National Academies Press: OpenBook

Representing Freight in Air Quality and Greenhouse Gas Models (2010)

Chapter: Chapter 4 - Conceptual Model

« Previous: Chapter 3 - Evaluation of Current Methods
Page 116
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 116
Page 117
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 117
Page 118
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 118
Page 119
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 119
Page 120
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 120
Page 121
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 121
Page 122
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 122
Page 123
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 123
Page 124
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 124
Page 125
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 125
Page 126
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 126
Page 127
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 127
Page 128
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 128
Page 129
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 129
Page 130
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 130
Page 131
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 131
Page 132
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 132
Page 133
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 133
Page 134
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 134
Page 135
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 135
Page 136
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 136
Page 137
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 137
Page 138
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 138
Page 139
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 139
Page 140
Suggested Citation:"Chapter 4 - Conceptual Model." National Academies of Sciences, Engineering, and Medicine. 2010. Representing Freight in Air Quality and Greenhouse Gas Models. Washington, DC: The National Academies Press. doi: 10.17226/14407.
×
Page 140

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.

116 This chapter includes the development of a Conceptual Model for freight transportation activity as it relates to emis- sions calculations. The Conceptual Model offers a structure for comprehensive representation of freight activity in the United States, covering all modes and relationships between modes. In order for this model to be effective in improving emissions estimates, the Conceptual Model captures factors in freight movement and freight equipment that most influ- ence emissions. The Conceptual Model is a functional model, which includes the specifications of an information system in the form of functional areas, business processes, and information flows between them. The Conceptual Model does not include a formal data model, so it does not contain the description of all data elements that are necessary for all business processes. It does, however, identify the information needs from busi- ness processes. The Conceptual Model serves several purposes, as follows: • Estimates multimodal emissions associated with specific supply chains, transportation corridors, freight facilities, and geographic regions; • Assists shippers, carriers, and logistics providers in incor- porating emissions in the planning and operations of their logistics activities; • Assists public agencies in incorporating emissions in the planning of transportation infrastructure, transportation investment decisions, and development of transportation regulations and/or voluntary programs; • Identifies elements of freight activity that are not well represented by available data and methods; • Identifies how new and emerging freight data and methods relate to existing data and methods, and how they can pres- ent a comprehensive picture of freight movement; • Identifies opportunities to link modal-specific freight activ- ity data and tools in a unified framework that spans multiple modes and possibly geographic and temporal dimensions; • Identifies the major sources of potential errors in emission estimation parameters and the steps in emissions calcula- tions that warrant improvement; and • Tracks trends in freight emissions over time, and identifies which parameters were responsible for changes in emission outputs. 4.1 Model Overview and Uses The Conceptual Model provides the link between economic activity, freight transportation activity, freight-related emis- sions, and associated health effects. The Conceptual Model uses commodity flows derived from economic activity fore- casts to determine freight activity (Exhibit 4-1). Even though the Conceptual Model does not model dispersion of emissions or health effects, it provides the spatial and temporal allocation of emissions, which are necessary inputs for dispersion models and health risk assessments. The Conceptual Model is based on a “link and node” transportation network. The link and node framework is the basis for representing roadway networks in travel demand models, and it is also regularly applied to other modes (e.g., rail). The link and node concept provides an effective way to link different modes in one supply chain and to represent the intermodal connections and freight transloading that are common in urban freight systems but poorly represented in current modal-specific analyses. The Conceptual Model includes the definition of all functional areas and business processes necessary for the cal- culation, allocation, and evaluation of freight transportation- related emissions. Based on input parameters, the Conceptual Model includes a set of information flows (between business processes) that are needed to calculate freight transportation emissions. Some basic equations for emissions calculations are provided, but the Conceptual Model is not designed to replace existing emission models. Instead, it is designed to calculate and characterize transportation activity in such a way that it C H A P T E R 4 Conceptual Model

improves the accuracy of freight transportation emissions. The emission outputs are associated with either a product (or quantity of a given commodity), vehicle activity (e.g., VMT), freight activity (e.g., measured in ton-miles), link, node, or a geographic area. The Conceptual Model includes processes for the spatial and temporal allocation of emissions in order to support dispersion models and health risk assessments. Lastly, it includes processes for the evaluation of emissions, including scenario analysis and uncertainty analysis. The broad nature of the assignment warrants the use of an established methodology to ensure that the Conceptual Model will lead to the development of an actual model in the future. Information engineering allows a hierarchical and structured analysis of a business area (i.e., freight trans- portation emissions), and uses simple but comprehensive modeling/diagrammatic techniques. There are four main phases in information engineering, typically depicted in a pyramid (Exhibit 4-2). • Strategy planning is the phase that addresses how technol- ogy can be used to achieve specific goals such as creating new opportunities, creating a competitive advantage, or advancing environmental stewardship. Strategy planning creates a high-level overview of the information needs of the stakeholders and the system functional areas that will fulfill such needs. In this case, this entails a closer look at the information needs of the four main categories of stakehold- ers (private industry, transportation agencies, environmen- tal regulatory agencies, and environmental organizations), which are addressed in Section 4.3.1. Based on these needs, five types of applications are created as discussed in Sec- tion 4.3.2 (global/national, freight corridor, metropolitan, facility, and supply chain). To fulfill the objectives of these five types of applications, six functional areas are discussed in Section 4.3.3 (transportation network design, planning of transportation services, execution of transportation opera- tions, calculation of freight emissions, allocation of freight emissions, and evaluation of freight emissions). • Analysis is the phase that examines the business processes needed to run a functional area, how these processes inter- relate, and which input parameters are needed. Section 4.3.4 describes the business processes included in the Conceptual Model, and Section 4.3.5 provides a simplified process flow describing how information flows between processes. • System Design is the phase that determines how selected processes in the selected business area are implemented in procedures and how these procedures work. This phase is outside the scope of this project. • System Construction is the phase when the system is con- structed with the assistance of programming tools and code generators, and coupled with system design tools. This phase is not included in the scope of this project. 4.2 Freight Modeling The design of the Conceptual Model must take into consid- eration current developments in freight modeling. As presented in NCHRP Synthesis 364: Forecasting Metropolitan Commercial and Freight Travel, (185) good freight models should incorpo- rate the following attributes: • Ability to depict local characteristics: the model should consider the appropriate spatial resolution to capture the unique characteristics of the specific region for which it is 117 Ec on om ic A ct iv ity GDP Fr ei gh t A ct ivi ty Commodity flows Ton-miles TEUs En er gy O ut pu t Fuel consumption Hp-hr Em is si on s GHGs CAPs HAPs Co nc en tra tio ns H ea lth E ffe ct s CONCEPTUAL MODEL Exhibit 4-1. Conceptual framework. Plan Strategy Analysis Systems Design Systems Construction Conceptual Model Future Developments Exhibit 4-2. Four phases of information engineering.

developed, including the appropriate characterization of freight facilities and transportation links with their partic- ular restrictions to freight vehicles. • Link with national and regional databases and models: freight flows into, out of, or through the region of interest should be related to the “outside world.” The model should distinguish between external trips serving the region, inter- nal trips, and through trips. • Link to economic forecasts and trends: independently of the geographic scale of an application, consideration should be given to the interaction between freight flows, commod- ity flows, and economic forecasts. • Ability to consider technological changes: a freight model should be flexible enough to incorporate the effects of new technologies on logistics patterns. Freight modeling applications are typically divided by geo- graphic scale and application purpose. First, national-level applications are based on high-level national and international economic flows and trade activity, providing a framework for different geographic regions to assess overall freight activity generated by economic trends. Second, state-level or corridor applications are generally based on commodity flows, which are linked to economic activity and translated to freight flows by allocation to specific transportation modes. Third, metropolitan/urban applications tend to focus on truck movements modeled through traditional travel demand models, which are based on the four-step process. Finally, shipper/carrier applications focus on modeling goods distri- bution through tour-trip optimization to maximize delivery efficiency and minimize logistics costs. The Conceptual Model incorporates these four levels of application into a common framework, but it recognizes the inherit differences in objectives and analytical procedures to model freight activity. The Conceptual Model also considers a fifth application that models activity at a specific freight facility (e.g., port terminal, railyard, airport), which is an important application for the analysis of local emissions and air quality. 4.2.1 National/International Models The Freight Analysis Framework (FAF), developed by FHWA, is a commodity origin-destination database that esti- mates tonnage and value of goods carried by mode of trans- portation within 114 domestic and 7 international regions. Different methods are used to disaggregate interregional flows into flows among individual counties over specific transporta- tion facilities. These methods are based on geographic dis- tributions of economic activity rather than on accurate depictions of local conditions. 4.2.2 State/Intercity Corridor Models Statewide models, typically developed by state DOTs, rep- resent a variation of the four-step process used by MPOs to model transportation activity at the metropolitan level described in the following subsection. The main difference to MPO models is that their focus is on (multimodal) commod- ity flows rather than truck flows. The primary source of this information is the TRANSEARCH database and FAF, as well as other roadside data sources. Flows are assigned geograph- ically through the application of an economic input-output model that links commodity flows to land use and employ- ment activity in traffic analysis zones (TAZs). Most models then allocate the assigned commodity flows to modes by using estimates of vehicle payload. Methods for mode split, as well as treatment of non-truck modes, range widely in terms of levels of sophistication. 4.2.3 Metropolitan/Urban Models Very few MPOs attempt to do freight modeling, and most travel demand models are limited to truck movements derived from the four-step process. In the first step, trip generation, the models estimate trip production and consumption based on economic activity (i.e., land uses) by truck type and TAZ. The models also estimate external trips. In the second step, trip distribution, the model combines internal and external truck trips by truck type onto an origin-destination matrix. The third step, mode split, is not performed since only trucks are considered. Finally, in the fourth step, network assignment, the model assigns truck trips by truck type to highway links, usually at specific time-of-day periods. MPO models have three main pitfalls when applied to freight. First, there are concerns regarding whether truck activity can be modeled effectively without a direct link to the economic activity that is creating the demand for a particular commodity. Second, the standard trip generation/attraction methods are based on unique trip tours (one origin and one destination), but many truck shipments are based on multi- stop pick-ups/deliveries. Third, even though trucks account for the majority of urban freight movements, consideration should also be given to other modes of transportation. This is important in emission studies, given that the local air quality impacts of freight facilities can be substantial. 4.2.4 Shipper/Carrier Models Shippers, carriers, and logistics operators are responsible for the logistics of goods movements, and use processes and mod- els to manage these operations. These models, usually in the realm of the private industry, range from strategy-level models that handle supply chain design (e.g., facility location) to tacti- 118

cal models that design routes and multistop deliveries to oper- ational models that handle day-to-day operations such as car- rier selection. Commercial vehicles engaged in distribution operations typically travel in multistop tours, rather than a one origin–one destination trip. As a result, such movements are not well modeled by MPO models. Previous research suggests the following three core ideas about how logistics organizations should be handled in a good freight model: • Logistics organizations focus on total logistics costs (trans- portation and inventory) when making decisions on how to ship materials across the supply chain. As a result, models should also account for how changes in transportation pat- terns could affect inventory costs. • Logistics costs are heavily influenced by how supply chains are designed, especially by how facilities are located in com- parison with the locations of suppliers and final consumers. Therefore, a freight model should take supply chain design into consideration. • As shipment sizes decrease in exchange for increased fre- quency (to minimize inventory costs), carriers increasingly combine shipments in vehicles using cross-dock operations, use special routing software to optimize routes with multi- ple stops, and reduce empty equipment repositioning costs. EPA’s SmartWay Transport Partnership has developed a number of tools that are directly relevant to the private indus- try, including the Carrier FLEET Model, the Shipper/Logistics FLEET Model, and the DrayFLEET Model. The Carrier FLEET Model allows firms to estimate the environmental perfor- mance of their fleet using different technologies and opera- tional strategies. The Shipper/Logistics FLEET Model allows shippers to score their operations on the use of different oper- ational strategies, including the use of SmartWay carriers. The DrayFLEET model is focused on ports, providing the ability to measure the air emissions impacts of employing such strategies as container chassis pools, off-peak gate hours, etc. EPA also has developed the Diesel Emissions Quantifier to assist firms in estimating the benefits of retrofits. These tools tend to be limited in scope, focusing on single modes individually or on the assessment of individual emission reduction strategies. There is also a rapidly growing number of private software products intended to help businesses measure their carbon footprint, including from transportation operations and supply chains. Examples include Microsoft Dynamics AX’s Environmental Sustainability Dashboard, (186) CSRWare’s Enterprise Sustainability Management Platform, (187) Revo- lution ID’s Foundation Footprint, (188) and Enverity’s ghg- Track. (189) The main issue with these current tools is the use of simplistic methodologies to calculate fuel consumption that do not consider factors that might be relevant in the evalua- tion of freight transportation emissions. For example, emis- sions occurring at nodes are not always considered, nor are idling emissions. Calculated fuel consumption in these mod- els is generally not sensitive to changes in equipment payload, nor are other parameters such as terrain grade or congestion considered. Additionally, only CO2 emissions can be esti- mated since the software tools calculate emissions from fuel consumption. 4.2.5 Other Models, Methods, and Data Sources This section discusses alternative or developing models, methods, and data sources used to measure emissions from freight transportation. In contrast to the models described previously, these new or developing models represent freight emissions in novel ways. Because these models, methods, and data sources are currently in development, they are not explored in depth, but are briefly summarized. GIFT The Geospatial Intermodal Freight Transportation (GIFT) model, developed by researchers at the Rochester Institute of Technology, assists users in understanding the environmen- tal impacts of shipping routes and choosing routes that minimize fuel use and emissions. Compared to models dis- cussed previously that focus on modal activity, the GIFT model offers a more complete supply-chain analysis of freight movement. (190) Instead of using modal activity and emission factors, GIFT models a supply chain as a collection of links and nodes, in which each link represents a trip by specific mode, and each node represents freight handling locations, including railyards, intermodal centers, and warehouses. Each link and node incorporates properties of the selected mode, such as truck emissions and fuel economy, as well as properties of the trip, including grade, distance, and congestion. In addition, each link and node includes freight cost information. By analyzing this structure, model users can make an informed choice about freight shipment routes, depending on their shipping needs. For example, a shipper could use GIFT to select the most economical route, the “greenest” route, or a route that includes specific modes or waypoints. Alternatively, a shipper could apply weighted preferences to each characteristic to optimize the shipping route consider- ing all parameters. GIFT is currently in development with support from fed- eral and state governments, as well as private industry. 119

CARB Freight Efficiency Program In 2006, California enacted ambitious statewide greenhouse gas reduction goals, aiming to reduce state GHG emissions to 1990 levels by 2020, and 80% below 1990 levels by 2050. In response to these targets, CARB enacted a detailed scoping plan that determined the share of GHG reductions by sector and source, as well as mitigation programs to be used to reach each goal. To reduce freight emissions, CARB is implementing a Freight Transport Efficiency Measure to increase the fuel econ- omy and decrease the carbon footprint of freight modes. The goal of this measure is to reduce CO2 emissions by 3.5 million metric tons (equivalent) by 2020. (191) The CARB Freight Efficiency Program is patterned after EPA’s SmartWay Program, and is intended to reduce fuel economy by introducing system and technology improvement across modes. The program has identified several near-term opportunities for reducing fuel consumption, including vessel speed reduction; on-road and nonroad anti-idling, transport refrigeration unit programs, and freight truck efficiency initia- tives. As of September 2009, CARB is working with industry stakeholders to craft an implementation plan for the Freight Efficiency Program. NCFRP 12: Specifications for Freight Transportation Data Architecture The goal of NCFRP Project 12 (192) is to develop a structure for storing freight data from existing data sets and new data collections. Prior studies by TRB and the Cooperative Research Programs have identified challenges in applying freight data stored in disparate forms by several agencies, and the opportu- nities to improve freight analysis by uniting these data sources. The desired outcome of NCFRP 12 is to create a unified data architecture and evaluate the costs and benefits of implement- ing the structure in industry. The results of this project will complement the results of the NCFRP 12 report, and inform decisions about optimal storage methods for data sources iden- tified here. International Fuel Tax Agreement The International Fuel Tax Agreement (IFTA) is an agree- ment between the continental U.S. states and Canadian provinces to measure cross-border truck activity and appor- tion fuel tax revenues to the appropriate jurisdiction. Under this agreement, implemented by the International Fuel Tax Association (also referred to as IFTA) truckers submit quar- terly fuel tax reports, in which operators report the miles driven and fuel consumed in each state and province. These data are used to correctly apportion fuel tax revenue to each jurisdiction and to determine if operators are due a fuel tax surcharge or refund. Although the primary goal of IFTA is to correctly distrib- ute tax revenue, the reports collected by the association con- tain a wealth of information about truck activity, geographic distribution, and fuel consumption. Although these data are only collected for trucks involved in freight movement across the U.S.–Canadian border, they could be extrapolated to rep- resent the entire trucking industry. However, the information stored by IFTA is currently unavailable to freight researchers and practitioners. 4.3 Model Scope and Structure This section describes the scope and structure of a variety of Conceptual Model components. The discussion starts with a description of the target audience, whose needs drive the development of five applications (global/national, freight cor- ridor, metropolitan, facility, and supply chain) and functional areas. Business processes fulfill the requirements of the func- tional areas, and the process flows describe information flows between processes. 4.3.1 Target Audience The Conceptual Model targets four types of stakeholders, each with different needs: the private industry (shippers, carri- ers, and logistics providers), environmental regulatory agen- cies, transportation agencies, and environmental organizations (Exhibit 4-3). Private Industry The private freight transportation industry consists of man- ufacturers, carriers, and logistics providers, as well as others responsible for the storage and distribution of parts and fin- ished products. The private industry’s modus operandi has been to provide the right product at the right place at the right time at the lowest possible cost. Typically, consideration of environmental criteria has been related to compliance costs. In recent years, however, many firms have started to address envi- ronmental considerations to capture and keep new markets that are environmentally conscious, to fulfill the needs of cor- porate social responsibility requirements, and to address con- cerns of potential new regulations. Additionally, firms have realized that GHG emission reductions are often associated with cost reductions (because of the direct correlation between CO2 emissions and energy consumption), so they can develop leaner and more cost-effective supply chains while promoting environmental stewardship. Private firms will use the model to understand how choices in terms of supply chain design, facility location, mode choice, route choice, inventory levels, packaging, and delivery patterns affect the environmental performance of their supply chains. 120

They could also compare their operation’s performance against best-in-class performance through a benchmarking analysis. Environmental Regulatory Agencies Public agencies responsible for environmental regulations include U.S. state and local environmental regulatory agen- cies. Freight transportation emissions estimates can be pre- pared in response to federal or state regulations. These include the National Environmental Policy Act (NEPA) and similar state laws, the Clean Air Act, and federal conformity regula- tions. In other cases, freight emissions estimates are used in non-mandatory studies that serve to educate stakeholders and guide government programs or policy. At the federal level, EPA is responsible for setting criteria pollutant emission and ambient air quality standards. Most states follow these guide- lines, although others set their own standards (notably, Cali- fornia via the California Air Resources Board). Air quality districts set local/regional policy to meet federal and state air quality guidelines. In many applications, freight emissions are combined with emissions from other mobile and local sources to identify the net impact on local populations and, in the case of nonattainment areas, plan progress toward meeting air quality standards. Environmental regulatory agencies also sponsor studies of public health effects of air pollution, and many of these studies begin with estimates of emissions, including freight emissions. For example, the National Air Toxics Assessment (NATA) pro- duces screening-level estimates of cancer and non-cancer health effects of air toxics by census tract for the entire United States. Additionally, under the Clean Air Act, EPA is required to peri- odically review the National Ambient Air Quality Standards (NAAQS) and, if warranted, modify them to protect public health and welfare. This review typically includes an assessment of human exposure at various concentration thresholds, which is combined with results from epidemiological studies in the decision to modify the NAAQS. EPA also compiles nationwide GHG emissions in the official EPA GHG Inventory. (1) This national GHG inventory pro- vides a common and consistent mechanism for all nations to estimate emissions and compare the relative contribution of individual sources, gases, and nations to climate change. Com- plementary studies to the GHG inventory influence federal pro- grams that, in turn, leverage programs targeting the freight sector (e.g., EPA’s SmartWay program). Transportation Agencies Transportation agencies include metropolitan planning organizations (MPOs), as well as state and federal DOTs. Trans- portation agencies have a key role in influencing transportation emissions. Transportation infrastructure investment can result in traffic flow improvements (that typically reduce emissions), as well as in mode shifts due to capacity improvements in cer- tain modes. Transportation agencies also can enact finance mechanisms such as taxes, fees, and tolls that can have a direct influence on freight transportation behavior through policies and transportation infrastructure investments. Transportation agencies will be interested in analyzing the environmental performance of different transportation corri- dors to inform infrastructure investment decisions. The Con- ceptual Model provides a framework to analyze freight activity and emissions along potential goods movement corridors. Environmental Organizations Environmental organizations include those groups inter- ested in public health and environmental justice. These groups 121 Private Industry (shippers/carriers) Evaluate environmental performance of supply chains Environmental Regulatory Agencies (EPA/ARB) & Air Quality Districts Evaluate environmental implications of regulations/voluntary programs Environmental Organizations Evaluate health effects of freight system State/Local Transportation Agencies Evaluate environmental performance of transportation corridors and facilities Conceptual Model Exhibit 4-3. Stakeholders.

examine transportation decisions from a health impact per- spective and environmental justice framework. They tend to make sure existing environmental laws are upheld when new transportation investments are made so that public health is not adversely impacted or toxic hot spots are not created. As such, these groups might use the model to deter- mine the incremental emissions impact of a new transporta- tion project. 4.3.2 Model Applications Depending on the analysis objectives and available input parameters, the Conceptual Model allows emissions estima- tion for five different categories of analysis. Four of these are geographic scales and one describes a business enterprise per- spective, as shown in Exhibit 4-4. Global/National The objective of this application is to calculate freight emissions inventories associated with large geographic areas, typically at the state, national, or global level. Because this application considers all transportation facilities where freight moves or is transloaded, all freight modes are included. The main users for this application will be government agencies aiming to estimate and track freight emissions over time, as well as to compare the environmental performance of freight systems in different geographic regions. The model input will be generators of freight activity (i.e., commodity flows) from which vehicle activity can be esti- mated, or direct vehicle or freight activity if statistics of vehicle-miles traveled or freight ton-miles are available. Alter- natively, fuel consumption data can be used to estimate freight emissions of particular pollutants (especially GHGs) if there is a reasonable way to allocate them to freight sources. Outputs from this application include freight emissions associated with particular modes on a large geographic scale. In instances where this application is intended to provide the necessary inputs for air quality models, the spatial and tempo- ral allocation of emissions also will be required to properly characterize emissions released to the atmosphere. Freight Corridor This application calculates freight emissions from a trans- portation corridor, which could fall within a single state or across multiple state boundaries. Objectives of this application include the following: • Analyze current environmental performance of freight corridors; • Analyze how future capacity improvements could affect environmental performance of a corridor (this could include environmental improvements from congestion relief, as well as from mode shift due to investments in a given mode); • Identify corridors that are particularly energy efficient (possibly for benchmarking purposes, or as candidates for further investment) or inefficient (as candidates for future improvements); • Compare environmental performance of different corridors in order to understand the correlations between corridor capacity, commodity mix, mode share, and environmental performance; 122 Type of Analysis Objective Modes Audience Global/National Calculate freight emissions inventories associated with large geographic areas. All Environmental Regulatory Agencies Freight Corridor Calculate freight emissions associated with a specific corridor. Typically truck and rail Transportation Agencies Private Industry State/Local Environmental Agencies Metropolitan Calculate freight emissions inventories within a metropolitan area. Typically truck only, but other modes can be included Transportation Agencies Air Quality Districts Facility Calculate emissions from freight activity at a specific facility (truck terminal, railyard, port, and airport). Varies, depending on the facility Air Quality Districts Private Industry Environmental Organizations Supply Chain Calculate freight emissions associated with the logistics of a product. Varies, depending on the supply chain Private Industry Exhibit 4-4. Types of model applications.

• Analyze how different freight modes compare in terms of environmental performance on specific corridors; and • Analyze environmental effects of mode shift on specific corridors. Typically, a freight corridor application will evaluate land- based modes, particularly truck and rail. However, other modes also could be compared against truck and rail in some freight corridors, including inland waterways, short-sea ship- ping, and air freight. It is also possible to use this application for intercontinental sea routes. Potential users of this application include transportation agencies interested in investigating the environmental conse- quences of different types of infrastructure investments. The private industry also could use this application to evaluate the effects of specific route choices (e.g., Chicago to Los Angeles via I-80 or I-40) on their environmental performance. Route length, mode availability, terrain grade, and availability of backhaul traffic could all affect the environmental perfor- mance of a freight corridor. The model input data sources used to calculate the amount of freight activity will depend on data availability. Ideally, esti- mates of vehicle activity and commodity flows are both avail- able; otherwise vehicle payload needs to be assumed. This can be problematic since payload can vary widely, and it has a strong effect on emissions. Other important input parameters include fleet characteristics (e.g., model year, vehicle technol- ogy, engine power, equipment capacity, emission controls), and network characteristics (e.g., link capacity, node capacity, congestion levels). Outputs from this application include freight emissions associated with particular modes under different scenarios that can be characterized by commodity mix, mode share, traffic capacity by mode, traffic volumes by mode, link characteristics (e.g., pavement quality, electrified railways), fleet characteris- tics, and timeframe. In instances where this application is intended to provide the necessary inputs for dispersion models, the spatial and tempo- ral allocation of emissions also will be required to properly characterize emissions released to the atmosphere. Metropolitan This application calculates freight emissions inventories with temporal and spatial resolution within a metropolitan region with the following goals: • Analyze current and future environmental performance of the freight system within a metropolitan region; • Analyze how future expansion/improvements in trans- portation infrastructure could affect the environmental performance of a metropolitan region (this could include environmental improvements from congestion relief, as well as from mode shift due to investments in a given mode); • Compare environmental performance of different metro- politan regions, which would identify benchmarking regions, as well as those that are particularly inefficient regarding emissions from moving freight (this type of analysis also would examine the correlations between infrastructure capacity, traffic volumes, mode share, fleet characteristics, and environmental performance); • Analyze the impact of freight emissions on local air quality and human health; and • Compare freight emissions with emissions from other sectors. A metropolitan application will evaluate and geographically situate all freight modes that are within metropolitan bound- aries, including all classes of heavy-duty trucks, rail, marine, and other intermodal facilities, as well as airports. Potential users include local government agencies that will find value in this type of application for planning purposes in order to identify how future improvements in transporta- tion infrastructure and/or freight forecasts will influence freight emissions, as well as to compare a local region with regions in the rest of the country. Air quality districts can use this application to support air quality analyses and health risk assessments. Input parameters to determine freight activity will differ by mode. Trucking activity likely will come from travel demand models, and it is important to understand how such estimates are determined. As indicated in Chapter 3, methods to estimate trucking activity in travel demand models can be somewhat inaccurate. Rail-related activity can be obtained directly from local railroads, or estimated from published statistics. Freight activity in terminals typi- cally needs to be calculated separately with facility-level analyses. Examples include truck terminals, warehouses, railyards, ports, and airports. Because of the high uncer- tainty in some of these input parameters, it is important that some indication of uncertainty levels be included in the cal- culations, in order to identify which data elements warrant further improvement to make the calculations of metropol- itan freight estimates more accurate. Outputs from this application include freight emissions associated with different scenarios characterized by traffic vol- umes, infrastructure capacity, mode share, link characteristics, node characteristics, fleet characteristics, and timeframe. Because this application also is intended to provide the nec- essary inputs for dispersion models, the spatial and temporal allocation of emissions is important. These allocations are nec- essary to determine where and when emissions are released to the atmosphere. 123

Facility A facility-level application calculates freight emissions from freight facilities including truck terminals, railyards, marine and inland ports, and airports. The application has the follow- ing objectives: • Develop current and future emissions inventories associated with a freight facility; • Optimize facility environmental performance; • Analyze how future expansion/improvements in the facility could affect its environmental performance (this could include environmental improvements from congestion relief, as well as from mode shift due to investments in a given mode); • Evaluate effects of different regulations/initiatives on the emissions from a freight facility (e.g., extended idling restric- tions, fleet renewal programs, chassis pools, and mode shift); • Compare environmental performance of (comparable) freight facilities, which could identify benchmarking regions, as well as those that are particularly inefficient regarding emissions from freight handling (this type of analysis would also examine the correlations between environmental performance and infrastructure capacity, operational characteristics, traffic throughput, and fleet characteristics); and • Analyze the impact of a freight facility on local air quality and human health. Different modes will be included depending on the facil- ity. The analysis of trucking terminals will involve only trucks, but the evaluation of railyards can include rail, CHE, and trucks, since most rail terminals are connected to the rest of the freight system by roadways. Marine and inland terminals could include OGVs, harbor craft, CHE, trucks, and possibly rail if on-dock or off-dock rail terminals exist. The evaluation of airports will include air freight and CHE as well as trucks. Users of this scale will include regulators involved in per- mitting facilities, owners, and operators seeking permits or improvement in operations, local air agencies considering facility contributions to local air emissions, and environmental organizations concerned with public health and environ- mental justice. Input parameters for the following modes will depend on the facility and the level of resources available for data collection: • Trucking terminals are likely to have records of the number of trucks entering and leaving the facility. Estimates of load- ing and unloading times can provide an estimate of idling time, which needs to be evaluated in conjunction with whether anti-idling programs exist. Although it is unrealis- tic to expect a detailed evaluation of the fleet characteristics, an indication of the general truck size, as well as fleet age, will be necessary. For example, emissions can be quite different if the truck fleet is a long-distance fleet versus a drayage fleet. If trucking activity on the surrounding roads is included, traffic levels also need to be estimated in order to provide accurate emission estimates (because congestion can have a strong effect on local emissions). • Railyards analysis typically can rely on detailed information about locomotive activity, including fuel consumed by switch locomotives. More sophisticated analyses include information about the share of time in each notch setting, which is an important determinant of average emission fac- tors. Cargo handling equipment information is necessary, including number and type of equipment. In the case of intermodal rail terminals, there are drayage trucks accessing the railyard. In this case, the same input parameters described for trucking terminals also apply. • Marine and inland port terminals’ emission calculations rely on amount of cargo moved by cargo type. More sophisti- cated analyses include information on individual ship and harbor craft movements, engine type, engine model year, fuel used, and geographic port information to calculate emissions, as well as information on amount and type of CHE, hours of use, and duty cycle. Truck and rail servicing ports also need to be accounted for by determining the amount of cargo moved by each, as well as general truck and rail characteristics. Similar information described for truck- ing terminals and railyards apply to trucks and rail that service ports. • Airport operations analysis would use detailed information on the number of air cargo aircraft and the fraction of weight associated with cargo movement when aircraft oper- ate in mixed modes. More sophisticated analysis would include detailed information about each aircraft TIM (approach, landing, taxi, takeoff, and climb out), along with specifics on the aircraft type (jet, turboprop, and piston) and engine type. Ground support equipment used to ser- vice air cargo also needs to be accounted for—this would include information on the hours of use, duty cycle, and fuel type. Outputs from this application include freight emissions associated with different scenarios characterized by traffic throughput, operational characteristics (e.g., idling times), infrastructure capacity, equipment characteristics, and timeframe. Because this application also is intended to provide the nec- essary inputs for dispersion models, the spatial and temporal allocation of emissions is important. These allocations are necessary to determine where and when emissions are released to the atmosphere. 124

Supply Chain This application calculates the emissions associated with a specific supply chain, including the supply of materials to manufacturing/assembly facilities, and/or the distribution of intermediate or finished products to other facilities, storage locations, distribution centers, or consumers. As follows, this application will: • Calculate the emissions associated with all freight trans- portation required to manufacture and distribute a product; • Optimize routing for best environmental performance; and • Evaluate the effects of mode, route, and equipment choice on the environmental performance of the transportation components of a supply chain (however, this application will not evaluate emissions embedded in materials or those emissions associated with the actual manufacturing and assembly of products). The modes included in this application will depend on the specific supply chain, and can potentially include all modes of transportation. This type of application will be most useful to shippers, carriers, or logistics providers interested in evaluat- ing the environmental performance of their supply chains, and in understanding the effects of mode, route, and equip- ment choice on emissions. Input parameters will include sup- ply chain design, facility location, mode choice, route choice, inventory levels, packaging, delivery patterns, and equipment characteristics. Outputs from this application include freight emissions associated with the transportation necessary to manufacture and distribute a product under different scenarios. These sce- narios can be characterized by product type, supply chain con- figuration (location of suppliers, manufacturing/assembly facilities, storage locations, and consumers), mode choice, route choice, fleet characteristics, and timeframe. 4.3.3 Functional Areas The Conceptual Model is divided into functional areas to fulfill the objectives of the five types of applications described in the previous section. These functional areas enable a user to define the freight movement framework, calculate freight emissions, and evaluate freight emissions. Functional areas can be thought as general categories of modules in a system, under which business processes run. Exhibit 4-5 illustrates the six proposed functional areas. The first three functional areas—transportation network design, planning of transportation services, and execution of trans- portation operations—are part of the system description. These three functional areas allow the user to configure the network and enter the necessary input parameters to describe commodity activity, vehicle activity, and equipment configu- ration. The following two functional areas—calculation of freight emissions and allocation of freight emissions—use the system setup information to calculate emissions and allocate them to specific geographic areas and points in time. The last functional area—evaluation of freight emissions—enables the comparison of different scenarios, as well as sensitivity and uncertainty analyses, to improve freight emission estimates. These six functional areas are described in detail in the follow- ing subsections. Transportation Network Design This functional area consists of inputs describing the simu- lated transportation network. Freight transportation activity and associated emissions will be calculated on a transportation network, which will be based on a link-node system. Nodes will represent freight facilities, including trucking terminals, railyards, marine/inland ports, and airports. Nodes also can be virtual points dividing two continuous links with different characteristics. For example, two consecutive sections of the 125 Transportation Network Design Planning of Transportation Services Execution of Transportation Operations System Setup Calculation of Freight Emissions Allocation of Freight Emissions Calculations Evaluation of Freight Emissions Evaluation Exhibit 4-5. Functional areas.

same roadway with different traffic volumes can be divided by a virtual node (e.g., freeway exit, interchange). Links will rep- resent transportation facilities where freight moves, including roadways, railways, inland waterways, ocean routes, and air routes. Freight activity and emissions will be allocated to either a link or a node. The Conceptual Model enables the creation of flexible net- works with different levels of aggregation that can fit the objectives and accuracy requirements of an emission analysis. Although virtual nodes can be created to divide one link into shorter links with different characteristics, the opposite is also possible in the case of more aggregate analysis. Compar- ative analyses could also be made to evaluate the loss in accu- racy by increasing the level of aggregation when defining links and nodes. Because the Conceptual Model will be setup to assist users in incorporating environmental criteria in the design of a trans- portation network, the Conceptual Model enables the creation of alternate nodes and links to test future potential transporta- tion networks. Allowance for changes in node structure (e.g., additional nodes to simulate the effects of a new (or modified) distribution center) will enable the user to compare emissions between scenarios. As described in the following section, there are three busi- ness processes that fall under this functional area—supply chain design, link characterization, and node characterization. The specific attributes of links and nodes are described under link characterization and node characterization, respectively. Planning of Transportation Services This functional area configures the necessary input param- eters for the determination of freight flows over a specified transportation network, including the determination of com- modity flows, service levels (i.e., requirements in terms of tran- sit time, and transit time reliability), mode choice, and route choice. For those applications that rely on commodity flows as input parameters, input data can be obtained from published sources, such as the Commodity Flow Survey, (193) or by internal sources of freight transportation demand in the case of private firms. In the latter case, requirements for transit time and transit time reliability also will assist in the selection of mode. After mode selection, one or more routes will be chosen for the analysis. Other types of applications will not require the determina- tion of commodity flows and, instead, transportation activity will be determined directly from measured (or estimated) vehicle activity. For example, the analysis of freight emissions in a metropolitan area is to likely rely on travel demand mod- els to estimate truck activity on a local transportation network. In this case, mode choice and route choice will already be determined. Users will be able to create different scenarios to test the effects of changes in freight demand, service levels, mode choice, and route choice on associated supply chain emissions. Execution of Transportation Operations This functional area takes the perspective of day-to-day transportation operations, and it collects inputs for three business processes. First, the equipment configuration deter- mines which type of equipment will handle the freight flows. Required input parameters for equipment configuration include model year, vehicle technology, engine power, emis- sion controls, equipment capacity, and fuel type. All of these parameters are important for the determination of emission factors associated with a specific equipment type. Second, loading patterns will be determined based on the specified commodity and equipment configuration. Based on commodity density and packaging requirements, payload will be determined. Loading patterns will also define require- ments for loading and unloading times, which will assist in the estimation of idling or hotelling times. Finally, vehicle activity will be determined based on com- modity flows, mode choice, route choice, equipment config- uration, and loading patterns. Alternatively, vehicle activity can be provided as a direct input parameter to the model. Calculation of Freight Emissions This functional area is responsible for calculating freight emissions. Emission factors are determined based on equip- ment characteristics and how the network is configured. Emis- sions can either be determined from vehicle activity directly by using emission factors in terms of freight activity (ton-mile, hp-hr, hour), or from fuel consumption. In the latter case, fuel consumption either can be a direct input parameter to the model, or it can be estimated from freight activity. The functional unit of the analysis determines how freight activity is measured. Typical functional units are VMT, ton- mile, horsepower-hour (hp-hr), kilowatt-hour, and hour. For example, truck activity is typically measured in terms of VMT, but vessel activity is measured in kilowatt-hours. Allocation of Freight Emissions After freight emissions are calculated, they need to be allo- cated to either a node or a link (i.e., spatial allocation). This functional area groups calculated emissions spatially and tem- porally. Because links and nodes are associated with geographic regions, this will provide the necessary information for air quality models and health risk assessments. Additionally, emis- sions also need to be allocated to a specific time (i.e., temporal allocation) since this is also an important input parameter for air quality models. Emissions also can be allocated to a specific product or commodity. 126

Evaluation of Freight Emissions This functional area allows comparisons between a variety of emission scenarios calculated by the Conceptual Model. The model may be used to calculate emissions under a range of scenario alternatives that may be compared against a baseline or a benchmarking target, allowing alternatives to be differen- tiated based on a variety of input parameters. This func- tional area also allows the effects of emission reduction strategies to be analyzed by the Conceptual Model, including the strategies affecting emission factors, freight activity, fuel efficiency, and congestion. The ability to perform sensitivity analysis of specific parameters is important for evaluating and improving the performance of supply chains and testing the effectiveness of transportation policies. For example, freight emissions can be evaluated over time to examine emission changes based on economic forecasts (which drive commod- ity flows), mode share forecasts, and advancements in vehicle fleet technology. Scenarios also can be modified based on spe- cific input parameters, enabling sensitivity analyses. Thus, users can create different scenarios to test the effects of changes in the level of network aggregation, freight demand, service lev- els, mode choice, route choice, and equipment configuration. 4.3.4 Processes Each of the six functional areas described in the previous section are essentially aggregations of related processes. Each of these processes is responsible for specific activities required to fully describe a functional area and, eventually, for the calcula- tion and evaluation of freight emissions. Exhibit 4-6 summa- rizes the processes included in each functional area. Some of these processes will apply only to some types of applications. For example, for those analyses that rely on travel demand models to estimate truck activity over specific links, all processes under the planning of transportation service 127 • Supply chain design • Link characterization • Node characterization 1. Transportation Network Design • Determination of commodity flows • Determination of service level • Mode choice • Route choice 2. Planning of Transportation Services • Equipment configuration • Determination of loading patterns • Determination of vehicle activity 3. Execution of Transportation Operations • Determination of emission factor • Calculation of fuel consumption • Calculation of emissions 4. Calculation of Freight Emissions • Spatial allocation of emissions • Temporal allocation of emissions 5. Allocation of Freight Emissions • Analysis of alternative scenarios • Sensitivity analysis of freight emissions based on changes in input parameters 6. Evaluation of Freight Emissions PROCESSESFUNCTIONAL AREAS Exhibit 4-6. Processes.

functional area will not be required. Processes also will need to be adapted depending on the application because of differ- ent analysis objectives, input parameters, calculation meth- ods, and accuracy needs. For example, fuel consumption can be a direct input to the model (i.e., facility-level applications where fuel consumed is available for participating carriers), it can be calculated from vehicle activity based on equipment fuel efficiency, or it might be disregarded altogether if emis- sion factors are not based on fuel consumed. Objects All entities in the Conceptual Model are considered objects. Objects may be either calculated from other objects or are external input parameters. Higher-order objects are inputs to lower-order objects. For example, emissions are a first-order object and are the product of two second-order objects: freight activity and emission factors. Emission factors are produced from third-order objects such as equipment model year, and so on. The discussion of processes sometimes refers to objects as input parameters; the terms are regarded as interchangeable in this report. Exhibit 4-7 provides a list of some of the most important objects in the Conceptual Model. Supply Chain Design No objects are involved in this process. This process enables users to define the facilities included in a product supply chain, as well as the possible material flows between facilities. Facilities can be divided into the following two types: • Logistics facilities where products are processed and/or stored, including suppliers’ locations, manufacturing and assembly plants, warehouses, distribution centers, whole- salers, retailers, and final consumers; and • Transportation facilities, such as trucking terminals, rail- yards, intermodal facilities, ports, and airports. The objective of this process is to determine the set of nodes involved in the analysis of freight emissions, as well as the flows that will move between these nodes. This process is conceptu- ally simple, and it requires only the determination of possible 128 Variable Code Description Activity ACT Freight activity is a measure of vehicle activity, cargo activity, or fuel consumption. Activity Profile PRO Activity profiles represent driving cycles, duty cycles, or any other distribution of vehicle activities that has an effect on emission factors. Area ARE Combination of links and nodes. Commodity COM In analyses in which vehicle (or freight) activity is not an external input parameter to the model, commodity flows will determine vehicle activity. Each commodity group will be assigned with ranges of possible densities for different types of equipment, so that a commodity can be converted into number of vehicles. Emission Factor EF Determines the amount of a given pollutant emitted as a function of freight activity, which can be measured in vehicle-miles traveled, idling hours, ton-miles, energy, fuel, etc. Emissions E Product of freight activity and emission factors. Transportation Equipment EQP This includes the information necessary to characterize transportation equipment (or a fleet), including vehicle type, weight class, engine technology, fuel type, power ratings, model year, and emission control technologies. Link LNK A link represents transportation facilities where freight moves, including roadways, railways, inland waterways, ocean routes, and air routes. Mode MOD Trucking, rail, water, cargo handling equipment, air. Node NOD At the local/project level, nodes represent freight facilities, including trucking terminals, railyards, marine/inland ports, and airports. Nodes can also be virtual points dividing two continuous links with different characteristics. At the regional and national level, nodes can represent cities or regions. Payload PAY Payload represents the amount of cargo that can be loaded into transportation equipment. Payload can be measured in terms of weight or volume. Pollutant POL Emissions are reported separately by pollutant. Route RTE A route is a series of links and nodes. Because links are mode-specific, a route is responsible for linking multiple modes into a single supply chain. Scenario SCE Scenarios can be differentiated by a variety of parameters, including year, equipment type, route choices, commodity flows, payload, emission reduction strategies, etc. Time Period TIM Time period represents the point in time at which emissions occur. Exhibit 4-7. Main objects.

facilities as well as flows between facilities. There also might be flows within the same facility, which can include the operations of drayage trucks within an intermodal terminal, switch loco- motives within railyards, waterborne vessels maneuvering at port terminal facilities, or aircraft taxiing on runways. There is a mutual dependency between supply chain design and other processes. Both mode choice and route choice depend on an initial selection of logistics facilities, while the selection of transportation-related facilities depends on mode and route selection. Required inputs for this process include freight transportation demand. For outputs, this process will determine the level of service required for a supply chain, given consumer preferences (e.g., fashion-related products require fast deliveries), production requirements (e.g., just-in-time systems require specific and reliable transit times), and commodity characteristics (e.g., high-value commodities require fast transit times because of inventory costs). Node Characterization The following object is involved in this process: • Node (NOD). Nodes represent freight facilities, including trucking termi- nals, railyards, marine/inland ports, and airports. Nodes also can be virtual points dividing two continuous links with dif- ferent characteristics. At the regional and national level, nodes also can be cities or regions. Nodes need to be characterized not only because they are the source of freight-related emissions, but because they pro- vide the connectivity between links, thus influencing mode and route choice. Exhibit 4-8 presents the input parameters that will charac- terize a node, the transportation modes to which a parameter applies, and the purpose of a parameter (e.g., determination of emission factor). Link Characterization The following object is involved in this process: • Link (LNK). A link represents a transportation facility connecting two nodes. Examples of links are roadways, railways, water routes, and air travel lanes. Links must be characterized based on a series of parameters required to determine freight emissions along a transportation link. Exhibit 4-9 includes the input parameters that will characterize a link, the transportation modes to which a parameter applies, and the purpose of the parameter (e.g., determination of emission factor). Link characterization is dependent on mode choice, since not all modes will be present between two nodes. The 129 Parameter Description Mode Purpose Type A node can be a freight facility (where transportation operations occur), or a virtual node (that exists to connect two links). All N/A Link connectivity Determines which links are associated with a specific node. All Determine nodes associated with a trip Mode availability Determines which modes can be associated with a specific node. For example, a marine terminal with road access but no on-dock rail access will be associated with truck transportation but not rail transportation. Consequently, node characterization will have an influence on mode choice because nodes will only be associated with certain modes. All Determine mode choice Equipment availability Because there are freight transportation-related operations taking place at certain logistics and transportation facilities, those can be associated with specific types of transportation equipment. For example, marine terminals are specifically associated with cargo handling equipment that does not leave the terminal’s premises. Similarly, switch locomotives can operate strictly within a rail terminal, and ground support equipment is confined to an airport. As a result, node emissions will depend on the characteristics of these types of equipment. All Estimate freight activity at nodes and estimate emission factor Geographic area Associates a node with a geographic region, which can be defined as a city, county, air basin, metropolitan region, state, country, continent, or another region defined by the user. All Allocate emissions to physical locations Exhibit 4-8. Parameters for node characterization.

characteristics of different links also will influence route choice. For example, a longer route with smoother grades might be preferable to a shorter (but steeper) route. Determination of Commodity Flows The following object is involved in this process: • Commodity (COM). Commodity flows define the weight and volume of com- modities between different origin-destination (O-D) pairs. In global/national and supply chain applications, commodity flows are the main drivers of freight activity and, consequently, of emissions. In the freight corridor and facility applications, freight activity might be determined by either commodity flows or direct estimates of freight activity. This process is not applicable for the metropolitan application because vehicle activity is estimated directly from travel demand models. Commodity flows will be determined by either the supply chain design process (in the case of the evaluation of specific supply chains), or by economic activity forecasts (in the case of national/regional analyses). Commodity flows will influ- ence the following processes: • Mode choice: O-D pairs will influence mode choice because not all modes are available for all O-D pairs; • Determination of service level: commodity selection will influence service level requirements because of commodity value (e.g., high-value commodities will require faster tran- sit times in order to minimize inventory levels in transit); and • Activity: in the global/national and supply chain applica- tions, commodity flows will determine freight activity. Determination of Service Level No objects are involved in this process. Service level is generally described as a combination of travel time (e.g., 2-day delivery), travel time reliability (± 4 hours), and delivery frequency. This process is only applicable for the supply chain application, in which users can determine the required service level for a given supply chain. This process depends on the supply chain design process, as well as on freight transportation demand (input parameter). Service lev- els affect the following three processes: • Mode choice: service levels will influence mode choice because certain modes can provide faster and/or more reli- able transit times; 130 Parameter Description Mode Purpose Mode By definition, a link is mode-specific because link attributes are also mode-specific. All N/A Length Measured in miles. All Calculate freight activity Initial node All Provide link with trip End node All Link capacity Generally measured in vehicles per hour. Truck Estimate congestion and average speed Number of lanes/tracks Truck, Rail Determine link capacity Facility type Can be the roadway classification (for trucks) or track class (for rail). Truck, Rail Traffic volume Generally measured in vehicles per hour. Truck Estimate congestion and average speed Average speed Measured in miles per hour, average speed either can be an input parameter as in the case of travel demand models, or it can be estimated based on link capacity and traffic volumes. Truck Estimate emission factor Congestion Road level of service, varying from A to F. Truck Estimate emission factor Equipment restrictions Determines any type of equipment restriction on a link, including size and weight restrictions, and emission control systems. Truck, Rail, OGV Equipment mix If the typical fleet operating at a link has different characteristics from the area’s average, the user can determine a customized equipment mix for a link. Configure equipment Terrain grade Terrain grade is an important attribute of a link since it has a strong influence on fuel consumption and emissions. Estimate emission factor Geographic area Associates a link with a geographic region, which can be defined as a city, county, air basin, metropolitan region, state, country, continent, or another region defined by the user. All Allocate emissions to physical locations Exhibit 4-9. Parameters for link characterization.

• Route choice: service levels will influence route choice because some routes are shorter or faster; and • Determination of loading patterns: load sizes are usually determined by the frequency of deliveries. Mode Choice The following object is involved in this process: • Mode (MOD). Based on a given O-D pair, mode choice will be determined by mode availability, as well as other criteria (e.g., service level, travel time, travel distance, cost, and emissions). The Concep- tual Model assumes that a user will evaluate these parameters outside of the model. For the applications that can estimate vehicle activity from commodity flows—national/global, freight corridor, and sup- ply chain—more than one mode might be necessary for a given flow. For example, a corridor analysis between Chicago and Los Angeles could require the use of a double-stack train, plus a drayage truck movement on each end of the trip. Mode choice will determine the following processes: • Equipment configuration: mode selection will determine the different types of vehicles involved in the analysis; and • Route choice: mode choice also will have an influence in the selection of a route. Route Choice The following objects are involved in this process: • Route (RTE), • Link (LNK), and • Node (NOD). A route is a series of links and nodes. Because links are mode-specific, a route is responsible for linking multiple modes along a single supply chain. The selection of a route is important because a route is associated with travel distance, and other characteristics specific to the links and nodes repre- sented in a route (e.g., terrain grade, average speed, conges- tion). For a given O-D pair and mode, more than one route might be available from origin to destination. In such cases, a route will be determined based on travel distance, travel time, travel time reliability (which depends on congestion), and cost. This process applies for three types of applications: national/ global, freight corridor, and supply chain. The metropolitan application does not require this process because routes are determined within a travel demand model. Because the facility application analyzes emissions at a node, route choice is not necessary. Initially, the Conceptual Model does not include an algo- rithm to assist users in route choice based on selection criteria. Instead, the user needs to consider all relevant criteria for route choice, and simply assign a route in the model. Route choice will influence equipment configuration because different equipment types might be associated with different regions. Equipment Configuration The following objects are involved in this process: • Transportation Equipment (EQP), and • Payload (PAY). This process is the determination of equipment charac- teristics for a specific route or combination of routes (for regional/national analyses). Exhibit 4-10 includes the parame- ters necessary for equipment configuration by mode. Some of these parameters are necessary for the calculation of payload 131 Mode Parameter Heavy-Duty Trucks Model year, mileage accumulation, truck weight class, payload, truck weight, fuel type, engine power, vehicle technology, emission control technology, truck capacity (weight and volume), commodity types Rail Locomotive type, engine power, locomotive tier (emission control technology) Ocean-Going Vessels Calls, ship type, engine type, engine model year, propulsion and auxiliary engine power, ship size (DWT or TEUs), fuel type Harbor Craft Population by engine type, number of engines per vessel, engine power by type, deterioration factor, growth factor, engine age, median life, scrappage, use of retrofit devices, fuel type Cargo Handling Equipment Population, engine power, deterioration factor, growth factor, engine age, median life, scrappage, use of retrofit devices, fuel type Air Freight Engine type, fuel type, fraction of payload used for air cargo, aircraft type, fuel flow rates, aircraft performance (throttle setting) Exhibit 4-10. Parameters for equipment configuration by mode.

(e.g., truck capacity), while others are used for the estimation of appropriate emission factors (not all parameters listed are always necessary for the determination of emission factors). Equipment configuration depends on commodity type, mode choice, and sometimes on route choice because some regions might have restrictions regarding which types of equip- ment are permitted. Equipment configuration also is influ- enced by loading patterns, which will determine payload. Equipment configuration will determine the following busi- ness processes: • Determination of loading patterns: load capacity influ- ences loading patterns; • Determination of emission factor: emission factors are dependent on equipment type, model year, engine charac- teristics, and equipment weight; and • Determination of vehicle activity: load capacity and equip- ment utilization determine how many vehicles are neces- sary to transport a given load. Determination of Loading Patterns The following objects are involved in this process: • Payload (PAY), and • Commodity (COM). Loading patterns consist of how commodities are loaded onto the transportation equipment. Loading patterns depend on the service level—which will drive delivery frequencies— and equipment capacity. The determination of loading pat- terns is important because it will influence vehicle activity and payload, which in turn has an effect on emission factors. This process is required for the supply chain and facility applications because vehicle activity might be determined from commodity activity. The determination of loading patterns is required for those applications in which vehicle activity is determined from commodity activity. This is the case in the supply chain and facility applications, and it is sometimes true for the national/global and freight corridor applications. This process is not required for the metropolitan application because vehicle activity for that application is determined directly by travel demand models. Loading patterns will influence the equipment configura- tion process because it will determine the payload, and con- sequently vehicle weight. Additionally, some supply chains prioritize delivery frequency over equipment capacity maxi- mization (e.g., just-in-time systems). In these cases, the normal decision to optimize capacity might not be a good decision given the specifics of supply chain requirements. Determination of Freight Activity The following objects are involved in this process: • Activity (ACT), • Scenario (SCE), • Mode (MOD), • Transportation Equipment (EQP), • Link (LNK), Node (NOD), • Time (TIM), and • Activity Profile (PRO). This process consists of the determination of freight activ- ity, which can be measured in terms of vehicle activity (e.g., vehicle-miles traveled), product activity (e.g., ton-miles), or fuel consumption (e.g., total gallons of fuel per functional unit). Vehicle activity either can be calculated from commod- ity flows, or it can be an external input parameter from travel demand models. Fuel consumption either can be estimated from vehicle activity or provided to the model as an input parameter. For example, the calculation of GHG emissions and the analysis of rail emissions commonly rely on fuel con- sumption. Exhibit 4-11 provides examples of activity metrics specific to each mode of transportation. 132 Mode Activity Metrics Activity Profile Parameters Heavy-Duty Trucks VMT, idling time, ton-miles Driving cycle, level of service, average speed, bin allocation Rail Train-miles, car miles, idling time, ton-miles Duty cycle Ocean-Going Vessels Calls, propulsion power Load factors, vessel speed Harbor Craft Annual activity, fuel consumption Load factors by engine type, duty cycle Cargo Handling Equipment Load factor, activity Emission factor, duty cycle Air Freight TIM (cruise, approach, taxi/idle, takeoff, climb out) Throttle setting (aircraft performance),emission indices, fuel flow rate Exhibit 4-11. Vehicle and freight activity by mode.

Activity profiles characterize freight activity based on param- eters that affect energy consumption and/or emissions from an activity. Exhibit 4-11 also summarizes the parameters that describe activity profiles. This process will be handled differently depending on the type of analysis. For those analyses that rely on commodity activity to determine vehicle and freight activity, this process provides the necessary formulas to make the calculations. Other types of analyses will rely on direct estimates of vehicle and freight activity as input parameters. An additional type of analyses relies on direct fuel consumption estimates as input parameters, in which case estimates of vehicle or freight activ- ity will not be necessary. In addition to emission factors, freight activity will be the most important input in the calculation of emissions. Freight activity will be calculated separately by scenario, mode, activ- ity profile, transportation equipment, link/node, and time period. The specific formulas that will be used to calculate freight activity will depend on the specific type of analysis and the exact input parameters. Equations 22 through 25 provide some examples of calculations of freight activity at the link level. Calculating vehicle activity from commodity activity (e.g., vehicle-miles traveled) is performed as follows: Calculating product activity from commodity activity (e.g., ton-miles) is performed as follows: Calculating fuel consumption from commodity activity (e.g., gallons) is performed as follows: Calculating fuel consumption from vehicle activity is per- formed as follows: ACT Vehicle Activity SCE MOD PRO EQP LNK TIM, , , , , _ = SCE MOD PRO EQP LNK TIM EQPFuel Efficiency , , , , , ,_ PRO LNK, ( )Equation 25 ACT COM SCE MOD PRO EQP LNK TIM SCE MOD PRO E , , , , , , , , = QP LNK TIM LNK SCE MOD EQP Link Length PAY Fu , , , , _× × el EfficiencyEQP PRO LNK_ ( ) , , Equation 24 ACT COMSCE MOD PRO EQP LNK TIM SCE MOD PRO E, , , , , , , ,= QP LNK TIM LNKLink Length , , _ ( )× Equation 23 ACT COM SCE MOD PRO EQP LNK TIM SCE MOD PRO E , , , , , , , , = QP LNK TIM LNK SCE MOD EQP Link Length PAY , , , , _ ( × Equation 22) Since empty equipment activity will affect emissions, they also will need to be included and allocated to the load in the case of the analysis of specific supply chains. Determination of Emission Factors The following objects are involved in this process: • Emission Factor (EF), • Pollutant (POL), • Mode (MOD), • Transportation Equipment (EQP), • Activity Profile (PRO), and • Link (LNK). Emission factors determine the amount of a given pol- lutant emitted based on a given functional activity unit, which can be related to vehicles (e.g., VMT, vehicle-hours, energy), or related to freight (e.g., ton-miles). Emission fac- tors can account not only for fuel combustion, but also for the refining and distribution of fuel if a full fuel cycle analy- sis is desired. Alongside vehicle/freight activity (or fuel con- sumption), this process is the main input for emissions calculations. Emission factors will be determined separately by pollu- tant, mode, transportation equipment, activity profile, and link (in the case of emissions at a link). Depending on the mode, emission factors can be determined from emissions models or based on guidance documents, as summarized in Exhibit 4-12. The Conceptual Model does not replace pre- vious models that estimate emission factors or guidance documents. Instead, it relies on emission factors from these sources. Factors related to cleaner fuels or emission control retrofits also should be used to adjust emission factors where needed. 133 Exhibit 4-12. Source of emission factors by mode. Mode Source of Emission Factors Heavy-Duty Trucks MOVES, Mobile6 Rail EPA guidance Ocean-Going Vessels EPA guidance Harbor Craft ARB NONROAD or EPA OFFROAD models, other EPA guidance, other studies Cargo Handling Equipment ARB NONROAD or EPA OFFROAD models, other EPA guidance Air Freight ICAO emissions certification databank and fuel flow rates

Calculation of Emissions The following objects are involved in this process: • Emissions (E), • Scenario (SCE), • Mode (MOD), • Link (LNK), • Node (NOD), • Activity (ACT), and • Emission Factor (EF). Although some methods and models are mode-specific, there are standard methods that can be applied to all modes. As illustrated in the Equation 26, freight emissions are generally the product of freight activity (e.g., fuel consumed, energy gen- erated, or vehicle miles traveled), and emission factors (in grams of pollutant per measure of freight activity). Depending on data availability and the complexity of ana- lytical methods, emissions might be calculated separately by vehicle type or other factors that affect emission factors (e.g., average speed, road level of service), and added up to a total by pollutant. With the exception of GHGs, which are summed by multiplying their respective emissions by their global warming potential, the emissions of other pollutants are always reported separately. The calculation of emissions will provide information for the following processes: • Spatial allocation of emissions: emissions will be allocated to specific links, nodes, and geographic areas; and • Temporal allocation of emissions: emissions can be allo- cated to specific times during the day, days of the week, or months of the year. Emissions will be calculated for each pollutant, scenario, mode, link/node, and time period, as shown in Equations 27 and 28. Calculating mode emissions at a link is performed as follows: Calculating mode emissions at a node is performed as follows: E ACTSCE POL MOD NOD TIM SCE MOD PRO EQP NOD, , , , , , , ,= , , , , , , ( TIM PRO EQP MOD PRO EQP NOD POLEF ∑ × Equation 28) E ACTSCE POL MOD LNK TIM SCE MOD PRO EQP LNK, , , , , , , ,= , , , , , , ( TIM PRO EQP MOD PRO EQP LNK POLEF ∑ × Equation 27) Emissions Freight Activity Emission Factor= × (Equation 26) Spatial Allocation of Emissions The following objects are involved in this process: • Area (ARE), • Emissions (E), • Scenario (SCE), • Link (LNK), and • Node (NOD). Freight emissions will always be associated with specific links and nodes, which in turn are linked to geographic areas. As a result, freight emissions can always be allocated spatially to specified geographic areas, thus supporting dispersion models and health risk assessments. This process is only applicable for the metropolitan and facility applications because of their nar- row geographic scope. The user will be able to define different geographic areas, which are defined as a combination of links and nodes. A GIS interface also can be created to provide a visual representation of emissions. Emissions at an area are calculated as shown in Equation 29. Temporal Allocation of Emissions The following objects are involved in this process: • Time (TIM), • Emissions (E), • Scenario (SCE), • Link (LNK), and • Node (NOD). Freight emissions can be allocated to specific times during the day, days of the week, or months of the year in order to support dispersion models and health risk assessments. Because the dispersion of pollutants relies on variables that are time-dependent (e.g., temperature, winds), the tempo- ral allocation of emissions also is an input for dispersion models and health risk assessments. This process is applica- ble to any spatial scale for which air quality modeling might be applied. Emissions at an area at a given time are calculated as shown in Equation 30. E ESCE ARE TIM POL SCE POL MOD LNK TIM LNK MO , , , , , , , , = D SCE POL MOD NOD TIM NOD MOD E ∑ ∑+ , , , , , (Equation 30) E ESCE ARE POL SCE POL MOD LNK TIM LNK MOD TI , , , , , , , , = M SCE POL MOD NOD TIM NOD MOD TIM E ∑ ∑+ , , , , , , (Equation 29) 134

Model Calibration This process allows the calibration of the Conceptual Model based on results or input parameters from other studies or models. Invariably, there will be instances where surrogate input parameters will be used due to a lack of information about a given project, or a lack of resources to collect project- specific data. If input parameters from surrogate studies are available, they can be used directly in the Conceptual Model. If only final results are available, however, the Conceptual Model can be calibrated so that the final results can “match” the results from surrogate studies. The Model Calibration process will let the users select one or more input parameters that will need to be modified to enable the adjustment of final results. Analysis of Scenarios This process allows the creation of alternative scenarios that can be compared against a baseline or a benchmarking target. Scenarios can be differentiated based on any parameter in the model. For example, freight emissions can be evaluated over time to examine emission changes based on economic fore- casts (which drive commodity flows), mode share forecasts, and advancements in vehicle fleet technology. Scenarios also can be modified based on specific input parameters, which will enable sensitivity analyses. For example, users can create differ- ent scenarios to test the effects of changes in the level of network aggregation, freight demand, service levels, mode choice, route choice, and equipment configuration. The effects of emission reduction strategies also are captured by the Conceptual Model, including the strategies affecting emission factors, freight activ- ity, fuel efficiency, and congestion. The ability to perform sen- sitivity analysis of specific parameters is important to evaluate and improve the performance of supply chains and to test the effectiveness of transportation policies. The emissions associated with a mode in one scenario are calculated as shown in Equation 31. Subsequently, total emissions associated with one scenario are calculated as shown in Equation 32. Sensitivity/Uncertainty Analysis The evaluation of uncertainty associated with methods to estimate freight emissions needs to consider that error propa- E ESCE POL SCE POL MOD MOD , , , ( )= ∑ Equation 32 E E E SCE POL MOD SCE POL MOD LNK TIM LNK TIM , , , , , , , = + ∑ SCE POL MOD NOD TIM NOD TIM , , , , , ( )∑ Equation 31 gates as freight activity is converted into emissions, which are then used in air quality models and health risk assessments. Uncertainty in the emissions calculations can generally be attributed to either process or parameter uncertainty. Process uncertainty is taken to be the degree to which algorithms used in the calculations represent the actual emissions processes. These include uncertainties in the models themselves, as well as uncertainties in choices made during parameterization, such as choice of models and geographic boundaries. Param- eter uncertainty is the uncertainty in the individual elements of the equations utilized. This includes uncertainties in emis- sion factors, populations, activity, and other inputs. In cases of both process and parameter uncertainty, any known biases should be corrected before calculations are made; it is assumed here that any calculations will be made with the best available information and methods. However, unknown bias and uncertainty may still influence resulting estimates. In some cases, this may only be estimated qualita- tively. In others, quantitative estimates of uncertainty may be made. Particularly, if the uncertainty (for example, the stan- dard deviation, error, or other measure for various input parameters) is known, then a quantitative estimate of the resulting uncertainty can be made using standard error prop- agation methods. A full discussion of error propagation methods is available elsewhere. (194) Generally, overall uncertainty is derived from a Taylor’s Series expansion of the controlling equation, such that if emissions can be described by f(x1, x2, . . . xn), then the variance of emissions is as shown in Equation 33. Where σ2i represents the variance on variable i and σ2ij represents the covariance between variables i and j. In many cases, the fluctuations between two input variables are uncorrelated, such that the cross-terms average to zero. In that case, the error equation is simplified, as shown in Equa- tion 34. This equation may be used to approximate overall uncertainty in emissions from a quantified set of parameter uncertainties. Another method to estimate parameter uncertainty is the use of Monte Carlo simulation. By specifying probability distributions for selected input parameters, a Monte Carlo analysis simulates real-world conditions in order to assess σ σ σ2 1 2 2 1 2 2 2Emissions f x x f x = ∂ ∂ ⎛⎝⎜ ⎞⎠⎟ + ∂ ∂ ⎛⎝⎜ ⎞⎠⎟ x2 + . . . ( )Equation 34 σ σ σ2 1 2 2 1 2 2 2emissions f x x f x = ∂ ∂ ⎛⎝⎜ ⎞⎠⎟ + ∂ ∂ ⎛⎝⎜ ⎞⎠⎟ x f x f x x x 2 1 2 2 1 22 + + ∂ ∂ ⎛⎝⎜ ⎞⎠⎟ ∂ ∂ ⎛⎝⎜ ⎞⎠⎟ + . . . . . . (σ Equation 33) 135

the uncertainty in emissions outputs. The biggest challenge remains in the selection of the most influential parameters and the determination of their probability distributions. Lit- erature research, data availability, and expert judgment can be used. It is important to emphasize that an uncertainty assessment does not make emission outputs more accurate. However, probabilistic simulation models (e.g., Crystal Ball) can determine the contribution of each parameter to the final outcome. Based on that information, priority can be given to find more reliable sources of data for those parameters, and suggest the use of ranges, instead of point estimates, for results. 4.3.5 Process Flows Process flows, or the way data and calculations flow into and between analytical process steps, will vary depending on the type of application. Some of these processes can apply to all types of applications, including equipment configuration, determination of freight activity and emission factors, calcu- lation of emissions, scenario analysis, and uncertainty analy- sis. However, other processes do not apply to all applications. Exhibit 4-13 summarizes how each process applies to the five types of applications. Variations among the applications are described in the following subsections. Global/National This application calculates freight emission inventories associated with geographic areas at the state, national, or global level. Supply chain design is not relevant because the application does not intend to model a specific supply chain. The level of link and node characterization will need to be commensurate with the level of detail and accuracy required by the analysis. Because freight activity will be determined from commodity flows, the processes regarding commodity flows, mode choice, and route choice are required. The deter- mination of service levels however, is not applicable because of the aggregate nature of the analysis (i.e., at an aggregate level, it is not possible to determine requirements such as transit times and delivery frequencies). All of the subsequent processes are necessary, including equipment configuration, 136 Facility Type Global/National Corridor Metropolitan Facility SupplyChain Supply Chain Design Link Characterization Node Characterization Determination of Commodity Flows Determination of Service Level Mode Choice Route Choice Equipment Configuration Determination of Loading Patterns Determination of Freight Activity Calculation of Fuel Consumption Determination of Emission Factors Calculation of Emissions Spatial Allocation of Emissions Temporal Allocation of Emissions Analysis of Scenarios Uncertainty Analysis Mandatory Applicable Not Applicable Key: indicates that a parameter is analyzed in the way denoted by the column: indicates that the parameter is not discussed in the way denoted by the column. Exhibit 4-13. Relationship between processes and applications.

determination of loading patterns (to calculate payload), freight activity, and emission factors, as is the calculation of fuel consumption (if emissions are calculated from fuel con- sumption) and emissions. The spatial and temporal alloca- tion of emissions is not relevant for this type of application because the input parameters do not offer the appropriate level of detail to support dispersion modeling. Freight Corridor This application calculates freight emissions from a trans- portation corridor, which can fall within one jurisdiction (state), or cross multiple jurisdictional boundaries. Supply chain design is not relevant because the application does not intend to model a specific supply chain. Link and node charac- terization are critical because links and nodes along a corridor can have unique characteristics in terms of capacity, traffic vol- umes, congestion levels, and grade. Because freight activity can be determined from commodity flows, the processes regarding commodity flows, mode choice, and route choice are all rele- vant to this application. The determination of service levels also can be relevant because of the logistics requirements from different commodity types (e.g., higher-value commodities demand faster transit times). As in other applications, the subsequent processes are required, including equipment con- figuration, determination of loading patterns (to calculate pay- load), freight activity, and emission factors, as is the calculation of fuel consumption (if emissions are calculated from fuel con- sumption) and emissions. The spatial and temporal allocation of emissions is not relevant for this type of application because the input parameters do not offer the appropriate level of detail to support dispersion modeling. Metropolitan This application calculates freight emissions inventories within a metropolitan region. Supply chain design is not rel- evant because the application does not intend to model a specific supply chain. Link and node characterization are important because links and nodes within a metropolitan region can have unique characteristics that affect emis- sions. Because vehicle activity is provided as an external input parameter, the processes regarding commodity flows, service level, mode choice, route choice, and loading patterns are not relevant to this application. As in other applications, the sub- sequent processes are required, including equipment con- figuration, freight activity, and emission factors, as is the calculation of fuel consumption (if emissions are calculated from fuel consumption) and emissions. Spatial and temporal allocation of emissions can be relevant for this type of appli- cation because the input parameters can offer the appropriate level of detail to support dispersion modeling. Facility This application calculates freight emissions from freight facilities including truck terminals, railyards, marine and inland ports, and airports. Supply chain design is not rele- vant because the application does not intend to model a spe- cific supply chain. Node characterization is possibly one of the most important processes given that the analysis is asso- ciated with a node itself. Link characterization can be used if the system boundaries associated with the analysis include surrounding transportation links (or if links within the facil- ity can be identified). The determination of service levels is not relevant to this application because mode choice is more a function of infrastructure availability. If freight activity is determined from commodity flows, the processes regarding commodity flows, service level, and mode choice are relevant to this application. Route choice is generally not relevant because the analysis is done at a facility level. As in other applications, the subsequent processes are required, includ- ing equipment configuration, freight activity, and emission factors, as is the calculation of fuel consumption (if emis- sions are calculated from fuel consumption) and emissions. Spatial and temporal allocation of emissions can be relevant for this type of application because the input parameters can offer the appropriate level of detail to support dispersion modeling. Supply Chain This application calculates the emissions associated with a specific supply chain. Supply chain design is required to deter- mine the location of the relevant facilities involved in the sup- ply chain. The level of link and node characterization will need to be commensurate with the level of detail and accuracy required by the analysis. Because freight activity will be deter- mined from commodity flows, the processes regarding com- modity flows, service levels, mode choice, and route choice are required. All of the subsequent processes are necessary, including equipment configuration, determination of loading patterns (to calculate payload), freight activity, and emission factors, as is calculation of fuel consumption (if emissions are calculated from fuel consumption) and emissions. The spatial and temporal allocation of emissions is not relevant for this type of application because the effects of an individual supply chain are not likely to have significant local impacts. 4.4 Case Study This section presents a case study that illustrates a possible application of the Conceptual Model. The case study involves the comparison of different supply chain configurations for importing products from Asia to Chicago. 137

Many product supply chains—from automotive to retail— rely on imports of parts or finished products from Asia. These shipments are typically consolidated before reaching an Asian outbound marine port, then shipped to an inbound marine port in North America. Most ocean containers are then either transloaded directly onto double-stack trains, or deconsoli- dated at transloading facilities, where shipments are trans- ferred to trucks for final delivery. In this specific case study, the goal is to quantify emissions associated with transporting 100 lbs of product X from Shanghai to Chicago via three supply chains: ocean/rail via Long Beach, ocean/truck via Seattle, and ocean/rail via Prince Rupert, BC. Other objectives of the analysis are as follows to: • Assist in incorporating emissions in the planning and oper- ations of logistics activities, • Identify which parameters are responsible for changes in emission outputs (e.g., facility location, mode choice, route choice, equipment configuration), • Track trends in freight emissions over time, and • Compare company performance against best-in-class through a benchmarking analysis. The most likely audience for this type of analysis will be manufacturers sourcing raw materials, parts, or finished prod- ucts from Asia. The results of the analysis are likely to be one of the criteria for designing or modifying a supply chain, given that other considerations such as economics and reliability also need to be taken into account. Input parameters include facility location, shipment charac- teristics, mode choice, route choice, inventory levels, packaging, delivery patterns, equipment characteristics, and timeframe. Outputs from this analysis include freight emissions associ- ated with the transportation necessary to manufacture and distribute product X under different scenarios in each of the three supply chains. All objects described in Exhibit 4-7 will be used in this analysis. The following sections define the processes required for this analysis. Supply Chain Design Users need to define the logistics facilities involved in a product supply chain, as well as the product flows between these facilities. In this case study, the following supply chains will be considered: • Shanghai to Chicago via Long Beach, with double-stack intermodal service from Los Angeles to Chicago; • Shanghai to Chicago via Seattle, with trucking service from Seattle to Chicago; and • Shanghai to Chicago via Port of Prince Rupert, with double- stack intermodal service from Port of Prince Rupert (PPR) to Chicago. Exhibit 4-14 illustrates the logistics facilities (nodes) and the product flows between facilities. For freight transportation demand, it can be assumed that calculations will be based on a product that weighs 100 lbs and weighs out. It also will be assumed that the user has enough volume to fill an entire ocean container. Because the functional unit for this analysis is one product and the modes are already pre-selected, the processes for deter- 138 Supply Chain Logistics Facilities/Nodes Product Flows Long Beach Port of Shanghai Port of Long Beach (POLB) Intermodal facility in Los Angeles Intermodal facility in Chicago Port of Shanghai to POLB (ocean) POLB to intermodal facility in Los Angeles (rail) Intermodal facility in Los Angeles to intermodal facility in Chicago (rail) Seattle Port of Shanghai Port of Seattle Trucking distribution center in Seattle Trucking distribution center in Chicago Port of Shanghai to Port of Seattle (ocean) Port of Seattle to trucking distribution center in Seattle (drayage truck) Trucking distribution center in Seattle to trucking distribution center in Chicago (long-distance truck) PPR Port of Shanghai Port of Prince Rupert (PPR) Intermodal facility in Chicago Port of Shanghai to PPR (ocean) PPR to intermodal facility in Chicago (rail) Exhibit 4-14. Logistics facilities and flows by supply chain.

mination of commodity flows, determination of service level, and mode choice are not required for this analysis. Node Characterization Nodes represent freight facilities, including trucking ter- minals, railyards, and marine/inland ports. The Exhibit 4-15 characterizes all nodes included in this analysis. For the sim- plest analysis, all nodes can be characterized as freight facilities (i.e., no virtual nodes). However, virtual nodes can be used to separate links on the same route with different activity profiles (e.g., road grade, rail grade, congestion levels). Nodes will not be characterized in terms of equipment availability (due to lack of detailed information from a shipper’s perspective) and geo- graphic area (because shippers are not interested in that type of information). Link Characterization A link is a transportation facility connecting two nodes. In this analysis, the links considered will be the following: • Ocean routes from the Port of Shanghai to the ports of Long Beach, Seattle, and Prince Rupert; • Alameda (rail) corridor between the Port of Long Beach to a rail intermodal terminal in downtown Los Angeles; • Rail corridor between a rail intermodal terminal in down- town Los Angeles to a rail intermodal terminal in Chicago; • Rail corridor between PPR and a rail intermodal terminal in Chicago; • Truck corridor between the Port of Seattle and a trucking distribution center in Chicago; • Truck corridor between trucking distribution centers in Seattle and Chicago. Depending on the level of detail required for the analysis, these corridors can be broken down in multiple sublinks to reflect different operational characteristics of different ocean, rail, and road sections. For example, ocean routes can be bro- ken down depending on the ships’ activity profiles: cruise, speed reduction zones, and maneuvering (hotelling emis- sions should be associated with a node). Truck and rail routes can be subdivided into multiple links, if detailed informa- tion about capacity, grade, average speed, and congestion level are available. All links need to be characterized as out- lined in Exhibit 4-9. Equipment Configuration This process consists of the determination of equipment characteristics for all routes included in this analysis. The fol- lowing equipment types should be characterized based on the parameters included in Exhibit 4-10: OGVs, double-stack trains, drayage trucks, and long-distance trucks. Depending on the level of sophistication of the analysis, users can either rely on industry defaults for vehicles or they can customize to the specific vehicles they utilize. For example, if a firm is a SmartWay partner, they might choose to configure a long- distance truck that has a better-than-average rating for fuel efficiency due to the use of aerodynamic devices. Determination of Loading Patterns The main importance of this process is to determine the payload associated with each type of equipment on each link. This will determine the share of vehicle emissions that need to be allocated to the product. Because the product in question weighs out, the equipment utilization (payload as a share of total weight capacity) needs to be determined. For example, if the capacity of a truck trailer is 80,000 lbs, the user can assume that a truck would carry 72,000 lbs (i.e., 90% utiliza- tion), and that 1/720 of total vehicle emissions would be allo- cated to a product that weighs 100 lbs. Determination of Freight Activity Freight activity can be calculated separately by scenario, mode, activity profile, transportation equipment, link/node, 139 Node LinkConnectivity Mode Availability Equipment Availability Geographic Area Port of Shanghai (POS) SEA, PPR, LBE Ocean, truck, rail N/A N/A Port of Long Beach (LBE) POS, LA_INT Ocean, truck, rail N/A N/A Port of Prince Rupert (PPR) CHI_INT Ocean, rail N/A N/A Intermodal facility in Los Angeles (LA_INT) LBE, CHI_INT Rail, truck N/A N/A Intermodal facility in Chicago (CHI_INT) PPR, LA_INT Rail, truck N/A N/A Trucking distribution terminal in Seattle (SEA_TRK) SEA, CHI_TRK Truck N/A N/A Trucking distribution terminal in Chicago (CHI_TRK) SEA_TRK Truck N/A N/A Exhibit 4-15. Parameters for node characterization.

and time period. The specific formulas that will be used to cal- culate freight activity will depend on the type of analysis and the exact input parameters. The following provide some exam- ples of calculations of freight activity at the link level. • Intermodal rail service: rail activity can be initially measured in ton-miles of revenue freight and then converted into fuel consumption. In this example, the product weighs 100 lbs, the rail link is 50 miles long, and rail activity will be equal to 100 × 50/2000 = 2.5 ton-miles. Rail activity in ton-miles will be divided by a fuel efficiency factor (ton-miles/gallons) that is representative of the rail link and equipment in question to determine the fuel consumption allocated to the product on that specific link. In this example, the fuel consumed to transport this load on this link will be 2.5 ton-miles/400 ton- miles/gallons = 0.00625 gallons. • Drayage and long-distance trucks: truck activity can be measured in VMT on each link allocated to the specific product. For example, if a product weighs 100 lbs, the link is 50 miles long, and the amount that can be loaded onto a truck is 72,000 lbs (90% of 80,000 lbs), the VMT allocated to this product on this link will be 100 lbs × 50 miles/ 72,000 lbs/vehicle = 0.0694 VMT. Since empty equipment activity will affect emissions, they will also need to be included and allocated to the load. Determination of Emission Factor The determination of emission factors needs to be commen- surate with the level of detail required by the analysis. In the ACT COM SCE MOD PRO EQP LNK TIM SCE MOD PRO E , , , , , , , , = QP LNK TIM LNK SCE MOD EQP Link Length Pay , , , , _ ( × Equation 36) ACT COM SCE MOD PRO EQP LNK TIM SCE MOD PRO E , , , , , , , , = QP LNK TIM LNK EQ Link Length Fuel Efficiency , , _ _ × P PRO LNK, , ( )Equation 35 simplest analysis, a user can rely on default emission factors by mode independently of the vehicle activity profile. For exam- ple, a single emission factor can be used for an entire ocean, rail, or truck route. For more sophisticated analyses, emission factors can be determined separately by transportation equip- ment, activity profile, and link. For example, different emission factors will be determined for different ship types for the fol- lowing operational modes: cruise, reduced speed zone, maneu- ver, and hotelling. Calculation of Emissions As previously indicated, freight emissions are generally the product of freight activity (e.g., fuel consumed, energy gener- ated, or VMT), and emission factors (in grams of pollutant per measure of freight activity). Emissions will be calculated for each pollutant, scenario, mode, link/node, and time period, as shown previously in Equations 27 and 28. This analysis does not involve the spatial or temporal alloca- tion of emissions. Model Calibration It is possible that the user might have information from car- riers (on fuel consumption, for example), which will enable the application of user-specific fuel efficiency factors instead of model defaults. Analysis of Scenarios Scenarios can be differentiated based on any parameter in the model. For example, freight emissions can be evaluated over time to examine emission changes based on changes in facility locations, production outputs, and service levels, as well as mode choice and/or equipment decisions. Sensitivity analyses can be performed to evaluate the effects of given parameters on emissions, and this can assist users in their decision-making process. The emissions associated with a mode in one scenario are calculated as shown previously in Equation 31. Subsequently, total emissions associated with one scenario are calculated as shown previously in Equation 32. 140

Next: Chapter 5 - Recommended Research Areas »
Representing Freight in Air Quality and Greenhouse Gas Models Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB’s National Freight Cooperative Research Program (NFCRP) Report 4: Representing Freight in Air Quality and Greenhouse Gas Models explores the current methods used to generate air emissions information from all freight transportation activities and their suitability for purposes such as health and climate risk assessments, prioritization of emission reduction activities, and public education.

The report highlights the state of the practice, and potential gaps, strengths, and limitations of current emissions data estimates and methods. The report also examines a conceptual model that offers a comprehensive representation of freight activity by all transportation modes and relationships between modes.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!