City Logistics Research: A Transatlantic Perspective (2013)

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.

77 APPENDIX A: COMMISSIONED WHITE PAPERS Modeling Approaches to Address Urban Freight’s Challenges A Comparison of the United States and Europe Michael Browne, University of Westminster, London, United Kingdom Anne V. Goodchild, University of Washington, Seattle, Washington, USA 1. IntroductIon The rise in urbanization at a global level has reinforced the need to understand complex city growth patterns and rapidly changing urban systems. These urban envi- ronments present special challenges to the movement of people and goods. The flow of freight is essential to the growth and functioning of cities but also contributes to problems such as congestion, air pollution, and degrada- tion of the urban environment. Researchers bring insight to these challenges through their work. Analytical models support a better under- standing of urban freight and constitute an important tool in addressing these problems. This paper identifies the problems that urban freight research aims to address in Europe and the United States; provides a better under- standing of existing data, analytical tools, and methods; and lays out some gaps and challenges in addressing these problems with existing resources. The past 10 years have seen a significant increase in research activity regarding issues of urban freight. While passenger travel has been well studied for some time, goods movement presents a different and arguably more complex set of challenges than personal travel. This is in part due to the great diversity of products, business models, and actors involved. The increased research activity is illustrated by the number of researchers work- ing in the field, the increase in freight-specific confer- ences and seminars, and the increase in the publication of papers. The United States funded the National Coopera- tive Freight Research Program (NCFRP), and the Euro- pean Union research and demonstration programs have funded many studies and pilot projects concerned with urban freight. The recent announcement by the Volvo Research and Education Foundations (VREF) that they will support two new centers of excellence focused on urban freight and sustainability issues is also significant. This increase in academic and research activity has been mirrored by the greater interest shown by policy makers at municipal, regional, and national levels (and, indeed, the international level in terms of the European Union’s European Commission). This also extends beyond agen- cies responsible for transportation to other sectors, such as the environment (e.g., the U.S. Environmental Protec- tion Agency’s Smartway program), and economic devel- opment. The private sector has also seen the opportunity and need to address urban freight issues in a more coher- ent and active manner, which has lead to many private- sector initiatives happening in cities around the world, both at the enterprise level (e.g., Coca-Cola’s sustain- ability initiatives and DHL’s program of city logistics work) and through organizations such as the Forum for the Future. Research developments have been supported by researchers becoming more actively engaged in interna- tional research networks that provide increasing oppor- tunities to compare and contrast developments and challenges between different regions; the U.S.-EU com- parative study syposium for which this paper was com- missioned is an example. This paper addresses the use of models to analyze urban freight problems. The term “model” is used to describe a tool, which includes a system of mathema- tical relationships that help explain a system, study the

78 city logistics research: a transatlantic perspective effects of different components, and make predictions about travel or travel behavior. While there are many urban freight problems of interest and concern, such as the safety of pedestrians around freight vehicles or noise irritation, discussion is confined to pressing problems most amenable to modeling approaches, including air pollution and congestion. The recent attention to the issues of freight trans- portation has brought new interest in freight modeling research. Currently, significant development is under way to improve and implement freight models for a vari- ety of applications. It is not the intent of this paper to capture these most recent developments but, rather, to describe the tools currently used to support analysis and decision making. Discussion of models is limited to the following: 1. Models applied at the urban or metropolitan spa- tial scale. While models exist on the multistate, national, and international scales, this paper examines models that assist with urban-scale freight transportation challenges. 2. Models designed to model road freight. Modal split will not be considered because the overwhelming major- ity of urban regions rely solely on road vehicles for dis- tribution of freight. 3. Freight-specific models or freight-specific compo- nents of models. 4. Transport models, rather than economic models such as input–output or computable general equilibrium models. Finally, models for planning applications are considered. This inclusion implies strategic models rather than real- time operational models. Enterprise-level tools, such as those used for facility location, vehicle assignment, routing, and scheduling, are important elements of a firm’s logistics planning. These models often use optimization or heuristic algo- rithms to find the solution that minimizes–maximizes or greatly reduces–increases the desired objective. While some insights can be gleaned from knowledge of these tools—for example, where firms may locate ware- houses and distribution centers or which routes they may use—these models are not used to analyze urban freight systems. The choice was made to consider the research issues of congestion, energy, and air quality because they are important research themes that are receiving attention in both the European Union and the United States. From an initial review of 110 academic papers, Lammgård and Hagberg (2013) made a detailed review of 76 papers published since 2000 to identify the key research themes for those papers regarding work on urban freight and city logistics topics. The most important research topics identified were the following: 1. Congestion, 2. Emissions, and 3. Safety (although safety lagged behind the first two topics in terms of the number of papers that address the topic). The rest of the paper is structured as follows: Section 2 discusses research on congestion and urban accessibility; Section 3 examines research in energy use, greenhouse gas (GHG) emissions, and air quality. Section 4 considers the availability of data to support analysis, and Section 5 presents the strengths and weaknesses of the research methods presented. Section 6 outlines the authors’ view of the challenges to urban freight analytical research, and Section 7 outlines some of the differences between U.S. and European approaches. The paper concludes with recommendations in Section 8. 2. congestIon And urbAn AccessIbIlIty Moving goods and people around urban environments is a great challenge. Congested streets cause delay to travelers and goods, making travel times unreliable and scheduling difficult. Congestion increases fuel con- sumption and emissions for the same amount of travel on an uncongested network. Lack of parking, loading, and unloading places causes vehicles to park illegally and leads to a disruption of traffic flow and decreasing capacity. Transportation planning and a range of trans- port policies have been implemented in an attempt to address these difficulties. For example, vehicle weight and access time restrictions and congestion charging schemes attempt to reduce or shift demand. Analysis and modeling are often used to support planning and policy evaluation through estimation of future traffic flows and a better understanding of travel and travel behavior. Travel demand forecasting is used to predict future traffic conditions within a variety of operating, infrastructure, and demand scenarios. In addition, traf- fic modeling work has explored ways in which to redis- tribute flows. 2.1 Topics in Freight Congestion Research Despite these actions, congestion remains a major chal- lenge. Freight transport activity contributes to this con- gestion through vehicle volumes, but also through traffic disruptions from loading and unloading. Freight vehicles also suffer delays resulting from general urban traffic congestion, and this has direct impacts on their costs and the efficiency of their operations. Research on this topic has addressed the following:

79appendix a Browne and Goodchild 1. The scope to reschedule deliveries (off-hours, out of hours, or night delivery); 2. Ways to allocate, manage, or regulate use of curb space; 3. Pricing and charging: charging vehicles for enter- ing zones within the city or the city itself and shifting demand through parking pricing; 4. Controlling or altering land use; and 5. Using regional consolidation centers. 2.1.1 Rescheduling Deliveries One way to make better use of road capacity is to deliver outside normal working hours (referred to in the United States as off-hours deliveries and in the United Kingdom as out of hours delivery, perhaps partly to avoid the use of the term “night delivery,” which has negative conno- tations of noise and disturbance for residents). However, it has proved very difficult to change deliveries to off- peak hours. From a societal and environmental perspec- tive, making vehicle deliveries and collection journeys at noncongested times can help reduce the contribution of freight transport to traffic congestion; potentially reduce fuel consumption and pollutant emissions; and improve safety due to fewer goods vehicle operations at times when most pedestrians, cyclists, and other vulnerable road users are on the roads. There are many research questions examining the scope for rescheduling deliveries, and a wide variety of research approaches have been adopted. For example, work in the Netherlands has concentrated on research into technological improvements for delivery opera- tions to reduce noise (the PIEK Program: Ainge et al. 2007; NICHES, n.d.; Klaasse et al. 2002); whereas, in the United Kingdom, most research has been in the form of pilot projects and small-scale demonstrations. Lamm- gård and Browne (2012) have summarized a range of research studies carried out on the opportunity for time shifting in EU countries. In many instances, researchers model existing flows and then consider the implications of changing the time of day of an operation. The result- ing costs and benefits can then be calculated. The behav- ior of businesses that ship and receive freight affects the ability to gain a temporal shift in freight movements. One may find that while carriers are willing to shift operations to off-peak hours, users of the service may not want to staff their facilities during these times just to ship or receive freight. Some benefits have already been mentioned, but examples of additional costs may be more expensive labor for nighttime working and the imposition of wider costs for additional noise. The most widely reported U.S. research on this topic took place in New York (see Holguín-Veras et al. 2011). This research built on an extensive series of surveys followed by a number of trials that have received considerable atten- tion beyond the academic community; see, for example, a recent reference in TIME Magazine (Sanburn 2013). To address the costs and benefits, the pilot studies have been supported with extensive modeling exercises; see, for example, Holguín-Veras et al. (2007, 2008, 2010). 2.1.2 Allocating, Managing, or Regulating Curb Space Congestion at the curbside is also important from an urban freight perspective. Freight transport opera- tors suffer from congestion at the curbside when try- ing to make deliveries and collections at busy times. Urban freight surveys show that demand for curb space is intense on some of the busiest urban streets during peak times (e.g., see Cherrett et al. 2012; NICHES, n.d.; Transport for London 2009). Many deliveries take place from the curbside because retail stores and offices do not have separate off-street loading bays that can be used by carriers. Research into the use of curb space and how to optimize its use is limited. Surveys provide insights, but only limited modeling work appears to have been under- taken to consider the consequences of changing the curb- side loading regime on both freight operations and the wider general traffic flows. Traffic models can be used to illustrate the impact of double-parking on loading and unloading operations. 2.1.3 Pricing and Charging A few EU cities have instigated road user charging (e.g., London, Stockholm, and Gothenburg). In addition, some cities apply tolls to parts of the network, such as bridge crossings that may be important in access trips to, from, and within cities (e.g., New York City). The Ports of Los Angeles and Long Beach charge a traffic mitiga- tion fee between 3:00 a.m. and 6:00 p.m. to address congestion, security, and air quality. With the revenue, the ports established five new shifts per week outside of these hours. The consequences for freight transport have been considered, but research is limited. The research typically shows that freight carriers do not respond to urban road pricing by changing the times at which they operate (Golob and Regan 2000; Hensher and Puckett 2005). Instead, they either absorb the charge or seek to pass the additional cost on to their customers. Whether they can pass the costs on will depend upon relative bar- gaining or negotiating strengths between the carrier and their customer. Urban freight operations may be carried out by small companies in a weak bargaining position with customers, and in these circumstances they likely will simply carry on operating as before and be forced

80 city logistics research: a transatlantic perspective to absorb the extra costs. The design of the charging system may also affect whether carriers are able or will- ing to respond by changing their behavior. For example flat-rate charges for cordon-based systems may have dif- ferent implications compared with charges that vary by road type, location, and time of day. Supply-chain deci- sion making is clearly a complicated mix between carrier and shipper–receiver, and the consequences may be quite varied among different urban supply chains. 2.1.4 Controlling or Altering Land Use Urban freight movements are strongly influenced by land use patterns, and many of the trends in recent years have forced urban distribution and storage activ- ity further from the metropolitan centers as land use values and changes of use work against having tradi- tional stockholding and transshipment points within the more central areas of the city. At a detailed level, specific land use patterns will influence trip generation and attraction rates. These topics have been the sub- ject of research on logistics sprawl in Paris and Atlanta (e.g., Dablanc and Rakotonarivo 2010; Dablanc and Ross 2012). This sprawl means that vehicles run more mileage in urban areas, thereby contributing to and being affected by congestion to a greater extent than in the past. Land use trip generation studies have been carried out by many researchers (e.g., see Debauche 2008; Fischer and Han 2001; Hunt et al. 2006; Jessup et al. 2004; Kriger et al. 2007; Lau 1995; Lawson and Strathman 2002; Patier and Routhier 2008; Shimuzu et al. 2008). However, land use patterns and urban freight have a complicated relationship. For example, retail land use is heterogeneous, leading to very differ- ent trip patterns (number of trips, average consignment size, time of delivery) for specific types of retailers, such as grocery compared with fashion. The research at the Laboratoire d’Economie des Transports (LET) based on the FRETURB model and urban surveys has been of particular interest here (see Routhier and Toilier 2007; Ambrosini et al. 2010; Patier and Routhier 2009). Ownership patterns can also be influential. Studies have shown that independent retail outlets often receive many more deliveries per 100 square meters of floor- space than the branches of retail chains, which reflects multiple retailers’ greater control of the upstream sup- ply chain (Anderson et al. 2005; Cherrett et al. 2012). So to understand the congestion implications of different types of land use, a detailed picture is required. These detailed issues could be modeled, but the underpinning data are often weak, particularly those relating to logis- tics systems and supply chain structures; therefore, it can be very difficult to incorporate these features into freight models. 2.1.5 Using Regional Consolidation Centers Commercial supply chains (including those serving urban areas) have evolved to become more efficient and responsive to customer requirements. This has allowed transport companies (carriers) and third-party logistics companies to group (or consolidate) flows to improve truck utilization and reduce costs for movement in the supply chain. Policy makers have become interested in the scope for consolidation activity to influence conges- tion within cities. Clearly reducing the number of vehi- cles required to achieve a given level of service should reduce miles traveled and have beneficial impacts. But it is possible that supply chains may be optimized and yet the miles run within the city (especially in the center) may not be reduced by this optimization; hence, there has been continued research interest, particularly in Europe, in the opportunity for area-based consolidation, in which operators would concentrate the flows of traf- fic for a given location through a small number of well- located terminals (e.g., see Allen et al. 2012; Browne et al. 2011; Gonzalez-Feliu and Morana 2010; van Duin et al. 2010). To understand the impact of such centers it is necessary to model the impact of the changes, while tak- ing account both of the new flows that arise and of what happens to the changed trip patterns of companies that now deliver to a consolidation point instead of direct to the final customer. In addition, with the rising interest in the issue of clean vehicles (see Section 3), many studies attempt to combine an understanding of changes in miles traveled with an assessment of the implications for new types of vehicles, such as electrically powered vans and cargo cycles. The use of models is an important way of understanding the potential impact of such schemes and systems, but the modeling is challenging because changes in the behavior of carriers and the way in which a sup- ply chain responds is often poorly understood. Thus, although a carrier may now deliver to a consolidation point for one customer or even several customers (offer- ing an opportunity for a reduction in urban miles trav- eled), they may continue to deliver into the city center for other customers. This scenario could mean that a con- solidation center might have the effect of increasing miles traveled in the city rather than reducing them, although this requires further investigation. 2.2 Forecasting Models Many of the previously mentioned research approaches rely on the ability to forecast demand for freight trans- portation under future scenarios. The models used for this forecasting are called travel demand forecasting models. The nature of these models and their character- istics are briefly described here.

81appendix a Browne and Goodchild Urban forecasting models that account for trucks are relatively common in large urban areas in the United States, with many of the modeling programs operated by metropolitan planning organizations (MPOs). A report published by the Transportation Research Board of the National Academies (TRB), Special Report 288: Metro- politan Travel Forecasting: Current Practice and Future Direction (2007), surveyed MPOs about travel modeling and noted the following: “Truck trips are modeled in some fashion by about half of small and medium MPOs and almost 80% of large MPOs.” This percentage has certainly increased since the survey, and there is particu- lar emphasis since the adoption of MAP-21 transporta- tion legislation in the United States. While many MPOs have some accommodation for trucks in their modeling efforts, they do not adequately explain the impacts on trucks from land use patterns. An overview of the state of truck modeling is provi- ded in TRB’s NCHRP Synthesis 384: Forecasting Metro- politan Commercial and Freight Travel (Kuzmyak 2008). This report identified urban goods movement forecasting methods used in professional practice (pri- marily in the United States) and completed a survey of organizations with active urban goods movement mode- ling programs. The report provided supplementary case studies highlighting more innovative goods movement forecasting methods and approaches. NCHRP 384 noted that almost all metropolitan planning organizations and urban areas that model goods movement are actually forecasting trucks using an adaptation of the traditional four-step process common in passenger forecasting. The four-step process estimates trip productions and attrac- tions, matches these productions and attractions into origin–destination pairs, assigns trips between origin– destination pairs to modes, and then selects routes for each set of trips (Virginia Department of Transporta- tion, n.d.). Following are the four steps (Kuzmyak 2008) adapted for truck forecasting; while some locations are considering improvements to this approach, the authors have yet to see a novel framework being consistently used at the metropolitan level for policy analysis. 1. Trip generation. For trucks trip generation is usu- ally an estimate of production or consumption linked to the economic activity represented within zones. Truck trips between internal locations or between locations external and internal to the study area can be factored in at this point. A number of studies have found the linkage between several land use variables (specifically employ- ment) and truck trips to be weak, and better data are needed (Fischer and Han 2001). 2. Trip distribution. Truck data are often integrated into the overall model during this step by the use of a zone-to-zone trip table (origin–destination matrix), which accounts for truck travel between zones. For a truck model, the external and internal trips are added, and flows are often sorted by truck size or type. This process creates a correspondence between actual and forecast link counts. Validating this step requires truck classification counts and survey data. 3. Mode choice. The mode choice step is not com- monly used for urban goods movement models because most goods move on trucks, and freight rail, shipping, and pipelines are not usually included. Mode choice could be used to select the type or size of trucks used, but this is not often done in practice. 4. Trip assignment. All vehicles, including passenger vehicles and trucks, are assigned by type or class to the roadway network, typically using shortest-path or low- est-cost travel times, often by time of day. The network is typically a limited set of the roadways (e.g., residential streets may be excluded). Reviews of these adapted four-step truck models reveal that they do not work well in dense urban environments. Comi et al. (2012) provide a review of the state of urban freight transport demand modeling (primarily in Europe), taking a broad view, and including strategic, tactical, and operative models. They categorize models into four types: truck based, commodity based, tour based, and mixed, and discuss their advantages and disadvantages. They categorize four-step models as truck based. One significant limitation is that the four-step process fails to account for the trip and tour (chaining) behavior of truck activity in urban areas,1 thus creating the existence of a separate class of tour-based models. Four-step models that fail to address multiple-stop tours cannot capture the number of, or the routes associated with, this type of travel. For example, this type of model may do a poor job of capturing the impact of the growth in large con- solidation and distribution centers and their impact on the pattern of urban truck travel (Kuzmyak 2008; Don- nelly et al. 2010). These four-step models are also not capable of account- ing for the impacts of truck parking or the impact other transportation modes have on truck routing choices. In addition, they rely on fixed vehicle trip rates by land use type and industry sector. This paper does not consider commodity-based mod- els because such models are generally applied at a larger- than-metropolitan scale due to data characteristics. There are several alternative approaches to the four- step, commodity, and tour-based models. These include activity-based models, simulation, and agent-based approaches, among others (Abdelgawad et al. 2011; Andreoli and Goodchild 2012; Samimi et al. 2012). TRB’s Second Strategic Highway Research Program 1 In Europe the term “multidrop round” or “multistop round” may be used for these more complicated chained trips.

82 city logistics research: a transatlantic perspective held a workshop in 2012 on freight demand modeling and data improvement. Their report summarizes many state-of-the-art research approaches (Chase et al. 2013). Tavasszy (2008) provides an earlier overview of interna- tional approaches. Activity-based models, which can be considered an extension of trip-based or tour-based models, use a demand-based approach. Unlike the traditional four- step model that uses single trips as the basic modeling step, these models forecast flow based on travel demand derived from activities that people (or goods) need to perform. Travel is based on the activities to be comple- ted and modeled in tours. Activity models may offer a more effective approach to modeling because trips made by trucks are not independent of each other and can be connected for efficiency or convenience (PB Consult Inc. 2007). In their review, Comi et al. (2012) point out that in the current urban freight demand modeling literature, the relation between policies–measures and stakehold- ers’ behaviors is not sufficiently represented; in particu- lar, urban-scale shopping trips and the siting of freight centers–platforms and shopping centers. Within current models, industrial employment or land use data are typi- cally used to estimate truck trip generation. Increasingly, however, goods deliveries are being made to residential locations, which has the effect of substituting freight vehicles for passenger vehicles on the last link in the sup- ply chain. This change in retail activity in urban areas and the increase in home shopping has been the subject of much urban freight research, including Feliu et al. (2012), who provide a review of recent trends in urban goods movement and suggest a framework for modeling the changes due to e-commerce and home delivery. Oth- ers ask whether these new models present an increase or decrease in vehicle miles traveled (VMT) and emissions, and how travel demand models should be restructured to reflect this change; for example, see Wygonik and Good- child (2012). 3. energy use, ghg eMIssIons, And AIr pollutIon The contribution of transport to GHG emissions has been widely recognized. Within the European Union and the United States there have been many research stud- ies to look at opportunities to reduce energy consump- tion in transport, some at the level of specific operating strategies for individual fleets (Wygonik and Goodchild 2012). At the urban freight level, the scope to reduce energy consumption has been featured in research (e.g., Kanaroglou and Buliung 2008; Sorrell et al. 2009; Yan- nisa et al. 2006; Zanni and Bristow 2010; Figliozzi 2010). At the EU level, attention has been increased by the EU white paper on transport (European Commission 2011). A number of challenging goals were set, including the aim of achieving essentially carbon dioxide (CO2)- free city logistics in major urban centers by 2030. The white paper makes the point that achieving essentially CO2-free city logistics would also substantially reduce other harmful emissions (see discussion below about air quality issues). London provides a clear example of the importance of transport within total CO2 emissions. Ground-based transport is responsible for 22% of total CO2 emissions in London, and within that freight transport by trucks and vans accounts for 23% (Allen et al. 2010). Among 18 European city regions reported in the GRIP Project (Carney et al. 2009), transport emissions accounted for 66% of total emissions in Oslo (the highest among the 18 regions), and for only 7% in Rotterdam (the lowest), reflecting the very different patterns of energy use influ- enced by the varied city economies. The proportion of CO2 accounted for by freight transport was not shown. In a U.S. context, the city of Boulder, Colorado, reported that transport accounted for 21% of CO2 (Boulder 2010), while Seattle reported that transport represented 71% of core CO2 split between 41% passenger transport and 30% freight. However, this calculation included residential and commercial CO2 from buildings but not CO2 from industry or the port (see Lazarus et al. 2011). Research into energy consumption and emissions reduction in urban freight has focused on several ques- tions, including the following: • How large is the scope to use alternative fuels (e.g., electric vehicles)? • Can new organizational approaches play an impor- tant role in minimizing freight transport demand (e.g., improving vehicle utilization and consolidating flows)? • What are the impacts of vehicle performance improvements and changes to behavior (such as driver training), and how can these be achieved? It is important to note that the research examining energy use in supply chains since this includes an assessment of energy use (and CO2 emissions) in freight transport including transport in urban areas. Much of this research adopts a life-cycle approach (e.g., Böge 1995; Browne et al. 2005; Jespersen 2004; McKinnon 2010; Rizet et al. 2012). The life-cycle assessment (LCA) approach is concerned with tracking a product from origin to consumption (and beyond, including waste and recy- cling). At each stage calculations are made about the energy used. The research is usually underpinned by data collection from surveys. However, to extend the research to consider (a) total energy use and (b) the impact of diffe- rent supply chain configurations, it becomes necessary to scale up the findings from the surveys. This may require

83appendix a Browne and Goodchild modeling of typical goods vehicle flows and assump- tions about the energy consumption for different types of vehicles that are making different types of trips (e.g., trips in urban areas). Emissions factors can be obtained from environmental agencies such as the U.S. Environmental Protection Agency’s MOVES model. The LCA approach can be complicated by the need to make decisions about the boundary of the system to be considered—for ex- ample, does it include the energy used in manufacture of the equipment that made the machinery for producing the item being considered? In addition, at the transport level there are questions about how to allocate energy use between different products when goods are deliv- ered using shared vehicle space—for example, in a par- cel operation. Because many products are consumed in urban areas, there is a need to consider urban freight trips within the LCA approach. Edwards and McKin- non (2010) compare the carbon footprint of the conven- tional and online retail supply chain, thereby including an assessment of urban freight transport energy use. Yet this type of modeling is not usually linked to the urban freight modeling approaches discussed so far. Given the existence of the link between energy use, emissions, and air quality, some of this research into energy consumption overlaps with research into air qual- ity standards within cities and the impact of urban freight transport. Urban regions have concentrated transporta- tion activity, increasing emissions and pollution expo- sure to those who live and work in the region. Often background levels of pollutants are also higher than in nonurban regions because of manufacturing and indus- trial activities that may also be concentrated in urban areas. Many urban regions have conducted emissions inventories (e.g., EMFAC), which will not be discussed here. In the EU context, much of the research concerned with air pollution in cities has been framed by the desire to create low-emissions zones in urban areas. As noted by the Low Emission Zone in Europe Network (LEEZEN), [A]ir pollution is responsible for 310,000 prema- ture deaths in Europe each year . . . more than caused by road accidents. Air pollution particularly affects the very young and the old and those with heart and lung diseases—both common causes of death in Europe. It also triggers health problems like asthma attacks and increases hospital admis- sions and days off sick. The human health damage that air pollution causes is estimated to cost the European economy between €427 and €790 billion per year. Because of this danger to health, many countries around the world, as well as the Euro- pean Union (EU), have set air quality targets to be met. In the EU, it is in order to meet these targets that LEZs are being implemented. (LEEZEN 2008) There are now more than 160 such zones in Europe, and they apply to both small and large cities. A key ele- ment in the creation of such zones is the underpinning research to demonstrate how large the zone needs to be, how stringent the emission standards, and how the enforcement can be managed, all of which influence the cost–benefit analysis. Research into the results of the implementation of low-emission zones (Allen and Browne 2009; Johansson and Burman, n.d.; Joint Expert Group on Transport and Environment 2005; Transport for London 2008) suggests that they are a useful policy measure to help improve local air quality conditions to achieve threshold air qua- lity values, especially in relatively small areas that suffer from regularly exceeded conditions. However, there are problems and questions that require more research. For example, low-emission zones can result in vehicle detours to avoid the zone, thereby imposing additional pollu- tion from noncompliant vehicles on locations outside the zone, as well as the redeployment of more polluting vehicles to operate in other locations without such emis- sions restrictions. Low-emission zones also impose vehicle replacement and retrofitting costs on some vehicle oper- ators, which can lead to higher freight transport costs for supply chain partners. Thereby such zones can result in additional operational, financial, and administrative burdens. Further research is required to determine whether countrywide or even European-wide measures concerning vehicle pollutants might be easier to implement, have a greater geographical coverage, and be more cost-effective overall than a proliferation of low-emission zones in urban areas. Research also needs to take into account that even though emissions from electric vehicles are low or zero, the method of generating the electricity may not be, and this may temper the benefits of switching to such vehicles. Many urban regions (e.g., Berkeley, Calif; and Seattle, Wash.) have conducted emissions inventories (e.g., using EMFAC); indeed, some of the research concerned with air quality and emissions is probably best considered as a process of accounting rather than as modeling (Boswell et al. 2012). In this accounting or inventory approach, it is important to consider the number of vehicles, the num- ber of trips, and then find agreed-upon and transparent assumptions about typical emissions values (e.g., using the U.S. Environmental Protection Agency’s MOVES model). The trips and distance run (and therefore the emissions) can then easily be summed. Different scenarios can be explored without reference to modeling simply by mak- ing different assumptions about the fleet mix, distance run, and typical emissions values. However, this ignores some of the important feedback loops that would ideally be considered—for example, as the mix of vehicles or the time of day of operation changes, so does the aver- age vehicle speed and, therefore, emissions values change. To accommodate these more complex patterns, modeling

84 city logistics research: a transatlantic perspective becomes essential. When these interactions are considered, the models need to be able to estimate existing and future traffic flows (including issues such as average speed). This information can then be linked to data on the vehicle fleet operating in urban areas to estimate changes in energy use and pollutant emissions. This approach applies both to freight and nonfreight trips. However, in the case of freight-related trips, the approach is complicated by fac- tors such as uncertainty over the types of freight vehicle in use, their relative age and fuel efficiency, and a lack of data about trip patterns and trip distances for freight. To estimate human exposure to pollution, detailed population modeling must be combined with emissions models, and the spatial nature of these data retained. However, with GHG emissions, the concern is not local- ized human exposure, but the planetary impacts of the release of the GHG into the atmosphere. This means the spatial component of the release may not need to be tracked in the analysis or modeling effort. In addition, CO2 production is often estimated directly from fuel con- sumption, significantly altering the modeling approach or data inputs required for the analysis. To reduce emissions, many urban regions are consider- ing how to move people and freight onto lower-carbon transport modes. This requires both an understanding of the contributions of various modes to regional air pollu- tion and an understanding of the financial and behavioral aspects of travel. This question applies to freight opera- tions as well as passenger travel. For example, what are the emissions consequences–benefits of shifting freight onto smaller vehicles that may be powered by cleaner fuels? With increasingly available computing power and data management tools, some research is moving toward very detailed models applied on larger geographic scales to model detailed vehicle speed, weight, and performance data at the regional level (Boriboonsomsin et al. 2012). 4. dAtA reQuIreMents And AvAIlAbIlIty Models can be applied at smaller and larger spatial extents of the urban freight transportation system, from an intersection, where pedestrian behavior may be exam- ined, to the entire urban region and beyond. Similarly, models may be designed to capture short-term variations in travel demand or predict flows over decades. Finally, the models may represent origins and destinations as individual parcels or as aggregate parcels in larger zones. These variations determine the data requirements and should be aligned with the objectives of the modeling effort. Table 1 (below) identifies the criteria on which models may vary in their scope. For example, most mod- els represent urban regions as a series of zones. These zones may be small, representing each parcel or estab- lishment and, therefore, the travel behavior associated with each parcel; or they may be large, aggregating travel to or from the zone. Model structures and logic are based on relationships observed from empirical data. Data are collected to build an understanding of the urban freight transportation sys- tem and serve as inputs to models once developed and comparisons for validation of model outputs—and, of course, model outputs are also used as data to support decision making and knowledge development. Much of the data used for urban freight transportation modeling and analysis are collected for a purpose other than urban freight research. For example, truck-count data are gen- erally collected for pavement design and maintenance. This presents significant challenges with respect to sta- tistical sampling and spatial or temporal relevance. For example, while Global Positioning System (GPS) data provide an opportunity to track truck trips, the market penetration of GPS or how this varies across subsectors of the trucking industry are unknown. In addition, sur- rogate measures must often be used in place of metrics of interest. For example, because total truck trip generation is unknown, employment or establishment size is often used as a surrogate. Table 2 (page 85) describes categories of data, exam- ples of data sets of this type in the United States, and some of their qualities, with an emphasis on transporta- tion data for urban freight analysis. The utility of national data sources is limited by the granularity they provide for urban scale analysis. TABLE 1 Scale of Model Application Descriptor Range Spatial extent Urban region Neighborhood Intersection Spatial granularity Large zones Establishment System components All traffic All truck traffic Enterprise Temporal granularity Typical day Day divided into 3 to 5 periods Hourly or by minute Temporal extent 40 years Seasonal Hours Network granularity Represent only highways and arterials Detailed representation of lanes and all roadway links

85appendix a Browne and Goodchild Because of this, many regions in the United States find that national data sources must be either disaggregated or complemented with local data collection efforts, including vehicle counts at specific locations. This disag- gregation is often accomplished using demographic or economic data such as employment or economic activity. Some guidance has been published on how to accom- plish this disaggregation, for example, the report from TRB’s NCFRP Project 20, Guidebook for Developing Suboperational Commodity Flow Data (NCFRP 2013). Some spot-count data are available because of col- lection for alternative purposes, including roadway and pavement design and maintenance, and they usually cat- egorize trucks in categories relevant for these purposes. However, this is not always relevant for urban freight demand modeling because vehicles of less than 14,000 lbs (6,350 kg) are combined into light-duty vehicles, making it difficult to differentiate smaller vans from pas- senger vehicles. The situation is broadly similar within the European Union, albeit that the maximum weight limit for light-duty vehicles (also referred to as light- goods vehicles or vans) is set at 3,500 kg. More recently, fleet activity data have become avail- able in both the United States and the European Union, either through GPS data records or partnerships with individual fleets or groups of fleets. These data are col- lected for fleet management or, in some cases, tax pur- poses, and to protect confidentiality are stripped of any identifying information prior to being shared with researchers. These data are well suited for applications where GPS records serve to represent general traffic flow, but the use of these data for other applications is limited by a lack of information on the population represented by this data set. This is also a concern for third-party data providers. In the United States, data with sufficient spatial reso- lution for many urban freight applications are generally collected at the urban scale and paid for by local and regional agencies. This includes localized surveys or data collection efforts. This practice presents several chal- lenges to urban freight researchers in that data are (1) not collected on a regular basis, which causes difficulty evaluating changes over time; (2) collected for a short period, which makes them unrepresentative of seasonal or other temporal factors; and (3) not collected using the same methodology in different locations, which makes them difficult to compare over space. The need to sup- plement national data at the city level is also apparent in the European Union. The most comprehensive urban freight surveys to provide better data were carried out in France in the 1990s (Patier et al. 1997, 2000). These sur- veys are now being repeated in several French cities and will be an important contribution to our understanding of changes during the past 15 years. A review of data requirements and availability at an EU level was carried out within the BESTUFS Project (BESTUFS 2006). Com- modity or industry-specific data can be useful for analy- ses of individual sectors of the freight system, but cannot be used to represent all freight flows. 5. strengths And WeAknesses oF current ApproAches In the United States, travel demand models are used extensively by MPOs to inform planning and support TABLE 2 Data Sources for Model Applications Scale Example Description National Commodity Flow Survey, Freight Analysis Framework 3 Limited spatial detail, must disaggregate for urban scale Spot data Truck counts and speeds, highway perfor- mance monitoring system or state department of transportation Most highways and Interstates, less available at arterial or residential street level; limited truck categories, if truck counts at all Fleet activity data GPS, partnerships with firms or fleets Sampling often unknown, offers challenges with representation Third-party data providers INRIX, TRANSEARCH Details of interpretation not transparent; not publicly available Surveys Detailed, tailored data Expensive, not repeated Public metropolitan, not transportation data Parking, land use, employment Must estimate models to convert to trip or travel data Commodity or industry-specific data Some federal agencies, some industry groups or organizations Potentially good for industry-specific analysis, difficult to estimate entire traffic stream from this approach Demographic (not transport) or economic data Census Must estimate models to convert to trip or travel data

86 city logistics research: a transatlantic perspective decision making. Generally, these models are accepted as useful tools and have been integrated into the deci- sion-making process. Many organizations have built truck modeling capacity into these travel demand mod- els, with use of the same four-step process. While some organizations have developed significantly more than others, as discussed, the structures of most four-step freight travel demand models significantly limit their ability to address some of the challenges to urban freight research. Typically, truck trip generation is most often a function of employment or establishment size and therefore independent of system performance or land use mix; therefore, the models cannot capture responses to access regulation, service improvement, land use mix, or street design. This means trip rates are inelastic to changes in performance from infrastructure modifications or tolling. In addition, available regional modeling tools have limited ability to address some of the more detailed changes or effects of dense urban environments. These models typically use a zone as their basic spatial unit, capturing trips from, to, and within each zone. Trips within each zone are not applied to the road network. These model results therefore do not comment on the impact of the last mile of travel because it occurs within one zone. This information can be incorporated into the model as an input by modifying the number manually, but adequate data are not currently available to ensure that changes to terminal times are appropriate. Models with zones larger than a city block were not designed to, and cannot in any detailed way, model loading areas, mixed-use development, or on-street parking. Micro- simulation models would be more appropriate for this type of evaluation. A further issue that limits the ability of travel demand models to address the contemporary challenges of urban freight is the lack of explicit tours for trucks in most truck models and the limited handling of intermediate locations. Again, reference is made to the models cur- rently in use by MPOs to address urban freight chal- lenges. Because a four-step model represents individual trips, it cannot account for the synergies apparent in routing and trip planning available through the sophisti- cation of logistics firms. Unfortunately, in the absence of considerable data development and research to validate improved models, tour-based models are unlikely to be operational in the near term for the majority of MPOs. Emissions modeling is a more recent development for metropolitan planners. Most agencies take advantage of emissions factors provided by other sources and can then consider the net emissions from transportation activity when these emissions factors are tied to transportation activity from a travel demand model. Because of this, the limitations of the travel demand models also apply to the emissions modeling. Also, emissions modeling is sensitive to the additional data requirements, such as the emissions factors available, transportation activity data quality, and fleet data. In addition to having these current data, the values must also be estimated for the future. Most travel demand models were not designed to cap- ture detailed transportation activities; for example, all vehicles travel the same speed on a link, and speed varia- tions are not considered. This makes the travel model- ing frameworks somewhat inappropriate for emissions modeling, which is more sensitive to speed variations. A significant limitation to both travel demand model- ing and emissions modeling is validation. If the intent is to improve, or make the models better, how are they defined better? How is the efficacy and utility of these modeling approaches judged? Will making the model better produce better policy outcomes? These questions can be separated into two issues: 1. Can the model be demonstrated to replicate observed data? 2. Is the model output used to influence decision making? Rules of good practice suggest that any model output should, to the extent possible, be validated, or compared against observations of the phenomema being modeled. Travel demand models are designed to estimate system- wide travel, but the validation must be conducted with time- and location-specific data. However, travel demand model outputs are usually compared with available data, including travel surveys and traffic counts (PSRC Travel Demand Model Documentation 2007). While limited, this is an important step; however, this process cannot comment on the predictive abilities of the model. If travel demand model output is used as an input into another process, multiple errors will be propagated through the modeling system, and the secondary estimate (e.g., emis- sions) could be extremely uncertain. With emissions modeling, one challenge is that ambient air quality mea- surements include background emissions but the models do not, and ambient air quality is a function of many additional features of the environment (e.g., see Fried- man et al. 2001). The issue of whether the models are effective at affect- ing decision making is more difficult to address, but discussion is encouraged around this topic at the sym- posium. Often these observations are spatially or tem- porally limited. 6. chAllenges It is clear from the discussion in the previous sections that research into urban freight has many analytical and

87appendix a Browne and Goodchild modeling challenges. This section summarizes these chal- lenges by considering the following: • Challenges driven by complexity and rapid change, • Challenges driven by a lack or limitation of knowl- edge or data, and • Gaps in communication. 6.1 Challenges Driven by Complexity and Rapid Change Researchers concerned with urban freight have increas- ingly sought to incorporate a supply chain approach into their analysis of urban freight. Adding the supply chain approach contributes to an understanding of the scope for changes in vehicle flows that result from changes in supply chain strategies and operations (e.g., changes in inventory management strategies by retailers). But this approach inevitably leads to greater complexity, and the current ana- lytical tools and freight models are not easily manipulated or adapted to include the supply chain. This challenge is magnified when consumer behavior and trips are consid- ered. Yet, this becomes increasingly important with the rise in e-commerce and home delivery of products leading, in turn, to changes in the flow of goods and vehicles (both for passengers and freight) within urban areas. Indeed, work on behavioral issues in general is weaker in freight transport than in passenger travel. Stated pref- erence surveys allow some understanding of trade-offs made by freight buyers, but much of this work is hard to incorporate into existing modeling frameworks. At the micro level of driver behavior, far less is known about the behavior of truck and van drivers (and the complex interactions with trip planners and managers) than is known about car users and travelers generally. Trip generation studies that underpin many planning initiatives also have flaws. Many of them are based on assumptions derived from relatively limited surveys, and it is not clear how the trip generation work can readily be adapted to changes in the wider economy—for exam- ple, more service activity, the rise of internet shopping, mobile working, and so on. Complexity is also influenced by the number of stakeholders engaged in urban freight decision making, whether at the public policy planning level or the supply chain level. Structuring problems in a way that allows for decisions can be difficult. In some cases there is pressure for simple solutions and a reluctance to spend sufficient time and money on detailed research of the problems and issues. The speed of technology development (especially in communications) has produced major opportunities for research in urban freight. For example, the scope to acquire information from new technologies related to vehicle and cargo tracking is enormous. However, the rapid pace of the change also creates some problems and challenges. One opportunity relates to the vast amounts of travel data becoming available (also referred to as the challenge of “Big Data”). Among the most important challenges are the following: • How to capitalize on the vast amount of data that are available but remain in private hands, • How to prevent data confidentiality and privacy from inhibiting the use of data from some of the com- mercial systems implemented by companies, and • How to understand the statistical sampling and sta- tistical confidence of the analytical results. 6.2 Challenges Driven by a Lack of or Limitations of Knowledge or Data A major weakness in the understanding of urban freight flows concerns smaller vehicles (below 3.5 tons in the European Union and 14,000 pounds in the United States). These vehicles make up a significant part of the vehicle flow in urban areas (in France, 50%) and are responsible for many of the deliveries and collections and for almost all of the service trips. Yet very little data are collected about these vehicles, and few models address the activity of the van sector. National data collection at the EU level for trucks of more than 3.5 tons is a statu- tory requirement; but for smaller vehicles, it is optional. And with cuts in public spending there are fewer efforts to collect such data than in the past. In the United States, delivery vans are not trackable in any existing data set; this means the magnitude of the VMT is hard to estimate, origin–destination and time of day patterns are not well understood, and therefore public-sector involvement in the field is limited. In urban freight it is apparent that sometimes research is directed at the vehicle, sometimes at the flow of freight, and sometimes at both factors. However, there is a lack of data matching cargo with vehicles. While vehicle flows and counts are reasonably well measured and modeled, very little data are available demonstrating which com- modities or products use which elements of the infra- structure. This particular lack of data limits the ability to understand the economic value of elements of the infra- structure and make investments based on these criteria. It also inhibits the analysis of freight-related environmental impacts at sector, industry, or commodity levels. At the EU level there are clear challenges when trying to make comparisons between the member states (and this can be extended to consider countries that are not EU members). Some comparisons are difficult because data are not collected in a consistent and common man- ner. But there are also challenges in terms of definitions

88 city logistics research: a transatlantic perspective (some likely caused by questions of language). For exam- ple, within urban freight surveys published in Europe, there are many terms used to describe delivery rounds (including tours, trips, and journeys), and even informa- tion about the delivery itself can be difficult to compare when terms such as drop, delivery, consignment, and item are all used without any standard definition for their meaning. The largest and most detailed surveys on urban freight movement (in Europe) have been carried out in France (see Patier et al. 1997, 2000), and yet the results and approaches are not very well known in the UK, pri- marily because few of the results and approaches (at any detailed level) have been widely disseminated in English. 6.3 Gaps in Communication Before discussing whether there is a gap between research and practice in urban freight it is relevant to mention the gaps that can be found within the urban freight research community. Perhaps one of the most striking that cer- tainly existed until the past few years is the gap between those engaged primarily in urban freight modeling and those working on policy or business-related research issues. This gap has been closing in recent years, and conferences addressing both perspectives are now more common. Improvements in modeling urban freight flows do not necessarily have to take account of the policy perspective. This lack of policy relevance may not be a problem from a scientific perspective, but it may make it much harder to use the models in a more pragmatic way to support decisions. Researchers focusing on policy questions do not seem to have framed some of these ques- tions in terms of the possible role of models in providing relevant answers, and this simply continues the divide. There would seem to be a gap between practitioner requirements for simple solutions that can be imple- mented relatively quickly and the requirement for rigor- ous research that often highlights complexity. As argued at the start of this paper, the urban freight system is com- plex, with multiple stakeholders and difficult decision- making trade-offs to consider. This complexity may be very interesting from a research perspective, but for the practitioner, given the task of resolving problems and improving the urban area, it can make it hard to achieve progress. And research often seems to identify the many different possibilities without giving a clear direction as to which of them is most appropriate. 7. other u.s.-eu dIFFerences This white paper provides a review of two areas of urban freight research where modeling approaches are used. The paper compares the approaches and their support- ing data between U.S. and European cities. While model- ing is one aspect of urban freight research, it cannot be forgotten that decisions regarding how to address urban freight challenges are made in a political environment. The political cultures across the Atlantic, but also within Europe and the United States, are varied, meaning that some policy solutions are acceptable in some locations but not in others. The roles of municipal governments in relation to regional or national governments vary, as do political ideologies and cultural values. In Europe, multinational cooperation has been strongly encouraged by the European Union. To encour- age this, funding opportunities have explicit require- ments on member state collaboration. While there is not an equivalent funding mechanism in the United States, one could argue that the lack of one reflects a lack of need, that collaborative partnerships between states are more implicit or occurring naturally. Data programs vary. Freight data collected for Euro- pean cities are very hard to compare, individual cities often have discretion over the monitoring and analysis of freight data, and there are also national differences in definitions and approaches. As noted in Section 6, there is little consistency or commonality in terms of the data collected about urban goods and vehicle flows. This can make comparisons between cities difficult and expensive (if new data need to be captured). This lack of a com- mon data approach may also inhibit the development of models. Models have been developed, but at the Euro- pean level those used in urban freight analysis very often remain national in context and application. Infrastructure in historical European cities is more constrained, and the development more dense than any location in the United States. Historical European cen- ters have tourist attractions whose integrity is threatened by vehicle emissions and vibrations. This has made many cities sensitive to the problems imposed by urban freight. In some cases, this problem-centered approach may stimulate research but it also leads to conflict between the freight transport community and those engaged in policy making and regulation enforcement. The nature of many urban areas has clearly influenced research top- ics in Europe, with many studies carried out on the use of small vehicles (and even cargo cycles in recent years). Within Europe there is a strong emphasis on the idea of transferability. Many European-funded projects explic- itly seek to address this question. There have also been projects where there are lead cities and follower cities to test whether it is possible to speed up the learning and implementation cycle for urban freight–related improve- ments. While of course this is valued in the United States, projects have not been set up in this fashion. Private-sector participation has been encouraged on both sides of the Atlantic. In Europe, it seems that trans- port companies and vehicle manufacturers have been

89appendix a Browne and Goodchild engaged—and to a lesser extent, retailers and suppli- ers. This is reversed in the United States, where shippers and receivers (and to a lesser extent, carriers) have been engaged in planning efforts. 8. recoMMendAtIons In this review, several urban challenges to which freight transport is a significant contributor have been identi- fied. It is intended that these identified challenges will serve to raise issues for discussion. These include the fol- lowing questions: 1. Should additional behavioral aspects be built into models? Will this allow them to be more policy sensitive? 2. To what extent can and should supply chain behav- iors be built into freight models? Given their tendency to change over time, what modeling approaches are appro- priate? Can a probabilistic approach be used? 3. How can the effectiveness of these models at addressing urban challenges be measured? 4. Why has there been little attention to small-scale problems, such as the impact of on-street parking and loading zones, and the impact of bicycle and pedestrian facilities on traffic flow and capacity? How can greater research attention be brought to these problems? The authors feel there is an opportunity to improve the communication and transfer of knowledge between urban freight researchers and policy makers. Why does this communications gap persist? How does this relate to analytical tools? To what extent do these tools address the urban freight problems that policy makers are most concerned with? This may be more than a communica- tion problem; for example, are academic projects not suf- ficiently relevant and therefore not considered by policy makers? Are the data requirements or modeling knowl- edge too onerous? Does academic research sufficiently capture the key features of private-sector activities? To further explore these questions, it is recommended that the following activities be considered by symposium participants: 1. Conduct a more systematic review of the state of modeling and analytical work completed with joint sup- port from the United States and Europe; 2. Fund joint projects, especially those that address the modeling–policy gaps; 3. Organize a showcase for some examples where the analytical and policy gap has been overcome or nar- rowed; 4. Organize an annual meeting aimed at encouraging European–United States cooperation between research- ers; and 5. Consider a journal special issue built around the workshop addressing the questions raised by the papers and presentations. AcknoWledgMents The authors thank the paper’s reviewers, whose com- ments have strengthened the paper considerably. reFerences Abdelgawad, H., B. Abdulhai, G. Amirjamshidi, M. Whaba, C. Woudsma, and M. J. Roorda. Simulation of Exclusive Truck Facilities on Urban Freeways. ASCE Journal of Transporta- tion Engineering, Vol. 137, No. 8, 2011, pp. 547–562. Ainge, M., P. Abbott, C. Treleven, P. Morgan, P. Nelson, G. Watts, and R. Staite. Alternative Methods for the Man- agement of Night-Time Freight Noise in London. Project Report PPR 286. TRL, Camberley, United Kingdom, 2007. Allen, J., and M. Browne. Environmental Zones in European Towns and Cities and the Implications for Freight Move- ment. In Supply Chain Management and Logistics in a Volatile Global Environment (E. Sweeney, ed.), Blackhall, Dublin, 2009. Allen, J., M. Browne, and A. Woodburn. London Freight Data Report 2010. Transport for London, 2010. Allen, J., M. Browne, A. Woodburn, and J. Leonardi. The Role of Urban Consolidation Centres in Sustainable Freight Transport. Transport Reviews, Vol. 32, No. 4, 2012, pp. 473–90. Ambrosini, C., D. Patier, and J.-L. Routhier. Urban Freight Establishment and Tour-Based Surveys for Policy-Oriented Modelling. Procedia—Social and Behavioral Sciences, Vol. 2, No. 3, 2010, pp. 6013–6026. Anderson, S., J. Allen, and M. Browne. Urban Logistics—How Can It Meet Policy Makers’ Sustainability Objectives? Jour- nal of Transport Geography, Vol. 13, No. 1, 2005, pp. 71–81. Andreoli, D., and A. Goodchild. A Supply Chain Analysis of Truck Trip Generation: A Case Study in Washington Potatoes. Transportation Letters, Vol. 4, No. 3, 2012, pp. 153–166. BESTUFS. BESTUFS Best Practice in Data Collection, Model- ing Approaches, and Application Fields for Urban Com- mercial Transport Models. Deliverable D3.1, 2006. www. bestufs.net. Böge, S. The Well-Travelled Yoghurt Pot: Lessons for New Freight Transport Policies and Regional Production. World Transport Policy and Practice, Vol. 1, No. 1, 1995, pp. 7–11. Boriboonsomsin, K., M. Barth, W. Zhu, and A. Vu. Eco-Rout- ing Navigation System Based on Multisource Historical and Real-Time Traffic Information. IEEE Transactions on

90 city logistics research: a transatlantic perspective Intelligent Transportation Systems, Vol. 13, No. 4, 2012, pp. 1694–1704. Boswell, M., A. Greve, and T. Seale. Local Climate Action Planning. Island Press, Washington, D.C., 2012. Boulder. Boulder’s Community Greenhouse Gas Inven- tory. 2010. http://www.bouldercolorado.gov/index .php?option=com_content&view=article&id=15354&Ite mid=5158. Accessed May 14, 2013. Browne, M., J. Allen, and J. Leonardi. Evaluating the Use of an Urban Consolidation Centre and Electric Vehicles in Cen- tral London. International Association of Traffic and Safety Sciences Research, Vol. 35, No. 1, 2011, pp. 1–6. Browne, M., C. Rizet, S. Anderson, J. Allen, and B. Keïta. Life Cycle Assessment in the Supply Chain: A Review and Case Study. Transport Reviews, Vol. 25, No. 6, 2005, pp. 761–782. Carney, S., N. Green, R. Wood, and R. Read. Greenhouse Gas Emissions Inventories for 18 European Regions. Report. EUCO2 80/50 Project, 2009. http://www.euco2.eu/ resources/EUCO2-Results-Summary-Small.pdf. Accessed May 14, 2013. Chase, K. M., P. Anater, and T. Phelan. SHRP 2 Report S2-C20-RR1: Freight Demand Modeling and Data Improve- ment. Transportation Research Board of the National Academies, Washington, D.C., 2013. Cherrett, T., J. Allen, F. McLeod, S. Maynard, A. Hickford, and M. Browne. Understanding Urban Freight Activity— Key Issues for Freight Planning. Journal of Transport Geog- raphy, Vol. 24, 2012, pp. 22–32. Comi, A., P. Delle Site, F. Filippi, and A. Nuzzolo. Urban Freight Transport Demand Modelling: A State of the Art. European Transport, Vol. 51, 2012. Dablanc, L., and D. Rakotonarivo. The Impacts of Logistics Sprawl: How Does the Location of Parcel Transport Termi- nals Affect the Energy Efficiency of Goods’ Movements in Paris and What Can We Do About It? Procedia—Social and Behavioral Sciences, Sixth International Conference on City Logistics (E. Tanguchi and R. G. Thompson, eds.), Vol. 2, No. 3, 2010, pp. 6087–6096. Dablanc, L., and C. Ross. Atlanta: A Mega Logistics Center in the Piedmont Atlantic Megaregion (PAM). Journal of Transport Geography, Vol. 24, 2012, pp. 432–442. Debauche, W. An Investigation into the Delivery of Goods to the City Centre of Liege. In Innovations in City Logistics (E. Taniguchi and R. Thompson, eds.), Nova Science Pub- lishers, New York, 2008, pp. 261–276. Donnelly, R., G. D. Erhardt, R. Moeckel, and W. A. Davidson. NCHRP Synthesis 406: Advanced Practices in Travel Fore- casting. Transportation Research Board of the National Academies, Washington, D.C., 2010. Edwards, J. B., and A. C. McKinnon. Comparative Analysis of the Carbon Footprints of Conventional and Online Retail- ing: A “Last Mile” Perspective. International Journal of Physical Distribution and Logistics Management, Vol. 40, No. 1, 2010. European Commission. Roadmap to a Single European Trans- port Area—Towards a Competitive and Resource Efficient Transport System. White paper. COM/2011/0144 final. Brussels, Belgium, 2011. Feliu, J., C. Ambrosini, and J. L. Routhier. New Trends on Urban Goods Movement: Modelling and Simulation of E-Commerce Distribution. European Transport, Vol. 50, 2012. Figliozzi, M. A. Vehicle Routing Problem for Emissions Mini- mization. In Transportation Research Record: Journal of the Transportation Research Board, No. 2197. Transporta- tion Research Board of the National Academies, Washing- ton, D.C., 2010, pp. 1–7. Fischer, M. J., and M. Han. NCHRP Synthesis 298: Truck Trip Generation Data. TRB, National Research Council, Wash- ington, D.C., 2001. Friedman, M. S., K. E. Powell, L. Hutwagner, L. M. Graham, and W. G. Teague. Impact of Changes in Transporta- tion and Commuting Behaviors During the 1996 Summer Olympic Games in Atlanta on Air Quality and Childhood Asthma. Journal of American Medical Association, Vol. 285, No. 7, 2001, pp. 897–905. Golob, T., and A. Regan. Freight Industry Attitudes Towards Policies to Reduce Congestion. Transportation Research Part E: Logistics and Transportation Review, Vol. 36, No. 1, 2000, pp. 55–77. Gonzalez-Feliu, J., and J. Morana. Are City Logistics Solu- tions Sustainable? The Cityporto Case. Territorio Mobilità e Ambiente, Vol. 3, No. 2, 2010, pp. 55–64. Hensher, D., and M. Puckett. Refocusing the Modelling of Freight Distribution: Development of an Economic- Based Framework to Evaluate Supply Chain Behaviour in Response to Congestion Charging. Transportation, Vol. 32, No. 6, 2005, pp. 573–602. Holguín-Veras, J., K. Ozbay, A. Kornhauser, M. Brom, S. Iyer, W. Yushimito, S. Ukkusuri, B. Allen, and M. Silas. Overall Impacts of Off-Hour Delivery Programs in New York City Metropolitan Area. In Transportation Research Record: Journal of the Transportation Research Board, No. 2238, Transportation Research Board of the National Academies, Washington, D.C., 2011, pp. 68–76. Holguín-Veras, J., K. Ozbay, A. Kornhauser, A. Shorris, and S. Ukkusuri. Integrative Freight Demand Management in the New York City Metropolitan Area. Final report. U.S. Department of Transportation, 2010. Holguín-Veras, J., M. Silas, J. Polimeni, and B. Cruz. An Inves- tigation on the Effectiveness of Joint Receiver–Carrier Poli- cies to Increase Truck Traffic in the Off-Peak Hours, Part I. The Behavior of Receivers, Networks, and Spatial Econom- ics, Vol. 7, 2007, pp. 277–295. Holguín-Veras, J., M. Silas, J. Polimeni, and B. Cruz. An Inves- tigation on the Effectiveness of Joint Receiver–Carrier Poli- cies to Increase Truck Traffic in the Off-Peak Hours, Part II. The Behavior of Carriers, Networks, and Spatial Econom- ics, Vol. 8, 2008, pp. 327–354.

91appendix a Browne and Goodchild Hunt, J., K. Stefan, and A. Brownlee. Establishment-Based Sur- vey of Urban Commercial Vehicle Movements in Alberta, Canada: Survey Design, Implementation, and Results. In Transportation Research Record: Journal of the Transpor- tation Research Board, No. 1957, Transportation Research Board of the National Academies, Washington, D.C., 2006, pp. 5–83. Jespersen, P. H. The Transport Contents of Products. World Transport Policy and Practice, Vol. 10, No. 3, 2004, pp. 28–35. Jessup, E., K. Casavant, and C. Lawson. Truck Trip Data Col- lection Methods. Final Report SP343. Oregon Department of Transportation, Salem, 2004. Johansson, C., and L. Burman. Swedish Experience with Low Emission Zones. Environment and Health Protec- tion Administration, Air and Noise Analysis, Institute of Applied Environmental Research, Stockholm University, Stockholm, Sweden, no date. Joint Expert Group on Transport and Environment. Report from the Working Group on Environmental Zones: Exploring the Issue of Environmentally Related Road Traffic Restrictions. European Commission, Brussels, Belgium, 2005. Kanaroglou, P., and R. Buliung. Estimating the Contribution of Commercial Vehicle Movement to Mobile Emissions in Urban Areas. Transportation Research Part E, Vol. 44, 2008, pp. 260–276. Klaasse, M., C. Maaskant, and R. de Lang. PIEK Project Pie- ktrekker. DAF Rapport No. 51212/02-140. 2002. Kriger, D., E. Tan, and A. Clavelle. Results of the TAC Project on the Framework for the Collection of High-Quality Data on Urban Goods Movement (Phase 1). Resource paper. Annual Conference of the Transportation Association of Canada, Saskatoon, Saskatchewan, 2007. Kuzmyak, V. NCHRP Synthesis 384: Forecasting Metropolitan Commercial and Freight Travel: A Synthesis of Highway Practice. Transportation Research Board of the National Academies, Washington, D.C., 2008. Lammgård, C., and M. Browne. Night-Time Deliveries: Sup- ply Chain Dream or Policy Nightmare? Logistics Research Network Conference, Cranfield, United Kingdon, 2012. Lammgård, C., and J. Hagberg. Designing for Sustainable Logistics in Urban Areas: What Do We Know? World Con- ference on Transport Research, Rio de Janeiro, Brazil, 2013. Lau, S. Truck Travel Surveys: A Review of the Literature and State of the Art. Metropolitan Transportation Commission, Oakland, Calif., 1995. Lawson, C., and J. D. Strathman. Survey Methods for Assess- ing Freight Industry Opinions. Final Report SPR 328. Ore- gon Department of Transportation, Salem, 2002. Lazarus, M., P. Erickson, C. Chandler, M. Daudon, S. Donegan, F. Gallivan, and J. Ang-Olson. Getting to Zero: A Pathway to a Carbon-Neutral Seattle. Report. City of Seattle Office of Sustainability and Environment, Seattle, Wash., 2011. http://www.seattle.gov/environment/documents/CN_Seat tle_Report_May_2011.pdf. Accessed May 14, 2013. LEEZEN. Low Emission Zones in Europe Network (LEEZEN). Europe-wide Information on LEZs. 2008. http://www .lowemissionzones.eu. Accessed Aug. 25, 2009. McKinnon, A. Product-Level Carbon Auditing of Supply Chains: Environmental Imperative or Wasteful Distraction? International Journal of Physical Distribution & Logistics Management, Vol. 40, No. 1–2, 2010, pp. 42–60. National Cooperative Freight Research Program. NCFRP Report 26: Guidebook for Developing Subnational Com- modity Flow Data. Transportation Research Board of the National Academies, Washington, D.C., 2013. http://www. trb.org/Main/Blurbs/169330.aspx. NICHES. Innovative Approaches in City Logistics: Inner-City Night Delivery. NICHES, no date. Patier, D., and J.-L. Routhier. How to Improve the Capture of Urban Goods Movement Data. 8th International Con- ference on Survey Methods in Transport, Annecy, France, 2008. Patier, D., and J.-L. Routhier. Une Méthode d’enquête du Transport de Marchandises en Ville pour un Diagnostic en Politiques Urbaines. Cahiers Scientifiques du Transport, Vol. 55, 2009, pp. 11–38. Patier, D., J.-L. Routhier, and Ch. Ambosini. Transport de marchandises en ville: Enquête quantitative réalisées à Dijon et Marseille. Rapport finaux MELT-DRAST. Labo- ratoire d’Economie des Transports, France, 2000. Patier, D., J.-L. Routhier, Ch. Ambosini, P. Bossin, P. Gelas, and M. Le Nir. Transport de Marchandises en ville: Enquête Quan- titative réalisée à Bordeaux. Rapport final MELT-DRAST. Laboratoire d’Economie des Transports, France, 1997. PB Consult Inc. An Independent Expert Review of the Cal- gary Regional Transportation Model. Forecasting Division, Transportation Department, City of Calgary, Alberta, Can- ada, 2007. PSRC Travel Demand Model Documentation. Cambridge Systematics. http://www.psrc.org/assets/1511/model_doc_ final_.pdf. 2007. Rizet, C., M. Browne, E. Cornelis, and J. Leonardi. Assess- ing Carbon Footprint and Energy Efficiency in Competing Supply Chains: Review—Case Studies and Benchmarking. Transportation Research Part D, Vol. 17, No. 4, 2012, pp. 293–300. Routhier, J.-L., and F. Toilier. FRETURB V3: A Policy Ori- ented Software of Modelling Urban Goods Movement. Proc., 11th World Conference on Transport Research, Berkeley, Calif., 2007. Samimi, A., A. Mohammadian, and K. Kawamura. Behavioral Freight Movement Modeling. In Travel Behaviour Research in an Evolving World (C. R. Bhat and R. M. Pendyala, eds.), Lulu Publishers, 2012. Sanburn, J. Deliver a Fix for Traffic Jams. TIME Magazine, March 25, 2013, p. 40. Shimuzu, M., T. Hyodo, H. Takebayashi, H. Kuse, and Y. Hagino. Study of Delivery Distribution in the Central Area by Tokyo Metropolitan Freight Survey. In Innovations in

92 city logistics research: a transatlantic perspective City Logistics (E. Taniguchi and R. Thompson, eds.), Nova Science Publishers, New York, 2008, pp. 231–248. Sorrell, S., M. Lehtonen, L. Stapleton, J. Pujol, and T. Cham- pion. Decomposing Road Freight Energy Use in the United Kingdom. Energy Policy, Vol. 37, 2009, pp. 3115–3129. Tavasszy, L. A. Freight Modeling: An Overview of Interna- tional Experiences. In Conference Proceedings 40: Freight Demand Modeling: Tools for Public-Sector Decision Mak- ing. Transportation Research Board of the National Acad- emies, Washington, D.C., 2008, pp. 47–55. Transport for London. London Low-Emission Zone Impacts Monitoring Baseline Report. 2008. Transport for London. Regent Street: Delivery and Servicing. Report. ARUP, 2009. Transportation Research Board. Special Report 288: Metro- politan Travel Forecasting: Current Practice and Future Direction. Transportation Research Board of the National Academies, Washington, D.C., 2007. van Duin, R., H. Quak, and J. Munuzuri. New Challenges for Urban Consolidation Centres: A Case Study in The Hague. Procedia—Social and Behavioral Sciences, Sixth International Conference n City Logistics, Vol. 2, 2010, pp. 6177–6188. Virginia Department of Transportation. What Is Travel Demand Modeling? http://www.virginiadot.org/projects/ resources/vtm/What_is_Travel_Demand_Modeling.pdf. Wygonik, E., and A. Goodchild. Evaluating the Efficacy of Shared-Use Vehicles for Reducing Greenhouse Gas Emis- sions: A U.S. Case Study of Grocery Delivery. Journal of the Transportation Research Forum, Vol. 51, No. 2, 2012, pp. 111–126. Yannisa, G., J. Goliasa, and C. Antonioua. Effects of Urban Delivery Restrictions on Traffic Movements. Transporta- tion Planning and Technology, Vol. 29, No. 4, 2006, pp. 295–311. Yu, R., Y. Wang, and T. Larson. Quantifying the Relation- ship Between Near-Road Concentrations of Fine Particulate Matter and Traffic Flow Observations. Zanni, A., and A. Bristow. Emissions of CO2 from Road Freight Transport in London: Trends and Policies for Long-Run Reductions. Energy Policy, Vol. 38, 2010, pp. 1774–1786.