it provides positive or negative impacts on achieving other goals) is indicated by + (positive), – (negative), or 0 (neutral). TABLE A.4. SYSTEM EFFICIENCY STRATEGIES Strategy GHG Reduction (2030)a Cost per Tonne Cobenefits Key Federal Policy Options Direct Net Included Highway Operations and Management Traffic management Low <0.1% to 0.5% Moderate to high Net savings to high + Funding for project implementation, technical support, and institutional coordination Real-time traveler information Low <0.1% High Low to high + Bottleneck relief Low <0.1% to 0.3%b NA NA +/– Project funding Reduced speed limits Moderate 1.1% to 1.8% Low Net savings – Federal speed limit policy, funding incentives for enforcement Truck Operations and Management Truck idling reduction Low 0.1% to 0.2% Moderate Net savings + Federal anti-idling law Truck size and weight limits Low <0.1% Low Net savings 0 Revise federal policy on truck size and weight limits Urban consolidation centers Low <0.1% Moderate Net savings + Feasibility studies and demonstration projects Freight Rail and Marine Operations Freight modal diversion Low <0.1% to 0.2% High Net savings to moderate 0 Funding for rail and intermodal capacity improvements Marine modal diversion Low <0.1% High High 0 Capital investment in inland waterways; subsidies for short-sea shipping Rail and intermodal terminal operations Low <0.1% Unknown Unknown + Funding for rail and intermodal capacity improvements Ports and marine operations Low <0.1% Unknown Unknown + Tools to assist in GHG assessment; regulations or voluntary partnerships to promote GHG emissions reduction practices (continued on next page)

116 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Figure A.9 shows a typical scenario analysis of the impact of varying levels of GHG emissions mitigation strategies (Greene and Plotkin 2011). The assumptions concern- ing different mitigation options are shown in Table A.5. Note that the percentages in the low-, mid-, and high-mitigation scenario columns of Table A.5 are the incremental Strategy GHG Reduction (2030)a Cost per Tonne Cobenefits Key Federal Policy Options Direct Net Included Infrastructure Construction and Maintenance Construction materials Moderate 0.7% Unknown Unknown 0 Continue R&D on warm- mix asphalt and recycled materials; construction material requirements Other transportation agency activities Low 0.1% Unknown Unknown 0 Model practices and assessment tools; regulations to reduce GHG emissions in construction; funding incentives for GHG reduction Combined benefits 2.9% to 6.1% Source: U.S. Department of Transportation (2010). aThe estimated benefits of traffic management, traveler information, and bottleneck relief all reflect offsetting effects of induced demand. Increased demand resulting from improved travel conditions is not reflected in other strategies in which it may be significant, such as aviation operations. bDoes not include emissions from construction activities or from additional delay during construction. TABLE A.4. SYSTEM EFFICIENCY STRATEGIES (CONTINUED) Figure A.9. Bundle strategy results from Moving Cooler study (Cambridge Systematics 2009).

117 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS changes in the 2010 Annual Energy Outlook (AEO) reference case projected impact. Thus, for example, for change in energy efficiency for the 2035 low-mitigation sce- nario, implementing fuel economy and emissions standards over those assumed in the reference case would increase the energy efficiency by 15 percentage points over the 39% projected in the reference case. More detailed information on the combination of different actions that can make up a mitigation strategy is available in the Moving Cooler report (Cambridge Systematics 2009) and Greene and Plotkin (2011). TABLE A.5. CHANGES IN MITIGATION OUTCOMES FROM DIFFERENT STRATEGIES Policy and/or Mitigation Option AEO 2010 (2010 to 2035) 2035 2050 Low Mid High Low Mid High Change in Energy Efficiency for Total Stock 39% Fuel economy and emissions standards 15.0% 30.0% 40.0% 35.0% 60.0% 80.0% Driver behavior and maintenance 2.5 5.0 10.0 2.5 5.0 10.0 Improved traffic flow 0.0 1.0 2.0 0.0 1.0 2.0 Pricing policies Carbon price 2.4 2.4 2.4 3.6 3.6 3.6 Road user tax on energy 0.0 1.6 1.9 2.2 2.2 2.2 Pay-at-the-pump insurance 0.0 4.4 4.4 0.0 5.2 5.2 Feebates 0.0 10.0 10.0 0.0 10.0 10.0 Automated highways 0.0 0.0 1.0 0.0 0.0 5.0 Change in Vehicle Miles Traveled (billions) 54% Road user tax on energy –0.2% –0.05% –0.6% –0.4% –0.8% –1.0% Carbon price –1.2 –1.2 –1.2 –1.7 –1.7 –1.7 Pay-at-the-pump insurance 0.0 –1.0 –1.0 0.0 –1.0 –1.0 Trip planning and route efficiency 0.0 –2.0 –4.0 0.0 –5.0 –10.0 Ridesharing 0.0 –0.7 –1.4 0.0 –1.0 –2.0 Land use and infrastructure development –0.5 –1.0 –2.0 –1.5 –3.0 –5.0 Freight Trucks 16% Fuel economy and emissions standards: Long haul 15.0% 25.0% 30.0% 25.0% 35.0% 40.0% Fuel economy and emissions standards: Local 15.0 25.0 30.0 25.0 35.0 40.0 Carbon price 1.2 1.2 1.2 1.8 1.8 1.8 Road user tax on energy 0.9 1.5 1.8 2.1 2.1 2.1 Pay-at-the-pump insurance 0.0 4.4 4.4 0.0 5.2 5.2 Traffic flow improvement 0.0 1.0 2.0 0.0 1.0 2.0 Automated highways 0.0 0.0 0.0 0.0 5.0 10.0 Rail –2% Change in energy intensity of all trains (1,000s Btu per ton-mile) –10.0% –15.0% –20.0% –25.0% –30.0% –40.0% Source: Adapted from Greene and Plotkin (2011).

118 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Tables A.6 and A.7 provide information from the literature regarding the effec- tiveness, cost-effectiveness, and feasibility of transportation GHG emissions reduction strategies. Table A.6 shows transportation system strategies directed at the design and operation of the transportation system itself and the behavior of users of the system. This table includes infrastructure planning and investment decisions; construction and maintenance practices; highway, transit, and freight operations; land use; taxation and pricing; travel demand management; and other public education. With some excep- tions (such as land use, many of the pricing strategies, and rail and port investment), the strategies shown in Table A.6 can largely be implemented by state and metropoli- tan transportation agencies. Table A.7 shows vehicle and fuel technology strategies that seek to reduce GHG emissions through the use of low-carbon fuels and/or more fuel-efficient vehicles. This table includes strategies that are primarily under the control of federal or state legisla- tive bodies and regulatory agencies rather than transportation agencies. The strategies included in these tables represent strategies for which informa- tion on GHG impacts and cost-effectiveness were identified in one or more litera- ture sources. Estimates were reviewed for reasonableness of assumptions, and in some cases, results were not presented if the assumptions were deemed to be too unrealis- tic. For example, one study’s estimates of carpooling reductions assumed that vehicle occupancies could be increased substantially (e.g., adding one person per vehicle to every commute trip) (International Energy Agency 2005). The context of the study was to provide information relevant to what might be achieved in response to a major oil supply disruption, in which case dramatic increases in fuel prices might be expected that could lead to or support significant changes in travel behavior. However, the esti- mate was not deemed realistic as an assessment of carpooling potential in the absence of such a major disruption. Tables A.6 and A.7 contain the following information: Key deployment assumptions: A description of the key strategy deployment assumptions in the underlying study. Percentage fuel and GHG emissions reductions: Potential reductions in total trans- portation fuel consumption and GHG emissions, generally in 2030. Table A.7 also shows 2050 savings for advanced technology strategies that will take many years to fully develop. The percentage reductions are based on reported GHG reductions from most sources except for the International Energy Agency (2005) report, which reports fuel (petroleum) use reductions. In some cases, the percentage reduction was taken directly from the source document. In others, the reduction was calculated based on absolute GHG emissions reductions reported in the source document. In these cases, absolute reductions were converted to percentage reductions based on the U.S. Depart- ment of Energy’s 2009 AEO reference case. The AEO adjusted 2030 transportation sector baseline is 2,171 million metric tons (MMT) of CO2e. Direct cost-effectiveness: Cost-effectiveness, expressed in dollars per tonne CO2e reduced, considering implementation costs only (typically public-sector costs for infra- structure, services, or programs; not shown for strategies in Table A.6). The estimates (continues on page 127)

119 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS TABLE A.6. TRANSPORTATION SYSTEM GHG REDUCTION STRATEGIES Strategy Name Key Deployment Assumptions Fuel/GHG Reduction in 2030 (%) Direct Cost- Effectiveness Data Source Feasibility Te ch n ic al In st it u ti on al P ol it ic al Transportation System Planning, Funding, and Design Highways Capacity expansiona, b, c 25% to 100% increase in economically justified investments over current levels 0.07%–0.29% [0.25%–0.96%] N/A Cambridge Systematics 2009 M H L–H Bottleneck reliefa, b Improve top 100 to 200 bottlenecks nationwide by 2030 0.05%–0.21% [0.29%–0.66%] N/A Cambridge Systematics 2009 M H L–H HOV lanes Convert all existing HOV lanes to 24- hour operation 0.02% 0.00% $200 International Energy Agency 2005; Cambridge Systematics 2009 H H H Convert off-peak direction general- purpose lane to reversible HOV lane on congested freeways 0.07%–0.18% $3,600– $4,000 Cambridge Systematics 2009 M H L–M Construct new HOV lanes on all urban freeways 0.05% $1,200 International Energy Agency 2005 L H L–M Truck-only toll lanes Constructed to serve 10% to 40% of VMT in large and/or high- density urban areas 0.03%–0.15% $670–$730 Cambridge Systematics 2009 L H L–M Transit Urban fixed- guideway transit Expansion rate of 2.4%–4.7% annually 0.17%–0.65% $1,800–$2,000 Cambridge Systematics 2009 M H M High-speed intercity rail 4 to 11 new HSR corridors 0.09%–0.18% $1,000–$1,400 Cambridge Systematics 2009 M M M (continued on next page)

120 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Strategy Name Key Deployment Assumptions Fuel/GHG Reduction in 2030 (%) Direct Cost- Effectiveness Data Source Feasibility Te ch n ic al In st it u ti on al P ol it ic al Non-motorized Pedestrian improvements Pedestrian improvements implemented near business districts, schools, transit stations 0.10%–0.31% $190 Cambridge Systematics 2009 H L–M M Bicycle Improvements Comprehensive bicycle infrastructure implemented in moderate to high-density urban neighborhoods 0.09–0.28% $80–$210 Cambridge Systematics 2009 M L M Freight Rail freight infrastructure Aspirational estimates of potential truck– rail diversion resulting from major program of rail infrastructure investments 0.01%–0.22% $80–$200 Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 M M L–H Ports and marine infrastructure and operations Land and marineside operational improvements at container ports 0.01%–0.02% NA Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 M M M–H Construction and Maintenance Practices Construction materialsd Fly-ash cement and warm-mix asphalt used in highway construction throughout U.S. 0.7%–0.8% $0–$770 Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 M–H M M–H TABLE A.6. TRANSPORTATION SYSTEM GHG REDUCTION STRATEGIES (CONTINUED) (continued on next page)

121 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Strategy Name Key Deployment Assumptions Fuel/GHG Reduction in 2030 (%) Direct Cost- Effectiveness Data Source Feasibility Te ch n ic al In st it u ti on al P ol it ic al Other transportation agency activitiesd Alternative fuel DOT fleet vehicles, LEED-certified DOT buildings 0.1% NA Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 H M M–H Transportation System Management and Operations Traffic management Deployment of traffic management strategies on freeways and arterials at rate of 700 to 1,400 miles/nationwide in locations of greatest congestion 0.07%–0.08% [0.89%–1.3%] $40 to >$2,000 Ramp meteringa Centrally controlled 0.01% [0.12%–0.22%] $40–$90 Cambridge Systematics 2009 H H M Incident managementa Detection and response, including coordination through traffic management center 0.02%–0.03% [0.24%–0.34%] $80–$170 Cambridge Systematics 2009 H M H Signal control managementa Upgrade to closed loop or traffic adaptive system 0.00% [0.01%–0.10%] $340–$830 Cambridge Systematics 2009 H M H Active traffic managementa Speed harmonization, lane control, queue warning, hard shoulder running 0.01%–0.02% [0.24%–0.29%] $240–$340 Cambridge Systematics 2009 M M H Integrated corridor managementa Multiple strategies 0.01%–0.02% [0.24%–0.29%] $240–$340 Cambridge Systematics 2009 M M H TABLE A.6. TRANSPORTATION SYSTEM GHG REDUCTION STRATEGIES (CONTINUED) (continued on next page)

122 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Strategy Name Key Deployment Assumptions Fuel/GHG Reduction in 2030 (%) Direct Cost- Effectiveness Data Source Feasibility Te ch n ic al In st it u ti on al P ol it ic al Real-time traffic informationa 511, DOT website, personalized information 0.00% [0.02%–0.07%] $160–$500 Cambridge Systematics 2009 M M H Transit Service Fare reductionse 25%–50% fare reduction 0.02%–0.09% NA Cambridge Systematics 2009 H H H 50% fare reduction 0.3% $1,300 International Energy Agency 2005 Improved headways and LOS 10%–30% improvement in travel speeds through infrastructure and operations strategies 0.05%–0.10% $1,200–$3,000 Cambridge Systematics 2009 L–M L–M M–H Increase service (minimum: add 40% to off peak; maximum: also add 10% to peak) 0.2%–0.6% $3,000–$3,300 International Energy Agency 2005 H H H Intercity passenger rail service expansion Minimum: Increase federal capital and operating assistance 5% annually versus trend. Maximum: Double federal operating assistance, then increase 10% annually 0.05%–0.11% $420–$1,500 Cambridge Systematics 2009 H H H TABLE A.6. TRANSPORTATION SYSTEM GHG REDUCTION STRATEGIES (CONTINUED) (continued on next page)

123 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Strategy Name Key Deployment Assumptions Fuel/GHG Reduction in 2030 (%) Direct Cost- Effectiveness Data Source Feasibility Te ch n ic al In st it u ti on al P ol it ic al Intercity bus service expansion 3% annual expansion in intercity bus service 0.06% NA Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 H M H Truck Operations Truck idling reductionc 30%–100% of truck stops allow trucks to plug in for local power 0.02%–0.06% $50 Cambridge Systematics 2009 H L–M M–H 26%–100% of sleeper cabs with on-board idle reduction technology 0.09%–0.28% $20 Cambridge Systematics 2009 H M M Truck size and weight limits Allow heavy/ trucks for drayage and noninterstate natural resources hauls 0.03% $0 Cambridge Systematics 2009 H M L–M Urban consolidation centers Consolidation centers established on periphery of large urbanized areas; permitting of urban deliveries to require consolidation 0.01% $40–$70 Cambridge Systematics 2009 M L L–M Reduced speed limitsf 55 mph national speed limit 1.2%–2.0% $10 Cambridge Systematics 2009; Gaffigan and Fleming 2008; International Energy Agency 2005 H M–H L TABLE A.6. TRANSPORTATION SYSTEM GHG REDUCTION STRATEGIES (CONTINUED) (continued on next page)

124 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Strategy Name Key Deployment Assumptions Fuel/GHG Reduction in 2030 (%) Direct Cost- Effectiveness Data Source Feasibility Te ch n ic al In st it u ti on al P ol it ic al Land Use Codes, Regulations, and Policies Compact development 60%–90% of new urban growth in compact, walkable neighborhoods (+4,000 persons/ mi or +5 gross units/) (Cambridge) 25%–75% of new urban growth in compact, mixed- use developments (Special Report 298) 0.2%–1.8% 0.4%–3.5% 1.2%–3.9%a $10 Cambridge Systematics 2009 Special Report 298 2009 Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 M L L Parking management All downtown workers pay for parking ($5/ average for those not already paying) 0.2% NA Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 H L L Taxation and Pricing Cap-and-trade or carbon tax Allowance price of $30–$50/tonne in 2030, or similar carbon tax 2.8%–4.6% NA Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 M M L–M VMT fees VMT fee of 2¢ to 5¢/mile 0.8%–2.3% $60–$150 Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 L H L TABLE A.6. TRANSPORTATION SYSTEM GHG REDUCTION STRATEGIES (CONTINUED) (continued on next page)

125 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Strategy Name Key Deployment Assumptions Fuel/GHG Reduction in 2030 (%) Direct Cost- Effectiveness Data Source Feasibility Te ch n ic al In st it u ti on al P ol it ic al Pay-as- you-drive insurance Require states to permit PAYD insurance (low)/ require companies to offer (high) 1.1%–3.5% $30–$90 Cambridge Systematics 2009 L–M L–M M Congestion pricing Maintain level of service D on all roads (average fee of 65¢/mile applied to 29% of urban and 7% of rural VMT) Areawide systems of managed lanes 1.6% 0.5%–1.1% $340 Cambridge Systematics 2009 Energy and Environmental Analysis 2008 L H L Cordon pricing Cordon charge on metro area CBDs (average fee of 65¢/mile) 0.1% $500–$700 Cambridge Systematics 2009 M––H M L Travel Demand Management Workplace TDM (general) Widespread employer outreach and alternative mode support 0.1%–0.6% $30–$180 Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 H L–H H Teleworking Doubling of current levels 0.5%–0.6% $1,200–$2,300 Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 M L M–H TABLE A.6. TRANSPORTATION SYSTEM GHG REDUCTION STRATEGIES (CONTINUED) (continued on next page)

126 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Strategy Name Key Deployment Assumptions Fuel/GHG Reduction in 2030 (%) Direct Cost- Effectiveness Data Source Feasibility Te ch n ic al In st it u ti on al P ol it ic al Compressed work weeks Minimum: 75% of government employees; maximum: double current private participationa 0.1%–0.3% NA International Energy Agency 2005 Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 H L L–H Requirement to offer 4/40 workweek to those whose jobs are amenable (IEA) 2.4% <$1 Ridematching, carpool, and vanpool Extensive rideshare outreach and support 0.0%–0.2% $80 Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 H L–M H Mass marketing Mass marketing in 50 largest urban areas 0.14% $270 Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 H M H Individualized marketing Individualized marketing reaching 10% of population 0.14%–0.28% $90 Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 M M H Carsharing Subsidies for start- up and operations 0.05%–0.20% <$10 Cambridge Systematics 2009 H M H TABLE A.6. TRANSPORTATION SYSTEM GHG REDUCTION STRATEGIES (CONTINUED) (continued on next page)

127 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Strategy Name Key Deployment Assumptions Fuel/GHG Reduction in 2030 (%) Direct Cost- Effectiveness Data Source Feasibility Te ch n ic al In st it u ti on al P ol it ic al Other Public Education Driver education/ eco-driving Reach 10%–50% of population + in-vehicle instrumentation 0.8%–2.3% 3.7% NA Cambridge Systematics 2009 International Energy Agency 2005 L L H Information on vehicle purchasea Expansion of EPA SmartWay program (freight- oriented) and consumer information 0.09%–0.23% NA Cambridge Systematics, Inc., and Eastern Research Group, Inc. 2010 H H H Notes: L, M, and H = low, medium, and high, respectively; LOS = level of service. aTop range (smaller reductions) includes induced demand effects as analyzed in Moving Cooler (Cambridge Systematics 2009); bottom range in brackets (larger reductions) does not. Cost-effectiveness estimates include induced demand effects. b Cost-effectiveness for capacity expansion and bottleneck relief strategies calculated from Moving Cooler data are undefined because net 2010–2050 GHG benefits were negative (2009). c Economically justified capacity expansion based on analysis using the FHWA Highway Economic Requirements System (HERS) model. dMost of the emissions reduced are from other (nontransportation) sectors. Reductions are shown as a percentage of transportation sector emissions for comparison. eFare reductions are considered as a transfer in the Moving Cooler study and therefore have no net implementation cost (2009). The IEA study considers costs to the public sector (lost fare revenues). fPercentage reduction from Gaffigan and Fleming (2008). Direct cost-effectiveness from International Energy Agency’s Saving Oil in a Hurry (2005). Net included cost-effectiveness from Moving Cooler (2009). TABLE A.6. TRANSPORTATION SYSTEM GHG REDUCTION STRATEGIES (CONTINUED) are usually based on a string of annual cost and benefit estimates (including capi- tal costs, annual operating costs, and annual operating GHG benefits) over a 20- to 40-year analysis horizon. Data sources: References providing the source(s) of effectiveness and cost-effec- tiveness data for the strategy. The data source is cited on the same line as its respective cost and/or cost-effectiveness estimate. Feasibility: Feasibility is assessed using a high, moderate, or low rating for three dimensions of feasibility: (continued from page 118)

128 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS • Technological: Is the technology well-developed and proven in practice? What is the likelihood that the technology could be implemented in the near future at the de- ployment levels assumed in the analysis? • Institutional: To what extent do the authority and resources exist for government agencies to implement the strategy, and what is the administrative ease of running a program and the level of coordination required among various stakeholders? • Political: Is the strategy generally popular or unpopular with any interested stake- holders, elected officials, and the general public? What is the political clout of those supporting versus those opposed to the strategy? Feasibility is assessed without respect to cost (which is evaluated in the cost-effec- tiveness measure). TABLE A.7. VEHICLE AND FUEL GHG REDUCTION STRATEGIES Strategy Name Key Market Penetration and Per Vehicle Benefit Assumptions Fuel/GHG Reduction (%) Net Included Cost- Effectiveness Feasibility 2030 2050 Te ch n ic al In st it u ti on al P ol it ic al Low-Carbon Fuels Ethanol (corn)a Maximum near-term corn ethanol production capacity; 68% increase to 60% benefit per E85 vehicle (1.1%)– 0.9% $90–∞ M H M Ethanol (cellulosic) Maximum cellulosic ethanol production capacity in 2030 (33% of LDV market at E85); 57%–115% GHG reduction per vehicle 11%– 23% $10–$30 L L ? Biodiesela Full substitution of diesel with B20 biodiesel blend from soy; 13% GHG reduction to 10% increase per vehicle (1.9%)– 2.9% $130–∞ M M ? Natural gas 2.5%–5% of total U.S. natural gas use diverted to transportation; 15% GHG reduction per vehicle 0.3%– 0.6% ($130) M M ? Electricityb 2030: 18% LDV market penetration, 40%–55% GHG reduction per vehicle 2050: 60% LDV market penetration, 79%–84% GHG reduction per vehicle 2.4%– 3.4% 18%– 22% ($160)–$70 L M ? Hydrogenb 2030: 5% LDV market penetration, 68%–80% GHG reduction per vehicle 2050: 56% LDV market penetration, 78%–87% GHG reduction per vehicle 2.2%– 2.5% 26%– 30% ($20)–($110) L L ? (continued on next page)

129 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Strategy Name Key Market Penetration and Per Vehicle Benefit Assumptions Fuel/GHG Reduction (%) Net Included Cost- Effectiveness Feasibility 2030 2050 Te ch n ic al In st it u ti on al P ol it ic al Advanced Vehicle Technology: Light-Duty Advanced conventional gasoline vehiclesb, c 8%–30% efficiency benefit per vehicle; 60% market penetration in 2030, 100% in 2050 2.5%– 9.0% 4.4%– 16% ($180)–($30) L–H H H Diesel vehiclesb 0%–16% efficiency benefit per vehicle; up to 45% market penetration in 2030, 100% in 2050 0%–4.1% 0%–9.9% ($240)–$660 H H M Hybrid electric vehiclesb 26%–54% efficiency benefit per vehicle; 28% market penetration in 2030, 56% in 2050 2.9%– 5.9% 7.4%– 15% ($140)–$20 M H H Plug-in hybrid electric vehiclesb 46%–70% efficiency benefit per vehicle, 15% market penetration in 2030; 49%–75% per vehicle, 56% market penetration in 2050 3.9%– 5.9% 16.4%– 26% ($40)–($110) L M M Advanced Vehicle Technology: Heavy-Duty On-road trucksc Fleetwide deployment of engine/and resistance reduction technologies, as appropriate for type of vehicle: 17%– 42% per vehicle efficiency benefit 4.4%– 6.4% ($140)–$40 L–H L–M M Vehicle Air Conditioning Systems Refrigerants Replacement of current a/c refrigerant with low global warming potential refrigerant 2.6% $40–$90 M M M Engine load reduction Reflective window glazings, secondary loop a/c systems, and improved a/c system efficiency 0.6%– 1.4% M M M Notes: The use of a “?” indicates that the feasibility of a particular strategy is unknown or is subject to political factors that could be either positive or negative depending on circumstances. Data are from the 2010 report Transportation’s Role in Reducing U.S. Greenhouse Gas Emissions. Estimates are original estimates based on data from numerous literature sources. aCorn ethanol and biodiesel estimates account for indirect effects, such as indirect land use change associated with agricultural production practices, based on analysis by the EPA in support of the proposed Renewable Fuel Standard (RFS2) rulemaking in 2009. The estimates show a wide range of impacts, depending on feedstock source, production methods, and analysis assumptions, and suggest that these fuels may increase GHG emissions under some circumstances. bMarket penetration estimates represent the high end of estimates found in the literature and assume that technology will be developed to the point of marketability in the analysis time frame. cFor advanced gasoline LDV and on-road truck technology, some strategies are proven or well-advanced, but others are not. TABLE A.7. VEHICLE AND FUEL GHG REDUCTION STRATEGIES (CONTINUED)

130 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Combined Strategy Impacts and Interactive Effects Many GHG emissions reduction strategies interact to produce different outcomes for total GHG reductions. The benefits of each strategy (or group of strategies) are not addi tive, and in fact may be reduced depending on other strategies that are imple- mented. However, some strategies are complementary or synergistic, and their effec- tiveness is likely to be enhanced if they are implemented in combination with each other. As an example of synergistic effects, transit, nonmotorized improvements, land use, and pricing strategies would be expected to be most effective when applied in combination. For example, a study by the Center for Transit-Oriented Development compared CO2 emissions per household based on characteristics including access to rail transit and neighborhood land use characteristics to characterize location effi- ciency. Compared with the average metropolitan area household, households in transit zones that fell into the two middle categories of location efficiency produced 10% and 31% lower transportation emissions, and households in the highest location-efficient category produced 78% lower transportation emissions than the average metropolitan area household (Haas et al. 2009). The Moving Cooler study also found that transit and nonmotorized improvements were more effective in areas of higher population density (Cambridge Systematics 2009). It may also be expected that strategies (such as road pricing) that encourage the use of alternative modes would have a greater impact when applied in conditions under which better alternatives exist (as would be found with increased transit investment and more compact land use patterns). Quantitative evidence on the interactive effects among various strategies in combination is limited, and existing evidence is generally based on simplified analysis. More sophisticated analysis of combined effects would require the use of an enhanced regional modeling system and careful selection of comparison scenarios. Three research studies have made assumptions concerning the synergistic effects of implementing different GHG emissions mitigation actions as part of a GHG mitiga- tion strategy. The Moving Cooler study created six strategy bundles and combined the individual benefits of strategies in each bundle in a multiplicative fashion (Cambridge Systematics 2009). For example, if Strategy A results in a 10% GHG emissions reduc- tion, and Strategy B results in a 10% reduction, the combined effect will be (1 – 0.10) × (1– 0.10) = 0.90 × 0.90 = 0.81, or a 19% combined emissions reduction, rather than a 20% reduction if they were simply added. The study also accounted for synergies among certain strategies; in particular, transit, bicycle, pedestrian, and carsharing strategies were assumed to be more effective in areas of greater population density, and therefore more effective under more aggressive land use scenarios. The six bundles resulted in reductions in GHG emissions versus the surface transportation baseline ranging from 3% to 11% in 2030 at aggressive levels of implementation, increasing to as much as 18% in 2050. Reductions under a maximum implementation scenario ranged as high as 17% in 2030 and 24% in 2050. Cost-effectiveness was also provided for each bundle. The estimated cost- effectiveness, including implementation costs only, ranged from a low of $80 per tonne for the low-cost bundle to more than $1,600 per tonne for a facility pricing bundle

131 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS that combines infrastructure improvements with local and regional pricing measures to pay for these improvements. The study concluded that a net savings would be real- ized for most bundles if vehicle operating cost savings were counted against the direct implementation costs. Based on information included in a U.S. DOT report to Congress (2010), Cambridge Systematics, Inc. developed combined GHG emissions reduction estimates for five categories of strategies: pricing carbon, low-carbon fuels, vehicle fuel efficiency, system efficiency, and travel activity. Mutually exclusive or redundant strategies were excluded from the combined estimates. The results showed that in the long term the most effective strategies for reducing GHG emissions were introducing low-carbon fuels, increasing vehicle fuel efficiency, and reducing carbon-intensive activity. The most rigorous attempt to consider the combined effects of different mitigation actions (or perhaps more correctly to avoid double-counting of energy reduction due to strategy implementation) is found in the Pew Center report Reducing Greenhouse Gas Emissions from U.S. Transportation (Greene and Plotkin 2011). This study used equations that decomposed the contributing factors that determined emissions from different modes, vehicle types, and fuels. The analysis also accounted for the rebound effect, which occurs when energy efficiency strategies reduce the use of energy. This reduction in energy use lowers the cost of energy, leading to increased consumption of energy and in some portion offsetting the benefits of increased efficiency. Readers interested in this approach are encouraged to read the Pew report. Other Studies Other studies have examined the potential for transportation sector GHG emissions reductions, but primarily for vehicle and fuel technology rather than travel activity and system efficiency. Bandivadekar et al. (2008) conclude that a 30%–50% reduction in fuel consumption is feasible over the next 30 years. In the short-term, this will come as a result of improved gasoline and diesel engines and transmissions, gasoline hybrids, and reductions in vehicle weight and drag…Over the longer term, plug-in hybrids and later still, hydrogen fuel cells may enter the fleet in numbers sufficient to have significant an impact on fuel use and emissions. Lutsey (2008), considering costs and effectiveness from a cross-sectoral perspec- tive, concludes that Transportation technologies are found to represent approximately half of the “no regrets” mitigation opportunities and about one-fifth of the least-cost GHG mitigation measures to achieve the benchmark 1990 GHG level. With the adoption of known near-term technologies, GHG emissions by 2030 could be reduced by 14% with net-zero-cost technologies, and emissions could be reduced by about 30% with technologies that each have net costs less than $30 per tonne of carbon dioxide equivalent reduced.

132 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Top-down, aspirational or scenario estimates of potential travel activity and sys- tem efficiency benefits have also been developed. These estimates make assumptions regarding what percentage vehicle miles traveled (VMT) reduction is needed or can be obtained to contribute to certain GHG emissions reductions in conjunction with other (non-VMT) strategies, rather than building from the bottom up according to individual strategy effects. As an example, an EPA wedge analysis of the transporta- tion sector assumes that a 10% to 15% reduction in VMT from travel demand man- agement strategies can contribute to GHG reductions along with vehicle efficiency and low-carbon fuel improvements (Mui et al. 2007). Another example of such a scenario approach is provided by the National Coop- erative Highway Research Program (NCHRP) Project 20-24, Task 59 study, which examines transportation GHG emissions through 2050 (Burbank 2009). This study makes assumptions about the reduction in carbon intensity of the vehicle fleet (58% to 79% reduction in carbon emissions per vehicle mile), reduction in growth of VMT (to 0.5% to 1.0% annually), and improvements in system operating efficiencies (pro- viding a 10% to 15% GHG emissions reduction). The resulting GHG emissions are compared against 2050 goals as established in various national and international cli- mate change proposals or initiatives. The various scenarios result in transportation GHG emissions levels from 44% to 76% below a 2005 baseline. Lutsey (2008) considers the VMT reductions needed to achieve aggressive GHG emissions reduction targets (80% reduction below 1990 levels by 2050) even after vehicle and fuel technology strategies have been fully realized. After deploying the level of GHG reduction technology for vehicles and fuels as described in this study (and no further advances), the travel demand reduction to achieve the 2050 target would be quite severe. For this amount of GHG reductions to come from travel reductions, national light-duty vehicle travel would have to be reduced annually by approximately 4%, instead of the fore- casted increase of about 1.8% annually from 2010 on. . . . Even after a new crop of vehicle and fuel technologies (e.g., plug-in hybrid-electric vehicles) emerges, it appears safe to speculate that some significant amount [of] reduc- tion in vehicle-miles-traveled will be needed to augment technology shifts to achieve deeper, longer-term GHG reductions. Summary There are no simple answers to the question of what are the most and least cost- effective strategies. The cost-effectiveness of most transportation system strategies depends greatly on what is included in the assessment of costs and cost savings. One way to look at cost-effectiveness is simply from the public agency perspective of the direct implementation costs. Including vehicle operating cost savings generally provides a much different picture, because consumers save money on fuel, maintenance, and so forth. However, even this is an incomplete accounting in that it does not consider fac- tors such as travel time savings, other welfare gains or losses (due to accessibility and increased or decreased convenience), or equity (incidence of costs and benefits across population groups). These factors represent important impacts of transportation

133 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS projects, but they are rarely quantified in GHG cost-effectiveness analysis. Therefore, the cost-effectiveness estimates shown in Table A.4, in particular, are incomplete and may not accurately represent full social costs and benefits. Furthermore, there is considerable uncertainty in the estimates for many strategies. Existing knowledge of both costs and benefits is in many cases limited, with estimates based on only a single study. In addition, drawing blanket conclusions about any par- ticular strategy is risky. Many individual projects or policies may be very cost-effective in one context but not at all cost-effective in another (e.g., a transit project in an area of high versus low population density). The cost-effectiveness estimates for the vehicle and fuel technology strategies shown in Table A.7 are much closer to a full social cost representation, because the non- monetary impacts of these strategies are for the most part relatively minor (there may be some impacts on vehicle performance, such as reduced range for electric vehicles). However, many of these estimates reflect considerable uncertainty over technological and economic factors, such as the time frame for technology advancement, future cost of the technology, future fuel prices, indirect effects of biofuels, and other factors. With these caveats in mind, the following conclusions can be drawn from the cost-effectiveness data. The largest absolute GHG benefits in the transportation sector are likely to come from advancements to vehicle and fuel technologies. Particularly promising technologies in the short- to midterm include advancements to conventional gasoline engines, truck engine improvements and drag reduction, and hybrid elec- tric vehicles. In the longer term, ethanol from cellulosic sources, battery-powered electric vehicles, plug-in hybrid electric vehicles, and hydrogen fuel cell vehicles all show great promise for reducing GHGs, but only if the technologies can be advanced to the point of being marketable and cost-competitive. Most of these strategies show the potential for net cost savings to consumers. The U.S. DOT (2010) estimates that hydrogen fuel cell vehicles could reduce per vehicle GHG emissions by 80% by relying on low-carbon sources for hydrogen production. Advanced gasoline vehicles could reduce per vehicle emissions by 8% to 30%, hybrid vehicles by 26% to 54%, and plug-in hybrids by 46% to 75%. The impacts of any single transportation system strategy (system efficiency and travel activity) are generally modest, with most strategies showing impacts of less than (and usually considerably less than) 1% of total transportation GHG emissions in 2030. A few strategies, including reduced speed limits, compact development, various pricing measures, and eco-driving, show larger impacts (greater than 1%); but the ability to implement these strategies at sufficiently aggressive levels is uncertain due to institutional and/or political barriers. For example, decreasing GHG emissions per VMT could reduce transportation GHG emissions by 3% to 6% through a combination of strategies such as the enforcement of lower speed limits, traffic signal synchronization, ramp meter- ing, and truck idle reduction (U.S. Department of Transportation 2010). Strategies that decrease carbon-intensive travel activity could reduce transportation GHG emissions by 5% to 17% in 2030. This approach includes measures to reduce VMT growth through pricing, compact development, improved public transportation, enhancements to bike and pedestrian facilities, and the promotion of eco-driving through driver education and

134 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS in-vehicle feedback technology. Thus, despite the modest individual strategy impacts, the combined effects of all transportation system strategies may be significant, on the order of 5% to 20% of transportation GHG emissions. Transportation infrastructure investment, whether highway or transit investment, is generally high cost. Based on limited evidence, bicycle and pedestrian improvements may be relatively lower cost (in the range of $200 per tonne), although the magnitude of impacts is likely to be very modest. Although major infrastructure investments are not among the most cost-effective GHG emissions reduction strategies, they may be worthwhile for other purposes, such as mobility, safety, or livability, or as part of a package of strategies that is collectively more cost-effective (e.g., transit with land use, bottleneck relief with congestion pricing). The Federal Highway Administration (FHWA) is currently funding research into the GHG benefits of highway capacity expansion and bottleneck relief when combined with congestion pricing. Although rail and marine freight are considerably more energy efficient than truck travel on average, the absolute magnitude of reductions from freight mode shifting is limited because only certain types of goods (particularly long-haul, non-time-sensitive goods) can be competitively moved by rail. One estimate of the cost-effectiveness of rail freight infrastructure improvements falls in the range of $200 per tonne, but this is based on highly aspirational estimates of truck–rail mode shift. Improved estimates are needed to assess the GHG emissions reductions and cost-effectiveness of rail and marine freight investments to encourage freight mode shift. Transportation system management strategies that reduce congestion and improve traffic flow may provide modest GHG emissions reductions at lower cost than capacity and/or system expansion (typically between $50 and $500 per tonne, with lower costs if operating cost savings to drivers are included). As with highway capacity strategies, however, there is considerable uncertainty in the GHG reduction estimates for these strategies because of uncertainty regarding the magnitude and treatment of induced demand. Like transit infrastructure improvements, urban and intercity transit service improvements have high direct (public sector) costs, generally more than $1,000 per tonne, although they provide similar nonmonetary (mobility) benefits and in some circumstances they may yield net savings to travelers as a result of personal vehicle operating cost savings. The GHG benefits of any particular transit project will vary depending on ridership levels, and they could be negative if ridership is insufficient. Truck operations strategies, in particular idle reduction, can provide modest total benefits with a low public investment cost while yielding net cost savings to truckers. The most effective strategy is to require on-board idle reduction technology, which would require harmonization of state regulations. Speed limit reductions or greater enforcement of existing speed limits can provide significant benefits at modest cost, although they have mobility disadvantages and are not likely to be popular. Studies that have examined land use strategies show a large range of potential reductions in GHG emissions. For example, the Moving Cooler report estimated a 6% to 21% GHG emissions reduction impact for its land use bundle (Cambridge

135 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Systematics 2009); a Transportation Research Board study estimated a 6% to 12% reduction with significant changes in land use policies and investments in transit (Spe- cial Report 298 2009); and the Pew Center study assumed a 5% GHG emissions reduction from land use strategies in 2050 (Greene and Plotkin 2011). Potentially important GHG emissions reductions over the long term could occur from land use strategies, at very low public-sector cost, if stringent public policies are enacted to encourage compact development, and supporting investments are made in transit and nonmotorized transportation options. Modest to moderate changes in land use pat- terns can probably be accomplished without significant loss of consumer welfare, but more far-reaching changes may not be popular and may be difficult to achieve in the current political and economic environment (Special Report 298 2009). Pricing strategies, especially those that affect all or a large portion of VMT, such as VMT-based fees or congestion pricing, can provide significant GHG emissions reduc- tions, but only by pricing at levels that may be unacceptable to the public (the 2- to 5-cent per mile fee analyzed in Table A.6 is equivalent to a gas tax increase of $0.40 to $1.00 per gallon at today’s fuel efficiency levels). Implementation costs are mod- erate (less than $100 per tonne to $300 per tonne or more) for most mechanisms, because of the technology and administrative requirements for VMT monitoring. Cost- effectiveness improves with higher fee levels, because the same monitoring and administration infrastructure is required regardless of the amount of the fee. Pricing strategies will have significant equity impacts unless revenues are redistributed or reinvested to benefit lower-income travelers. A gas tax increase or carbon tax could be implemented at much lower administrative cost, but these strategies are not currently politically acceptable at a national level or in most states. Transportation demand management strategies have a modest GHG emissions reduction potential at moderate public cost (typically in the range of $100 to $300 per tonne), but they require widespread outreach efforts combined with financial incen- tives. Furthermore, the public sector has so far demonstrated little ability to influence strategies such as telecommuting and compressed work weeks, and adoption of these strategies has primarily been driven by private initiative. Studies have suggested that eco-driving may significantly reduce GHG emissions while providing a net savings to travelers. However, these results are based on limited European experience and have not yet been tried in any significant way in the United States. GHG ANALYSIS TOOLS Travel Demand and Related Models Travel Demand Models Travel demand models are a commonly used tool to forecast traffic conditions based on future socioeconomic and demographic projections by traffic analysis zone and alter native transportation networks. All MPOs are required to maintain travel demand forecasting models for use in transportation planning and, if needed, air quality analy- sis. Some regional planning agencies located outside of metropolitan areas may also

136 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS maintain travel demand models, and some state DOTs have developed statewide models. These models have varying capabilities for GHG analysis. All produce traffic vol- umes and speeds for each link in the modeled roadway network that can be used in conjunction with an emissions factor model such as MOVES or EMFAC to develop GHG emissions estimates from highway vehicle travel. They are best suited for analyz- ing changes in the transportation network such as capacity expansion or new road- ways. In addition, it may be possible to analyze the following GHG strategies using some regional travel demand models (see also Sun et al. 2009): • Transit capacity expansion or service improvements: Some models, especially for those used in larger metropolitan areas, have a transit component, including a transit network and mode choice model, which can be used to forecast VMT reductions from transit improvements. The sensitivity of the model for the particular transit improve- ments of interest should be evaluated by the analyst. • Regional land use patterns: These models can be used to test changes in regional land use patterns (e.g., focus on infill, transit corridors, or activity centers) by changing the distribution of future population and employment among traffic analysis zones. • Bicycle and pedestrian improvements and land use design: Some models include nonmotorized mode choice, although only a few have been enhanced to be sensitive to the effects of nonmotorized infrastructure improvements. Techniques such as 4-D postprocessors can be used in conjunction with travel model output to estimate the travel and resulting emissions impacts of changes to the various land use–related D metrics (e.g., density, diversity, design, destination accessibility). • Pricing: Travel demand models generally forecast travel based on generalized travel cost, which is based on both the cost and the time of making a trip. However, the ability to model the effects of pricing measures depends on the particular measure (e.g., tolling, congestion pricing, parking pricing) and the model structure and calibra- tion. Most models will need some level of enhancement to reasonably capture effects such as time-of-day shifting from congestion pricing or changes in the number of total trips taken. For a discussion of travel demand model strengths, limitations, and enhancements, with a specific emphasis on smart growth and nonmotorized travel, see Assessment of Local Models and Tools for Analyzing Smart-Growth Strategies (DKS Associates and University of California 2007). Donnelly et al. (2010) consider advanced practices in travel demand forecasting, including integration with emissions models. Integrated Transportation–Land Use Models A method of forecasting the land use impacts of transportation investments, and the subsequent feedback to VMT and transportation network conditions, is necessary if the induced demand effects of transportation improvements are to be fully captured

137 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS (see “Indirect Effects and Induced Demand” below). Many MPOs have a land use forecasting model for developing future projections of population and employment. Only a few, however, have developed integrated transportation and land use forecast- ing models (such as UrbanSim or PECAS) that are highly sensitive to both transporta- tion improvements and various land use policies. In most cases, the use of these models for GHG analysis will not be an option because they are highly resource intensive to develop. The existing model applications are most appropriate for regional and systems-level analysis and have not yet been proven for use in analyzing individual transportation projects. Less resource-intensive methods, however, have been applied to assess the land use impacts of transportation investments and capture the resulting feedback through travel demand models. For basic information on transportation and land use modeling, including inte- grated models as well as other methods, refer to NCHRP Report 466 (Louis Berger Group 2002) for a core guidance document that provides information and guidance on the various methods available for land use forecasting. This report is complemented by Forecasting Indirect Land Use Effects of Transportation Projects (Avin et al. 2007). Interim Guidance on the Application of Travel and Land Use Forecasting in NEPA identifies land use forecasting methods across a range of levels of effort suitable for use in project-level analysis (Federal Highway Administration 2010). Intelligent Transportation Systems Deployment Analysis System (IDAS) FHWA developed IDAS as a sketch planning tool to estimate the impacts, benefits, and costs resulting from the deployment of intelligent transportation system (ITS) compo- nents. IDAS interfaces with regional travel demand model output and can be used to estimate CO2 emissions as well as other impacts. CO2 factors are sensitive to speed and facility type (freeway or arterial). However, as of 2010 the factors had not been updated to account for federal fuel economy and GHG emissions standards adopted in 2010 or for California Pavley GHG standards. Traffic Simulation Models Traffic simulation models are used to evaluate the impacts of changes in transportation network characteristics (e.g., capacity, roadway geometry, signal timing, or ITS strate- gies) on traffic flow patterns (e.g., vehicle speeds, acceleration, and delay). Examples include TSIS-CORSIM, VISSIM, Paramics, SimTraffic, TransModeler, and Aimsun. Most of these models have internal data sets and modules for calculating changes in fuel use and air pollutant emissions resulting from changes in traffic characteristics (speed and acceleration). Although most traffic simulation models do not currently produce GHG emissions estimates, fuel CO2 emission factors (see Table A.8) can be applied to fuel use changes to determine changes in GHG emissions. Model output on traffic conditions can also be used in conjunction with EPA’s MOVES to incorpo- rate the most up-to-date relationships between vehicle characteristics, operations, and emissions.

138 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS TABLE A.8. CO2 EMISSION FACTORS BY FUEL TYPE Fuel CO2 Emission Factor (kg/gal) Gasoline 8.81 Diesel 10.15 E10 (gasoline with 10% ethanol) 7.98 Source: General Reporting Protocol, Version 1.1 (The Climate Registry 2012); Fuel Emission Factors (Energy Information Administration 2012b). Traffic simulation models can be divided into two general classes: mesoscopic and microscopic. Mesoscopic models, including Dynasmart, TransModeler, Dynus-T, and VISTA, are based on deterministic relationships between roadway and intersec- tion characteristics and traffic flow; microscopic models simulate the movement of indi vidual vehicles through the network being modeled. Many software packages are capable of modeling both mesoscopically and microscopically, and some can run both simulations within the same model. Microscopic models can be further divided into deterministic models and dynamic models. Deterministic models, including Synchro/SimTraffic, CORSIM, and FREQ, simulate predetermined traffic volumes and turning movements that are input by the user. Dynamic simulation models, including Paramics, VISSIM, TransModeler, Aimsun, and Dynasim, microsimulate origin–destination patterns that allow vehicles to dynamically reroute from origin to destination based on real-time congestion in the system, driver information, and alternative routes. The predetermined volumes in deterministic models make them difficult to use for network analysis because reassigning vehicles is not possible. These packages are more suited for individual intersection analysis or signalized arterial corridors. Dynamic models capable of simulating origin–destination tables are designed to per- form network-level analysis of mixed facility types (freeways and arterials), as well as transit and pedestrian operations. GHG Inventory and Policy Analysis Tools This category includes tools with a wide variety of characteristics. Their common thread is that they are all specifically designed to assist transportation agencies in cal- culating GHG emissions and/or reductions from transportation sources. CCAP Transportation Emissions Guidebook The purpose of this guidebook (Center for Clean Air Policy 2012) is to assist state and local officials in understanding the extent to which policy decisions affect air pollu- tion, energy use, and GHG emissions. The guidebook provides guidance, rough esti- mates, and a spreadsheet calculation tool to assess the impact of various strategies and technologies. The guidebook is now a few years old, and Part 2 (Vehicle Technology and Fuels), in particular, may contain some information that is out-of-date regarding vehicle technologies and emissions impacts.

139 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS CACP 2009: Clean Air and Climate Protection This software tool, available from ICLEI–Local Governments for Sustainability, was developed in partnership with the National Association of Clean Air Agencies and EPA (ICLEI 2012a). The software is used to develop local communitywide or internal government GHG emissions inventories, quantify emissions reductions from various emissions reduction measures, project future emissions levels, and set reduction tar- gets and track progress toward meeting these targets. The model includes CO2, CH4 (methane), and N2O (nitrous oxide) emissions, as well as criteria pollutants. Inputs include fuel use or VMT by a government vehicle fleet or the community as a whole. Therefore, this tool is best suited to translating VMT and/or fuel consumption (or changes in these) into GHG emissions (or changes), rather than directly estimating the VMT or fuel consumption impacts of strategies. The calculations are based on the principles and methods included in the Local Government Operation Protocol devel- oped by ICLEI in collaboration with the California Air Resources Board (CARB) and The Climate Registry. Many cities have used the ICLEI software to develop baseline estimates of GHG emissions from transportation and other sectors. Climate and Air Pollution Planning Assistant (CAPPA) This software tool, also developed by ICLEI, is a simple spreadsheet-based tool to estimate the GHG benefits of a wide variety of transportation-related policies and strategies, including travel reduction and vehicle and fuel technology strategies, as well as nontransportation strategies (ICLEI 2012b). Its focus is on measures that can be implemented at a local (municipal) level. It includes more than 100 municipal actions (e.g., vehicle fleet purchases and light-emitting diode traffic signal replacement) and community actions (e.g., transit-oriented development and bicycle programs). The purpose of the application is to help decision makers choose a suite of measures that when combined would get them to their jurisdiction’s reduction goal, rather than to model the impact of any particular measure in a detailed way. A limitation of the tool is that it generally requires user inputs of traveler response factors (such as increased transit ridership or nonmotorized travel) rather than predicting response. Users should carefully review the default assumptions embedded in the model. Climate Leadership in Parks (CLIP) Tool Developed by EPA and the National Park Service, the CLIP tool allows for GHG and criteria pollutant emissions estimation at the local level for all highway and non highway transportation and mobile sources, including off-road sources such as construction equipment (National Park Service 2012). Although default vehicle char- acteristics are geared toward travel situations at national parks, CLIP allows users to enter additional data to reflect local conditions. The user must estimate activity parameters such as VMT reduction, fuel use reduction, or percentage idle time reduced, and the tool converts these inputs to CO2 emission reductions. The tool includes six strategies: (1) reduce visitor VMT; (2) reduce fuel consumption among park, conces- sionaire, and other vehicles; (3) reduce fuel consumption among nonroad equipment; (4) replace existing park, concessionaire, and other vehicles with more fuel-efficient

140 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS vehicles; (5) replace existing park, concessionaire, and other vehicles with alternative fuel vehicles and hybrids; and (6) reduce vehicle idling. FHWA Carbon Calculator Tool FHWA-sponsored work was underway in the fall and winter of 2010 to develop a carbon calculator tool. The tool will use GreenSTEP (see description below) as a foun- dation for broader use by state DOTs and MPOs when analyzing various GHG-related scenarios, and it is likely to be a major contribution to the range of tools available to practitioners. Readers are encouraged to obtain the latest information on this cal- culator for their analysis. GHG Calculator for State DOTs (GreenDOT) This software tool, developed for NCHRP Project 25-25, Task 58, calculates CO2 emissions from the operations, construction, and maintenance activities of state DOTs (National Cooperative Highway Research Program 2012). GreenDOT is designed to calculate emissions for geographical areas ranging from a single project to an entire state, and over time periods ranging from one day to several years. The two most likely uses of the tool are calculating annual agencywide emissions and emissions related to a specific project, covering a period of days or years. The tool’s four modules calcu- late emissions from on-road vehicles, off-road equipment, electricity used in trans- portation facilities, and construction materials. In addition, an auxiliary calculator for traffic-smoothing strategies estimates changes in GHG emissions on a roadway segment based on changes in average traffic speed. GreenDOT calculates a baseline scenario and a mitigated scenario for all modules and includes a number of common mitigation strategies, often with default percentage reductions built in. However, the tool requires detailed inputs, such as gallons of fuel for off-road equipment, metric tons of concrete and asphalt, and megawatt-hours of electricity usage. GreenSTEP The Greenhouse Gas Statewide Transportation Emissions Planning model ( GreenSTEP) is a tool originally developed by the Oregon DOT for estimating the GHG emis- sions reduction potential of policy proposals for the land use and transportation sub committee of Oregon’s Global Warming Commission. GreenSTEP is designed to estimate the effects of policy changes on factors that influence GHG emissions, in- cluding metropolitan population densities and relative amounts of urban and rural development; capacity and use of transit service and highways; use of alternative fuel or technology vehicles, vehicle fuel efficiency, and future market share of efficient auto- mobiles; the carbon content of fuels and fuel costs; potential VMT-based fees and other vehicle charges that may be levied; and GHG emissions from electrical power generation. GreenSTEP also allows modeling of several types of travel demand man- agement and the potential for switching more travel to bicycles and other light-weight vehicles (e.g., electric bicycles). Version 2 of the model, developed in the fall of 2010, focuses on LDV travel, but Oregon DOT plans to add long-distance travel and freight models and to develop a metropolitan-area version of GreenSTEP.

141 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS New York State DOT Draft Guidance on Transportation GHG Analysis The New York State Department of Transportation (NYSDOT) developed a series of draft guidance documents to assist in calculating the fuel consumption and GHG im- pacts of transportation projects for project alternatives analysis and for MPOs’ long- range transportation plans and transportation improvement programs. The methods account for the direct impacts of vehicle speeds on fuel consumption and indirect impacts from construction and maintenance activities, relying on procedures summa- rized in the 1983 Caltrans Energy and Transportation Systems manual. NYSDOT has also developed the Motor Vehicle Emission Simulator–Roadway and Rail Energy and Greenhouse Gas Analysis Extension (MOVES-RREGGAE), an interface designed for NYSDOT that provides a platform for estimating energy and GHG emissions associ- ated with transportation projects, plans, and improvement programs in New York State. MOVES-RREGGAE extends EPA’s MOVES-HVI Demo model by enabling analyses of energy and GHG emissions from the operation of roadway projects, plans, and programs. MOVES-RREGGAE also includes modules for calculating energy and GHGs from the construction, maintenance, and rail components of a project, accord- ing to NYSDOT’s guidance documents. State Inventory Tool (SIT) EPA-developed SIT is a spreadsheet-based tool designed to develop comprehensive GHG inventories at the state level using a combination of state-specific inputs provided by the user and default data preloaded for each state (U.S. Environmental Protection Agency 2012c). SIT covers all sectors of the economy, including all on-road and off- road transportation modes. Multiple calendar years can be modeled simultaneously. To estimate CO2 emissions, SIT uses fuel consumption data (measured in British ther- mal units), which can be a user input or default data. Default fuel consumption data come from the Energy Information Administration’s State Energy Data. Estimates of N2O and CH4 emissions from marine vessels, aircraft, and locomotives also use fuel consumption data. To estimate N2O and CH4 emissions from highway vehicles, state- level VMT data are required for each vehicle type; users can apply their own data or use SIT’s preloaded default vehicle mix data, which come from FHWA’s Highway Statistics. Inputs of emissions factors for each fuel and vehicle type are also required. SIT has been used for many GHG inventories and forecasts developed for state climate action plans. This model does not estimate highway vehicle CO2 emissions separately from total transportation CO2 emissions or allocate CO2 emissions to spe- cific highway vehicle types. For example, the transportation estimate of diesel CO2 from SIT includes diesel fuel used by highway vehicles, locomotives, and commercial marine vessels. Methods for allocating the transportation fuel consumption and emis- sions by transportation category have been developed by various analysts, but these are not included with SIT. Emissions forecasts can be developed using SIT baseline emissions. On-road vehi- cle emissions can be projected based on total VMT growth rates by vehicle type at the state level, if available. If state-level VMT growth rates by vehicle type are not avail- able, they can be developed from the national vehicle type VMT forecasts reported

142 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS in the Annual Energy Outlook 2009 (Energy Information Administration 2009). If CO2 emissions are projected based on VMT growth rates, they should be adjusted to account for anticipated improvements in fuel efficiency. URBEMIS URBEMIS (Urban Emissions) is environmental management software that was origi- nally developed by CARB as a modeling tool to assist local public agencies with estimating air quality impacts from land use projects when preparing a California Environmental Quality Act analysis (Urbemis 2012). The model was developed as a user-friendly computer program that estimates construction, area source, and opera- tional air pollution emissions from a wide variety of land use development projects, including residential neighborhoods, shopping centers, and office buildings. The model also identifies mitigation measures and emissions reductions associated with specific mitigation measures. The mobile source mitigation component allows the user to estimate the potential vehicle travel and emissions reduction benefits from various land use and transportation-related strategies within the project site and in the surrounding area. These strategies include pedestrian and bicycle facilities, public transit facilities and service, the design and mix of land uses, on-site services, and other measures such as telecommuting and alternative work schedules. The model uses the Institute of Transportation Engineers’ Trip Generation Manual and CARB’s EMFAC2007 model for on-road vehicle emissions and OFFROAD2007 model for off- road vehicle emissions (Urbemis 2012). Nearly all the model defaults can be modified if more accurate information is available. The outputs of URBEMIS include total trips, total VMT, and annual tons of volatile organic compounds, oxides of nitrogen, carbon monoxide, sulfur dioxide, CO2, and 2.5- and 10-µm particulate matter. Other Travel Demand Analysis Tools COMMUTER The EPA-developed COMMUTER model is designed to analyze the impacts of employer- or worksite-based transportation demand management programs and tran- sit improvements on VMT, criteria pollutant emissions, and CO2 (U.S. Environmental Protection Agency 2012e). The model can also be adapted for sketch-level analysis of general responses to pricing policies or to measures that affect travel time. The CO2 calculations are simple and based on default emission factors from MOBILE6. Because the emission factors are MOBILE6-based, this model will not show the impacts of changes in speeds. This model was most recently updated in 2005 and reflects fleet- wide average fuel economy at that time. TRIMMS Developed by the Center for Urban Transportation Research at the University of South Florida, TRIMMS (Trip Reduction Impacts for Mobility Management Strategies) is a spreadsheet model to predict trip, VMT, fuel, and emissions impacts for worksite-based transportation demand management programs (University of South Florida 2012). It has many similarities to the COMMUTER model but uses different methodologies.

143 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS The model is also intended for cost-benefit assessment and incorporates damage costs for various pollutants. As with COMMUTER, emissions factors are not speed-based. Version 2.0, released in 2009, reflects fleetwide average fuel economy at that time. Land Use Scenario Planning Tools These geographic information system–based tools (including INDEX, Smart Growth INDEX PLACE3S, CommunityViz, CorPlan, and others) are primarily designed to assist planners with the development and analysis of alternative land use scenarios at a site, community, or regional level. Tool outputs include a wide variety of community indicators related to transportation, land use, the environment, and other issues such as VMT per capita or household, fuel consumption, and GHG emissions. The models typically estimate changes in VMT based on elasticities, or relationships between fac- tors such as population density, land use mix, and pedestrian design and vehicle travel. The estimates therefore tend to be relatively simplistic because they usually do not account for the regional context of the development, which tends to have a greater impact on vehicle travel and GHG emissions than the characteristics of an individual development. However, these tools can be of value in creating inputs (i.e., in the form of land use changes) to a regional travel demand model that can be used for GHG emissions analysis purposes. They also can estimate energy use and GHG emissions from buildings, taking into consideration factors such as building density, orientation, floor space, and mix of housing types. They are relatively data intensive to set up; in particular, they require detailed land use data, and (except for EPA’s Smart Growth INDEX) are not intended for evaluating transportation network changes. FHWA’s Tool Kit for Integrating Land Use and Transportation Decision- Making, although a few years old, includes several case studies and examples of scenario plan- ning and visioning projects using these and other geographic information system–based tools (Federal Highway Administration 2005). Smart Growth INDEX is available free from the EPA. PLACE3S is available from the California Energy Commission, INDEX from Criterion Planners, CommunityViz from Placeways, and CorPlan from the Renaissance Planning Group. Emissions Factor and Fuel Economy Models GlobeWarm GlobeWarm is a tool developed by the Washington State DOT to help easily esti- mate GHG emissions at a planning level using either transportation systemwide sum- mary travel data or link-by-link travel model data. It incorporates emissions data from MOVES but does not require the user to run MOVES. Inputs related to the vehicle fleet and technology include vehicle age distribution, fuel types and market shares, vehicle fuel efficiency, vehicle emissions control technology, and GHG emissions reduction factors for alternative fuels. Defaults are provided (primarily from EPA data) for all of these inputs. System-level GHG emissions estimates can be developed by providing data on average trip length, percentage of trips beginning with a cold start, VMT, and vehicle hours traveled. Link-level estimates can be developed using link-level travel demand model output of VMT and speeds by vehicle type. The tool estimates three

144 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS primary GHGs: CO2, CH4, and N2O. The summary output from the tool includes es- timated quantity of GHGs in the base and alternative cases and percentage change of GHG emissions from the base to the alternative case. Motor Vehicle Emission Simulator EPA’s Motor Vehicle Emission Simulator (MOVES2010) model was released in December 2009 and is officially approved for use in state implementation plans and for transportation conformity analyses outside of California. MOVES replaces EPA’s previous MOBILE6 and NMIM models. MOVES can estimate CO2, N2O, and CH4 from on-road vehicles and accounts for the impacts of vehicle speeds, driving cycles, age, and vehicle stock on emissions. MOVES can be used to develop GHG emissions estimates at project, county, regional, statewide, and national levels. The model is best suited for evaluation of GHG emissions reductions from measures that would change travel characteristics on roadways (e.g., speeds, congestion levels, or idling times). Emission Factors Model CARB developed the Emission Factors (EMFAC) model as the California counterpart to EPA’s MOBILE (now MOVES) model. Using emission factors and vehicle activity inputs, EMFAC develops emissions estimates for on-road vehicles to be used in devel- oping emissions inventories, projections, and other project-level analyses. The CO2 emission rates vary by vehicle speed. According to the EPA, EMFAC CO2 predictions are close to the output from MOVES. EMFAC combines locally specific emission rates and vehicle activity to generate hourly or daily total emissions for geographic areas (statewide, air basin, air pollu- tion control district, or county) in California (California Air Resources Board 2010b). EMFAC estimates fuel consumption for gasoline and diesel, as well as emissions of CO2 and CH4 (but not N2O). The model performs separate calculations for each of 13 classes of vehicles by fuel usage and technology group. EMFAC contains local data for each county in California; however, the user can edit inputs such as VMT, vehicle population, technology fractions, speed fractions, and other factors. EMFAC can be run in three modes: Burden, Emfac, and Calimfac. The Burden mode is used for calculating emissions inventories and reports total emissions as tons per weekday using emissions factors, corrected for ambient conditions and speeds, combined with vehicle activity. The Emfac mode generates emissions factors as grams of pollutant emitted per vehicle activity and can calculate a matrix of emissions factors at specific values of temperature, relative humidity, and vehicle speed. One important use for the Emfac mode is to generate files for use with the DTIM model and other air quality models such as AIRSHED, CALINE, and URBEMIS. The Calimfac mode is used to calculate detailed emissions rates for each vehicle class and model years from 1965 to the scenario calendar year. CARB made several adjustments to EMFAC output data in developing the California statewide GHG inventory (California Air Resources Board 2010a). EMFAC estimates do not include effects of the federal CAFE standards or other GHG emis- sions standards. However, CARB has developed the Pavley I + Low-Carbon Fuel Stan- dard postprocessor to adjust CO2 emissions from EMFAC output to account for the

145 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS reductions caused by the adopted Pavley I regulation and the Low-Carbon Fuel Stan- dard in the light-duty fleet (California Air Resources Board 2010c). EMFAC differs from MOVES in how it estimates emissions. MOVES calculates emission rates associated with vehicle operating modes (e.g., cruise and acceleration). These emission rates are based on the second-by-second power demand placed on a vehicle when operating in various modes and speeds. The activity data in MOVES are vehicle operating times. In contrast, EMFAC, like MOBILE, calculates emissions esti- mates from trip-based travel activities. EMFAC quantifies running exhaust emissions factors in grams per mile for a specific speed bin. The emissions factors are compos- ite emission rates aggregated from base rates by vehicle class, technology group, and model year. EMFAC uses VMT for activity data. Other differences between these two models include the following: • EMFAC does not distinguish roadway links, and thus is better suited to regional- scale than link-level applications. MOVES can be used for regional- down to link- level inventories. However, unlike EMFAC, MOVES does not contain county-specific default activity data. For county-level runs, the user must enter county-level activity data. MOVES can derive state and county activity data by applying spatial allocation factors to national data, although this is not recommended for county-level analyses. • EMFAC calculates hourly or daily inventories for an average weekday by month, season, and year; MOVES provides hourly, daily, monthly, or annual emissions for weekdays, weekends, months, or years. • MOVES identifies vehicle class based on the classification used by the federal Highway Performance Monitoring System (HPMS). EMFAC uses a different vehicle classification scheme. • MOVES can be used for any geographic area in the United States, but EMFAC only contains activity data for California counties and California-specific emission rates. If using MOVES to model California rates, additional inputs are needed to model the California-specific emission rates correctly. Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation Model Developed by the Argonne National Laboratory and sponsored by the U.S. Department of Energy, the Greenhouse gases, Regulated Emissions, and Energy use in Transporta- tion (GREET) model is designed to fully evaluate the energy and emissions impacts of advanced vehicle technologies and new transportation fuels, considering the fuel cycle from wells to wheels and the vehicle cycle from material recovery to vehicle disposal (Argonne National Laboratory 2012a). GREET can estimate emissions of three GHGs (CO2, CH4, and N2O) and five criteria pollutants (oxides of nitrogen, sulfur dioxide, 10-µm particulate matter, carbon monoxide, and volatile organic compounds), as well as total energy use. Inputs related to the fuel, fuel processing and refining, and vehicle technologies are needed to estimate emissions factors with this model. GHG emis- sions rates (in grams per mile) are produced for three categories of light-duty vehicles (LDVs): passenger cars, light trucks 1, and light trucks 2.

146 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS This model is most appropriate in cases for which life-cycle emissions for differ- ent types of vehicle and fuel technology are of interest. It is also useful for obtaining emissions factors for fuels other than gasoline and diesel (e.g., biofuels, electricity). The model can be used with default factors or with a wide variety of user inputs rep- resenting different fuel production processes. GREET Version 1.8c, released in March 2009, does not incorporate the latest EPA or CARB research on the life-cycle impacts of biofuels, although the U.S. Department of Energy plans to integrate this informa- tion in a future model update. VISION The VISION model estimates the potential energy use, oil use, and carbon emission im- pacts of various light- and heavy-duty vehicle technologies and alternative fuels through the year 2100 (Argonne National Laboratory 2012b). It also provides total VMT by technology and fuel type by year. This model compares the market penetration of vari- ous alternative fuel and advanced vehicle technologies to a baseline scenario in which these technologies have not been implemented. The simulation is based on a set of input parameters that includes vehicle market penetration and fuel economy ratios by technol- ogy, fuel types (including alternative fuels) and price, VMT, future vehicle sales, popula- tion and gross domestic product growth, and vehicle costs. Default values come from the Annual Energy Outlook for the baseline scenario, and all input values can be changed by the user to customize the simulation and show the sensitivity of various assumptions. VISION is updated annually to reflect changes in energy consumption according to the most recent Annual Energy Outlook. VISION outputs include energy use by fuel type, full fuel-cycle carbon emissions (million metric tons [MMT] carbon equivalent), full fuel-cycle GHG emissions (MMT CO2e), fuel expenditures (billions of dollars and as a percentage of gross domestic product), and light-vehicle miles per gallon gasoline equivalent. The VISION model works exclusively with highway vehicles on 10-year increments. The primary use of VISION for transportation planners is for long-term policy analyses of state- and regional-scale vehicle and fuel technology strategies. OFF-MODEL METHODS In many cases, an appropriate analysis tool may not exist for a particular strategy, or may have data and resource requirements that are beyond what are available for the study. Common off-model techniques include elasticities and case examples. Other existing tools for travel analysis that do not directly produce GHG emissions estimates can also support GHG estimation (e.g., by taking changes in VMT forecast using these tools and applying emissions factors). Elasticities Elasticities are expressions of a relationship between two factors (e.g., between the price of travel by a given mode and the amount of travel by that mode). Specifically, the elasticity value is the ratio of a percentage change in one factor to a percentage change in the other. For example, a VMT price elasticity of –0.4 means that if the price of travel increases by 10%, VMT will decline by 4% (10% × –0.4).

147 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Elasticities are commonly used to analyze strategies that affect the cost of travel (e.g., road pricing, transit fares, commuter incentives) and strategies that affect travel time (e.g., reduction in bus headways or running time). They have also been developed for other relationships, such as VMT versus land use density. Many sketch plan methods, such as the TRIMMS model and the Clean Air and Climate Protection tool, incorporate elasticities. Caution should be used in applying elasticities developed from data in one particular location to another location, because conditions in the second location may differ from the situation for which the elasticity was developed. Useful sources of transportation elasticities include • Transit Cooperative Research Program (TCRP) Report 95: Traveler Responses to Transportation System Changes—This series of reports includes chapters provid- ing evidence on the travel and mode shift impacts of a variety of strategies, including parking and transit pricing, ridesharing and vanpooling, transit promotion and service improvements, and land use and site design; and • The Victoria Transport Policy Institute’s Online TDM Encyclopedia, which pro- vides a summary of research, examples, and evidence on a variety of travel demand management and land use strategies. Case Examples A case example simply refers to using data on the impacts observed in other situa- tions to predict impacts in the situation of interest. For example, a transit agency may observe that the use of hybrid electric buses has reduced fuel consumption by 30% compared with their standard diesel buses. Case examples are usually applied in con- junction with scaling factors (e.g., size of bus fleet) to transfer percentage impacts to the situation in which the strategy is being applied. Case examples must be used with caution to ensure that conditions in the situation of interest will result in GHG reduc- tions similar to those observed elsewhere, and that the data from the case example are valid. Case examples can be found in the TCRP and Victoria Transport Policy Institute sources referenced above. Other Tools A variety of other tools and resources do not directly provide GHG emissions esti- mates, but can assist in estimating the VMT or traffic flow impacts of many GHG reduction strategies. Recommended Practice for Quantifying Greenhouse Gas Emissions from Transit This document provides guidance on estimating the GHG emissions reduction benefits of transit projects (American Public Transportation Association 2009). It identifies three types of benefits: mode shifting, congestion relief, and indirect benefits associated with more efficient land use patterns. The guidance addresses issues such as analysis boundaries, emissions factors, and emissions from electricity generation. The guidance describes how to estimate GHG emissions based on project ridership estimates and operating data; describes how to calculate congestion relief benefits; and discusses the

148 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS land use multiplier, which represents additional benefits from land use changes related to transit investments. FHWA Highway Economic Requirements System The Highway Economic Requirements System (HERS) is an analysis tool that uses en- gineering standards to identify highway deficiencies and then applies economic criteria to select the most cost-effective mix of improvements for systemwide implementation. HERS is designed to evaluate the implications of alternative programs and policies on the conditions, performance, and user cost levels associated with highway systems. HERS-ST is a version of the model provided for state-level use (Federal Highway Ad- ministration 2012a). While primarily intended for economic analysis, HERS can be used to support GHG analysis by providing information on changes in vehicle speeds, volumes, and fuel consumption as a result of capacity or operational improvements to the statewide highway network. Unlike many models, HERS has explicit procedures to account for induced demand effects. FHWA IMPACTS IMPACTS is a series of spreadsheets developed to help screening-level evaluation of multimodal corridor alternatives, including highway expansion, bus system expan- sion, light rail transit investment, high-occupancy vehicle lanes, conversion of an ex- isting highway facility to a toll facility, employer-based travel demand management, and bicycle lanes (Federal Highway Administration 2012b). Inputs are travel demand estimates by mode for each alternative. The impacts estimated include costs of imple- mentation; induced travel demand; benefits including trip time and out-of-pocket cost changes such as fares, parking fees, and tolls; other highway user costs such as accident costs; revenue transfers due to tolls, fares, or parking fees; changes in fuel consump- tion; and changes in emissions. FHWA Screening Tool for ITS The Screening Tool for ITS (SCRITS) is a spreadsheet tool developed by FHWA for evaluating ITS strategies at a screening level when a more sophisticated evaluation using a tool such as IDAS cannot be performed (Federal Highway Administration 2012c). SCRITS includes changes in energy consumption as an output. It requires users to sup- ply a set of baseline data, including a study area and associated travel statistics or other parameters that are used in a variety of the ITS applications. For example, the user must define the area or facilities covered and supply an estimate of VMT. The fuel efficiency factors in the model may not be up-to-date, but they can be adjusted by the user. FHWA Sketch-Planning Analysis Spreadsheet Model The Sketch-Planning Analysis Spreadsheet Model (SPASM) is designed to assist plan- ners in sketch planning analyses of packages of transportation actions at the system and corridor level, including transit system improvements, highway capacity improve- ments, high-occupancy vehicle lane improvements, and auto use disincentives (Federal Highway Administration 2012d). Reported benefits include changes in energy use. The model takes into account congestion-related effects of changes in VMT on speeds during peak and off-peak periods, diversion of traffic among parallel highway facilities

149 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS in a corridor, induced (or disinduced) traffic occurring as a result of changes in high- way congestion levels, and effects of speed and cold starts on motor vehicle emissions and fuel consumption. The fuel efficiency factors in the model may not be up-to-date, but they can be adjusted by the user. Florida DOT Transit Mode Shift Measures The Florida DOT developed a report to produce measurable criteria that can be used by the agency to determine where and under what circumstances an investment in transit service and facilities will reduce energy consumption and realize the associated health benefits of transit (Florida State University 2009). The report provides mode shift factors for different transit modes (i.e., the percentage of people who use transit despite having another option available to them to make the trip). The factors are based on surveys of transit riders in Florida and other states. These factors on prior and/or alternative modes of travel can be combined with trip length and ridership data to estimate GHG emissions. Guidelines for Analysis of Investments in Bicycle Facilities This NCHRP report includes methodologies and tools to estimate the cost of vari- ous bicycle facilities and for evaluating their potential value and benefits (National Cooperative Highway Research Program 2006). The report is designed to help trans- portation planners integrate bicycle facilities into their overall transportation plans and on a project-by-project basis. The research described in the report has been used to develop a set of web-based guidelines that provide a step-by-step worksheet for estimating the costs, demands, and benefits associated with specific facilities under consideration (Active Communities 2012). Using Trend Analysis to Project Future VMT Areas with a regional travel demand model will typically have generated 20-year VMT forecasts that consider a variety of factors influencing travel, such as population, em- ployment, household size, income levels, land use patterns, and planned transporta- tion system improvements. These VMT forecasts can serve as a basis for GHG emis- sions forecasts when they are used in conjunction with VMT-based emissions factors or fuel economy projections and the carbon content of fuel. To estimate future GHG emissions in areas without a travel demand model, VMT must be projected in other ways, such as by an extrapolation of historic trends. In most cases it is inappropriate to project future GHG emissions based only on historic emis- sions trends because of changes in vehicle fuel economy and carbon content of fuels over time. If a VMT trend extrapolation using a linear or other function is not reasonable for a certain area, a more detailed approach can be used that considers the factors that influ- ence VMT. Both trend extrapolation and more detailed projections are discussed below. Trend Extrapolation of Historic VMT FHWA’s HPMS is an ideal source for historic VMT because it offers data for all parts of the country in a standardized format. HPMS provides estimates of VMT data by county using HPMS urban and rural roadway functional classifications for every state in the country on an annual average basis.

150 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS A regression analysis using a linear, logistic, or other function can be applied to a group of recent years of HPMS VMT data to generate a VMT estimate for a planning horizon year or years. Local areas must decide the appropriate number of years to include in their regression analysis, but in general 10 to 20 years is appropriate for the longer-term forecasts often used in GHG projections. For example, in projecting VMT at the state level for GHG forecasts, several states have performed a VMT forecast based on historic VMT for each year from 1990 to the latest available year of VMT data. A shorter time period (10 years) may be more appropriate if it appears that struc- tural factors have led to a significantly different trend in the past decade compared with previous decades, and if this new trend is expected to continue. These regression analyses should be performed at the county level for each functional roadway classification and for each HPMS vehicle type. For rural areas along Interstate highways, a large portion of the county’s VMT can be made up of through traffic. For these rural areas the VMT for the interstate facility should be forecast separately from the VMT for the rest of the county. Linear regression can result in some VMT trends for individual functional classifications becoming negative, either as a result of historically decreasing traffic counts or as a result of changes to urban–rural HPMS designations that shift the functional class bin in which traffic counts are reported. For situations in which this occurs, VMT is recommended to be held constant at the level of the latest year for which HPMS data are collected to provide a conservative estimate of VMT. VMT Projection Using Contributing Factors A simple extrapolation of historic VMT trends is not ideal because the various drivers of VMT (e.g., regional population, income, vehicle trip rates, trip lengths, vehicle occupancy) may change over time in ways that do not reflect the regression equation chosen. An improved approach that does not involve significantly more effort is to forecast population and VMT per capita separately and then combine them as follows: VMT = population × (VMT/capita). Forecasts of the individual factors may be available from the following sources: • Population growth forecasts are often generated by state or local planning agen- cies for use in a variety of applications, such as transportation planning, comprehen- sive planning, and economic development; • Future VMT per capita can be projected using historic VMT and population data to calculate historic VMT per capita, which can be trended forward using linear pro- jection methods as described above or professional judgment; and • State or local programs and policies, such as travel demand management or land use planning programs, may influence future VMT per capita and may have already set specific goals for this metric. An even more detailed projection could involve analysis of multiple factors, such as trip rates by household income level and size, trip lengths, mode shares, vehicle occupancy, and/or economic activity. For example, historic data could be used to model per capita VMT as a function of population, household size, and area employ- ment. The model would then be used to develop future per capita VMT projections, using forecasts of these driving variables.

151 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS As a hypothetical example, VMT in Massachusetts is projected to illustrate how projections of population and VMT per capita can be used to forecast future VMT. Historic data from Massachusetts from 1990 through 2008 are used to project VMT through 2030. The steps are as follows (excerpts of the data are shown in Table A.9): 1. Historic population estimates for 1990 through 2008 were taken from the U.S. Bureau of the Census. (Note that the years between the decennial censuses are esti- mates that are not as accurate as the decennial census years; 2001 to 2009 values will be retrospectively adjusted to match the 2010 census totals.) The population is pro- jected to increase by 6% in 2030 compared with 2008. 2. Population projections for 2010, 2020, and 2030 were taken from state sources, and intermediate years were obtained using linear interpolation. 3. Annual VMT for the state was obtained from the FHWA Highway Statistics, Table VM-2. 4. VMT per capita was calculated for 1990 to 2008. 5. Trendlines were used to extrapolate VMT per capita (Figure A.10). Two trendlines were established: 1990 to 2008 and 1999 to 2008. It may be observed that the growth in VMT per capita was considerably lower over the past decade than over the previous decade. 6. As a sensitivity analysis, two VMT projections were developed: one using the higher rate of VMT growth, and one using the lower rate observed over the past decade (Figure A.11). These estimates show total VMT increasing by 8% to 18% over the 2008 to 2030 time frame. TABLE A.9. MASSACHUSETTS POPULATION AND VMT DATA Year Population (Historic and Projected) Annual VMT (millions) Historic VMT per Capita VMT per Capita Annual VMT (millions) Low Projection High Projection Low Projection High Projection 1990 6,016,425 46,177 7,675 na na 46,177 46,177 1991 6,049,692 46,537 7,692 na na 46,537 46,537 ↓ 2008 6,623,273 54,505 8,229 na na 54,505 54,505 2009 6,636,136 8,238 8,274 54,666 54,906 2010 6,649,000 8,246 8,318 54,828 55,308 ↓ 2020 6,856,000 8,330 8,762 57,107 60,076 ↓ 2030 7,012,000 8,413 9,207 58,992 64,558 2030 versus 2008 5.9% 2.2% 11.9% 8.2% 18.4%

152 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS y = 47.057x + 7626.7 y = 8.8439x + 8200.6 6,500 7,000 7,500 8,000 8,500 9,000 9,500 19 90 19 92 19 94 19 96 19 98 20 00 20 02 20 04 20 06 20 08 20 10 20 12 20 14 20 16 20 18 20 20 1990-2008 trendline 1999-2008 trendline VMT per Capita Figure A.10. Massachusetts VMT per capita. - 10,000 20,000 30,000 40,000 50,000 60,000 70,000 19 90 19 92 19 94 19 96 19 98 20 00 20 02 20 04 20 06 20 08 20 10 20 12 20 14 20 16 20 18 20 20 20 22 20 24 20 26 20 28 20 30 Low Projecon High Projecon Total VMT (millions) Figure A.11. Hypothetical VMT projections for Massachusetts.

153 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Converting Highway and Nonhighway VMT into Emissions Overview of Vehicle Emissions GHG emissions estimates should always include CO2 and will usually include CH4 and N2O. Hydrofluorocarbons (HFCs) may also be included to account for emissions from leaks and repairs related to air conditioning. GHG emissions from highway vehicles are dominated by CO2. In 2005, on a CO2e level, CO2 accounted for about 95% of transportation emissions, with HFCs accounting for just more than 3%, N2O slightly less than 2%, and CH4 about 0.1% of national transportation GHG emissions (U.S. Environmental Protection Agency 2010b). Black carbon is an additional pollutant with climate change implications, but the current state of scientific knowledge does not support expression of these emissions in terms of global warming potential. Different methods are generally used for calculating CO2 than are used for calcu- lating CH4 and N2O emissions from highway vehicles. This is because emission factors for CO2 are generally expressed in terms of fuel consumed (e.g., grams per gallon of gasoline), but emission factors for CH4 and N2O are expressed as a function of vehicle activity (e.g., grams per VMT). Although the CO2 emission factors vary only by fuel type (gasoline versus diesel), CH4 and N2O emission factors vary significantly accord- ing to vehicle technology. For example, N2O emission factors range from as low as 0.001 g/mi for a diesel passenger car to as high as 0.2 g/mi for an older gasoline heavy- duty vehicle. HFC emissions are not usually calculated because they are more related to vehicle maintenance practices than VMT. All emissions estimates are derived as the product of vehicle activity and emission rates that reflect the vehicle activity. Emission rates can be derived based on mass per time (grams per second), mass per distance (grams per mile), or mass per unit of fuel consumed (grams per gallon). Mass per time and mass per distance are related by speed. Time-based emission rates can be defined for both idling and moving vehicles. Fuel-based emission rates are useful if fuel consumption rates are known (some traffic simulation models provide outputs of fuel consumption but not GHG emissions). Two approaches are identified that can be used to generate emission rates: • The regulatory emissions factor models can be used (MOVES outside of California and EMFAC within California). These models generate activity-based emission rates. See below for additional guidance on using the MOVES model; or • Fuel-based emission rates can be used. Both EPA and CARB have identified grams per gallon CO2 emission rates for gasoline- and diesel-powered vehicles, consistent with Intergovernmental Panel on Climate Change (IPCC) protocols. This approach is covered in detail in this section. If the necessary resources or data needed to run MOVES or EMFAC to estimate GHG emissions are not available, other more simplified methods are available that can be accomplished in a spreadsheet. CO2 emissions are calculated as a function of fuel consumption and CH4 and N2O emissions are calculated based on VMT, so separate approaches are used for these two sets of pollutants, as described below. CO2 emis- sions are essentially a direct function of fuel consumption, depending only on fuel

154 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS type. In contrast, CH4 and N2O are strongly determined by the vehicle’s emissions control technology, and therefore emissions factors from a model that accounts for vehicle technology must be used for these gases. CO2 Emissions Historic or future CO2 emissions from vehicles can be determined based on VMT and fuel efficiency using the following general equation: emissions = ∑[(VMTabc/FEabc) × EFb] where FE = fuel economy (mi/gal), EF = emissions factor (g/gal), a = vehicle type (e.g., passenger car, light-duty truck), b = fuel type (e.g., diesel or gasoline), and c = analysis year. The specific steps to fill in each of the variables in the equation are described below. Step 1: Split VMT into Vehicle Types VMT data from HPMS should already be available by vehicle type, such as passenger cars, light-duty trucks, and heavy-duty vehicles. If such data are not available, VMT can be divided using national percentage distributions from the following national sources. Historic Years. The percentage distribution by vehicle type for historic years can be derived from FHWA’s Highway Statistics, Table VM-1, or from a compilation of this table for a series of years in the Bureau of Transportation Statistics’ National Transportation Statistics, Table 1-32. Both of these tables provide VMT by vehicle type, which can be converted to percentage distributions. These data are available for six vehicle types, but can be combined into three vehicles types as shown in the example percentage distributions in Table A.10. Future Years. The percentage distribution by vehicle type for future years can be derived from two tables from the AEO (Energy Information Administration 2010). Table 7 from the AEO reference case provides VMT for LDVs, commercial light trucks, and freight trucks. This can be combined with Supplemental Table 58 to divide the LDVs into passenger cars and light-duty trucks. Table A.11 gives an example of dis- tributions derived from the VMT and vehicle stock estimates from these AEO tables. TABLE A.10. EXAMPLE VEHICLE TYPE PERCENTAGE DISTRIBUTION FOR HISTORIC YEARS Vehicle Type 2000 2001 2002 2003 2004 2005 2006 2007 Passenger cars 58.48% 58.41% 58.27% 58.04% 57.53% 57.35% 56.31% 55.40% Light-duty trucks 33.73% 33.84% 33.94% 34.16% 34.76% 34.95% 36.06% 36.84% Heavy-duty trucks 7.79% 7.75% 7.78% 7.80% 7.70% 7.70% 7.63% 7.76% Source: Bureau of Transportation Statistics (2009), Table 1-32.

155 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS TABLE A.11. EXAMPLE VEHICLE TYPE PERCENTAGE DISTRIBUTIONS FOR FUTURE YEARS Vehicle Type 2010 2015 2020 2025 2030 2035 Passenger cars 51.64% 49.17% 49.85% 52.07% 54.34% 56.27% Light-duty trucks 41.51% 43.18% 42.32% 40.22% 37.96% 35.96% Heavy-duty trucks 6.84% 7.66% 7.83% 7.70% 7.70% 7.77% Source: Energy Information Administration (2010), Table 7 and Supplemental Table 58. Step 2: Calculate Fuel Consumption by Dividing by Fuel Economy Fuel economy should be collected for each vehicle type into which VMT was split. Care must be taken to understand whether the stated fuel economy represents the fuel economy of a specific vehicle model year or whether it represents the fleetwide aver- age of all vehicles in use. The fleetwide average of all vehicles in use should always be used. Also, care should be taken to note which fuel the fuel economy provides for so that the appropriate carbon contents can be applied in the next step. Most passenger cars and light-duty trucks use gasoline and most heavy-duty trucks use diesel, but exact gasoline–diesel splits can be obtained from the EPA if a more detailed esti- mate is desired. If available, local data on fuel economy can be used; otherwise, the follow ing national sources are recommended. If using these national sources, analysts should note whether the area they wish to use for comparison uses reformulated gasoline as a result of being an ozone nonattainment area; if so, the fuel efficiency should be adjusted to be 1% to 3% worse than the amounts given here to account for the poorer fuel economy from reformulated gasoline (U.S. Environmental Protection Agency 2010d). Historic Years. The fuel economy by vehicle type for historic years can be derived from FHWA’s Highway Statistics, Table VM-1, or a compilation of this table from a series of years in the Bureau of Transportation Statistics’ National Transportation Statistics, Tables 4-13, 4-14, 4-15, and 4-23. Although these tables provide fuel econ- omy for six vehicle types, the data shown in Table A.12 combine single-unit trucks, combination trucks, and buses into the heavy-duty trucks category by using a VMT weighted average. TABLE A.12. EXAMPLE FUEL ECONOMY FOR 2000 TO 2007 Vehicle Type 2000 2001 2002 2003 2004 2005 2006 2007 Passenger cars 21.9 22.1 22.0 22.2 22.5 22.1 22.5 22.5 Light-duty trucks 17.4 17.6 17.5 16.2 16.2 17.7 17.8 18.0 Heavy-duty trucks 6.01 6.12 6.01 6.91 6.85 6.28 6.17 6.20 Source: Bureau of Transportation Statistics (2009), Tables 4-13, 4-14, 4-15, and 4-23. Note: Fuel economy is expressed in miles per gallon of gasoline equivalent. Future Years. Fuel economy by vehicle type for future years can be derived from the AEO. Data are available for passenger cars and light trucks (AEO Supplemental Table 59) and medium- and heavy-duty vehicles (AEO Supplemental Table 67); these data are summarized for select years in Table A.13. Note that this table reports fuel

156 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS economy in miles per gallon of gasoline equivalent. A gallon of gasoline equivalent is the amount of fuel that has the same energy content as a gallon of gasoline. TABLE A.13. EXAMPLE FUEL ECONOMY FOR FUTURE YEARS Vehicle Type 2010 2015 2020 2025 2030 2035 Passenger cars 23.67 25.29 27.73 29.79 31.46 32.68 Light-duty trucks 18.37 19.47 21.05 22.57 23.93 25.08 Heavy-duty trucks 6.05 6.30 6.62 6.82 6.95 7.03 Source: Energy Information Administration (2010), Supplemental Tables 59 and 67. Note: Fuel economy is expressed in miles per gallon of gasoline equivalent. Step 3: Apply CO2 Emissions Factors Once the amount of fuel consumed is calculated, it should be multiplied by a fuel- specific CO2 emission factor (kilograms per gallon) to calculate the total amount of direct CO2 emitted. The emissions factors shown in Table A.8 are for gasoline and diesel fuel. Because alternative fuels often have significant CO2 emissions in the rest of the life cycle beyond fuel consumption, emissions factors for alternative fuels are covered below in “Vehicle and Fuel Life-Cycle Emissions.” It should be noted that gasoline and diesel also have additional life-cycle emissions; using the factors shown in Table A.8 only calculates the direct emissions from burning the fuel in vehicles. Many areas of the country that are in nonattainment for ozone air quality standards use reformulated gasoline, which has slightly different CO2 emission factors because of the addition of ethanol or other additives to the gasoline. Alternate CO2 emission factors for gasoline with 10% ethanol (E10) are available in Table A.8. Step 2 contains details for adjusting the fuel economy to account for reformulated gasoline. CH4 and N2O Emissions CH4 and N2O collectively represent only a small fraction of GHG emissions from motor vehicles: about 2% for LDVs and about 0.3% for diesel-powered heavy-duty vehicles as measured in CO2e. In 2010, CH4 represented about 2.0% of GHG emis- sions from LDVs and N2O represented about 0.1%. CH4 emissions from trucks were negligible, and N2O emissions represented 0.3% of total GHG emissions from these vehicles. These figures are from Cambridge Systematics (2009) calculations based on AEO Table 2-15, which shows historic data through 2006 extrapolated through 2010 (Energy Information Administration 2009). CH4 and N2O emissions do not offer large opportunities in terms of mitigation potential and should be considered a lower priority to calculate. The simplest approach to including these emissions in a GHG emissions inventory is to scale up CO2 emissions by 2% for LDVs and 0.3% for heavy- duty vehicles, or to use alternative scaling factors using future year emissions factors. If a more detailed approach to estimating CH4 and N2O emissions is desired, the general methodology employed in the EPA GHG Inventory (U.S. Environmental

157 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Protection Agency 2010b), described in the following equation and The Climate Reg- istry’s General Reporting Protocol, can be used: emissions = ∑(EFabc × activityabc) where EF = emissions factor (e.g., g/mi), activity = activity level measured in the units appropriate to the emission factor (e.g., mi), a = fuel type (e.g., diesel or gasoline), b = vehicle type (e.g., passenger car, light-duty truck), and c = model year. Once CH4 and N2O emissions are calculated it is often useful to sum them with CO2 emissions to get total GHGs. However, because each of these GHGs has differ- ent abilities to trap heat in the atmosphere, it is necessary to weight the emissions of CH4 and N2O relative to the ability of CO2 to trap heat. A global warming potential factor is used to provide this weighting and to convert CH4 and N2O emissions into grams of CO2e. The global warming potential values currently used by the EPA in the Inventory of U.S. Greenhouse Gas Emissions are provided in Table A.14. These values are based on IPCC’s second assessment report (Intergovernmental Panel on Climate Change 1996); the 2007 fourth assessment report values are shown for comparison. TABLE A.14. 100-YEAR GLOBAL WARMING POTENTIAL OF SELECT GHGS Gas EPA Inventory/IPCC 2nd Assessment IPCC 4th Assessment CO2 1 1 CH4 21 25 N2O 310 298 Source: March 2009 public review draft of EPA’s Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990–2007; Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor, and H. L. Miller, eds.). Cambridge University Press, Cambridge, United Kingdom, 2007. CH4 and N2O emissions depend heavily on the type of emissions control technol- ogy used in the vehicle, and the type of control technology used generally correlates with the year of vehicle manufacture. Control technologies include uncontrolled, non- catalyst, oxidation catalyst, Tier 0, Tier 1, Tier 2, and low-emission vehicles. Because the introduction of these control technologies and the emission rates for N2O and CH4 both vary by model year for each vehicle type, VMT should be divided accordingly. The following steps should be used to calculate N2O and CH4 emissions. Step 1: Split VMT into Vehicle Types This can be done using the details in Step 1 from the CO2 emissions section above.

158 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Step 2: Split VMT into Vehicle Model Years State registration databases, which give the number of vehicles for each model year, can be used in combination with mileage accumulation assumptions by model year to distribute VMT among all model years. For nonattainment areas, this data should be readily available from transportation conformity analyses. For other areas, national defaults from the MOVES emissions model can be used. Step 3: Obtain N2O and CH4 Emission Factors Emission factors on a gram per mile basis for specific vehicle types and model years are available from the EPA’s GHG Inventory or The Climate Registry’s General Reporting Pro- tocol. These rates are shown in Table A.15. Updated versions of these documents should be consulted for emissions factors for more recent model years as they become available. Step 4: Multiply VMT by N2O and CH4 Emission Factors To calculate the N2O and CH4 emissions, multiply the VMT distributed by vehicle type and model year by the corresponding emissions factors. Then sum all N2O emissions and all CH4 emissions. The mobile combustion module of EPA’s State Inventory Tool may also be of use in helping to perform these calculations at a regional level (U.S. Environmental Protec- tion Agency 2012c). TABLE A.15. N2O AND CH4 EMISSION RATES Vehicle Type and Model Year N2O (g/mi) CH4 (g/mi) Gasoline Passenger Cars 1984 to 1993 0.0647 0.0704 1994 0.056 0.0531 1995 0.0473 0.0358 1996 0.0426 0.0272 1997 0.0422 0.0268 1998 0.0393 0.0249 1999 0.0337 0.0216 2000 0.0273 0.0178 2001 0.0158 0.011 2002 0.0153 0.0107 2003 0.0135 0.0114 2004 0.0083 0.0145 2005 0.0079 0.0147 (continued on next page)

159 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Vehicle Type and Model Year N2O (g/mi) CH4 (g/mi) Gasoline Light Trucks (Vans, Pickup Trucks, SUVs) 1984 to 1993 0.1035 0.0813 1994 0.0982 0.0646 1995 0.0908 0.0517 1996 0.0871 0.0452 1997 0.0871 0.0452 1998 0.0728 0.0391 1999 0.0564 0.0321 2000 0.0621 0.0346 2001 0.0164 0.0151 2002 0.0228 0.0178 2003 0.0114 0.0155 2004 0.0132 0.0152 2005 0.0101 0.0157 Gasoline Heavy-Duty Vehicles 1985 to 1986 0.0515 0.409 1987 0.0849 0.3675 1980 to 1989 0.0933 0.3492 1990 to 1995 0.1142 0.3246 1996 0.168 0.1278 1997 0.1726 0.0924 1998 0.1693 0.0641 1999 0.1435 0.0578 2000 0.1092 0.0493 2001 0.1235 0.0528 2002 0.1307 0.0546 2003 0.124 0.0533 2004 0.0285 0.0341 2005 0.0177 0.0326 Diesel Passenger Cars 1960 to 1982 0.0012 0.0006 1983 to 2004 0.001 0.0005 Diesel Light Trucks 1960 to 1982 0.0017 0.0011 1983 to 1995 0.0014 0.0009 1996 to 2004 0.0015 0.001 Diesel Heavy-Duty Vehicles All model years 0.0048 0.0051 Source: The Climate Registry (2012), Table 13.4. TABLE A.15. N2O AND CH4 EMISSION RATES (CONTINUED)

160 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Refrigerant Emissions Refrigerants used in air conditioning and refrigeration systems represent an additional source of GHG emissions from highway vehicles. Emissions from mobile air condition- ers and refrigerated transport accounted for about 4% of GHG emissions from cars, 5% from light-duty trucks, and 0.6% from heavy trucks in 2006 (Cambridge System- atics 2009; calculations are based on AEO Table 2-15, April 2009 release). Modern refrigerants are potent GHGs with a high global warming potential. HFC-134a, the most commonly used refrigerant today, has a global warming potential of 1,300. HFCs are released into the atmosphere through leaks in mobile air conditioners or refriger- ated transport units during servicing, operation, and retirement. Refrigerants may or may not be included in transportation GHG inventories. For example, EPA’s State Inventory Tool does not attribute refrigerant emissions to the mobile sector, but rather includes them in the industrial sector (because they are dis- pensed and recovered at automobile repair facilities). Mobile source emissions factor models do not provide for refrigerant emissions, so if they are to be included in a transportation inventory, they will need to be scaled from nonrefrigerant emissions using AEO data. Black Carbon While less is understood about the effect of black carbon on climate change than the above-mentioned GHGs, there is increasing evidence that it causes direct positive radiative forcing (i.e., has a net warming effect on the earth). Black carbon differs from GHGs such as CO2 because it remains in the atmosphere for only days or weeks and dissipates before it reaches a global scale; in contrast, CO2 remains in the atmosphere for decades and has a global spatial scale. The exact definition of black carbon varies by the source consulted, but in general it is a component of particulate matter, or soot, produced from the incomplete combustion of fossil fuel, biofuels, and biomass (Diesel Technology Forum 2012). Black carbon is also referred to as elemental carbon. Black carbon has a warming effect because it absorbs light and turns it into heat. When black carbon is deposited on ice and snow, it reduces their ability to reflect light, which in turn reduces their global cooling effect and simultaneously heats the ice and snow to melt them. In climate change research, black carbon is often grouped with other aerosols, such as sulfate, organic carbon, nitrate, and dust, because of their similar characteristics as short-lived climate forcers. However, many of these aerosols have a cooling effect on the climate, as opposed to the warming effect of black carbon. Black carbon is not required to be included in official GHG emissions inventories, so little information is available on the amount of black carbon emissions from spe- cific transportation sources in the United States. However, some studies offer global inventories of black carbon. For example, Bond (2009) estimates that on-road trans- port sources contribute 16% of total black carbon emissions and off-road transport contributes 9%. The remaining sources are open biomass burning (39%), residential cooking and heating (25%), and industrial (11%).

161 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS The nature of black carbon as a subspecies of particulate matter suggests that among transportation sources, those that emit high levels of particulates, such as heavy-duty diesel vehicles, would also emit high levels of black carbon and would be a potential target of mitigation strategies. Mitigation strategies could include those already being promoted to control particulate emissions, such as clean diesel fuels, advanced engine designs, and control technologies (e.g., particulate filters). There would be cobenefits for human health because of a simultaneous reduction in particu- late matter from implementing these types of mitigation strategies. These mitigation strategies may also reduce other particulate subspecies that have a cooling effect on the climate, such as organic carbon. Therefore, care must be taken to calculate the overall net effect on cooling and warming when considering mitigation strategies. The MOVES model produces emissions outputs for both elemental carbon (black carbon) and organic carbon as particulate matter subspecies. However, because of the current lack of knowledge related to quantifying the warming effect of these emissions, it is not recommended that agencies calculate black carbon emissions until further research is available. Typically, emissions of GHGs are converted to CO2e emissions based on their global warming potential over 100 years. However, because the resi- dence time of black carbon in the atmosphere is much shorter than 100 years and the amount of its warming effect is still uncertain, using this conversion is not pos- sible. Similar uncertainty exists for the cooling particulate subspecies, which must be included in the calculation to take into account the net cooling and warming effects. EPA is currently funding research on black carbon’s role in global- to local-scale climate and air quality, including alternative conversion schemes (U.S. Environmental Protection Agency 2010a). GHG EMISSIONS FROM TRANSIT VEHICLES GHG emissions from transit can be calculated in a manner similar to highway vehicles as outlined above, with the exception of electrically powered transit vehicles. There are also different sources of data, such as the National Transit Database (NTD). NTD provides direct fuel consumption data for transit systems across the United States, which allows the analyst to omit the fuel consumption calculation using VMT and fuel economy. Historic GHG Emissions from Transit Transit Buses VMT data from HPMS include buses that travel on public roadways, but do not include other transit vehicles such as light rail, heavy rail, commuter rail, or buses that travel on a dedicated transit right-of-way. To avoid double-counting, one of the fol- lowing approaches should be used: • Remove transit bus VMT from the HPMS data before calculating transit bus GHG emissions separately. This approach is desirable if the transportation project or plan alternatives being analyzed will include different levels of transit service; or

162 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS • Leave transit bus VMT in the HPMS data and calculate transit bus GHG emis- sions as part of the process described above. If calculating transit bus emissions separately, the following equations can be used: CO2 emissions = fuel consumption × CO2 emission factor (g/gal), N2O emissions = annual distance driven × N2O emission factor (g/mi), and CH4 emissions = annual distance driven × CH4 emission factor (g/mi). NTD data provide annual fuel consumption to calculate CO2 and annual miles to calculate CH4 and N2O. The following steps describe the specific calculation processes. CO2 Emissions Step 1: Determine Transit Bus Fuel Consumption. Fuel consumption by mode for each transit system can be found in Table 17 (Energy Consumption) from the NTD Annual Databases (Federal Transit Administration 2012). Table 17 provides gallons of liquid fuels (diesel, gasoline, compressed natural gas) used and kilowatt-hours of electricity used (see “Electric Transit Vehicles” below). Table A.16 uses the Los Angeles County Metropolitan Transit Authority (LACMTA) as an example of the data format avail- able in NTD. NTD has more types of fuel listed in additional columns, but they are deleted from this example for simplicity because they were zero for LACMTA. TABLE A.16. EXAMPLE OF FUEL CONSUMPTION DATA AVAILABLE FOR LACMTA FROM NTD Mode Sources of Energy (in thousands) Gallons Kilowatt-Hours Diesel Gasoline Compressed Natural Gas Electric Propulsion Heavy rail 0.0 0.0 0.0 84,828.0 Light rail 0.0 0.0 0.0 92,637.2 Motor bus 1,439.5 0.0 41,327.5 0.0 Source: Federal Transit Administration (2012), Table 17. Step 2: Apply CO2 Emission Factors. The amount of fuel consumed can be multiplied by the fuel-specific emission factor to calculate the CO2 emissions. Table A.17 pro- vides CO2 emission factors from The Climate Registry’s General Reporting Protocol. For compressed natural gas, care should be taken to convert the units, because NTD provides gallons and the General Reporting Protocol provides kilograms of CO2 per standard cubic foot (scf). The compressed natural gas industry has adopted a standard measurement that states that 135 scf of natural gas is equal to 1 gallon of liquid diesel fuel (or 124 scf for 1 gallon of gasoline).

163 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS TABLE A.17. CO2 EMISSION FACTORS BY FUEL TYPE Fuel Type CO2 Emission Factor Unit Diesel 10.15 kg CO2/gal Gasoline 8.81 kg CO2/gal Liquefied petroleum gas 5.79 kg CO2/gal Liquefied natural gas 4.46 kg CO2/gal Methanol 4.1 kg CO2/gal Ethanol (E100) 5.56 kg CO2/gal Compressed natural gas 0.054 kg CO2/scf Kerosene 9.76 kg CO2/gal Biodiesel (B100) 9.46 kg CO2/gal Source: The Climate Registry (2012), Table 13.1. N2O and CH4 Emissions Step 1: Determine Transit Bus VMT. VMT by mode for each transit system can be found in Table 19, Transit Operating Statistics: Service Supplied and Consumed, from the NTD Annual Databases. Table 19 provides annual vehicle miles for each mode of directly operated or purchased transportation in the transit system. Annual vehicle revenue miles and annual scheduled vehicle revenue miles are also available from this table, but they should not be used because they do not cover all VMT (they exclude miles traveled to and from storage and maintenance facilities). Table A.18 uses LACMTA as an example to show the format of data available in NTD. NTD has more metrics for service supplied in additional columns, but they are deleted from this example for simplicity. This example shows 107,955,500 vehicle miles traveled for all motor buses. TABLE A.18. EXAMPLE OF VMT DATA AVAILABLE FOR LACMTA FROM NTD Mode Type of Service Service Supplied (in thousands) Annual Scheduled Vehicle Revenue Miles Annual Vehicle Miles Annual Vehicle Revenue Miles Heavy rail Directly operated 6,034.1 6,200.7 6,003.5 Light rail Directly operated 8,928.6 8,940.4 8,812.5 Motor bus Directly operated 85,105.7 99,732.9 83,530.0 Motor bus Purchased transportation 6,786.3 8,222.6 6,751.7 Vanpool Purchased transportation 0.0 13,065.2 13,065.2 Source: Federal Transit Administration (2012), Table 19.

164 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Step 2: Apply N2O and CH4 Emission Factors. VMT can be multiplied by a fuel-specific emission factor to calculate N2O and CH4 emissions. Table A.19 provides N2O and CH4 emission factors from The Climate Registry’s General Reporting Protocol. These emissions are almost negligible for diesel vehicles because of their lack of catalytic converters, but CH4 emissions from compressed natural gas vehicles can be significant because of incomplete combustion of natural gas (methane). TABLE A.19. N2O AND CH4 EMISSION FACTORS FOR BUSES Bus Fuel Type N2O (g/mi) CH4 (g/mi) Diesel 0.0048 0.0051 Methanol 0.175 0.066 Compressed natural gas 0.175 1.966 Ethanol 0.175 0.197 Source: The Climate Registry (2012), Tables 13.4 and 13.5. Commuter Rail Vehicles and Ferry Boats CO2 emissions for commuter rail vehicles (considered locomotives) and ferry boats should be calculated using the same fuel consumption–based method described above for transit buses. However, for N2O and CH4 emissions, a fuel consumption–based method is normally used, as opposed to a mileage-based method as described for tran- sit buses. NTD Table 17 for fuel consumption and the emission factors in Table A.20 can be used to calculate N2O and CH4 emissions for commuter rail vehicles and ferry boats. TABLE A.20. N2O AND CH4 EMISSION FACTORS FOR COMMUTER RAIL AND FERRY BOATS Vehicle and Fuel Type N2O (g/gal) CH4 (g/gal) Ships and Boats Residual fuel oil 0.3 0.86 Diesel fuel 0.26 0.74 Gasoline 0.22 0.64 Locomotives (Rail) Diesel fuel 0.26 0.8 Source: The Climate Registry (2012), Table 13.6. Electric Transit Vehicles To calculate GHG emissions from electric transit vehicles, Table 17, Energy Consump- tion, from the NTD Annual Databases can be used. This table provides electric con- sumption in kilowatt-hours for a variety of transit modes including heavy rail, light rail, and trolley bus. Table A.21 provides electric emission factors for the three major GHGs and the combined CO2e amount from EPA’s eGRID database. The most recent

165 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS rates available as of this writing are for 2007, but the EPA’s website should be con- sulted for updated rates (U.S. Environmental Protection Agency 2012a). These emis- sions factors are organized by the subregions of the country shown on the map in Fig- ure A.12. Depending on the mix of electric generation sources, rates for local utilities may vary from these regional rates. To calculate CO2, N2O, and CH4 emissions, each of the emission factors can be multiplied by the kilowatt-hours of electricity consump- tion from the NTD data. TABLE A.21. ELECTRIC EMISSION FACTORS (2007) BY eGRID SUBREGION eGRID Subregion Acronym eGRID Subregion Name CO2 (lb/mW-h)a CH4 (lb/gW-h)b N2O (lb/gW-h) CO2e (lb/mW-h) AKGD ASCC Alaska Grid 1232.4 25.6 6.5 1234.9 AKMS ASCC Miscellaneous 498.9 20.8 4.1 500.6 AZNM WECC Southwest 1311.1 17.5 17.9 1317 CAMX WECC California 724.1 30.2 8.1 727.3 ERCT ERCOT All 1324.4 18.7 15.1 1329.4 FRCC FRCC All 1318.6 45.9 16.9 1324.8 HIMS HICC Miscellaneous 1514.9 314.7 46.9 1536.1 HIOA HICC Oahu 1812 109.5 23.6 1821.6 MROE MRO East 1834.7 27.6 30.4 1844.7 MROW MRO West 1821.8 28 30.7 1831.9 NEWE NPCC New England 927.7 86.5 17 934.8 NWPP WECC Northwest 902.2 19.1 14.9 907.3 NYCW NPCC NYC/Westchester 815.5 36 5.5 817.9 NYLI NPCC Long Island 1536.8 115.4 18.1 1544.8 NYUP NPCC Upstate NY 720.8 24.8 11.2 724.8 RFCE RFC East 1139.1 30.3 18.7 1145.5 RFCM RFC Michigan 1563.3 33.9 27.2 1572.4 RFCW RFC West 1537.8 18.2 25.7 1546.2 RMPA WECC Rockies 1883.1 22.9 28.8 1892.5 SPNO SPP North 1960.9 23.8 32.1 1971.4 SPSO SPP South 1658.1 25 22.6 1665.7 SRMV SERC Mississippi Valley 1019.7 24.3 11.7 1023.9 SRMW SERC Midwest 1830.5 21.2 30.5 1840.4 SRSO SERC South 1489.5 26.3 25.5 1498 SRTV SERC Tennessee Valley 1510.4 20.1 25.6 1518.8 SRVC SERC Virginia/Carolina 1134.9 23.8 19.8 1141.5 Source: U.S. Environmental Protection Agency (2012a), Year 2005 GHG Annual Output Emission Rates. amW-h = megawatt-hour (1,000 kilowatt-hours). bgW-h = gigawatt-hour (1,000,000 kilowatt-hours).

166 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Future Year Transit GHG Emissions To calculate transit emissions for future years it is necessary to estimate future year annual distance driven by mode. This can be done in three ways: • First, for buses included in the HPMS inventory, use general roadway VMT fore- casts and assume that buses continue to make up the same fraction of future VMT; • Second, for other modes, or for buses not included in the HPMS highway inven- tory, extrapolate future service levels (total vehicle miles, as illustrated in Table A.6) from recent historic trends using NTD data; or • Finally, develop future estimates based on transit service growth assumptions provided by local transit agencies or other transportation planning agencies. This is usually the most appropriate approach; the previous extrapolation approaches should be used only if consultation with local agencies is not possible. It is also necessary to forecast emissions factors for future years. There are several options for accomplishing this. For transit buses, adjust base year emission rates to account for any known effi- ciency standards or improvements planned to be implemented by the local transit agency through the purchase of more efficient or alternative fuel vehicles. This can be done based on emissions data for the specific types of transit vehicles as provided by the manufacturer or obtained from other transit agencies or research studies. For transit vehicles operating on electricity, current electricity emission factors from the eGRID database can be adjusted downward based on applicable state, regional, or national initiatives to reduce GHG emissions from electricity generation. Figure A.12. EPA eGRID subregions.

167 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS For example, the Regional Greenhouse Gas Initiative in the northeastern and mid- Atlantic states aims to reduce GHG emissions from electricity-generating sources by 2.5% per year between 2014 and 2018. The Annual Energy Outlook can be used to develop projections of electric genera- tion GHG intensity. It is recommended that the AEO’s regional forecasts be used (the AEO provides forecasts for nine regions of the country). Table 1 of the AEO can be used to identify total electric energy consumption, and Table 21 can be used to iden- tify total CO2 emissions from the electric sector; these can be combined to calculate carbon intensity. It may be desirable to factor current emissions factors from eGRID by the future trend in carbon intensity from the AEO because eGRID provides greater geographic detail on the electricity generation mix. If it is believed that GHG emission rates from electricity generation will decline at a different rate than identified from the above sources, emission rates may be assumed from a credible scenario analysis that most closely resembles future conditions for the area under consideration. For example, the Electric Power Research Institute provides emission rates for low, medium, and high CO2 intensity scenarios based on assump- tions about the price of CO2 emissions allowances, the rate at which older power plants are retired, the availability and performance of new generation technologies, and the annual growth in electricity demand (Electric Power Research Institute 2007). Table A.22 shows the assumptions for each scenario and the CO2e emission rates for each. These scenarios represent national average conditions and would need to be adjusted for local differences. TABLE A.22. EPRI 2050 SCENARIOS FOR ELECTRIC SECTOR CO2E EMISSION RATES Scenario Definition CO2 Intensity High Medium Low Price of GHG emissions allowances Low Moderate High Power plant retirements Slower Normal Faster New generation technologies Unavailable: Coal with CCS New nuclear New biomass Available: IGCC coal with CCS New nuclear New biomass Advanced renewables Available: Retrofit of CCS to existing IGCC and pulverized coal plants Lower performance: SCPC, CCNG, GT, wind, and solar Nominal EPRI performance assumptions Higher performance: Wind and solar Annual electricity demand growth 1.56% per year on average 1.56% per year on average 2010 to 2025: 0.45% 2025 to 2050: None 2050 electric sector average CO2e intensity (g/kW-h)a 412 199 97 Source: Electric Power Research Institute (2007). Note: IGCC = integrated gasification combined-cycle; SCPC = supercritical pulverized coal; CCNG = combined-cycle natural gas; GT = gas turbine (natural gas); CCS = carbon capture and storage; EPRI = Electric Power Research Institute. aFor comparison, the average CO2e intensity of the electric sector in 2005 was 612 g/kW-h; the EPRI-predicted rate was 573 g/kW-h in 2010.

168 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Improvements in GHG emission rates for all transit modes can also be estimated based on local analysts’ expectations regarding technology improvement. For example, a recent study of nationwide GHG reduction measures developed its own estimates of efficiency improvements based on NTD energy consumption trends, transit mode shares, transit trip lengths, improved bus technology, and decreased power generation emissions to estimate GHG emissions per passenger mile in 2050 for each transit mode (Cambridge Systematics 2009). The results assume that transit load factors remain constant. Percentage per year reductions derived from this analysis are provided in Table A.23. These reductions represent aggressive improvements that were developed to correspond to similarly aggressive improvements in LDV efficiency. TABLE A.23. ESTIMATED ANNUAL IMPROVEMENTS IN GHG EMISSION RATES Transit Mode Annual Reduction in GHG Emission Rates Through 2050 (%) Commuter rail 1.07% Heavy rail 1.46% Light rail 1.25% Bus 0.54% Othera 0.57% Source: Cambridge Systematics (2009), Appendix B, Table 4.8. aAutomated guideway, cable car, ferry, incline, trolley bus, and vanpool. GHG EMISSIONS FROM NONROAD SOURCES Although the majority (about 80%) of GHG emissions from the transportation sector is from on-road sources, nonroad sources (including air, rail, and marine) also con- tribute to GHG emissions. The statewide and metropolitan transportation planning process usually focuses exclusively or primarily on surface transportation, especially highway and transit. However, freight rail and marine transport may be included, especially at a statewide level, but also in metropolitan plans. Aviation, particularly airport facilities, may also be included in these plans. This section provides basic information on developing GHG inventories for non- road sources. EPA’s State Inventory Tool (SIT; discussed above) provides a framework for developing nonroad GHG inventories at the state level. The tool calculates CO2 emissions based on fuel sales by fuel type, which for the most part is not useful for developing mode-specific inventories. However, the mobile combustion module for calculating CH4 and N2O provides a framework for assembling mode-specific activity data that could also be used for calculating CO2 emissions. Some fuel types in SIT, notably aviation fuel and residual fuel, correspond with specific submarkets (aviation and large ships, respectively) and can be used to support inventories for these modes.

169 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Rail Transportation Urban rail Emissions from urban rail transportation (light and heavy rail, commuter rail) can be calculated using the data sources and methods described above in “GHG Emissions from Transit Vehicles.” Intercity Passenger Rail Emissions can be estimated for a state-level inventory based on train counts from Amtrak schedules and estimates of route mileage within the state or metropolitan area based on data from Amtrak or user-generated estimates. Activity should be estimated for both diesel and electric locomotives. Table 4-26 in the National Transportation Sta- tistics provides historic information on the average energy intensity of passenger rail services in the United States, including diesel and electric services combined ( Bureau of Transportation Statistics 2009). Future improvements in efficiency may be estimated by using either the AEO reference case projections for freight rail or other assumptions as determined appropriate by the analyst. More detailed analysis of existing and future efficiencies by locomotive type may be warranted for analysis of specific intercity rail policies, such as implementation of high-speed rail or electrification. Freight Rail Estimates and forecasts of ton-miles carried by rail within a state may be available in summary form from a state rail or freight plan. Current estimates are usually derived from data obtained directly from the rail carriers or from FHWA’s Freight Analysis Framework. Forecasts may be obtained from the Freight Analysis Framework or other analysis conducted for the state rail or freight plan. Current and forecast energy effi- ciency (measured in ton-miles per thousand Btu) may be obtained from Table 7 of the AEO. These must be converted to ton-miles per gallon using Btu/gallon factors for diesel fuel such as those provided by the U.S. Department of Energy (Energy Informa- tion Administration 2012a). Marine Transportation Marine transportation includes domestic and international shipping, ferries, and rec- reational boats. National GHG inventories do not include the international bunker fuels used for international shipping, so these fuels are typically not included in state inventories. Ships typically use residual fuel (a less refined form of diesel fuel). The fuel-based estimates for diesel and residual fuel contained in EPA’s SIT may correspond fairly closely with domestic and international shipping. However, the data may be of questionable quality, showing substantial variation from year to year. State-level gaso- line and diesel fuel use in smaller boats can also be found in SIT (although again it may be of questionable quality), with the assumption that these fuels are used in smaller boats. Activity or port call data (typically expressed as horsepower-hours) representing the operation of both primary and auxiliary engines for the vessel fleet may also be available from the relevant port authority in an area. Projections may be developed by factoring state-level current estimates by AEO national-level forecast changes in U.S. shipping volume (ton-miles) and efficiency (ton-miles per thousand Btu) from Table 7.

170 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Activity data for ferries can be obtained from NTD for individual ferry transit operators. However, fuel consumption (gallons per vessel mile) is not available at this level and must be based on national average fuel consumption from NTD. If specific policies affecting marine operations are to be analyzed, a more detailed analysis of this sector based on locally obtained data (e.g., from a ferry operator or port authority) may be warranted. Aviation Data on sales of jet fuel and aviation fuel by state are included in SIT (although, like the marine fuel use data, this information can be of questionable quality, and interna- tional bunker fuels for aviation are not included in national inventories). GHG emis- sions from aircraft are typically allocated to the airport at which the aircraft’s flight originates. It is assumed that fuel sales by location correspond with the fuel used for flights originating from that location. The Federal Aviation Administration’s Emission Dispersion Modeling System (EDMS) is a modeling tool for estimating GHG (and other criteria pollutant) emis- sions from aircraft. EDMS relies on user-provided landing and takeoff data by aircraft type, which may be obtained directly from airports in a specific metropolitan area. An advantage to using EDMS is that it also estimates ground support equipment and auxiliary power unit emissions associated with the specified aircraft activity. An alter- native approach, and one that can be used for forecasting, is to obtain estimates of seat miles originating from major airports from the Bureau of Transportation Statistics and to multiply by efficiency (expressed as seat miles per gallon) from AEO Table 7. Future year emissions can be projected by adjusting base year emissions by forecasts of gen- eral aviation and commercial aircraft operations in the state, which may be available from the state or regional aviation authority, or from the Federal Aviation Administra- tion’s Terminal Area Forecasts. Future year aircraft efficiency can also be adjusted by seat mile efficiency projections from AEO. Emissions from Construction, Maintenance, and Operations Overview Construction, maintenance, and operations (CMO) emissions are GHG emissions asso ciated with constructing a transportation facility (such as material inputs and con- struction equipment operations), as well as with ongoing maintenance and operations activities for the facility or system (such as repaving, mowing, plowing, and installing and maintaining traffic signals). These emissions are not typically included in inven- tories for the transportation sector because they involve nontransportation mobile sources, such as construction equipment, mowers, snow removal trucks, and aircraft ground support equipment. Nevertheless, transportation agencies may be interested in GHG emissions associated with these activities, and state DOTs, metropolitan plan- ning organizations, and other transportation agencies may wish to examine ways to reduce GHG emissions from their own processes as part of an environmental manage- ment system or other commitments to reduce emissions.

171 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Washington State has developed Greenroads, a sustainability performance metric and rating system for roadway design and construction best management practices that uses information from pavement life-cycle analysis studies (Greenroads 2012). Greenroads is applicable to new, reconstructed, and rehabilitated roadways and works by awarding points for approved sustainable practices. Several European studies have looked directly at energy and emissions related to construction and maintenance (Commission of the European Communities 2006; Karlsson and Carlson 2010). These studies suggest that construction GHG emissions are on the order of 1 to 2 years of operational emissions, and annual maintenance is about 10% to 20% of construction emissions. Greenroads estimates that materials production accounts for about 75% of energy use and 60% to 70% of CO2 emissions associated with construction. Appropriate Situations in Which to Analyze CMO Emissions CMO emissions can be assessed at a project level to support environmental impact analysis and the development of construction and maintenance GHG mitigation strat- egies, such as the use of recycled materials or energy-conserving processes. They may be a lower priority for evaluation at a regional or corridor level, because it is difficult to estimate facility-related emissions and reduction opportunities at such a general scale. However, it still may be desirable to assess the relative magnitude of CMO emis- sions from various levels of plan investment. Data Sources and Calculation Methods Because CMO energy use and emissions have received little research attention until the past few years, there are limited data on which to base an assessment of CMO GHG emissions. However, there is considerable interest in the topic and hence a growing body of literature that is likely to drive rapid evolution of the state of the practice. At least two tools are available as of late 2010 for estimating emissions from construction and maintenance activities based on detailed vehicle activity data and/ or materials inputs: the Greenhouse Gas Calculator for State Departments of Trans- portation (GreenDOT), developed in 2010 through an NCHRP project, and EPA’s NONROAD model. GreenDOT calculates emissions from state DOT activities based on detailed inputs, such as gallons of fuel for off-road equipment, metric tons of concrete and asphalt, or megawatt-hours of electricity usage. For more information, see “GHG Calculator for State DOTS” above. The NONROAD model can calculate pollutant emissions, including CO2, for various equipment types relevant to construction activities (e.g., excavators, graders) and maintenance operations (e.g., mowers, paint sprayers) (U.S. Environmental Pro- tection Agency 2008). Emissions and associated energy use depend on equipment populations and characteristics such as fuel type, engine horsepower, and hours of use. NONROAD is primarily intended for generating state- or county-level emissions inventories, but subcounty inventories may also be produced. NONROAD provides county-level emissions estimates based on default values for populations and activity data, but for a specific area, users may provide local data to better reflect their fleet numbers and characteristics. Activity on a project basis can be estimated by making

172 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS adjustments to the state-level equipment population files and associated county allo- cation files to reflect the fraction of county populations in use for a given project. In addition, emissions for time periods less than 1 year can be developed, and hours of use can be changed to reflect the actual amount of time equipment is operating. The following actions may be considered as an alternative approach to estimating CMO emissions without gathering extensive data for materials or equipment activity inputs: • Gather data on basic infrastructure outputs (new lane miles of road by facility type, miles of rail transit by type, number of rail transit stations by type) associated with each plan or project alternative. • Apply emissions factors by activity and facility type. Some general rules derived from the European literature are provided in Table A.24. These may be updated as addi tional research from the United States becomes available. As of this writing, re- search on construction emissions is underway for Caltrans and the New Jersey DOT. • Optional: Refine with more detailed facility breakdowns (e.g., miles of surface versus tunnel versus bridge alignment). For transit, if a detailed breakdown of the amount of building materials or revenue vehicles purchased is known, the American Public Transportation Association provides default emission factors for transit capital projects; these factors are shown in Table A.25. TABLE A.24. TYPICAL CO2 EMISSIONS FOR ROAD CONSTRUCTION AND MAINTENANCE Facility Type Construction Emissions (tons/mi) Maintenance Emissions (tons/mi/year) Collector 330 2.8 Arterial 460 3.3 Divided highway 560 2.5 Source: Derived from Karlsson and Carlson (2010). Note: Assumes 60-year average for maintenance. TABLE A.25. DEFAULT EMISSION FACTORS FOR TRANSIT CAPITAL PROJECTS Reporting Year Input Default Emission Factor (metric tons of CO2e) Steel used 1.06 per metric ton of steel used Cement used 0.99 per metric ton of cement used Asphalt used 0.03 per metric ton of asphalt used Revenue vehicles purchased 85 per light rail train 42 per bus Source: American Public Transportation Association (2009).

173 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS VEHICLE AND FUEL LIFE-CYCLE EMISSIONS Overview The term life cycle is often used to refer to all emissions associated with the construc- tion, maintenance, and operation of the transportation system and the vehicles that use that system. In addition to the emissions associated with infrastructure construc- tion, maintenance, and operations as described in the previous section, these include the full range of emissions associated with vehicles and the fuels they consume. Life-cycle fuel emissions for transportation fuels include GHG emissions associ- ated with the production and distribution of the fuel used, in addition to the direct or tailpipe GHG emissions from vehicle operation. The full fuel cycle includes upstream emissions (sometimes called well-to-pump emissions) associated with drilling, explora- tion and production, crude oil transport, refining, fuel transport, storage, and product retail; and downstream disposal or recycling of oil products. Life-cycle analysis can also be expanded to include the full vehicle life cycle, including vehicle manufacturing (raw material extraction, processing, and transport; manufacture of finished materials; assembly of parts and vehicles; and distribution to retail locations), maintenance, and disposal. Fuel and vehicle life-cycle emissions are also known as embodied emissions. Fuel-Cycle Emissions Fuel-cycle emissions for fossil fuels are typically 5% to 29% higher than direct GHG emissions, varying by vehicle and fuel type. These percentages are based on a compari- son of the data in the GREET model, Version 1.8c, and per gallon emissions found in the General Reporting Protocol, Version 1.1 (The Climate Registry 2012). The fuel- cycle emissions of biofuels show much greater variability (when compared with direct emissions) than those of fossil fuels. Table A.26 compares CO2e emissions for LDVs (a mix of passenger cars and light-duty trucks) powered by different fuel sources. Direct combustion emissions come from The Climate Registry’s General Reporting Protocol, using the sum of emissions from CO2, CH4, and N2O for a model year 2005 vehicle. These are compared with fuel-cycle emissions from the GREET model for current av- erage fuel production conditions and vehicle efficiencies in the United States. The California Air Resources Board has also examined life-cycle emissions and developed carbon intensity factors for gasoline, diesel, and a variety of alternative fuels for use in implementing the state’s low-carbon fuel standard (California Air Resources Board 2012). These factors reflect production pathways specific to the California fuel supply. Table A.27 shows examples of direct and total emissions for conventional fuels versus biofuels with different production pathways. Biofuel emissions may vary significantly depending on production pathways. Fur- thermore, the emissions estimates for biofuels in Table A.26 do not reflect indirect impacts from the production of these fuels, such as conversion of other land to crop- land to make up for production lost to biofuels. EPA conducted an in-depth study of biofuel impacts in support of its March 2010 Renewable Fuel Standard (RFS2) rule- making. The study compared life-cycle GHG emissions for a variety of biofuels with conventional gasoline and diesel under different assumptions about production path- ways and other factors (U.S. Environmental Protection Agency 2010e). EPA provides

174 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS TABLE A.26. DIRECT AND LIFE-CYCLE EMISSION FACTORS BY TRANSPORTATION FUEL, AVERAGE FOR PASSENGER VEHICLES Motor Vehicle Fuel Direct Emissions, kg CO2e/gal (from The Climate Registry) Fuel Cycle Emissions, kg CO2e/gal (from GREET) Percentage Difference Fuel Cycle versus Direct Fuel Cycle versus Gasoline or Diesel Gasoline 8.85 10.45 18.2% Corn ethanol (E100)a 6.22 8.94 43.7% –14.4% Corn ethanol/gasoline (E10)a 8.59 10.30 19.9% –1.4% Corn ethanol/gasoline (E85)a 6.61 9.17 38.6% –12.3% Cellulosic ethanol (E100)a 6.22 1.47 –76.4% –85.9% Diesel 10.22 10.72 4.9% Biodiesel (B100)a 9.46 0.83 –91.2% –92.3% Biodiesel (B20)a 10.07 8.74 –13.2% –18.5% CNG 6.89b 8.92 29.5% –16.8%c Note: CNG = compressed natural gas. aE100 refers to 100% ethanol and B100 to 100% biodiesel. Fuels in use today include ethanol at up to a 10% blend (E10) in gasoline vehicles, ethanol at an 85% blend (E85) in dedicated or bifuel vehicles, and biodiesel at up to a 20% blend (B20) in diesel vehicles. Emission factors for the different fuel types are weighted according to the composition of the fuel. bThis direct emissions factor for CNG is from GREET. The Climate Registry emissions factor for CNG was expressed in different units, making the comparison with other fuels less clear. The CNG figure expressed here is for an amount of CNG with energy content equivalent to 1 gallon of gasoline. cCNG is compared with diesel fuel. TABLE A.27. COMPARISON OF SELECT BIOFUELS TO CONVENTIONAL FUELS IN CALIFORNIA Fuel Pathway Description Direct Emissions (gCO2e/MJ) Land Use or Other Indirect Effect (gCO2e/MJ) Total Emissions (gCO2e/MJ) Versus Average Gasoline or Diesel Gasoline California Average 95.86 0 95.86 Ethanol from corn Midwest average; 80% dry mill; 20% wet mill; dry DGS 69.40 30 99.40 3.7% California average; 80% Midwest average; 20% California; dry mill; wet DGS; NG 65.66 30 95.66 –0.2% Midwest; dry mill; wet DGS; 80% NG; 20% biomass 56.80 30 86.80 –9.5% California; dry mill; wet DGS; NG 50.70 30 80.70 –15.8% Ethanol from sugarcane Brazilian sugarcane using average production processes 27.40 46 73.40 –23.4% Diesel California Average 94.71 0 94.71 Conversion of waste oils to biodiesel when “cooking” is required 15.84 0 15.84 –83.3% Conversion of Midwest soybeans to biodiesel 21.25 62 83.25 –12.1% Source: California Air Resources Board (2009), Table 6. Note: DGS = distiller’s grain with solubles; NG = natural gas.

175 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS midpoint estimates of GHG emissions reductions of 22% for corn ethanol (E100) versus gasoline, and 57% for soy biodiesel (B100) versus diesel, assuming typical 2022 production pathways. The EPA’s midpoint 2022 values may not be representative of current or local production conditions. Selected findings from the EPA analysis, illustrating the uncer- tainty inherent in the estimates, are shown in Table A.28. This table shows both the midpoint estimate in GHG emissions reductions and the upper and lower bounds of a 95% confidence interval. For example, the table shows that one can say with 95% cer- tainty that corn ethanol from a new natural gas plant in 2022, using the identified pro- duction technology, will reduce GHG emissions by between 7% and 32% compared with conventional gasoline. The table also illustrates that ethanol production from a coal-fired plant has the potential to increase life-cycle emissions, but production from a biomass-fired plant may decrease emissions more significantly. TABLE A.28. CHANGE IN LIFE-CYCLE GHG EMISSIONS FROM BIOFUELS VERSUS CONVENTIONAL FUELS Fuel and Source 95% Confidence Interval, Upper Bound Midpoint Estimate 95% Confidence Interval, Lower Bound Ethanol from corn—natural gas planta –7% –21% –32% Ethanol from corn—coal plantb +19% +5% –7% Ethanol from corn—biomass plantb –24% –38% –49% Biodiesel from soybeans –22% –57% –85% Ethanol from sugarcanec –52% –61% –71% Ethanol from switchgrassd –102% –110% –117% Source: U.S. Environmental Protection Agency (2010e), Section 2.6. Note: Ethanol is compared with gasoline and biodiesel with diesel. Negative value denotes a net benefit for the biofuel. All results are for an average new 2022 plant. Results assume a 30-year time horizon and 0% discount rate. Future benefits with a higher discount rate will be lower, because future benefits are discounted, and therefore the initial investment (i.e., land clearing for crop production) takes longer to pay back. Using a longer time horizon will show greater benefits. aNatural gas plant, 63% dry, 37% wet distiller’s grain with solubles (DGS) with fractionation. (Fractionation separates the corn kernel into its pieces, including food-grade corn oil and protein, which are valuable coproducts that can be sold for agricultural use. DGS is the primary coproduct from a dry-mill plant. Wet versus dry DGS is the percentage of coproduct sold prior to drying.) bDry DGS with fractionation. cNo residue collection. dBiochemical process producing ethanol, excess electricity production.

176 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Vehicle-Cycle Emissions GREET and the Life-Cycle Emissions Model (LEM) provide estimates of GHG emis- sions from vehicle-cycle processes for on-road vehicles; additional estimates are provided by Chester (2008). With these estimates expressed relative to combustion emissions, the manufacture-cycle GHG emissions represent an additional 14% to 19% beyond gasoline combustion emissions, and manufacturing of freight trucks is 6% to 17% beyond diesel combustion emissions (U.S. Department of Transportation 2010). Vehicle life-cycle emissions may be considered in the context of transportation plan- ning if the impact of policies on vehicle ownership (e.g., transit, land use, and parking policies that encourage people to own fewer vehicles) can be assessed. Another reason to incorporate vehicle life-cycle effects is to quantify the effects of VMT increases on vehicle replacement rates. All else being equal, increases in VMT due to a transporta- tion plan or project will mean that vehicles will reach the end of their useful life sooner and need to be replaced. Appropriate Situations in Which to Analyze Life-Cycle Impacts A life-cycle assessment may be conducted when a more complete inventory of GHG emissions is desired. It also is important when evaluating transportation policies that affect vehicle fuels and technology types. For alternative fuel vehicle strategies (e.g., purchases of alternative fuel buses, incentives for consumer use of alternative fuel vehicles), the benefits of strategies on a life-cycle GHG basis may be markedly differ- ent than when only examining direct vehicle emissions. Many alternative fuels have relatively high life-cycle emissions compared with direct emissions. Life-cycle emis- sions must be considered in any analysis of alternative fuels such as biofuels (including biodiesel, ethanol, and biobutanol), as well as natural gas and hydrogen. Life-cycle emissions also must be considered when analyzing electrically powered vehicles, be- cause these vehicles do not have any tailpipe emissions. Probably the most important application for life-cycle analysis in transportation planning is at the project level, because a project may have a small impact on opera- tional GHG emissions but a large construction footprint. Life-cycle analysis can be used to calculate the payback period, the period over which the operational improve- ments are sufficient to offset the GHG emissions associated with construction. This is particularly important when GHG emissions reduction is part of the purpose and need (e.g., the project is being implemented as part of a climate action plan) or is one of the claimed benefits of the project. If the project is being implemented under a cli- mate action plan, decision makers should have information as to whether the project will have net benefits (beyond the payback period) by the target year(s) in the climate action plan. Data Sources and Calculation Methods Full fuel-cycle emissions factors (GHGs per vehicle mile) can be obtained from the GREET model or by adjusting gasoline or diesel emissions by a carbon intensity factor derived from a source such as EPA or the California Air Resources Board.

177 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS GHG emissions for electricity generation can be obtained from EPA’s eGRID database. These factors will vary by region of the country, depending on the regional electricity generation mix. For a discussion and examples, see “GHG Emissions from Transit Vehicles.” For developing future year life-cycle emission factors, it may be assumed that fuel-cycle emissions decline in proportion to direct (tailpipe) emissions as vehicle fuel economy improves. This assumes that the GHG emissions associated with producing a gallon of fuel remain constant. This factor may decline in the future if production and distribution practices become more efficient, or alternatively, may increase if more energy-intense methods are used (such as gasoline or diesel production from tar sands). To some extent, the GREET model may be used to evaluate different production path- ways and assumptions. Emissions factors for electricity generation may also change in the future as the mix of electricity generation methods changes. The GREET model can also be used to evaluate changes in the generation mix. See “GHG Emissions from Transit Vehicles” for additional discussion of potential assumptions regarding future emissions from electricity generation. A recent inventory of GHG emissions in the 13-county North Jersey region devel- oped by the North Jersey Transportation Planning Authority included full fuel-cycle (energy-cycle) emissions factors for comparison with direct emissions only. The inven- tory included estimates of both direct emissions (those occurring within the region’s boundaries) and consumption-based emissions, which were allocated based on trips that ended within the region. For a complete accounting of GHG emissions, energy- cycle GHG emissions associated with the production, refining, and transport of motor vehicle fuels were also calculated based on the consumption-based emissions estimate. Energy-cycle emissions factors were developed for gasoline, diesel, and ethanol- blend fuels using GREET, Version 1.8b. A MOVES run using default data for Bergen County, New Jersey, in 2006 was developed to obtain the output of energy consump- tion by fuel type and source type. A comparison of fuel combustion emissions from The Climate Registry’s General Reporting Protocol with energy-cycle emissions from GREET showed that energy- cycle emissions for gasoline were 23.0% higher than direct emissions (assuming that gasoline includes 10% corn ethanol by volume), and diesel energy-cycle emissions were 10.8% higher than direct emissions. (These energy-cycle emission estimates were developed using GREET Version 1.8b emissions factors, which differ from the GREET Version 1.8c factors cited above). In order to estimate energy-cycle emissions, the consumption-based GHG emissions estimates were multiplied by the appropri- ate energy-cycle multiplier, which varied between 11% and 23% depending on the amount of diesel versus gasoline used by vehicle type. For example, light commercial trucks use (84.7% gasoline * 23.0% increase) + (15.3% diesel * 10.8% increase). This results in an estimated increase in energy-cycle emissions for all light-duty commercial trucks of 21.2%. These percentages were then applied to the consumption-based emis- sions to estimate energy-cycle emissions from on-road vehicles. Table A.29 shows the resulting differences in consumption emissions versus full energy-cycle emissions.

178 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS TABLE A.29. ON-ROAD VEHICLE GHG EMISSIONS PROJECTIONS IN NORTH JERSEY Total Emissions (MMT CO2e) 2006 2020 2035 2050 Consumption 17.0 21.2 29.1 26.6 Energy cycle 20.8 25.9 35.5 32.4 The Columbia River Crossing Project energy and CO2e analysis for a large bridge replacement and freeway upgrade project in Portland, Oregon, and Vancouver, Washington, shows the type of analysis that can occur at a project level ( Columbia River Crossing Project Team 2008). Energy supply and demand in the states of Washington and Oregon were characterized by energy supply sources and use sectors. Specific data relating to fuel consumption rates were obtained from the state DOTs and the U.S. Department of Energy. The state DOTs also provided traffic volumes and vehicle classification, and transit vehicle energy consumption data were provided by local transit agencies. The energy analysis addressed four primary issues: • Energy consumed during construction of the project; • Energy consumed during operation of the project; • Measures to reduce or offset construction and operational effects on energy; and • CO2e emissions resulting from use of electricity, gasoline, and diesel. Emissions factors obtained from EPA were used to estimate CO2 and other GHGs produced from combusting gasoline or diesel in a motor vehicle. For petroleum-based fuels, the amount of fuel consumed by the project was multiplied by the applicable emission factor to estimate CO2 emissions, then multiplied by another conversion fac- tor to account for the global warming potential of other GHGs emitted by vehicles. A general equation used for estimating CO2 and CO2e emissions was EM = FC × EF × CDE where EM = emissions of CO2 or CO2e (lbs), FC = fuel consumed (gals or kW-h), EF = emission factor (lbs of CO2/gal or lbs of CO2/kW-h) (based on fuel type), and CDE = CO2e conversion factor (100/95). The fuel consumed was the amount used to operate the facility. The approach for determining energy use during construction was based on an input–output method developed by Caltrans. This method estimates energy requirements using energy fac- tors that were developed for a variety of construction activities (e.g., construction of structures, electrical substations, and site work). These energy factors relate project costs with the amount of energy required to manufacture, process, and place construc- tion materials and structures.

179 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS The general equation for estimating energy consumed during construction was E = C × EF × DC where E = energy consumed (Btu), C = cost of a particular construction activity (2007$), EF = energy factor (Btu/1973$), and DC = dollar conversion (1973$/2007$). Because the Caltrans energy factors were based on construction cost estimates in 1973 dollars, a dollar conversion was necessary because the project’s cost estimates were in 2007 dollars. The results of the analysis are shown in Tables A.30 and A.31. TABLE A.30. DIFFERENCES BETWEEN SYSTEM-LEVEL CHOICES OF DAILY ENERGY USE AND CO2E EMISSIONS Energy Consumed Electricity Consumed Gasoline Consumed Bio/Diesel Consumed CO2e Emissions System- level Choice mBtu Change (%) kW-h Change (%) gal Change (%) gal Change (%) tons Change (%) Build-replace Build- Supplement –506.6 — –8.8% — –8,018 — –5.0% — –147 — –1.5% — –3,343 — –11.6% — –43.5 — –8.8% — Bus Rapid Transit Light Rail Transit — –5.8 — –0.1% –9,435 — –5.8% — — 0 — 0.0% — –289 — –1.1% –0.1 — 0.0% — Vancouver alignment Interstate 5 alignment –27.5 — –0.7% — –649 — –0.4% — 0 — 0.0% — –182 — –0.8% — –2.4 — –0.7% — Full Length Clark College MOS Mill Plain MOS — –7.7 –13.2 — –1.4% –2.4% — –2,262 –3,854 — –1.4% –2.4% — 0 0 — 0.0% 0.0% — 0 0 — 0.0% 0.0% — –0.8 –1.4 — –1.4% –2.4% No Toll Standard Toll on I-5 Standard Toll on I-5 and I-205 — –102.0 –186.2 — –1.9% –3.5% — 0 0 — 0.0% 0.0% — –615 –1,256 — –6.0% –12.3% — –186 –220 — –0.7% –0.9% — –8.5 –1504 — –1.8% –3.3% Source: Columbia River Crossing Project Team (2008). Note: MOS = minimum operable segment. “—” indicates the highest amount of energy, electricity, fuel consumed, and CO2e emitted, a “0” indicates no differences between alternatives, and a negative number indicates the difference (amount less).

180 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS TABLE A.31. ALTERNATIVES SUMMARY OF CONSTRUCTION-RELATED ENERGY USE AND CO2E EMISSIONS, COLUMBIA RIVER PROJECT Alternative Energy Consumed (mBtu) CO2e Emissions (tons) Alternative 2 With 16th Street Tunnel 7,055,867 590,178 With McLoughlin Tunnel 6,997,372 585,536 Alternative 3 With 15th Street Tunnel 7,281,549 608,224 With McLoughlin Tunnel 7,221,671 603,472 Alternative 4 5,903,553 494,010 Alternative 5 6,084,734 509,171 Source: Columbia River Crossing Project Team (2008). INDIRECT EFFECTS AND INDUCED DEMAND Overview A transportation project or program may have indirect as well as direct effects on GHG emissions. Most notably for transportation analysis, GHG emissions can result from additional travel (induced demand) that would not have occurred in the absence of the proposed highway or transit project, corridor improvements, or proposed trans- portation system plan or program. Indirect or Secondary Effects In its regulation on the National Environmental Policy Act, the Council on Environmen- tal Quality defines indirect effects as effects that “are caused by the action and are later in time or farther removed in distance,” in contrast to direct effects, which are “caused by the action and occur at the same time and place.” Indirect effects “may include growth-inducing effects and other effects related to induced changes in the pattern of land use, population density, or growth rate,” as well as environmental and other ef- fects related to these land use changes (Council on Environmental Quality 1986). Induced Demand In a general sense, induced demand can refer to increased travel by any mode that results from improved travel conditions (e.g., reduced travel times or costs). As used here, induced demand refers specifically to induced vehicle travel, because that is the impact of interest for GHG analysis. An increase in vehicle travel may be a result of greater trip making, longer trips, and/or shifts from other modes. In the short term, this primarily reflects a redistribution of trip-making patterns among the same spatial distribution of activities. Over the longer term, additional induced demand may re- sult as land use patterns change in response to transportation system changes. For a particular transportation facility, an improvement may also result in more traffic due to route shifting from other facilities to the improved facility, or shifts in the time of day of travel. However, only net new travel (VMT) is truly characterized as induced demand. Induced demand is therefore largely an indirect effect.

181 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Induced demand effects are significant in GHG analysis because they can partially offset the GHG emissions reduction benefits of capacity or operational improvements that reduce congestion. However, as discussed below, induced demand effects are dif- ficult to measure and forecast accurately, and there is currently no consensus as to how the magnitude of these impacts compares with the benefits of capacity expansion and congestion relief over the long run. Also, there is little evidence regarding how induced demand effects might vary for operational improvements (such as signal synchroniza- tion) versus capacity expansion, or for strategies that primarily affect reliability (such as incident management) rather than average travel time. Measurement and Forecasting of Induced Demand Effects Induced demand is a widely recognized phenomenon, but its magnitude may vary considerably from context to context. Induced demand can be described in terms of elasticities, such as a change in VMT with respect to a given change in highway capacity or travel time. Induced travel elasticities are often expressed in relationship to highway capacity (e.g., lane miles) because it is easier to measure than travel time. However, it is preferable to determine elasticities in relation to travel time because reduced travel time (increased travel speeds) is generally the factor that leads to addi- tional travel. Induced demand may result from people making more trips, shifting modes (from nonauto to auto), or making longer vehicle trips (because of shifts in routes and/or destinations). In the long term, land use patterns may disperse, leading to more increases in vehicle trip lengths as destinations become farther apart (and per- haps less transit accessible). Attempts to measure induced demand elasticities, usually in terms of VMT with respect to lane miles, have produced widely varying estimates. The literature usu- ally distinguishes short-term (less than ~5 years) from long-term elasticities, because response tends to be greater over the long term as people have more opportunities to shift their activity locations and travel patterns. Short-term estimates have ranged from about 0.1 to 0.7, with most clustering in the range of 0.2 to 0.5. An elasticity of 0.5 means that a 10% increase in lane miles for a region or corridor would result in a 5% increase in VMT. Long-term estimates have ranged from about 0.3 to 1.1, with most clustering in the range of 0.4 to 0.9. Elasticities of VMT with respect to lane miles will be positive, and elasticities of VMT with respect to travel time will be negative. Only a handful of studies have examined elasticities of VMT with respect to travel time or speed, but the results tend to fall in the same absolute range (Cervero 2002). Cohen (2002) cautions that because of various methodological issues some of these studies have likely overstated actual elasticities; he concludes from a review of other studies and original analysis that travel time elasticities are probably in the range of –0.1 to –0.4. Measurement of induced demand is complicated by the need to separate spatial or temporal shifts in traffic patterns (e.g., to the facility) from net increases in traffic, and over the long term, by the need to separate the effects of other trends (e.g., background population and employment growth) from the effect of the facility or program itself. The magnitude of induced demand is likely to vary widely from situation to situation.

182 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Expansion of a highly congested freeway in Los Angeles is likely to result in a signifi- cant amount of induced demand, while expansion of a four-lane freeway in rural North Dakota is likely to have little, if any, effect on traffic volumes. Induced demand is likely to be greater over the medium to long term, but measuring induced demand becomes more difficult over a longer time frame. Travel demand forecasting models can help to separate these various effects and predict induced demand. However, most models in practice today have limitations in their ability to accurately reflect induced demand effects. In particular, • Many models do not include feedback from the traffic assignment to the trip dis- tribution step. As a result, longer trips due to reduced congestion will not be modeled; • Even fewer models include feedback to the trip generation or auto ownership steps (although evidence suggests that the primary impact of congestion is on trip lengths and mode choice, not total trip rates); • Few travel models are integrated with a land use forecasting model, meaning that growth effects will not be captured. Furthermore, land use forecasting methods must accurately consider the effects of transportation infrastructure and accessibility im- provements on development patterns if they are to capture induced travel effects; • Time-of-day shifts are not reflected in many models. While time-of-day shifts themselves do not represent true induced travel, they may affect the amount of induced travel by changing the level of congestion in peak and off-peak periods; • Static traffic assignment algorithms may not account for the impacts of queuing on route shifts; and • The interpretation of the doubly constrained gravity model commonly used in the trip distribution step is unclear, even though the results may mimic the impacts of induced demand. One study by Rodier et al. (2001) of the Sacramento, California, region suggested that about 50% of long-term induced travel is not captured by the use of travel demand models when they are used without land use feedback. In contrast, a review by the Metropolitan Washington Council of Governments (2001) found that the agency’s travel forecasting process generally captures induced travel, although it does not sepa- rate induced travel from other increases in travel. The review notes that the travel fore- casts were based on a cooperative land use forecasting process that addresses changes in development patterns predicted to occur as a result of major changes in transporta- tion system capacity. DeCorla-Souza (2000) also concluded that travel models may adequately capture induced demand, based on the observation that relatively large travel time elasticities were imputed from the use of a regional travel demand model in Memphis, Tennessee (–0.7 without feedback, –1.1 with feedback). An assessment of three case study models by Rodier et al. (2001) found that when travel times are fed back to a land use model and/or the trip distribution step, then models “can represent induced travel within the range documented in the empirical literature.”

183 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Appropriate Situations in Which to Conduct Additional Analysis of Induced Demand In some cases, the analytic tools used to estimate VMT and GHG emissions may already account, at least partially, for induced demand. As noted above, this is most likely to be the case when a regional travel demand model with appropriate feedback loops is being used. For projects that do not significantly affect the overall time and/ or cost of travel (e.g., projects that primarily improve safety), induced demand effects are likely to be minimal. For small projects such as traffic signal timing, there is some debate as to whether there is any measurable induced demand effect, although it may be that an areawide collection of smaller projects adds up to improvements in condi- tions large enough to affect travel choices. For major investment projects that significantly affect travel time and/or costs, it is desirable to specifically analyze potential induced demand effects and assess how they may affect the GHG emissions benefits of a project or program. Such situations may include • Project-level analysis in which a travel demand model that includes feedback effects is not already being used; • Project- or systems-level analysis in which the existing modeling system has limited or no ability to capture effects, including growth-inducing effects of transportation system improvements, and/or feedback from congestion to the trip distribution step; • Projects that generate interest in induced demand and growth effects for GHG emissions, air quality, land use impacts, and/or other analysis purposes; and • Projects for which resources are available to enhance modeling systems or apply sketch-level analysis methods. Data Sources and Calculation Methods Methods to account for induced travel and indirect effects include the following, listed in order of increasing level of effort: • Direct application of induced demand elasticities; • Sketch plan corridor-level models; • Enhancements to travel demand models to better incorporate feedback from con- gestion; and • Land use forecasting methods incorporating accessibility, with feedback to travel demand models. Direct Application of Induced Demand Elasticities Induced demand elasticities from the literature may be applied to changes in travel conditions to estimate changes in traffic volumes. For example, consider a 5-mile- long segment of a four-lane freeway with a 2-hour peak period, peak-direction traffic

184 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS volume of 8,000 vehicles. This facility operates at an average of 30 mph under peak conditions. The total peak period congested VMT on this facility is 8,000 vehicles × 5 miles × 2 time periods (a.m. and p.m.), or 80,000 VMT. The freeway is being expanded to six lanes, which is projected to increase the average peak period speed to 60 mph, representing a 50% reduction in travel time over this segment. Assuming the long-run elasticity of VMT with respect to travel time in this area is 0.4, this means that VMT will increase by 80,000 × 50% × 0.4, or 16,000 VMT per day within a 5- to 10-year period. The GHG emissions impacts of this additional VMT can then be compared with the emissions benefits of reduced congestion, as analyzed using other methods such as a traffic simulation model or MOVES emissions factors by vehicle speed. This is a crude assessment that does not account for factors such as the existing versus postproject level of congestion and travel times on the facility, differences in local versus regional traffic shifts, effects of traffic changes on nearby facilities, or feed- back from changes in congestion to changes in VMT. When choosing an appropriate elasticity, the analyst must be careful to consider whether the elasticity value includes route diversions and time-of-day shifts that do not actually represent new VMT; that is, some of the additional VMT and GHG emissions calculated for the expanded facility are actually removed from other facilities or off-peak time periods. Nevertheless, the elasticity approach may be useful for determining an order of magnitude assessment of how induced demand effects are likely to compare with congestion relief benefits. Sketch Plan Corridor-Level Models FHWA has developed a spreadsheet-based model known as Spreadsheet Model for Induced Travel Estimation (SMITE). This model is useful for sketch planning analysis, especially when four-step urban travel models are either unavailable or are unable to forecast the full induced demand effects. SMITE can be used to provide useful information to assist policy makers in evaluating proposals for specific additions to highway capacity for corridor studies. An accompanying paper discusses the concepts underlying SMITE and describes an application of SMITE to the evaluation of a typi- cal freeway capacity expansion project (Federal Highway Administration 2012f). SMITE has the following notable features: • It accounts for changes in traffic volumes and speeds on parallel arterials due to diversion, as well as the freeway corridor being improved; • It uses speed relationships to estimate the effects of congestion on speeds; • The speed estimates are sensitive to peak spreading and queuing under congested conditions; • It allows the user to provide travel demand elasticity estimates to obtain estimates of induced travel; and • It computes external costs of induced travel (mobility and other user and nonuser benefits and costs) using user-provided estimates of external costs per VMT and com- pares benefits and disadvantages over the life of the investment.

185 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS SMITE-ML, a variation of SMITE, was developed specifically to evaluate man- aged lanes (Federal Highway Administration 2012e ). SMITE and SMITE-ML do not estimate GHG emissions directly, but they produce outputs of changes in freeway and arterial VMT and travel speeds that could be used to develop estimates of GHG emis- sions using emissions factors from MOVES or another model. IMPACTS, another FHWA spreadsheet-based sketch plan model that incorporates induced demand, is a series of spreadsheets developed to help screening-level evalua- tion of multimodal corridor alternatives. The model is described in more detail above. Enhancements to Travel Demand Models Differences in parameters and methods make it difficult to generalize regarding how to ensure that travel demand models accurately reflect induced demand effects. However, it is apparent from the literature that the two most important enhancements are (1) feedback from traffic assignment to the trip distribution step and (2) feedback to land use patterns. COMSIS (1996) provides a general discussion of the incorporation of feedback into travel demand models. Feedback to the trip distribution step can be incorporated within the framework of a typical four-step model. Feedback to land use may be more problematic, as the availability of land use models incorporating accessibility measures varies widely by metropolitan area, and few are integrated with the transportation model. Furthermore, today’s land use models are suitable only for plan- and systems- level analysis and not for analyzing the specific impacts of individual transportation projects (Cambridge Systematics and Gliebe 2009). Consideration of the GHG emis- sions impacts of induced demand may therefore need to be modeled at a systems level (considering an overall systems plan) or done using a qualitative or sketch-level assess- ment of the likely land use impacts of a project, as described below. Land Use Forecasting Methods Land use forecasting methods span the range from qualitative, judgment-based assess- ments; to spreadsheet-based allocation models; to fully integrated travel demand and land use forecasting models. The most appropriate technique depends on a number of factors, including • Specific information needs, and the level of accuracy and precision needed or desired for decision making; • The magnitude of the transportation project and nature of its potential impacts; and • Availability of data and analysis resources. Several reports provide guidance on forecasting methods for land use and indi- rect effects of transportation projects. More recent publications include the Desk Reference for Estimating the Indirect Effects of Proposed Transportation Projects (Louis Berger Group 2002). This NCHRP report contains guidance and a framework for prac titioners to use for defining the indirect effects of proposed transportation projects, identifying tools for estimating these effects, and analyzing these effects. Avin

186 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS et al. (2007), in their report Forecasting Indirect Land Use Effects of Transporta- tion Projects, provide more detailed guidance and information on selected land use forecasting methods and assist practitioners in selecting an appropriate methodology. FHWA’s Interim Guidance on the Application of Travel and Land Use Forecasting in NEPA (2010) is intended to encourage improvement in how project-level forecasting is applied in the context of the NEPA process. It focuses on the procedural or process considerations in forecasting rather than on technical methods. In order to complete the loop of incorporating indirect effects into GHG emis- sions analysis, there must be feedback between the land use forecasting process and the travel demand analysis. At a minimum, the revised land use forecast considering a project’s effects must be used as a basis for redoing the travel demand forecast and obtaining updated VMT and speeds by facility type. Ideally, multiple iterations are conducted, so that the updated travel conditions are used to update the land use fore- cast, which is then used to derive new travel forecasts. This process should be repeated until the travel forecasts converge (i.e., until the difference between the previous fore- cast and the current forecast is small). GHG emissions under the final project scenario are then compared with GHG emissions projected under the no-build scenario with the baseline land use and travel forecasts. Land use forecasting models and methods vary considerably in how transportation improvements are assumed to affect land development patterns. An ideal model will be based on a sound, empirically supported relationship between land development and accessibility, which is a measure of the number of potential destinations that can be reached from a particular location within a given travel time. Approaches that rely on judgment or unvalidated relationships may still provide useful insights, but should be subject to sensitivity testing to determine how a range of plausible assumptions may affect conclusions about induced demand and the resulting GHG effects. USING MOVES TO ESTIMATE GHG EMISSIONS Overview EPA’s Motor Vehicle Emission Simulator (MOVES2010) model was released in December 2009. On March 2, 2010, EPA provided notice in the Federal Register that MOVES2010 is officially approved for use in state implementation plans and for transportation conformity analyses outside of California (U.S. Environmental Protec- tion Agency 2010c). MOVES allows the transportation planner to model GHG emis- sions from the project level to the regional level, and in a manner consistent with the way emissions of other pollutants would be calculated using MOVES for state imple- mentation plans or transportation conformity analyses. The primary uses of MOVES for transportation planners include estimation of GHG emissions inventories from the project level to the regional level using typically generated local data, as well as evalua- tion of GHG emissions reduction policies affecting vehicle speeds, activity, or fleet mix, and some biofuels. The MOVES model software and documentation is available from the EPA website (U.S. Environmental Protection Agency 2012b).

187 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS MOVES can be used as either an inventory model or an emissions rate (factor) model. When used as an emissions rate model, MOVES provides only the emission rates; total emissions must be calculated using VMT and vehicle population outside of MOVES, likely in a spreadsheet lookup table or in an air quality postprocessor to a travel demand model. When used as an inventory model, however, MOVES uses VMT and vehicle population to calculate total emissions internally. In order to explain all of the MOVES inputs, this document assumes the use of MOVES as an inventory model. EPA provides technical guidance for which inputs are unnecessary when using MOVES as an emissions rate model. This section summarizes the use of MOVES in three contexts: • Using MOVES at a regional scale in areas without a travel demand model, • Interfacing a state or regional travel demand model with MOVES, and • Using MOVES for project- or corridor-level analysis. Locally specific inputs, such as vehicle age distributions and characteristic ambi- ent meteorology, are also discussed. Inputs to MOVES that are dependent on vehicle activity (e.g., VMT distributions by speed and road type) are discussed in the indi- vidual sections, because these can vary depending on the type of MOVES application. Finally, inputs that are common across all types of MOVES applications are discussed. Methods to account for new fuel economy and/or low-carbon fuel standards are presented. Federal emissions and fuel economy standards that are proposed but not adopted or have been recently adopted may not be reflected in the emission rates built into the most current version of the MOVES model. MOVES will also not reflect any state-adopted GHG emissions or low-carbon fuel standards that go beyond federal requirements. One of the first steps in any GHG emissions analysis should be to define a protocol that explains what will be studied, the spatial and temporal scale of analysis, inputs, and any limitations that would be applicable to the results. Applying MOVES in Areas Without a Travel Demand Model MOVES can be used to develop GHG emissions inventories by using regional VMT forecasts developed outside of a travel demand model as inputs. For this level of analysis, less detailed data are available than in the other scales of analysis discussed in this document. In this scale of inventory, emissions in MOVES are calculated based on the regional distribution of activity by speed, road type, and vehicle type (known as source type in MOVES). MOVES users experienced with developing criteria pollutant emissions esti- mates for ozone or particulate matter analyses will note that the primary differences for GHG emissions analysis are the time scales needed and the pollutants to be mod- eled. GHG inventories are calculated at an annual level rather than the seasonal or daily level typically used for particulate matter or ozone inventories. Areas without travel demand models that have previously developed criteria pol- lutant emissions estimates are more likely to have travel data in the form of annual average daily travel rather than annual VMT, although in some cases, annual VMT

188 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS would have been necessary. Annual average daily travel can be scaled to the annual level by multiplying by 365. The information below assumes that the practitioner has access to the MOVES2010 model, user’s guide, and technical and policy guidance documents prepared by EPA. From the main MOVES screen, the following options should be selected for use in developing a GHG emissions inventory: • Scale: County (shown in Figure A.13); • Calculation type: Inventory (shown in Figure A.13); • Region: County, then select appropriate state and county. An example of this selec- tion is shown in Figure A.14; • Time aggregation level: Year (automatically checks all months, hours, weekends, and weekdays) (Illustrated in Figure A.15); • Fuels and source use types: Select all included in the inventory area, as shown in the example in Figure A.16; • Road types: Select all road types, including off-network, that are included in the inventory area, as illustrated in Figure A.17; and • Pollutants and processes: In order to calculate CO2e emissions, all of the following must be selected from the pollutant list: total energy consumption, methane, nitrous oxide, atmospheric CO2, and CO2e. This screen is illustrated in Figure A.18. Figure A.13. MOVES domain/scale screen.

189 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Figure A.14. MOVES geographic bounds screen. Figure A.15. MOVES time span screen.

190 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Figure A.16. MOVES fuel and source types selection screen. Figure A.17. MOVES road type selection screen.

191 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS After the analyst has made the appropriate selections on the screens illustrated in the figures above, the MOVES County Data Manager is used to enter local data for the area. Local data inputs are required for various parameters for MOVES to run success- fully. However, some of these inputs can be based on EPA default data if they are not important to the calculation of GHG emissions or if no locally derived data are avail- able. The local inputs for which data are needed for a successful MOVES run include vehicle age distributions, average speed distributions, fuel information, temperature and humidity information, road type distribution, source type population, VMT data, and inspection and maintenance program information. Of these, accurate local data are most important for VMT, vehicle age distributions, and road type distributions in creating a GHG emissions inventory. These inputs are described in greater detail below. From the MOVES County Data Manager, in most cases, a spreadsheet template can be exported that includes all combinations of the fields that need to be populated for use in MOVES. This template can then be populated by the user with local data and imported as part of the MOVES database for the selected county. A separate database (and MOVES model run) would need to be developed if sepa- rate runs for each county in the area are to be performed. If only areawide GHG emissions data are needed, then the County Data Manager should be populated using data representative of the entire area (e.g., including all VMT in the area), but using a single county code to represent the area. Alternatively, the Custom Domain option in MOVES can be used to represent a multicounty area. However, several additional Figure A.18. MOVES pollutants and processes selection screen.

192 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS inputs are needed with this option. The inputs related to vehicle activity are discussed below, and other inputs that apply to all scales of analysis are discussed at the end of this document. Local Area Vehicle Travel Inputs for MOVES in Areas Without a Travel Demand Model VMT For areas without travel demand models, it is expected that VMT will be available at the facility level for each county within the area. For other types of analyses, VMT would normally be collected at the annual average daily travel level (such as for an ozone analysis). However, for a GHG emissions analysis, VMT needs to be the estimated an- nual volume. The total VMT data input to MOVES must be broken down by Highway Performance Monitoring System (HPMS) vehicle types (i.e., passenger car; motorcycle; other two-axle, four-tire vehicles; single-unit and combination trucks; and buses). In addition to total county- or area-level VMT by HPMS vehicle category, MOVES requires VMT fractions by hour, day, and month. For areas without travel demand models, it is recommended that the user export the default data for these inputs from the MOVES County Data Manager and use those as input. However, the total VMT data (in the HPMSVtypeYear table) must be populated with local data, as VMT defaults in this table are zero. Speed Distributions To calculate emissions inventories for areas without travel demand models, the speed inputs typically used would be an average speed by roadway type and would gener- ally apply to the entire day rather than speeds developed for peak and nonpeak travel periods. If more detailed speed data are available for an area without a travel demand model, the section in this report on speed distribution inputs for areas with regional travel model outputs should be reviewed, and more specific speed distributions should be developed, as applicable. The average speed by roadway type data need to be en- tered in MOVES as speed distribution files via the MOVES County Data Manager. If the user has previously developed these speeds into speed distribution files for use with MOBILE6, EPA has provided a converter program, in Excel spreadsheet format, that can convert the MOBILE6 speed distribution files to the format needed by MOVES. If no previous speed distribution files exist, the user will need to develop the MOVES speed distribution data directly by exporting a template for the Average Speed Distribu- tion table through the MOVES County Data Manager. This will create a spreadsheet or text table that includes all combinations of the sourceTypeID codes (vehicle catego- ries), roadTypeID codes, hourDayID codes, and avgSpeedBinID codes that are needed for input to MOVES with the avgSpeedFraction field blank for all records. Using this template, the user would then populate the speed bin or bins that represent the desired speed with 1.0 for all occurrences of this speed bin or bins for the selected road type.

193 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS The speed distribution data in MOVES represent the fraction of time that a specific vehicle (source) type spends within a specific speed range. For example, if it has been determined that the average speed on urban restricted access roads is 45 mph, then the user would populate the avgSpeedFraction field of the AvgSpeedDistribution table with 1.0 for avgSpeedBinID equal to 10 (speeds at least 42.5 mph and less than 47.5 mph) and for roadTypeID equal to 4 (urban restricted access) for all sourceTypeIDs and all hourDayIDs. The avgSpeedFraction field for all other speed bins on roadTypeID equal to 4 would be populated with 0. Note that in order to represent speeds that are not the average of the end points of one of the speed bins, speed distribution fractions will split between two adjacent speed bins to correctly model the selected speed. The procedure for estimating the fraction of hours of travel in each bin is explained in the MOVES technical guidance document. Default speed distribution tables can be obtained from the MOVES County Data Manager. However, because the CO2 emission calculations are sensitive to speed, the analyst is advised to use speeds representative of travel in the local area rather than the MOVES default data. Road Type Distribution of VMT The Road Type Distribution table also requires local data inputs through the MOVES County Data Manager. For each of the MOVES source type IDs, the fraction of VMT that occurs on each of the MOVES road types must be entered. This data set is relevant to a GHG emissions inventory because the speed data affecting the CO2 emission rates are defined by road type. Thus, the assignment of VMT fractions to specific road types will determine the total amount of VMT that is represented at a specific speed. Again, a template can be exported that provides the necessary fields and combinations of data IDs. Applying MOVES in Areas with a Travel Demand Model Travel demand models are a commonly used tool to forecast traffic conditions based on future socioeconomic and demographic projections and alternative transportation networks. MOVES can be used in conjunction with travel demand model VMT out- put by vehicle type (light duty versus heavy duty), facility type (freeway, arterial, local street), and speed to estimate overall GHG emissions from on-road vehicle travel. Emissions for a particular plan alternative can then be compared with emissions for a no-build or existing + committed (E+C) forecast and/or base year emissions by apply- ing emissions factors to link-level model output, stratified by speed, road type, and vehicle type. E+C forecasts represent conditions if no further transportation improve- ments were implemented beyond what is already funded to complete construction within the last year of the transportation improvement program. E+C conditions typi- cally represent the no-build scenario for comparison of long-range transportation plan alternative scenarios. In using MOVES with a travel demand model there are often complications involving differences in vehicle and facility type definitions between the two modeling paradigms. The next three subsections discuss an approach to dealing with common interface issues.

194 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Correspondence Between Travel Demand Model Facility Classifications and MOVES Roadway Type Run the travel demand model for the no-build or E+C scenario; output link-level vol- umes and speeds by MOVES road type. The facility types within the travel demand model will likely not match up with the MOVES road types. However, an equivalency table of travel demand model facility type to MOVES road type can be created to ap- pend the MOVES road type code to each link in the model network. Allocation of VMT by Vehicle Type As a default, travel demand models do not output trips or VMT by the 13 MOVES source (vehicle) types. Typically, the models output VMT by passenger trips or pas- senger and truck trips. Sometimes truck trips are further disaggregated into light-, medium-, and heavy-duty truck trips. Because the 13 MOVES source (vehicle) types nest within the six HPMS vehicle types, HPMS data can be used to obtain the percent- age of VMT by the six HPMS vehicle types, and the percentage distribution can then be applied to the travel model VMT. Although default vehicle type distributions from MOVES can be used, EPA encourages the use of local data for the six HPMS vehicle types and discourages the use of national defaults unless they are used to further break down the six HPMS types into the 13 MOVES vehicle types. Mapping MOVES Emissions Factors to Travel Demand Model Link-Level Data When running MOVES to get emissions factors by vehicle type and speed, 2.5- or 5-mph speed increments covering the range of speeds observed on the roads to be modeled should be used. If emissions factors will be applied for the six HPMS catego- ries, the emissions factors output from MOVES for the 13 MOVES source (vehicle) type categories can be postprocessed using a weighted average. The emissions factors can then be mapped to links either by matching the link-level speed with the nearest incremental speed modeled in MOVES or by interpolating from a lookup table of emissions factors by speed increment. Applying MOVES in Project-Level Analyses MOVES can also be applied to estimate the GHG emissions associated with project- or corridor-level changes to the transportation system. As of 2010 there was limited EPA or FHWA guidance available for how to conduct a project-level GHG emissions analysis using MOVES. The EPA–FHWA MOVES training course (U.S. Environmental Protec- tion Agency 2012d) presents information on project-level MOVES analyses that is con- sistent with the information in the EPA MOVES user’s guide, but it has limited utility for GHG assessments. Similarly, the particulate matter hot-spot guidance (U.S. Environmen- tal Protection Agency 2010f) has potential crossover information, but the scope of the assessments for criteria pollutants or mobile source air toxins would be expected to be considerably different from those for GHGs, both in geography and time scale. Transportation projects in the context discussed here are generally road projects that do not directly cause new trip generation. In contrast, development projects result in the generation of new local trips. For example, a road-widening project or intersection

195 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS improvement would be considered a transportation project. Such projects lead to changes in GHG emissions as a result of changes in speed, flow of traffic, or traffic volumes. Use of the MOVES model now enables transportation modelers to capture GHG emissions changes that result from such projects. These changes result from changes in fuel con- sumption with speed, traffic flow, and volume changes. In order for these changes to be captured in a GHG analysis, detailed information at the link level is needed. Appendix G of the MOVES user’s guide provides some information about how to set up a MOVES run for a project-level analysis. However, the example provided is directed toward criteria pollutant analyses and needs to be adapted to the larger scale of analysis that is likely to be typical of a GHG project study. The information below summarizes some of the information in the MOVES user’s guide that is likely to be relevant to project-level GHG studies. It is suggested that weekday and weekend day analyses be developed separately and then weighted to reflect the prevalence of weekday versus weekend travel patterns during a year. For each roadway link, the user must specify the MOVES road type that best rep- resents it. Any of the four road types may be chosen to represent each project link. A link length must be specified for each roadway link to be modeled. Traffic volume estimates must be specified for each link. This is the total average traffic flow from all vehicle types on the link during the period being modeled. Any or all of the MOVES vehicle source types may be included at the same time in a project- level run. The average speed on each link must be specified. The drive schedule inputs should match the overall average speed(s) of the individual drive cycles as submitted in the LinksDriveSchedule tab. The average road grade must be specified for each link. This input represents the overall average grade of the entire link, not one specific link segment. It is used only if a drive schedule input is not provided. MOVES can use second-by-second driving schedules to model vehicle operation, such as those that might be obtained from a traffic simulation model. If drive sched- ules are not provided, MOVES uses the average speed and average grade inputs plus default MOVES driving cycles to model the driving behavior. The distribution of traffic by MOVES source type is another MOVES input. It is entered as the SourceTypeHourFraction in the LinkSourceType worksheet and the LinkSourceType input tab. If there is an expected difference in the distribution of traf- fic among source types, this difference should be reflected in the build versus no-build MOVES simulations. Emissions that occur off the transportation network (e.g., in parking lots) can be included in a project-level MOVES run via the extended idle and parked vehicle fraction parameters. The operating mode distribution needs to reflect the soak times in parking lots or parking garages. The soak time is the time since the vehicle engine was turned off. The source type age distribution for the vehicles in the project area needs to be input to MOVES. This age distribution would usually be expected to be the same as an area would use for any urban-scale analysis such as a state implementation plan

196 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS emissions inventory. However, if the project is expected to attract a nonrepresentative vehicle age distribution (e.g., a football stadium might attract a newer fleet than aver- age), then that age distribution should be used in MOVES. A younger fleet would be important for any forecast year analyses that will be affected by new fuel economy standards (i.e., lower GHG emission rates). Link-Specific Driving Schedules As discussed above, MOVES can use second-by-second driving schedules to model ve- hicle operation. This section addresses developing link-specific driving schedules using microscopic traffic simulation models as a preferred alternative to standard VMT with estimated average speeds. Microscopic transportation simulation models simulate the movement of individual vehicles at second or subsecond intervals through a repre- sentative travel network. In doing so, they keep track of the speed and acceleration of every second of the simulation. This approach offers a level of refinement beyond using average speeds, which does not capture the details of how congestion forms and dissipates in practice. Whether an analyst chooses to use a microsimulation model to prepare a GHG emissions analysis for a transportation project depends on whether there is expected to be a significant difference in the traffic delay characteristics with the transportation project, because GHG benefits accrue from reducing the excess fuel consumption that might occur without the project. If average speeds and times in modes (acceleration, deceleration, cruise, and idle) do not change with the project, then a microsimulation analysis may not be needed. Vehicle activity is significantly different under different regimes of congestion, such as • Queue-forming transitional flow, characterized by backward-forming shock waves; • Movement within the queue; and • Recovery from queuing conditions. The output of each simulation run is a vehicle trajectory file that for every second of the simulation indicates the speed and acceleration of every vehicle in the network; that is, the output consists of instantaneous speed and acceleration. Such voluminous data require some summarization before being input to MOVES. Table A.32 provides an excerpt from a typical driving cycle for MOVES input. There are other possible approaches to using the microsimulation model outputs than using the average speed by link. The average speed assumption is a reaction to long MOVES run times, which would be exacerbated by using the speed–acceleration trajectories for every vehicle. Another user option is to sample the second-by-second individual vehicle trajecto- ries and to then use that sample as MOVES inputs. If such an approach is considered, it is suggested that the microsimulation model results for each time period be reviewed to determine how much variability there is in the speed–acceleration–time traces in order to develop an appropriate sampling method. A significant amount of variability in the vehicle-to-vehicle speeds and accelerations during a time period suggests that a larger sample size is warranted. This approach is discussed in EPA’s particulate matter

197 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS hot-spot guidance: “For both free-flow highway and intersection links, users may directly enter output from traffic simulation models in the form of second-by-second individual vehicle trajectories” (U.S. Environmental Protection Agency 2010f). If this approach is used, then EPA recommends using data from a representative sample of links, in which each link represents an individual vehicle trajectory, as input to the MOVES model and then scaling the results based on the number of vehicles on the actual road links to the number of sampled vehicle trajectories modeled. The sampled vehicle trajectories should include idling, acceleration, deceleration, and cruise. TABLE A.32. EXAMPLE FILE STRUCTURE OF DRIVING CYCLE FILES PRODUCED FROM MICROSIMULATION MODEL OUTPUT SPEED (mph) SECOND TIME LINK HOUR 60.31 1 4:00:00 I-805 Lane 1 1600-1700 60.70 2 4:00:01 I-805 Lane 1 1600-1700 61.00 3 4:00:02 I-805 Lane 1 1600-1700 60.57 4 4:00:03 I-805 Lane 1 1600-1700 60.58 5 4:00:04 I-805 Lane 1 1600-1700 61.42 6 4:00:05 I-805 Lane 1 1600-1700 62.31 7 4:00:06 I-805 Lane 1 1600-1700 62.91 8 4:00:07 I-805 Lane 1 1600-1700 62.54 9 4:00:08 I-805 Lane 1 1600-1700 61.91 10 4:00:09 I-805 Lane 1 1600-1700 62.01 11 4:00:10 I-805 Lane 1 1600-1700 61.66 12 4:00:11 I-805 Lane 1 1600-1700 61.87 13 4:00:12 I-805 Lane 1 1600-1700 61.08 14 4:00:13 I-805 Lane 1 1600-1700 60.20 15 4:00:14 I-805 Lane 1 1600-1700 58.85 16 4:00:15 I-805 Lane 1 1600-1700 58.44 17 4:00:16 I-805 Lane 1 1600-1700 59.05 18 4:00:17 I-805 Lane 1 1600-1700 60.16 19 4:00:18 I-805 Lane 1 1600-1700 62.98 20 4:00:19 I-805 Lane 1 1600-1700 62.95 21 4:00:20 I-805 Lane 1 1600-1700 63.48 22 4:00:21 I-805 Lane 1 1600-1700 62.51 23 4:00:22 I-805 Lane 1 1600-1700 61.83 24 4:00:23 I-805 Lane 1 1600-1700 61.44 25 4:00:24 I-805 Lane 1 1600-1700 60.23 26 4:00:25 I-805 Lane 1 1600-1700 59.88 27 4:00:26 I-805 Lane 1 1600-1700 60.15 28 4:00:27 I-805 Lane 1 1600-1700 60.33 29 4:00:28 I-805 Lane 1 1600-1700 61.55 30 4:00:29 I-805 Lane 1 1600-1700

198 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Yet another option is to select percentiles from the speed distribution for analysis using MOVES. For practitioners without microsimulation modeling capability, an FHWA con- tract is underway in 2010 that will produce example vehicle-specific power profiles for MOVES under various congestion conditions. When these vehicle-specific power profiles are released by FHWA, users will be able to use the ones that best match their specific roadway configuration and will be able to perform a more complete analysis of their transportation project than would otherwise be possible. Local Area Vehicle Travel Inputs for MOVES for Project-Level Analyses This section discusses the individual input tables needed to model a project-level analysis in MOVES. Links The use of the Project Level within MOVES requires a complete definition of the project. All individual roadway links and the off-network area must be specified by the user. This can be done by exporting a template for the Links table through the MOVES Project Data Manager, which will create a spreadsheet or text table that in- cludes all combinations of the linkID, countyID (only one county may be chosen for a given project-level run), zoneID, roadTypeID, linkLength, linkVolume, linkAvgSpeed, linkDescription, and linkAvgGrade that are needed for input to MOVES. The roadTypeID is required for each roadway link and can be chosen from the four avail- able road types (unrestricted or restricted urban or rural roads). The linkLength is the length in miles of each of the road links. The user must also specify link volume for each modeled roadway link. Link volume is the total average traffic flow from all vehicle types on the link during the period being modeled (for the project level, the period can only be the hour). The average speed and average road grade represent the overall average of the entire link. If driving schedules are not provided, MOVES will use the average speed and grade inputs and default MOVES driving cycles to do the calculation. However, if a link driving schedule is provided, then the average speed and grade will be obtained through that input. Link Driving Schedule The Link Drive Schedules Importer defines the precise speed and grade as a function of time (seconds) on each roadway link. Exporting a template for the Link Drive Sched- ules table through the MOVES Project Data Manager will create a spreadsheet or text table that includes all combinations of the LinkID, secondID, speed, and grade that are needed for input. The speed variable is entered in miles per hour and the grade vari- able in percentage grade (i.e., vertical distance or lateral distance; 100% grade equals a 45-degree slope). For each MOVES run, only one driving schedule can be input by the user. It is important to note that for a given roadway link, a user-supplied driving schedule will take precedence over an average link speed or grade input.

199 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Link Source Type The Link Source Type Importer describes the distribution of VMT by MOVES source (vehicle) type. The percentage of the total traffic on each link needs to be allocated to spe- cific source types. This can be done by exporting a template table through the MOVES Project Data Manager, which will create a spreadsheet or text table that includes all necessary combinations of the LinkID, SourceTypeID, and SourceTypeHourFraction. Operating Mode Distribution The Operating Mode Distribution Importer is used to import operating mode fraction data for source types, hour combinations, roadway links, and pollutant and process combinations. By exporting a template for the table through the MOVES Project Data Manager, a spreadsheet is created that includes all combinations of the SourceTypeID, HourDayID, LinkID, PolProcessID, OpModeID, and OpModeFraction that are needed for input to MOVES. Operating modes are modes of vehicle activity that have distinct emission rates. For example, running activity has modes that are distinguished by their vehicle specific power and instantaneous speed. Start activity modes are distinguished by the time the vehicle has been parked prior to the start (soak time). For a given source type, hour and day combination, roadway link, and pollutant and process combination, the operating mode distribution must sum to one. The Operating Mode Distribution Importer is required for the Project Data Manager when modeling any nonrunning emissions-producing process (such as idling or start processes). It is also required for modeling running emissions processes when either the Link Drive Schedules Importer is not used, or the link average speed input is not entered in the Links Importer. It is important to note that Operating Mode Importer data will take precedence over data entered in the Link Drive Schedules Importer and the Links Importer if conflicting data are entered. Off Network The Off Network Importer provides information about vehicles that are not driving on the links, but still contribute to the project emissions (for instance, when starting or idling). Exporting a template through the MOVES Project Data Manager will create a spreadsheet that includes all combinations of the sourceTypeID, vehiclePopulation, startFraction, extendedIdleFraction, and parkedVehicleFraction that are needed for input. For each source, vehicle population is the average number of off-network vehicles during the hour being modeled. The startFraction field, a number from zero to 1.0, specifies the fraction of this population that has a start operation in the given hour. The extendedIdleFraction field is also a number from zero to 1.0; it specifies the fraction of the population that has had an extended idle operation in the given hour. Finally, the parkedVehicleFraction field is a number from zero to 1.0 that specifies the fraction of the population that has been parked in the given hour.

200 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS Local Area Inputs for MOVES Not Related to Vehicle Travel This section describes the local area inputs to MOVES that are needed for all analysis types. Vehicle Age Distributions It is important that the vehicle age distribution data be representative of the local area for a GHG emissions inventory (as it is for other emissions inventories). The data under lying the CO2 emissions calculations differ by model year group. Thus, areas with newer vehicle fleets should have lower CO2 emission rates (on a gram per mile basis) than areas with older vehicle fleets. For areas that have developed MOBILE6 reg- istration distributions, EPA has provided a registration distribution converter that will take the MOBILE6-based registration data and format the data for use with MOVES. Otherwise, the MOVES County Data Manager can be used to create a template for preparing the age distribution data, and data from the state’s Department of Motor Vehicles registration database can then be used to populate the template. Fuel Data For the fuel supply and formulation inputs, the primary concerns are obtaining the correct mix of gasoline and diesel and identifying the share of ethanol in gasoline. The MOVES defaults for these values should be exported from the MOVES County Data Manager for the selected county. The resulting values should be evaluated based on what is known about the area’s fuel supply. Any necessary changes should be made to the values, and the resulting tables should then be imported via the MOVES County Data Manager. Meteorology Data and Inspection and Maintenance Program Data Temperature and humidity data are not important in developing a GHG emissions inventory. The user can export the MOVES default data for these parameters for the selected county to include in the MOVES GHG runs. The same is true of the inspection and maintenance program inputs. Source Type Population The total number of vehicles in the selected county or area is needed for each of the 13 MOVES source types. EPA’s MOVES technical guidance document explains how the source type population data can be developed, because this is a new input for MOVES that was not required for MOBILE6. Accounting for New Fuel Economy and/or Low-Carbon Fuel Standards Data in the original (December 2009 release) MOVES2010 model represented in-use fuel economies based on the federal fuel economy standards through the corporate aver age fuel economy (CAFE) standards that were updated as a result of the Energy Independence and Security Act (EISA) of 2007. An update of the MOVES model released in August 2010, MOVES2010a, includes GHG emissions or fuel economy values representative of those included in the April 2010 joint EPA–National Highway Traffic Safety Administration (NHTSA) rulemaking that affected the light-duty vehicle

201 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS (LDV) GHG emissions standards and CAFE standards for model year 2012 through 2016 vehicles. These standards were developed to make the national standards equiva- lent to the California standards for those model years. There may be times in the future when the most recent release of MOVES lags behind the adoption of federal fuel economy standards. There may also be situations in which states have adopted standards more stringent than federal standards (e.g., consistent with California) or in which planners wish to consider the effects of pro- posed, but not yet adopted, standards. There also may be situations in which state or regional low-carbon fuel standards are adopted or proposed that are not reflected in the MOVES emission rates. The way that MOVES handles fuel economy and CO2 emissions is fairly com- plex. No input or database of fuel economy or direct CO2 emissions are included in the MOVES database. Instead, the MOVES EmissionRate table in the default MOVES2010 database includes energy consumption rates that vary by operating mode, model year group, engine size, and vehicle weight. These rates are then con- verted to the CO2 emissions values output by MOVES. These energy consumption rates are based on tested in-use vehicle fuel consumption as opposed to EPA’s vehicle fuel economy rating or CAFE standards. The formula used within MOVES to convert energy per unit activity to fuel econ- omy is essentially fuel economy (unit activity/gal) = fuel density (g/gal) × (energy content (kJ/g)/emission rate (kJ/unit activity)) Data on fuel density and energy content for gasoline and diesel fuel are contained in the MOVES Fuel Type table. EPA is looking for ways to simplify the fuel consumption data within MOVES, but at present the complexity of the MOVES fuel consumption data means there is no easy way to directly model changes to fuel economy or GHG standards. However, EPA provides a work-around method for estimating changes in CO2 emissions result- ing from changes in fuel economy in Appendix F of the MOVES User Guide. Using this approach, the user would need to develop a table of the baseline MOVES fuel economy values compared with the fuel economy values of the scenario to be analyzed by model year and vehicle type. The values used for comparison should be expressed in terms of the equivalent CAFE standards (as opposed to the on-road fuel economy values). Based on the percentage reduction in the inverse of the fuel economy values, and factoring in the diesel vehicle market share for the model year of interest, the user would then develop a table to be modeled in MOVES via the Alternate Vehicle Fuels and Technologies strategy inputs. The percentage reduction in fuel economy is essen- tially modeled as an electric vehicle penetration value. Because electric vehicles are modeled assuming zero CO2 emissions in MOVES, the end result, if modeled correctly, should give a reduced CO2 emission value that would be the same as the value that would be calculated if a different fuel economy value were used. The basic equation used for calculating the Electric Vehicle fraction for a given model year is as follows: scenario fraction electric vehicle = 1– {(1/scenario fuel economy)/(1/MOVES baseline fuel economy)}

202 PRACTITIONERS GUIDE TO INCORPORATING GREENHOUSE GAS EMISSIONS INTO THE COLLABORATIVE DECISION-MAKING PROCESS It may be possible to adjust output CO2 emissions in a similar manner, without the need for making an additional MOVES model run. However, to do this, the output emissions would need to be at the model year level of detail. The analyst may also consider a simpler option, which involves adjusting MOVES CO2 emissions factors by the ratio of on-road LDV fuel economy with versus without the new standards in a given analysis year. Such information may be available from a regulatory impact assessment by a federal or state agency. It may be possible to inter- polate results for interim years not analyzed. Similarly, a low-carbon fuel standard could potentially be modeled in a simplistic fashion by reducing CO2 emission rates in proportion to the average reduction required by the standard (e.g., 10% in year X). This approach would maintain the ability to apply emissions factors specific to vehicle type, facility type, and speed to VMT forecasts, while adjusting overall emissions to be consistent with reductions expected from the fuel economy or carbon standard.