Transportation Research Part D...Direct impacts of off-hour deliveries on urban freight emissions José Holguín-Verasa,⇑, Trilce Encarnacióna, Carlos A. González-Calderóna, James

Transportation Research Part D xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Transportation Research Part D

journal homepage: www.elsevier .com/ locate/ t rd

Direct impacts of off-hour deliveries on urban freight emissions

http://dx.doi.org/10.1016/j.trd.2016.10.0131361-9209/� 2016 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.E-mail address: [email protected] (J. Holguín-Veras).

Please cite this article in press as: Holguín-Veras, J., et al. Direct impacts of off-hour deliveries on urban freight emissions. TranspoPart D (2016), http://dx.doi.org/10.1016/j.trd.2016.10.013

José Holguín-Veras a,⇑, Trilce Encarnación a, Carlos A. González-Calderón a, James Winebrake b,Cara Wang a, Sofia Kyle a, Nilson Herazo-Padilla a,c, Lokesh Kalahasthi a, Wilson Adarme d,Víctor Cantillo e, Hugo Yoshizaki f, Rodrigo Garrido g

aVREF Center of Excellence for Sustainable Urban Freight Systems, Rensselaer Polytechnic Institute, 110 8th St, Troy, NY 12180, USAb Public Policy and Dean, College of Liberal Arts, Rochester Institute of Technology, Rochester, NY 14623, USAcUniversidad de la Costa, Barranquilla, ColombiadUniversidad Nacional de Colombia, Bogotá, ColombiaeUniversidad del Norte, Barranquilla, ColombiafUniversidade de São Paulo, São Paulo, BrazilgUniversidad Diego Portales, Santiago, Chile

a r t i c l e i n f o

Article history:Received 8 August 2016Revised 7 October 2016Accepted 29 October 2016Available online xxxx

a b s t r a c t

The most significant negative environmental impacts of urban trucking result largely fromtravel in congested traffic. To illustrate the potential of innovative solutions to this prob-lem, this paper presents new research on the emission reductions associated with off-hour freight deliveries (OHD). The paper uses fine-level GPS data of delivery operationsduring regular-hours (6 AM to 7 PM), and off-hours (7 PM to 6 AM), to quantify emissionsin three major cities in the Americas. Using second-by-second emissions modeling, thepaper compares emissions under both delivery schedules for: reactive organic gases, totalorganic gases, carbon monoxide, carbon dioxide, oxides of nitrogen, and particulate matter.The results show that the magnitude of the emission reductions depends on the extent ofthe change of delivery time. In the case of the ‘‘Full” OHD programs of New York City andSão Paulo—where the deliveries were made during the late night and early morning peri-ods (7 PM to 6 AM)—the emission reductions are in the range of 45–67%. In the case of the‘‘Partial” OHD used in Bogotá (where OHD took place between 6 PM and 10 PM), the reduc-tions were about 13%. The emission reductions per kilometer are used to estimate the totalreductions for the cities studied, and for all metropolitan areas in the world with more thantwo million residents. The results indicate the considerable potential of OHD as an effec-tive—business friendly—sustainability tool to improve the environmental performance ofurban deliveries. The chief implication is that public policy should foster off-hour deliver-ies, and all forms of Freight Demand Management, where practicable.

� 2016 Elsevier Ltd. All rights reserved.

1. Introduction

The threat of climate change and the health toll caused by air pollution have created enormous pressure to reduce green-house gases (GHGs) and local pollutants in urban environments. As one of the largest sources of such emissions, the trans-portation sector must play a leading role in finding technological, operational, and behavioral solutions to improve urban

rt. Res.

http://dx.doi.org/10.1016/j.trd.2016.10.013

mailto:[email protected]


http://www.sciencedirect.com/science/journal/13619209

http://www.elsevier.com/locate/trd


2 J. Holguín-Veras et al. / Transportation Research Part D xxx (2016) xxx–xxx

sustainability. Policy makers have recognized this reality, and have enacted numerous initiatives to reduce the environmen-tal pollution produced by mobile sources.

Understandably, a great deal of emphasis has been placed on the use of environmental regulations that seek to cap, andultimately lower, the amount of pollutants emitted by the various sectors of the economy. In mobile sources, this has takenthe form of increasingly strict vehicle emission or efficiency standards for new vehicles. Arguably, these stricter standardshave improved environmental conditions with tangible health benefits. However, in the case of freight vehicles—the mainsubject of this paper—the tightening of emission standards has generally been resisted by the trucking industry. Private-sector carriers argue that the incremental costs associated with the newer standards produce undue financial pressure onan industry characterized by chronic over-supply; an industry where a competitive market with marginal cost pricing is typ-ically the norm. In such markets, additional investments in fixed assets like vehicles cannot be recovered by the prevailingrates. As a result of its position, the trucking industry is generally perceived as an opponent of sustainability and environ-mental improvements. The authors believe that despite its position on emission standards, which is explainable withinthe realities of the market, the trucking industry can become full partners in the quest for sustainability. The key to progressis to design appropriate win-win opportunities for such collaborations to emerge.

The field of Freight Demand Management (FDM) provides such opportunities. FDM seeks to induce behavioral changes onthe generators of freight demand—the ‘‘receivers” of the cargo—to increase the sustainability of the freight traffic that is cre-ated. The importance of differentiating freight demand and freight traffic is central to FDM. The former is related to the needof commercial establishments and households to consume goods; the latter represents the flow of freight vehicles needed totransport those goods. The chief implication is that freight traffic is only a physical expression of the demand—albeit, onethat produces significant amounts of negative externalities. Some of these effects are the responsibility of carriers, whileothers are the responsibility of the receivers. For example, the emissions emitted by a poorly maintained truck are the clearresponsibility of its owner. In contrast, the congestion produced by a truck making deliveries during congested traffic hoursis ultimately the responsibility of the receiver, who typically dictates when deliveries must be made.

The role played by the interactions among shippers, carriers, and receivers in the generation of the various externalitiesproduced is not widely understood. For this reason, numerous programs aimed at increasing the sustainability of freightactivity focus on its most visible expression, the vehicles, and implicitly assume (such as in road pricing) that carriers decidewith complete freedom when and how they travel. Such programs ignore the constraints placed on carriers by the shippersand receivers. Such programs also rarely succeed, because receivers determine how and when they wish to receive deliver-ies; carriers generally cannot induce shippers and receivers to change their behavior. In contrast, FDM recognizes the pivotalrole played by the receivers, and accordingly seeks to change the receivers’ demand in order to indirectly change the behav-ior of shippers and carriers (Holguín-Veras and Sánchez-Díaz, 2016). FDM seeks to induce changes in the timing, number,destination, and/or the mode used to make deliveries to reduce the externalities produced by the associated freight traffic.Off-Hour Deliveries (OHD), the FDM program studied in the paper, induces receivers to change the time of deliveries.

A number of potentially transformative FDM programs could be considered (Holguín-Veras and Sánchez-Díaz, 2016),including, in addition to OHD: (1) Staggered Delivery Programs, where the receiver is incentivized to spread out deliveriesduring the work hours, helping to lower peak-hour freight traffic, thus reducing emissions during the peak hours; (2)Receiver-Led Consolidation Programs, which reduce the number of deliveries that arrive at a given location, lowering freightpeak-traffic and the congestion it generates; and (3) Destination Change Programs where deliveries are sent, not directly tothe receivers, but to another location where they are consolidated and sent to the receivers using environmentally-friendlyvehicles. These strategies were successfully used during the London Summer Olympics, to ensure a timely flow of the sup-plies required to satisfy the needs of the local economy, the millions of visitors to the games, and the local residents, at thesame time that large portions of the primary road network had been allocated to the exclusive use of the Olympic Games(COE-SUFS, 2013b; Browne et al., 2014). By estimating the emission reductions from OHD, the paper provides insight intothe potential impacts of FDM as a central component of a holistic strategy to foster the sustainability of freight activity.

The role that receivers could play in a sustainability strategy was recognized as the result of a large body of research thatused game theory and other forms of behavioral research. The focus of the research was the interconnections between theconstituent agents of an elementary production-consumption link: the shipper, the carrier, and the receiver (Holguín-Veraset al., 2007; Holguín-Veras, 2008; Holguín-Veras et al., 2008; Holguín-Veras, 2011). The analyses of these relations have con-clusively established that in a competitive market, receivers have significant influence over when, where, and how deliveriesare made. The reason is simple: shippers and carriers have to be responsive to the needs and desires of receivers. Since recei-vers are the customers, shippers and carriers that do not deliver the supplies in the manner desired by the receiver run therisk of being replaced by more accommodating competitors. The chief role played by receivers implies that public policiesthat induce beneficial behavioral changes in receivers—particularly those located in congested areas—are likely to lead tochanges in the behavior of shippers and carriers; systemic changes that will reduce freight externalities. The chief findingsof the behavioral research conducted can be summarized as follows: (1) receivers are the key decision-makers regardingtime-of-delivery because, generally speaking, carriers do what receivers say; (2) carriers do not need incentives to partici-pate in OHD, because the cost savings from OHD are more than enough incentive; (3) receivers do need incentives to acceptOHD, otherwise they would not become involved because of the perceived risk and/or business inertia; and (4) once a suf-ficient number of receivers agree to OHD, and carriers can justify at least one off-hours delivery route, a shift to OHD takesplace.

Please cite this article in press as: Holguín-Veras, J., et al. Direct impacts of off-hour deliveries on urban freight emissions. Transport. Res.Part D (2016), http://dx.doi.org/10.1016/j.trd.2016.10.013


J. Holguín-Veras et al. / Transportation Research Part D xxx (2016) xxx–xxx 3

Receivers have a vested interest in improving conditions in the areas surrounding their businesses, because increased liv-ability and quality of life can lead to rising land values and sales. Yet, receivers are often not cognizant of the externalitiesproduced by the deliveries they depend on, or of their own potential role in sustainability improvements. FDM seeks tochange this situation by making receivers aware of the impacts of their practices, and engaging them in the quest to changesupply chains for the better. Incentivizing receivers to adopt beneficial changes in delivery practices is a good way to promptthem into action.

The effectiveness of this shift of focus—from the carrier to the receiver—was demonstrated by the OHD program in NewYork City (NYC), where a one-time financial incentive of $2000 induced more than 400 establishments in the food, retail, andaccommodation sectors to accept OHD. The success of this implementation—where all stakeholders benefited—was widelyreported in the press (Journal of Commerce, 2009; Journal of Commerce, 2010; Transport Topics, 2010; Wall Street Journal,2010). This, in turn, led to pilot tests in such other cities as Copenhagen (Denmark), São Paulo (Brazil), and Bogotá (Colombia)(COE-SUFS, 2013a; Centro de Inovação em Sistemas Logísticos, 2015; COE-SUFS, 2016a; COE-SUFS, 2016b). The success of theOHD program is rooted in the collaborative involvement of the key stakeholders. The public sector provides financial incen-tives to the receivers that accept OHD; private-sector groups help with outreach to receivers; and carriers provide dis-counted shipping rates and implement low-noise delivery practices to mitigate noise impacts to communities. Thebenefits of OHD include reduced daytime congestion levels, safer roadways for pedestrians and bicyclists, superior reliabilityof deliveries, lower shipping costs, increased productivity of delivery operations, and reduced emissions. The quantificationof this final element is the main subject of this paper. The preliminary estimates of the environmental impacts of OHD sug-gest that major reductions in criteria pollutants are possible (Holguín-Veras et al., 2013b). These benefits result from trav-eling in what amounts to free flow conditions, without the stop-and-go driving and low speeds typical of congestion. Theemission savings are particularly noteworthy because they do not require additional expenses on the part of the industry,and since only operational changes need to be implemented, the savings could materialize in a matter of months. In contrast,technological improvements could take years or even decades to reach full penetration, as new technologies would need tobe developed, and gradually adopted by the companies. Moreover, since carriers directly benefit from OHD, these programscan garner their support. Instead of being part of the problem of urban congestion, through OHD programs the private sectorbecomes an active participant in its solution.

This paper’s main objectives are to quantify the emission reductions that could be obtained from inducing a switch ofdeliveries from regular-hours (6 AM to 7 PM) to off-hours (7 PM to 6 AM); and to gain insight into the potential benefitsattributable to this FDM sustainability strategy if deployed worldwide. The authors estimate and compare the direct envi-ronmental impacts (i.e., the emissions directly produced by the vehicles) of mid-size trucks operating during regular-hours and off-hours in the cities of Bogotá (Colombia), New York City (USA), and São Paulo (Brazil). The indirectimpacts—such as the reductions in emissions associated with the lowering of congestion produced by reduced truck trafficduring the regular hours, and the reductions in freight vehicle-kilometers-traveled (VKT) produced by the consolidation ofmultiple regular-hour-delivery (RHD) routes into a smaller number of OHD routes—though substantial, are not consideredhere. To estimate emissions, the authors use fine-level GPS data collected from trucks operating during both regular-hourand off-hour conditions. These data are post-processed to estimate the associated emissions for the cities studied. These esti-mates are then used to produce order-of-magnitude estimates of the emission reductions that could materialize from OHDprograms implemented in metropolitan areas with more than twomillion residents, in both developed and developing coun-tries. The results provide insight into the potential environmental benefits of FDM initiatives, and support to public-sectorprograms that foster OHD in congested environments. This paper adds an important and missing piece—the estimation ofemission reductions from OHD—to the large body of research already conducted on OHD.

The rest of the paper is organized as follows: Section 2 provides a basic literature review on subjects relevant to thispaper; Section 3 provides an overall description of the research approach used; Section 4 briefly describes the cities studied,and provides a general description of the delivery routes considered; Section 5 presents the analyses of the GPS data and theestimates of emission reductions for the delivery routes studied; and Section 6 discusses the estimates produced by theauthors of total emissions at the studied cities, and for all metropolitan areas in the world with more than two million res-idents. The final section discusses the policy implications that emerge from this research, and provides concluding remarks.

2. Literature review

This section discusses research publications associated with: OHD; emission modeling with an emphasis on freight vehi-cles; and the impacts of freight activity on emissions. OHD programs seek to foster a switch of freight delivery operationsfrom regular business hours to off-hours to reduce the congestion and environmental impacts produced by freight activity.Interest in OHD has recently increased because of its successful implementation in NYC. The NYC-related OHD research—which started at the request of the private sector—gained insight into the behavioral attitudes toward OHD of the variousagents involved, and the identification of effective policies to foster OHD (Holguín-Veras, 2006). The research analyzedthe strategic interactions between shippers, carriers, and receivers, and concluded that receivers are the ones that decideon delivery times (Holguín-Veras et al., 2007; Holguín-Veras, 2011). This insight explained the failure of time-of-day pricingto influence delivery times, first reported in Holguín-Veras et al. (2006b). Moreover, the research revealed that offering acombination of incentives to receivers—such as financial incentives, shipping discounts, and public recognition—in exchange




for their commitment to accept OHD, was the most effective policy (Holguín-Veras et al., 2006a; Holguín-Veras et al., 2007;Holguín-Veras et al., in press). In a clear demonstration of effectiveness, these incentives were found to be more impactfulthan road pricing (Holguín-Veras et al., 2008; Holguín-Veras et al., 2015b).

The incorporation of research findings into the design of the NYC OHD program was a key factor in the success of twolarge OHD projects in NYC. The first project was the pilot test, conducted in 2009, in which 35 companies participated(Holguín-Veras et al., 2010). The second project marked the launch phase of the OHD program in NYC, which, because ofthe success of the first pilot, involved more than 400 companies (Holguín-Veras et al., 2013b). Satisfaction surveys indicatethat all involved—shippers, carriers, receivers, local communities—benefit from OHD (Brom et al., 2011). Economic analysesconfirmed that OHD programs are beneficial in congested metropolitan areas. Holguín-Veras et al. (2010), Holguín-Veraset al. (2011), and Holguín-Veras et al. (2012b), simulated the effects of increases in OHD using the regional transportationdemand model for the NYC metropolitan area to quantify the economic impacts. The results indicated that increasingOHD generates benefits in the range of $100 to $200 million/year, in both emission reductions and travel-time savings. Sim-ilar results were produced by Holguín-Veras et al. (2013b) and Yannis et al. (2006), who assessed the traffic and environmen-tal impacts for OHD scenarios in Athens, Greece.

The NYC experience has led to pilot tests in other cities. A pilot study conducted in Belgium (Verlinde et al., 2010;Verlinde et al., 2011) concluded that OHD is beneficial for receivers and carriers, as the extra labor costs are compensatedfor by increased operational productivity. Of great relevance to this paper are the pilot tests conducted by the authors inBogotá and São Paulo, as they provide part of the data used in this research. São Paulo’s 2014 OHD pilot was a collaborationbetween the Government of São Paulo, the Union of São Paulo Freight Transport Providers (SETCESP), and the University ofSão Paulo’s Center for Logistics Innovation. It involved 11 companies that conducted OHD for about four weeks. Bogotá’s pilotprogram of OHD was implemented between December 2015 and February 2016 (COE-SUFS, 2016b). The program was apublic-private-academic effort between the Chamber of Commerce, the Secretariat of Mobility, Invest in Bogotá, the Univer-sidad Nacional de Colombia, and the 17 companies that conducted OHD for two weeks. In both Bogotá and São Paulo, theparticipants were highly satisfied with the results, and are currently expanding the program. Moreover, the ColombianGovernment decided to conduct pilot tests of OHD in the cities of Barranquilla, Medellin, and Cali.

Recent research on OHD has assessed the environmental impacts, the role of OHD in a sustainability strategy, and the useof novel policy measures to foster OHD. Palmer and Piecyk (2010) reported CO2 emission reductions of up to 11% with theimplementation of OHD programs, while Zanni and Bristow (2010) identified OHD as a feasible policy for long-term reduc-tions in CO2 emissions. Jaller et al. (2016) qualitatively assessed a number of potential sustainability programs and concludedthe largest benefits could be attained by OHD programs. Holguín-Veras et al. (2013a) discussed the role of low-noise deliverytechnologies and practices to support OHD. Holguín-Veras et al. (in press) used behavioral models to identify the best waysto foster unassisted OHD. Holguín Veras et al. (in press) evaluated the potential benefits of a program that helps receiversinterested in OHD find vendors that could be trusted to make unassisted OHD. Other publications discuss the various lessonslearned during the process of implementing OHD programs (Holguín-Veras, 2013; Holguín-Veras et al., 2014).

Transportation related emissions, and ways to reduce them, are subjects of great importance to the research reported inthis paper. These emissions—in the form of carbon monoxide (CO), nitrogen oxides (NOx), and non-methane hydrocarbons(NMHC)—are all precursors to ozone pollution, and of major concern because of their health impacts (Lena et al., 2002; Linand Vallamsundar, 2013). The Environmental Protection Agency (2014) has compiled numerous studies that show that expo-sure to ambient concentrations of ozone is associated with detrimental respiratory, cardiovascular effects, reproductive,developmental and central nervous system effects, and mortality in the case of long-term exposure. (Lepeule et al., 2012).Estimates of health impacts attributed to emissions of PM2.5 caused by congestion in 38 urban areas, estimate that the valueof the mortality caused by these emissions in 2000 reached US$37 billion (Levy et al., 2010). Carbon monoxide concentra-tions are also related to adverse health effects, such as cardiovascular morbidity, particularly increased emergency room vis-its and hospital admissions for coronary heart disease (Greco et al., 2016).

Given the importance of transportation-related emissions, a large body of research focuses on emission modeling., A greatdeal of this research is on idle emissions, a major issue at large generators such as ports and intermodal terminals, wheretrucks idle to generate power for air conditioning or heating units. McCormick et al. (2000) reported that CO idle emissionsranged from 5% to 75% of total emissions. Although there are several studies on idle emissions, a smaller number of publi-cations focus on the role of logistics and operational practices in reducing emissions. Brodrick et al. (2002) quantified theeffects of accessory loadings and increased speeds during idling, and reported that emissions could increase by 33–58% whenusing air conditioning and other accessories.

One of the first comprehensive research efforts regarding emission estimation using traffic flows and vehicle speeds isfound in Anderson et al. (1996). They integrated transportation and emission models to assess emission patterns and theimpact of congestion, demonstrating that emissions are highly sensitive to variations in vehicle speeds due to congestion,and that urban traffic produced major environmental impacts. Other researchers reached similar conclusions. Choudharyand Gokhale (2016) found that frequent accelerations and decelerations—of the kind found in congested traffic—sharplyincrease emission levels. Zhang et al. (2011) found similar results for light-duty and heavy-duty vehicles.

Research has already established that traffic dynamics—such as speed and acceleration changes—must be considered toaccurately quantify emissions (Poudenx, 2008). The Comprehensive Modal Emission Model, CMEM (Center forEnvironmental Research and Technology, 2016) is one of the earliest models that is able to capture traffic dynamics. CMEMestimates second-by-second tailpipe emissions and fuel consumption. An important step toward more accurate emission




models was the development of MOVES by the Environmental Protection Agency (2012). MOVES—more comprehensive thanMOBILE—accounts for various driving modes and estimates emissions using second-by-second activity and speed data(Vallamsundar and Lin, 2011). For comparisons between these models, see (Bai et al., 2009; Zhang et al., 2011; Lin andVallamsundar, 2013; Vallamsundar and Lin, 2013).

A minimal body of research has quantified the environmental impacts of freight activity at the urban level. One of the fewstudies is Segalou et al. (2004), which assessed freight’s environmental impacts in three French cities. They quantified thecontribution of urban freight activity to energy consumption, congestion, pollution, greenhouse gases, and noise. They foundthat on average, urban freight traffic contributes to more than 40% of peak concentrations for NOx and PM, and up to 20% forCO and HC. A similarly small number of studies have estimated freight emissions in Latin America. In most cases these stud-ies have used emission factors (Álvares Jr and Linke, 2001; Corvalán et al., 2002; Zárate et al., 2007; Cortés et al., 2008;Manzoli, 2009; Rodríguez et al., 2016).

The main insight gained from this review is that very little research has assessed the performance of OHD, or any otherform of FDM, in the context of a detailed, micro-scale emissions analysis. The research reported in this paper addresses thatgap. The next section describes the research approach implemented by the authors.

3. Research approach

The data used in this research were collected at the cities of Bogotá (Colombia); NYC (USA); and São Paulo (Brazil), as partof five different projects in which the authors were involved. The first two projects are the 2009 OHD pilot test (Holguín-Veras et al., 2010; Holguín-Veras et al., 2011) and its second phase that launched the implementation of OHD in NYC(Holguín-Veras et al., 2013b). The third and fourth projects are the pilot tests of OHD conducted by the authors in the citiesof Bogotá and São Paulo. The fifth source of data was a project funded by the Inter-American Development Bank that col-lected data about the impacts of congestion on supply chains in Latin America. Taken together, the data provide a uniquepicture of the emissions produced in a sample of important cities in the Americas. The cities selected exhibit a wide rangeof conditions in terms of traffic patterns, freight traffic regulations, congestion levels, logistical patterns, and driving cultures,among others. If not accounted for, this heterogeneity would make it difficult to compare and contrast results. Fortunately,throughout these projects the team used variants of the research approach shown in Fig. 1, which was designed to enhancecomparability. The approach uses receivers as the anchors of the data collection, enabling the authors to collect data fromsimilar locations, industry sectors, and congestion levels. This leads to more comparable results across cities and deliveryroutes, as the level of heterogeneity in the data is significantly smaller.

The first step of this approach was to identify large receivers in industry sectors whose primary economic activity involvedthe production or consumption of goods. These receivers had to be located in areas of congestion in each city, and they werethen asked to help convince their suppliers to participate in the research. The suppliers that agreed to participate were givenGPS data loggers to install in their trucks, to collect travel data to estimate congestion, delays, and ultimately freight emis-sions. These loggers collect the date, time, location, altitude, speed, and direction of the vehicle. The data were for two to fourweeks of RHD, and another two to four weeks of OHD. These data were ‘‘cleaned” to remove data for routes that were not partof the study, or anomalous observations produced by GPS signal errors caused by signals bouncing from buildings and othertall objects, among other factors. Table 1 provides basic information on the participating receivers and suppliers.

A total of 38 routes were chosen for study. Of those, 6 correspond to NYC, 15 from Bogotá, and 17 from Sao Paulo, account-ing for both RHD and OHD measurements. To the extent practicable, the team and participating companies tried to ensurethat RHD and OHD delivery routes were identical. This, however, proved to be a major challenge because quite frequentlysome of the receivers in the routes were different. In total, twenty OHD routes were included in the analysis (four fromNYC, ten from Sao Paulo, and six from Bogotá).

4. Case studies

Table 2 shows summary statistics for the distances traveled on the delivery routes considered in the research. As shown,the delivery routes exhibit a wide range of conditions. The shortest routes correspond to Bogotá, and the longest routes are

Engage receivers in geo-graphic areas of interest

Engage receivers’ suppliers; install GPS devices

Clean GPS data

Collect GPS data for dura-tion of study period

Estimate emissions and ana-lyze results

Fig. 1. Outline of research approach.



Table 1Participating receivers and suppliers.

City Receiver Suppliers

Bogotá Alkosto, a chain of stores that sells groceries, household products, home appliances,electronics and tires

Bavaria, a beer and beverage supplierNutresa, a food processing conglomerateCoimpresores, a printing and marketingbusinessEmpacor, a packaging facilityDoria, a pasta manufacturer

New YorkCity

Grocery stores in Manhattan Logistic company

Sao Paulo GPA (Groupo Pão de Acúcar), the largest private employer in Brazil Danone, a producer of dairy productsFEMSA, a manufacturer of soft drinks andwater


from São Paulo and NYC. Fig. 2 shows GIS maps depicting the delivery routes. In all cases, the freight vehicles used weremedium-sized diesel trucks. In the case of NYC, the fleet included vehicles manufactured between 2005 and 2008. The trucksin Bogotá and São Paulo were manufactured between 2000 and 2005.

The two sets of values for NYC, labeled ‘‘original” and ‘‘shortened” deserve explanation. When deciding which routes toanalyze, the authors selected those for which the GPS data seemed to be the best (both RHD and OHD). In NYC, the RHD andOHD routes selected contain the same set of receivers, which ensures maximum comparability. However, since these routesare atypically long (as shown in the ‘‘original” values, the range is between 128 and 448 km) they are not representative oftypical delivery routes, which are generally shorter. To correct this situation, while using the best data available, the authorsdeleted from the ‘‘original” data the first 100 km, from the real origin up to a ‘‘simulated” fictitious origin location near theintersection of I78 and I287 (see panes c and d in Fig. 2). The resulting data are labeled ‘‘shortened.” Since the rest of the GPSdata are exactly the same, the net effect of this deletion is a delivery route that is more representative of typical routes. Briefdescriptions of the cities, together with maps of the delivery routes studied, are provided next.

4.1. Bogotá, Colombia

Bogotá is Colombia’s capital city, and houses 23% of the country’s population. With 7.9 million inhabitants in themetropolitan area, the Bogotá region contributes an estimated 25% of the Gross Domestic Product, and houses 35% of thebusinesses in the country (DANE, 2014). Although traffic congestion is a major problem, most public-sector efforts havefocused on passenger transport, not freight. The participants were in the food, accommodation, and retail sectors. The mainresearch partner was the Universidad Nacional de Colombia at Bogotá.

4.2. New York City, USA

With about 20.63 million residents, the NYC metropolitan area is the largest in the US. Its economic importance cannot beoverstated; statistics show that the number of non-agricultural jobs in the region amounts to 9.5 million (US Bureau of LaborStatistics, 2016). It is an important financial center, and also plays a major role in high technology, real estate, and insurance.About 60 million tourists visit the city each year (McGeehan, 2016), stimulating the food and accommodation sectors. TheRensselaer Polytechnic Institute was in charge of the research.

4.3. São Paulo, Brazil

São Paulo, the largest metropolitan area in Brazil with a population of 20.36 million, generates about 18% of the country’sGDP (IBGE, 2015; SEADE, 2015), and suffers from heavy congestion. Of note is the city’s restriction of large trucks during thedaytime hours, which impacts about 100 km2 surrounding the downtown area (Cunha et al., 2016). Since receivers still insiston RHD, shippers and carriers are forced to use small 4-ton trucks (producing more congestion than the original large trucks).The University of São Paulo led the research.

Table 2Distance traveled in kilometers (km).

City Mean Standard deviation Minimum Maximum Number of routes

RHD OHD

Bogota, Colombia 19.73 14.33 4.56 62.08 10 5NYC, USA (original) 352.36 79.68 128.00 447.93 2 4NYC, USA (shortened) 234.70 71.25 109.61 336.87 2 4Sao Paulo, Brazil 51.82 45.30 23.71 176.79 5 10



(a) Bogotá, Colombia (b) São Paulo, Brazil

(c) New York City, US (original) (d) New York City, USA (shortened)

Simulated origin

Actual origin

Fig. 2. Routes studied per city.


As a prelude to the discussion of emission results, it is important to discuss data quality to help the readers interpret thefindings. As indicated in the introduction, the data used in the paper come from five different projects. In spite of the author’sefforts to ensure comparability, using data frommultiple sources and, in some cases, collected for entirely different purposes,leads to multiple levels of data quality. Table 3 shows a summary of the data used, and the authors’ assessment of dataquality.

The authors believe that the best data are those for NYC, for the simple reason that the data were collected for RHD andOHD routes with very similar sets of receivers. In the case of Bogotá, although the pilot collected data for the same receiversin both RHD and OHD, the OHD took place in the 6 PM to 10 PM timeframe, as opposed to the late night and early morningschedules in NYC and São Paulo (during the 7 PM to 6 AM period). The São Paulo data are ‘‘third-best,” because the receiversin the RHD and OHD routes do not always match, because the data were collected from two different projects for differentpurposes (the RHD were collected for a study of the impacts of congestion on supply chains, while the OHD were collected aspart of the OHD pilot). However, since the RHD data and OHD generally cover the same parts of the city, the authors decidedto use them.

As shown, the data represent two different modalities of OHD. In the case of São Paulo and NYC, the OHD were made dur-ing the entire off-hour period (7 PM to 6 AM), when overall congestion is lightest; in Bogotá, the OHDwere conducted duringthe initial part of the off-hour period (6 PM to 10 PM), when congestion is waning, though present. For ease of reference, theformer is referred to as ‘‘Full-OHD” and the latter as ‘‘Partial-OHD.”

It is also important to gain insight into the main travel patterns captured in the GPS data. To this effect, the authors con-ducted descriptive analyses of the instantaneous speeds measured by the GPS devices, typically, every three seconds. In addi-tion, the values of the instantaneous speeds for the cities with the best data (Bogotá and NYC) were used to producecumulative distribution functions (CDF) for RHD and OHD. These CDFs provide a way to assess the changes produced by

Table 3Data sources used and overall level of quality.

City RHD data OHD data Remarks Quality

NYC OHD projects OHD projects (7:00 PM–6:00 AM) Same receivers BestBogotá OHD pilot OHD pilot (6:00 PM–10:00 PM) Same receivers 2nd bestSao Paulo IADB/OHD pilot OHD pilot (7:00 PM–6:00 AM) Different receivers 3rd best




the change of time-of-travel. When producing the CDFs, the authors eliminated the observations with speeds equal to zero tofocus on speeds while actually traveling. See Fig. 3.

As shown in pane a of Fig. 3, the net effect of a Full-OHD operation is the lowering of the frequency of the speeds between30 kph and 80 kph, with a concomitant increase in the speeds between 80 kph and 110 kph. Most likely, these results reflectthe effects of the faster travel at the arterials and highways leading to, and inside, New York City. In the case of Bogotá,shown in pane b, the Partial-OHD leads to a shift toward faster speeds that is visible for the speeds between 10 kph and45 kph. The descriptive statistics for speeds are shown in Table 4.

Readers should interpret these descriptive statistics with caution. The comparison of the median and average values forRHD and OHD in NYC may lead readers to conclude that RHD is better than OHD (after all the average speeds are about thesame, and the median speed for OHD is lower than the one for RHD). In reality, the increased of the frequency of ‘‘high”speeds (the ones faster than 80 kph), which is accompanied by a lowering of the frequency of speeds between 30 kph and80 kph) leads to significant emission reductions, notwithstanding the fact that the statistics of central tendency do not reflectthese impacts.

5. Emission estimates by time-of-day

To estimate the emissions produced by the delivery activity, the authors developed a computer program that reads theGPS data and applies appropriate emission factors corresponding to truck type, model year, fuel type, and speed, over theentire course of the route. This software has two major components that focus on data preparation and emission estimation.As part of the data preparation, the program reads the GPS data, ensures that they have the right format, conducts a diagnosisto identify problems, and prepares the second-by-second data needed for the emission calculations. In cases where the GPSdata are not provided second-by-second, the program interpolates speeds and locations from the data available. Since inmost cases the data collected came in intervals of three seconds, this interpolation has minimal consequences on quality.

The emission factors used in the software come from the EMFAC Web Database (California Environmental ProtectionAgency, 2016). The authors selected the EMFAC factors because: (1) they could be directly used to estimate emissions ona second-by-second basis (a key consideration for cases such as this, which rely on the use of GPS data); and (2) they providean adequate level of detail in terms of vehicle type, fuel type, age, and the like. To expedite computations, the original emis-sion factors in a tabular format were replaced with statistical models using various non-linear forms. The models express thevarious pollutants (the dependent variables) as functions of vehicle type, speed, age of the trucks, and type of fuel. In mostcases, the models had a very high level of explanatory power with coefficients of correlation always exceeding 0.88. To con-sider imprecise information about the age of the truck, a group of models were estimated using five-year intervals of age. Thepollutants estimated are carbon dioxide (CO2), carbon monoxide (CO), nitrogen oxides (NOX), total organic gases (TOG), reac-tive organic gases (ROG), and particulate matter with aerodynamic diameters of less than 10 lm and 2.5 lm (PM10 andPM2.5, respectively). The emissions in idling conditions were estimated—in those cases where the team was able to identifythat idling did occur—following the guidelines suggested by the EMFAC Web Database (California Environmental ProtectionAgency, 2016).

The remainder of this section presents the results for both RHD and OHD for each city. In each case, total emissions bydelivery route and average emissions per kilometer for both RHD and OHD are shown. Although the total emissions per

a) New York City (Full-OHD, 6PM to 7AM) b) Bogotá (Partial-OHD,6PM to 10AM)

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 20 40 60 80 100 120 140Speed (kph)

RHD routes

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 20 40 60 80 100

Speed (kph)

OHD routesOHD routes

RHD routes

Fig. 3. Cumulative distribution function of instantaneous speeds.



Table 4Descriptive statistics of instantaneous speeds.

City Bogota NYC

OHD RHD OHD RHD

Average 24.21 22.23 69.80 69.67St. Dev 14.80 14.75 37.24 41.67Median 23.00 19.00 84.75 90.12

Min 1.00 1.00 1.00 1.00Max 68.00 83.00 125.00 118.88Count 10,642 35,245 56,245 14,264


delivery route provide an indication of the environmental impact of making deliveries and pick-ups to all of the customersalong the route, these results are not easily comparable across cities and routes because the distances traveled in each routeare different. Therefore, average emissions per kilometer for each route are provided for comparison purposes. However,readers should recognize that these averages include both low and high congestion segments, so the worst impacts maybe masked by the use of the average. Nevertheless, taken together, these measures provide a solid picture of the environmen-tal effects of RHD vs. OHD.

Tables 5 and 6 show the average total emissions for the delivery routes studied for both RHD and OHD, respectively.Table 7 shows the percentage reductions of these emissions due to a shift to OHD. Tables 8 and 9 show the average emissionsper kilometer for both RHD and OHD, respectively. Table 10 displays the differences in emissions per km between RHD andOHD.

The results shown in Table 10 deserve further discussion. To start, it is important to mention that these values estimatethe direct emission reductions attributable to OHD, or alternatively, the incremental emissions produced by congestion dur-ing regular-hours. The table shows the differences among the emission rates per kilometer between RHD and OHD. Asshown, the emission reductions for all cities are fairly similar, in spite of the significant differences in terms of traffic control,infrastructure, network topology, and urban form.

Table 11 shows the percentage reduction of these emissions per kilometer due to a shift to OHD. As shown, the emissionreductions are considerable, both in terms of total emissions and emissions per km. In the case of Bogotá, NYC and Sao Paulo,the emissions estimated from the actual GPS data indicate that potential savings are in the range of 13%, 64%, and 48%respectively.

To illustrate the geographic patterns of emissions, the authors created an emission-space diagram showing the cumula-tive emissions in the vertical axis, and the distance traveled in the horizontal axis. Although dozens of emission-space dia-grams could be produced for all of the combinations of pollutants and cities, it is not practical to do so in this paper. Theauthors decided to present the NYC results for CO2 because they are based on the best data available, and provide insighton an important pollutant. See the results in Fig. 4.

The results in Fig. 4 are noteworthy in numerous ways. First, there is the effect of the congested Hudson River crossings.As shown, both RHD and OHD vehicles travel without experiencing major delays from the origin up to about km 60, wherethey encounter (congested) traffic trying to cross the Hudson River. The figure shows that during the regular-hours morethan half of the CO2 is produced in the highways leading and exiting the crossings (the location of the river crossings seemsto shift between km 60 and km 110 because the distance traveled to the crossings depends on the route followed by thetruck). In the OHD routes, the congestion at river crossings is less severe, which leads to significant reductions in CO2. A sec-ond effect is that the RHD routes—in spite of being shorter in length—produce more CO2 than all but one of the OHD routes,which are, on average, more than twice as long. The factors behind this pattern deserve further discussion. The sharpincreases in CO2 emissions at river crossings are caused by the severe congestion that prevails in these areas. This congestiongreatly reduces the amount of time (and thus the distance traveled) that trucks have to make deliveries to customers; timethat is already limited by Hours-of-Service regulations.

In contrast, the faster travel during the off-hours enables the carriers to deliver to more customers (or deliver more cargo)than would be possible during the regular-hours. This leads to a reduction in the number of tours, and freight VKT during theregular-hours. For instance, a carrier with four delivery routes delivering to five customers in each route, during the regular-hours, may be able to replace two of the RHD routes with only one OHD route. Instead of four routes traversing the HudsonRiver crossings during the regular-hours, the combined OHD/RHD operation would only require two crossings during the

Table 5City average emissions per route (g/route) for regular-hour deliveries.

City/pollutant ROG (g) TOG (g) CO (g) CO2 (g) NOX (g) PM10 (g) PM25 (g)

Bogotá 17 19 242 90,809 37 6 6New York City 33 37 478 209,291 78 11 11Sao Paulo 19 21 194 78,056 31 6 5



Table 6City average emissions per route (g/route) for off-hour deliveries.

City/pollutant ROG (g) TOG (g) CO (g) CO2 (g) NOX (g) PM10 (g) PM25 (g)

Bogotá 12 14 172 63,794 26 4 4New York City 22 25 322 190,088 64 8 8Sao Paulo 5 6 51 23,996 9 2 2

Table 7Reductions in emissions per route due to OHD.

City/pollutant ROG TOG CO CO2 NOX PM10 PM25

Bogotá 29.33% 29.33% 28.86% 29.75% 29.10% 29.54% 29.54%New York City 32.92% 32.92% 32.68% 9.18% 17.58% 29.63% 29.63%Sao Paulo 74.00% 74.00% 73.75% 69.26% 70.09% 71.51% 71.51%

Table 8Average city regular-hour emissions per kilometer.

City/pollutant ROG (g/km) TOG (g/km) CO (g/km) CO2 (g/km) NOX (g/km) PM10 (g/km) PM25 (g/km)

Bogotá 1.084 1.235 15.475 5708.064 2.318 0.376 0.359New York City 0.247 0.281 3.607 1548.156 0.581 0.086 0.083Sao Paulo 0.313 0.356 3.394 1306.599 0.519 0.093 0.089

Table 9Average city off-hour emissions per kilometer.


Bogotá 0.938 1.068 13.385 4959.141 2.024 0.325 0.311New York City 0.081 0.092 1.190 694.450 0.236 0.030 0.028Sao Paulo 0.156 0.178 1.649 751.057 0.287 0.050 0.048

Table 10Off hour emissions reductions per kilometer.


Bogotá 0.146 0.166 2.090 748.923 0.294 0.050 0.048New York City 0.166 0.189 2.416 853.705 0.346 0.057 0.054Sao Paulo 0.156 0.178 1.746 555.542 0.232 0.043 0.041Grand average 0.156 0.178 2.084 719.390 0.291 0.050 0.048Standard deviation 0.010 0.011 0.335 151.260 0.057 0.007 0.007Number of cases 38 38 38 38 38 38 38CI Upper Limit 0.165 0.187 2.378 851.975 0.341 0.056 0.054CI Lower Limit 0.148 0.168 1.790 586.805 0.241 0.044 0.042

Note: ‘‘CI” denotes 95% confidence interval.

Table 11Off hour percentage of emissions reductions per kilometer.

City/pollutant ROG TOG CO CO2 NOX PM10 PM25

Bogotá 13.49% 13.49% 13.50% 13.12% 12.70% 13.41% 13.41%New York City 67.17% 67.17% 67.00% 55.14% 59.47% 65.53% 65.53%Sao Paulo 49.98% 49.98% 51.43% 42.52% 44.64% 45.90% 45.90%


regular-hours, plus one during the off-hours. OHD could lead to a reduction in total truck VKT because the OHD routes wouldbe able to deliver to more customers than the RHD routes. This important substitution effect is not accounted for in thispaper, as the estimates in Table 10—based on the emission rates per kilometer—implicitly assume that the total freightVKT is the same. The emission reductions obtained in this paper are a lower bound of the potential emission reductions thatOHD could produce.



0

50000

100000

150000

200000

250000

0 50 100 150 200 250 300 350

CO

2 (g

)

Distance (km)

OHD1

OHD2

OHD3

OHD4

RHD1

RHD2

Hudson river crossings

Fig. 4. CO2 cumulative emissions by distance for RHD and OHD (NYC).


6. Total emission reductions: Cities studied and global

The consistency of the emission reduction figures shown in Table 10 suggests that they could be indicative of the totalemission reductions that could materialize in these cities, and others around the globe, as a result of the implementationof OHD. These estimates are very important for policy-making purposes, because they provide an indication of the overallworthiness of an OHD program. Obviously, there are major differences in the ways deliveries are made from country to coun-try, and from city to city. In some countries, like India and Bangladesh, the relatively large role of human-powered transportreduces the reliance on diesel trucks; in developed countries, trucks tend to be newer and better maintained than those inmost developing countries. There are also major differences in traffic control systems and transportation infrastructures. Rel-atively simple estimation procedures, like the ones used in this paper, cannot account for these important differences, nor dothey have to, as the main objective of the analyses here is to produce order-of-magnitude estimates of global emission reduc-tions from OHD. In the absence of hard data, reasonable estimates could be produced assuming that:

(1) the number of deliveries in the cities is generated at a rate of 0.03 deliveries/person-day (except for NYC, which is 0.06deliveries/person-day);

(2) every delivery generates one additional kilometer of travel (i.e., a portion of the distance traveled from the warehouseto the delivery area, plus the average inter-customer distance);

(3) 30% of these deliveries are made with midsize trucks like those studied in this research;(4) 10% of the deliveries are switched to OHD;(5) The average reductions in emissions are equal to the differences between RHD and OHD average emissions estimated

in this paper; and(6) There are 260 days a year of average travel conditions.

It is important to explain the rationale for these assumptions. The freight trip generation rate for NYC comes from datacollected in NYC (Holguín-Veras et al., 2012a). The rate for cities in developing countries was assumed to be 50% of NYC’s. Itis worth noting that these estimates do not take into account Internet-related deliveries, which in NYC are generated at a rateestimated by the authors to be about 0.10 deliveries/person-day.

The estimation of the vehicle-km generated by one delivery—defined as the total tour distance divided by the number ofdeliveries made—is a challenge, as there are no publicly available data that could be used. The estimates produced by theauthors for NYC using the Behavioral Micro Simulation suggest a value of 9 km (Silas and Holguín-Veras, 2009; Holguín-Veras and Aros-Vera, 2014). The estimates from Holguín-Veras and Brom (2008) based on survey data reveal that, in NYC,the average tour distance and number of delivery stops are 120 km, and 4.54 stops/tour. Assuming two deliveries for eachstop, a single delivery would create 13.66 vehicle-km. However, since these estimates correspond to deliveries to Manhattan,they are not necessarily representative of the larger metropolitan area. Since Manhattan is at the center of the metropolitanarea, delivering or picking up cargo there requires longer travel distances from distribution centers located in the peripheryof the urban core. There are no such estimates for the other cities. Assuming that one km is generated by one delivery is aconservative assumption.




Similarly, there are no data about the percent of deliveries made using midsize trucks. In NYC, no freight origin-destination studies have been conducted in more than five decades. Most recent estimates (Wood and Leighton, 1969;Wood, 1970), place the number of trips by midsize trucks at 38.80%. Anecdotal evidence indicates that in the last severalyears there has been an increase in the number of deliveries made using light freight vehicles, such as delivery vans. Thus,assuming 30% as the market share of mid-size trucks seems reasonable.

The last three assumptions are straightforward. The shift to OHD is based on the estimates made by the authors (Holguín-Veras et al., in press). The assumption that the emission reductions produced in the paper are representative of overall con-ditions is reasonable, though the reductions could be overestimated for deliveries made to less congested areas. Using260 days of average congestion level is also conservative, as it ignores the emission reductions associated with weekendactivity. The combined effects of these assumptions are conservative estimates of the total reductions in emissions. Theresults are shown in Table 12, in descending order of impacts.

The admittedly approximate estimates of Table 12 suggest that OHD programs could have a tremendous impact on emis-sions. As shown, 13.5 million tons of CO2 could be eliminated every year, together with 39,051 tons of CO, 5484 tons of NOX,and 928 tons of PM2.5. By any account, these impacts are significant. The largest reductions would be seen in NYC, followedby São Paulo, Bogotá.

The results in Table 12 can be used to estimate the impacts that could be realized if an OHD program were to be imple-mented on a worldwide scale. The procedure used to produce these global estimates exploits an important relation betweenpopulation, employment, and freight activity. The data show that, absent government policies that prevent the free move-ment of individuals, population tends to be highly correlated with employment, and employment is highly correlated withfreight activity. As a result, population is a very good proxy to estimate the number of freight trips that are generated. Theseinsights suggest that the estimates in Table 12 could provide an idea of the potential emission savings that would accruewith a switch to OHD in similar cities throughout the world, under the assumption that the urban core populations are pro-portional to the population living in the metropolitan area. To operationalize this procedure, the estimates of emission sav-ings shown in Table 12 for NYC—the best data available—were divided by the total population in the NYC metropolitan area(20.63 million in 2016). Further, the authors assumed that the per capita emission reductions for developed countries wouldbe equal to the ones for NYC; and that the per capita emissions for developing countries would be equal to half the values ofthe developed countries (to account for the differences in FTG rates). The results are shown in Table 13.

The values in Table 13 provide an order-of-magnitude estimate of the emission reductions that could be achieved with anOHD program that shift 10% of the deliveries made by midsize trucks to the hours of 7 PM to 6 AM (Full-OHD). Multiplyingthese numbers by the population in millions of a metropolitan area would estimate the potential emission reductions. Indoing so, it is important to only consider metropolitan areas for which OHD makes sense. For OHD to be a reasonably goodalternative, there must be a significant difference in the levels of congestion between the regular and the off-hours. Other-wise, OHD will not help. In metropolitan areas with low levels of congestion, OHD may not be worthy of considerationbecause the incremental costs of the effort are not compensated by the incremental benefits in terms of congestion andemission reductions. Although there are no studies that could shed light into what is the ‘‘minimum” size for OHD to bea practical solution, the authors estimate that the threshold of 2 million is a conservative value of the minimum populationsize for which OHD is likely to be a good option. Obviously, this is not an exact rule. OHD could indeed be feasible inmetropolitan areas of less than two million people, if they experience high levels of congestion. However, for purposes ofthis paper, the analyses used the two million people threshold as it is both sensible and conservative. The scenarios consid-ered involve a Full-OHD (7 PM–6 AM) program for all metropolitan areas of the world with populations: (1) of more than 15million people; (2) between 10 and 15 million people; (3) between 5 and 10 million people; and (4) between 2 and 5 millionpeople. The population estimates, from U. S. Census Bureau (2016), were broken down by population size and type of coun-try. See a summary in Table 14.

As shown, about 15% of the world’s population lives in metropolitan areas of more than 2 million inhabitants. Amongthose, about three quarters are from developing countries, with 34% living in the largest metropolitan areas of more than15 million people. Worthy of mention is the large number of cities in the 2–5 million range—many of them experiencingrapid growth—as these could quickly become hotspots of environmental emissions.

The estimates of emission reductions for the Full-OHD program are shown in Table 15. The results in the table clearlyshown that the bulk of the emission reductions would take place in the largest metropolitan areas, those with more than15 million people, followed by those in the 5–10 million range, then the ones in the 2–5 million range, and finally the onesbetween 10 and 15 million. Collectively, the emission reductions of CO2 would be in the range of 300 million tons/year; aconsiderable reduction by any standard.

Table 12Reductions in emissions due to OHD in studied cities.

City Veh-km shifted/year (millions) ROG (tons) TOG (tons) CO (tons) CO2 (million tons) NOX (tons) PM10 (tons) PM25 (tons)

New York 10.5 1733.2 1973.2 25273.5 8.9 3616.3 591.8 566.2Sao Paulo 5.2 806.9 918.6 9015.2 2.9 1197.0 221.0 211.4Bogotá 2.3 333.3 379.4 4762.2 1.7 671.0 114.8 109.8TOTAL 17.9 2873.4 3271.2 39051.0 13.5 5484.3 927.6 887.5



Table 13Emission reductions per capita (tons-per-year/million-persons).

Type of country ROG (tons/106-persons)

TOG (tons/106-persons)

CO (tons/106-persons)

CO2 (106-tons/106-persons)

NOX (tons/106-persons)

PM10 (tons/106-persons)

PM25 (tons/106-persons)

Develo-ped 84.015 95.646 1225.086 0.433 175.296 28.688 27.447Develo-ping 42.008 47.823 612.543 0.216 87.648 14.344 13.723

Table 14Total population in metropolitan areas by size and level of development (2016).

Country >15 million 10–15 million 5–10 million 2–5 million Total

Population %

Population of the metropolitan areaDeveloped 114.46 31.27 73.56 61.63 280.91 24.72%Developing 272.38 161.23 200.40 221.38 855.38 75.28%Total 386.84 192.50 273.96 283.00 1136.30 100.00%% 34.04% 16.94% 24.11% 24.91% 100.00%

Number of citiesDeveloped 5 3 12 16 36 24.00%Developing 13 13 29 59 114 76.00%Total 18 16 41 75 150 100.00%% 12.00% 10.67% 27.33% 50.00% 100.00%

Source: U. S. Census Bureau (2016).


The top 10 countries in terms of potential emission reductions from a Full-OHD program are shown in Table 16. It isimportant to mention that China is listed as the country with the highest potential for emission reductions, in spite ofthe fact that in most large Chinese cities there are daytime truck bans. Such bans, in theory, are supposed to lead to OHD.In reality, the authors’ conclusion—based on interviews with carriers and analyses of postings at official websites—is thatthese daytime delivery bans have been counterproductive. Without the concurrence of the receivers (who have not beeninduced to accept OHD), the carriers cannot change their operations to the off-hours. Faced with the prospect of losing cus-tomers, most carriers have opted to use passenger vehicles to make deliveries in the daytime hours. Obviously, since mul-tiple passenger vehicles are needed to transport the same amount of cargo transported by a truck, the net result is actually anincrease in congestion and pollution (Holguín-Veras et al., 2015a). Readers interested in the impacts of nighttime deliveryrestrictions are advised to examine the analyses of Holguín-Veras et al. (2015a).

Fig. 5 shows the top 25 countries in CO2 emission reductions from Full-OHD programs. The most obvious result is that therole of European and African cities becomes evident. The case of European cities deserves discussion because many of them,e.g., Brussels, have enacted very strict noise regulations that effectively prevent OHD from taking place. As a result, the 4–5%of deliveries that naturally would take place at night (Holguín-Veras et al., 2007; Domínguez et al., 2012) is forced to takeplace during the congested regular hours. These policies—notwithstanding the fact that there are technological and opera-tional options to make low-noise deliveries—artificially increase regular-hour congestion. If these restrictions are lifted, and

Table 15Yearly direct emission reductions in metropolitan areas (Full-OHD).

Population Type of country ROG (tons) TOG (tons) CO (tons) CO2 (106tons) NOX (tons) PM10 (tons) PM25 (tons)

>15 million Developed 9616 10,947 140,217 49.54 20,063 3283 3141Developing 11,442 13,026 166,847 58.95 23,874 3907 3738Total 21,058 23,973 307,064 108.49 43,937 7190 6879

10–15 million Developed 912 1039 13,302 4.70 1903 311 298Developing 6773 7710 98,757 34.89 14,131 2313 2213Total 7685 8749 112,059 39.59 16,034 2624 2511

5–10 million Developed 6180 7036 90,116 31.84 12,895 2110 2019Developing 8418 9584 1,22,753 43.37 17,565 2874 2750Total 14,598 16,619 212,869 75.21 30,459 4985 4769

2–5 million Developed 5178 5895 75,502 26.68 10,803 1768 1692Developing 9299 10,587 135,602 47.91 19,403 3175 3038Total 14,477 16,481 211,104 74.58 30,207 4943 4730

Totals Developed 21,886 24,916 319,137 112.75 45,665 7473 7150Developing 35,933 40,907 523,959 185.12 74,972 12,269 11,739Grand Total 57,819 65,823 843,096 297.87 120,637 19,743 18,889



Table 16Top ten countries in terms of emission reductions from a Full-OHD program.

Country ROG (tons) TOG (tons) CO (tons) CO2 (106 tons) NOX (tons) PM10 (tons) PM25 (tons)

China 10,512 11,967 153,286 54.16 21,933 3589 3434United States 8620 9813 125,690 44.41 17,985 2943 2816Japan 5500 6261 80,199 28.33 11,476 1878 1797India 4724 5379 68,891 24.34 9858 1613 1543South Korea 2301 2619 33,550 11.85 4801 786 752Brazil 2229 2538 32,503 11.48 4651 761 728Indonesia 1893 2155 27,599 9.75 3949 646 618Pakistan 1501 1709 21,889 7.73 3132 513 490Mexico 1208 1375 17,610 6.22 2520 412 395Italy 1081 1231 15,766 5.57 2256 369 353


Full-OHD programs are implemented, the emission reductions would be even larger than what have been estimated in thispaper.

To provide a fuller picture of the potential emission reductions by world regions, the authors post-processed the esti-mates of the reductions in CO2 emissions. Fig. 6 shows the cumulative CO2 emission reductions (by region of the world)in the vertical axis and the number of metropolitan areas considered in the horizontal axis. As shown, regardless of the typeof metropolitan areas that are included in the program, the largest emission reductions would materialize in Asian and theAmericas, where the bulk of the largest metropolitan areas are located. In all scenarios of implementation, Chinese citiescome up as the biggest beneficiaries of widespread implementation of OHD. Implementing OHD programs there is boundto have a major impact in reducing CO2 and ameliorating the chronic air pollution that afflict their residents.

The results clearly show that developing countries could play a major role in the fight against climate change, as about60% of the possible emission reductions would come from their cities. International support for developing countries toimplement FDM programs such as OHD could help accomplish environmental and economic objectives simultaneously.On the environmental front, OHD could help reduce emissions of GHG, and other pollutants known to negatively impactthe health of communities. The adoption of OHD would also lead to more efficient and productive urban supply chains. Itsuffices to mention that the OHD pilots in Bogotá, NYC, and Sao Paulo revealed cost savings in the range of 35% duringthe off-hours, which are significant given the fact that the so called ‘‘last mile deliveries” account for 28–53% of the totallogistic cost (McCrea, 2016). OHD could indeed have a major impact in boosting economic efficiency and productivity, of par-ticular importance in developing countries that typically struggle to compete in a global economy. Taking into account thatmetropolitan areas play a primary role in manufacturing, OHD programs will also increase the economic competitiveness ofurban areas. The opportunities that OHD could provide to improve both the environment and the economy should not bemissed.

Fig. 5. CO2 emission reductions in top 25 countries.



-

50

100

150

200

250

300

Top 10 Top 25 Top 50 Top 100 All

Redu

ctio

ns in

CO

2( 1

06to

ns) p

er y

ear

Metropolitan Areas Considered

Australia-NZ

East Europe

North Europe

West Europe

South Europe

East Africa

South Africa

Middle Africa

West Africa

North Africa

Cent. Am. & Carib.

South America

North America

West Asia

South-East Asia

South Asia

East Asia

Fig. 6. Reductions in CO2 emissions vs. metropolitan areas considered.


These analyses provide a picture of great optimism in terms of the potential of OHD, and FDM by extension, to bring aboutsubstantial emission reductions. The possibility of achieving reductions in the range of 45–67% in the case of a Full-OHD(7 PM–6 AM), and about 13% for a Partial-OHD program (6 PM–10 PM), should motivate policy makers to undertake proac-tive FDM programs. The fact that these changes could be implemented with private-sector support provides another com-pelling reason to take full advantage of FDM. Collaborative private-public sector relations are foundational to FDM; the effortexpended to develop such collaborations can be leveraged to address other challenges.

7. Policy implications and concluding remarks

The research reported in this paper estimates the potential emission reductions attributable to a shift of deliveries fromthe regular business hours to the off-hours, for midsize freight vehicles making deliveries in three different cities in theAmericas: Bogotá (Colombia), NYC (USA), and São Paulo (Brazil). To this effect, the authors collected GPS data for RHDand OHD, and developed a computer program that reads these data, and applies the corresponding second-by-second emis-sion parameters, to estimate the emissions of CO2, CO, NOX, TOG, ROG, PM10 and PM2.5. The estimates for the off-hoursrevealed substantial reductions in emissions. The results indicate that OHD produces emissions that are 13% (Partial-OHD) to 45–67% (Full-OHD) lower than those produced by deliveries made in similar routes during regular hours.

Using a set of conservative assumptions, the authors estimated the total emission reductions that could be achieved byOHD programs in the cities studied. The assumptions are that: (1) the rate of generation of deliveries is 0.06 deliveries/person-day for NYC, and 0.03 deliveries/person-day for the other cities; (2) each delivery generates one additionalvehicle-km of travel; (3) 30% of the total number of deliveries are made by mid-size trucks like the ones analyzed in thepaper; (4) 10% of these deliveries switch to the off-hours in response to the OHD program; (5) the average emission reduc-tions are equal to the differences between the average emissions for RHD and OHD; and (6) there are 260 days/year withtravel conditions like the ones in the GPS data.

Using these assumptions, the authors estimated the emission reductions for the three cities studied, and for metropolitanareas of more than two million people, if a Full-OHD program (where the deliveries are conducted during the 7 PM to 6 AMtime period) were to be implemented. The results indicate that such a program could lead to emission reductions of: 57.8thousand tons of ROG, 65.8 thousand tons of TOG, 843.1 thousand tons of CO, 297.8 million tons of CO2, 120.6 thousand tonsof NOx, 19.7 thousand tons of PM10, and 18.9 thousand tons of PM2.5. The research also shows that about 62% of the totalemission reductions would be accrued in developing countries; 36% in metropolitan areas of more than 15 million people,and 25% in metropolitan areas in the range of five to ten million residents. The chief insight is that OHD programs in devel-oping countries could generate tremendous emission reductions. In terms of climate change, OHD programs can help reducetransportation-related CO2 emissions; in terms of human health impacts, significant benefits can also be expected, though




precise estimates of these benefits would be a function of atmospheric fate, and transport and urban meteorology (Sathayeet al., 2010).

FDM programs like OHD present both a challenge and an opportunity. The implementation of FDM programs could be achallenge because they require a sustained public-private-academic effort to bear fruit. In cities where there is no tradition ofsuch collaborations, starting and sustaining the effort could be challenging. This was certainly the case in NYC, where puttingtogether the informal coalition that pushed forward the OHD program required great effort and perseverance. However, thedemonstration effect provided by NYC facilitated the process of coalition building in Bogotá and São Paulo, and provided atangible, and successful, example of what could be accomplished if all stakeholders work together. Moreover, the successfulpilots in Bogotá and São Paulo demonstrated that arranging a successful OHD program is obviously doable in the context ofdeveloping countries. The experience of these cities has stimulated other pilots. In Colombia, in response to the successfulpilot test in Bogotá, the public sector decided to orchestrate pilots in the cities of Barranquilla, Cali, and Medellin.

OHD programs could also improve the economic competitiveness of metropolitan areas, benefitting shippers, carriers,and receivers. This is of great relevance, as these areas are estimated to produce the lion’s share of the manufactures. It suf-fices to say that 80% of the cargo handled in the United States has its origin or destination in one of the top 100 metropolitanareas (Tomer and Kane, 2014). Increasing the productivity of urban supply chains will help reduce the cost of the last leg ofdeliveries, which is estimated to represent 28–53% of total logistic costs. Receivers benefit from shipping discounts and thesuperior reliability of OHD, which enables them to reduce their inventories, and business hour interruptions. These impactsprovide a major boost to local companies.

These substantial environmental and economic benefits are achievable without major investments on engine technology,or any other major expense that could be resisted by the freight industry. They are also achievable now, without waiting forany new/green technology development. These benefits are achievable working within present-day contexts of technologyand urban supply chain dynamics, and transferable to any/most metropolitan areas. In fact, since the industry stands to ben-efit through higher productivity and lower delivery costs, these types of programs could be implemented with their support,as clearly demonstrated by the pilot tests conducted in NYC, São Paulo, and Bogotá, in which industry enthusiastically par-ticipated. In these three cities, the pilot tests were followed by calls from the industry to expand the OHD program.

Significant as these benefits are, the real-life impacts of a full-fledged OHD program are likely to be larger than the esti-mates presented in the paper for two key reasons: (1) the estimates produced do not consider light-duty and heavy-dutyfreight vehicles; and (2) the other indirect environmental effects produced by OHD—such as the reductions in emissionsproduced by the rest of the traffic stream, and the potential reductions of total VKT as a consequence of higher productivityduring the off-hours—are not included. As mentioned, the data only included midsize trucks, so emission reductions inheavy-duty and light-duty trucks could not be estimated. Considering these vehicles will significantly increase the emissionreductions attributable to OHD. Moreover, OHD provides more benefits than the direct environmental effects studied in thispaper. As delivery vehicles switch to the off-hours they free up road capacity, leading to travel time reductions for the rest ofthe traffic during the regular-hours. In addition, the estimates in the paper do not take into account that, because of thehigher productivity in the off-hours, carriers could deliver to more customers or deliver more cargo than they possibly couldduring the congested day hours. As a result, fully utilized OHD routes are likely to reduce total freight VKT by means ofreducing the number of routes needed to deliver to all customers or transport a set amount of freight. This substitution effectwill increase emission reductions above and beyond the reductions associated with the implementation of OHD, while main-taining freight VKT constant (which is how the emission reductions in the paper were estimated). These effects will increasehuman health benefits and lower climate change risks.

These results hint at the magnitude of emission reductions that could be generated by a comprehensive FDM program.After all, OHD is only one of many types of FDM that could be implemented. Other FDM programs could increase the sus-tainability of urban supply chains that serve receivers for which OHD is not a good option (Holguín-Veras and Sánchez-Díaz, 2016). Staggered Delivery Programs incentivize receivers to spread out deliveries during the work hours, helping toreduce peak-hour congestion and emissions. Receiver Led Consolidation Programs reduce the number of deliveries thatarrive at a given location, helping to reduce freight peak-traffic, as well as the associated congestion and emissions. In Desti-nation Change Programs the deliveries are sent, not directly to the receivers, but to another location where they are consol-idated and then sent to the receivers using environmentally-friendly vehicles.

Table 17Summary statistics for Bogota regular-hour routes.

Variable Mean Std Dev Min Max

Distance 23.0 16.3 4.6 62.1ROG 17.1 11.5 4.2 40.2TOG 19.4 13.1 4.8 45.7CO 241.6 165.0 56.2 581.7CO2 90809.1 59938.1 24173.1 207626.3NOX 36.6 24.6 9.7 85.8PM10 5.9 4.0 1.5 13.8PM25 5.7 3.8 1.4 13.2



Table 18Summary statistics for New York City regular-hour routes.



Table 19Summary statistics for São Paulo regular-hour routes.



Table 20Summary statistics for Bogota off-hour routes.



Table 21Summary statistics for New York City off-hour routes.


Distance 278.5 33.9 254.5 336.9qROG 21.9 4.7 15.3 27.7TOG 24.9 5.4 17.4 31.5CO 321.8 64.6 229.2 402.4CO2 190087.8 13979.2 175581.2 210591.5NOX 64.2 8.2 54.7 75.3PM10 8.1 1.6 5.9 10.0PM25 7.7 1.5 5.7 9.6

Table 22Summary statistics for São Paulo off-hour routes.







Given the lack of research on FDM, it is not possible to ascertain with great confidence the number of freight delivery tripsthat could be retimed, consolidated, eliminated, or modified in any other way to increase the sustainability of the supplychain. However, there are reasons to believe that FDM has considerable potential to induce changes in a large number ofdeliveries. If one combines the freight trip generation for the industry sectors that researchers have found to be inclinedto respond to FDM initiatives, the resulting number represents 40–70% of the total freight trip generation in American cities.Although not much is known about how many of these trips could change behavior in response to FDM, the potential toinfluence such a large portion of the total freight activity should provide enough stimulus to conduct further research in thispromising field.

The authors believe that the stage is set for a broadening of the public-sector initiatives used to reduce freight emissionsin congested non-attainment areas. It seems clear that FDM programs could indeed lower freight traffic emissions and,equally important, could do so with the enthusiastic support of the private sector. The opportunity presented to the publicsector by FDM is extremely unique, and should not be squandered. The multifaceted complexity of today’s environmentalchallenges, and the potential consequences of inaction require the use of all possible approaches towards the bettermentof environmental conditions. Designing and embracing such innovative programs will be key to improving the environmen-tal performance of freight systems.

Acknowledgements

This research is made possible by partial funding frommultiple sources, including: the VREF Center of Excellence for Sus-tainable Urban Freight Systems, the Inter-American Development Bank grant ‘‘Methodology to Assess the Impacts of Conges-tion on Supply Chains”, and the Commercial Remote Sensing and Spatial Information Technologies’ project ‘‘IntegrativeFreight Demand Management in the New York City Metropolitan Area: Implementation Phase,” which is part of the UnitedStates Department of Transportation’s Office of the Assistant Secretary for Research and Technology (USDOT’s OASRT).

A: Summary statistics for regular-hour deliveries

See Tables 17–19.

B: Summary statistics for off-hour deliveries

See Tables 20–22.

References

Álvares Jr, O.D.M., Linke, R.R.A., 2001. Metodologia Simplificada de Cálculo das Emissões de Gases do Efeito Estufa de Frotas de Veículos no Brasil. CETESB,São Paulo 14 (06).

Anderson, W., Kanaroglou, P., Miller, E., Buliung, R., 1996. Simulating automobile emissions in an integrated urban model. Transport. Res. Record: J.Transport. Res. Board 1520, 71–80. http://dx.doi.org/10.3141/1520-09.

Bai, S., Eisinger, D., Niemeier, D., 2009. MOVES vs. EMFAC: A Comparison of Greenhouse Gas Emissions Using Los Angeles County. Transportation ResearchBoard 88th Annual Meeting: 09-0692.

Brodrick, C.-J., Dwyer, H.A., Farshchi, M., Harris, D.B., King Jr., F.G., 2002. Effects of engine speed and accessory load on idling emissions from heavy-dutydiesel truck engines. J. Air Waste Manag. Assoc. 52 (9), 1026–1031. http://dx.doi.org/10.1080/10473289.2002.10470838.

Brom, M.A., Holguín-Veras, J., Hodge, S., 2011. Off-hour deliveries in Manhattan, New York City: experiences of pilot test participants. Transp. Res. Rec. 2238,77–85. http://dx.doi.org/10.3141/2238-10.

Browne, M., Allen, J., Wainwright, I., Palmer, A., Williams, I., 2014. London 2012: changing delivery patterns in response to the impact of the games on trafficflows. Int. J. Urban Sci. 18 (2), 244–261.

California Environmental Protection Agency, 2016. EMFAC Web Database, from <http://www.arb.ca.gov/emfac/> (Retrieved 07/11/2016).Center for Environmental Research & Technology, 2016. Comprehensive Modal Emission Model (CMEM), from <http://www.cert.ucr.edu/cmem/>

(Retrieved 07/11/2016).Centro de Inovação em Sistemas Logísticos, 2015. Avaliação do Projeto-Piloto de Entregas Noturnas no Município de São Paulo. Serie Cadernos Tecnicos, 18,

from <http://www.antp.org.br/website/biblioteca/>.Choudhary, A., Gokhale, S., 2016. Urban real-world driving traffic emissions during interruption and congestion. Transport. Res. Part D: Transp. Environ. 43,

59–70.COE-SUFS, 2013a. Webinar #1: Lessons from the Off-Hour Delivery Program in New York City. Peer-to-Peer Exchange Program, from <https://coe-sufs.

org/wordpress/peer-to-peer-exchange-program/webinar01/> (Retrieved 8/15/2013).COE-SUFS, 2013b. Webinar #2: Lessons for Urban Freight from the London 2012 Olympic Games. Peer-to-Peer Exchange Program Retrieved 12/9/2013, from

<https://coe-sufs.org/wordpress/peer-to-peer-exchange-program/webinar02/>.COE-SUFS, 2016a. Webinar #13: Updates on Off-Hour Delivery Pilots Part I: The Experiences of Sao Paulo, Brazil and Copenhagen, Denmark. Peer-to-Peer

Exchange Program, from <https://coe-sufs.org/wordpress/peer-to-peer-exchange-program/webinar13/> (Retrieved 4/14/2016).COE-SUFS, 2016b. Webinar #14: Updates on Off-Hour Delivery Pilots Part II: The Experience of Bogota, Colombia. Peer-to-Peer Exchange Program, from

<https://coe-sufs.org/wordpress/peer-to-peer-exchange-program/webinar14/> (Retrieved 6/15/2016).Cortés, C.E., Vargas, L.S., Corvalán, R.M., 2008. A simulation platform for computing energy consumption and emissions in transportation networks.

Transport. Res. Part D: Transp. Environ. 13 (7), 413–427. http://dx.doi.org/10.1016/j.trd.2008.07.006.Corvalán, R.M., Osses, M., Urrutia, C.M., 2002. Hot emission model for mobile sources: application to the metropolitan region of the city of Santiago, Chile. J.

Air Waste Manag. Assoc. 52 (2), 167–174. http://dx.doi.org/10.1080/10473289.2002.10470774.Cunha, C.B., Yoshizaki, H.T.Y., Almeida, F.G.V., Hino, C.M., Kako, I.S., Laranjeiro, P., Arbex, R.O., 2016. Truck GPS data for in-depth characterization of urban

logistics in the Megacity of São Paulo, Brazil. In: Proceedings of the 2016 World Conference on Transport Research – WCTR 2016, Shanghai, China.DANE, 2014. Departamento Administrativo Nacional de Estadística-Colombia, from <https://www.dane.gov.co/> (Retrieved 07/06/2014).Domínguez, A., Holguín-Veras, J., Ibeas, Á., Dell’Olio, L., 2012. Receivers’ response to new urban freight policies. Procedia-Soc. Behav. Sci. 54, 886–896.


http://refhub.elsevier.com/S1361-9209(16)30471-0/h0005


http://dx.doi.org/10.3141/1520-09

http://dx.doi.org/10.1080/10473289.2002.10470838

http://dx.doi.org/10.3141/2238-10



http://www.arb.ca.gov/emfac/

http://www.cert.ucr.edu/cmem/

http://www.antp.org.br/website/biblioteca/



https://coe-sufs.org/wordpress/peer-to-peer-exchange-program/webinar01/






http://dx.doi.org/10.1080/10473289.2002.10470774

https://www.dane.gov.co/




Environmental Protection Agency, 2012. Motor Vehicle Emission Simulator (MOVES): User guide for MOVES2010b, from <http://www.epa.gov/otaq/models/moves/documents/420b12001b.pdf>.

Environmental Protection Agency, 2014. Control of Air Pollution From Motor Vehicles: Tier 3 Motor Vehicle Emission and Fuel Standards; Final Rule, from<https://www.gpo.gov/fdsys/pkg/FR-2014-04-28/pdf/2014-06954.pdf> (Retrieved 81, 79).

Greco, S., Belova, A., Huang, J., 2016. Benefits of decreased mortality risk from reductions in primary mobile source fine particulate matter: a limited dataapproach for urban areas worldwide. Risk Anal. http://dx.doi.org/10.1111/risa.12612.

Holguín-Veras, J., 2006, December 21, 2006. Potential for off-peak freight deliveries to congested urban areas, from <http://www.rpi.edu/~holguj2/OPD/OPD_FINAL_REPORT_12-18-06.pdf>.

Holguín-Veras, J., 2008. Necessary conditions for off-hour deliveries and the effectiveness of urban freight road pricing and alternative financial policies incompetitive markets. Transport. Res. Part A: Pol. Pract. 42 (2), 392–413.

Holguín-Veras, J., 2011. Urban delivery industry response to cordon pricing, time-distance pricing, and carrier-receiver policies in competitive markets.Transp. Res. Part A 45, 802–824.

Holguín-Veras, J., 2013. The good, the bad, and the ugly: lessons from the off-peak delivery project in New York. In: Wolmar, C. (Ed.), Urban Freight forLivable Cities. Goteborg, The Volvo Research and Educational Foundations, VREF, pp. 118–125.

Holguín-Veras, J., Amaya-Leal, J., Wojtowicz, J., Jaller, M., González-Calderón, C., Sánchez-Díaz, I., Wang, X., Haake, D., Rhodes, S., Hodge, S.D., Frazier, R.J.,Nick, M.K., Dack, J., Casinelli, L., Browne, M., 2015a. Improving freight system performance in metropolitan areas. National Cooperative Freight ResearchProgram Retrieved NCFRP 33, from <http://onlinepubs.trb.org/onlinepubs/ncfrp/ncfrp_rpt_033.pdf>.

Holguín-Veras, J., Aros-Vera, F., 2014. Self-supported freight demand management: pricing and incentives. EURO J. Transport. Logist. 3 (1), 1–24. http://dx.doi.org/10.1007/s13676-013-0041-1.

Holguín-Veras, J., Aros-Vera, F., Browne, M., 2015b. Agent interactions and the response of supply chains to pricing and incentives. Econ. Transport. http://dx.doi.org/10.1016/j.ecotra.2015.04.002.

Holguín-Veras, J., Brom, M.A., 2008. Trucking Costs in a Metropolitan Area: A Comparison of Alternative Estimation Approaches. Annual Meeting of theTransportation Research Board, Washington, DC.

Holguín-Veras, J., Jaller, M., Sanchez-Díaz, I., Wojtowicz, J.M., Campbell, S., Levinson, H.S., Lawson, C.T., Powers, E., Tavasszy, L., 2012a. NCHRP Report 739/NCFRP Report 19: Freight Trip Generation and Land Use. National Cooperative Highway Research Program/National Cooperative Freight ResearchProgram, from <http://onlinepubs.trb.org/onlinepubs/nchrp/nchrp_rpt_739.pdf>.

Holguín-Veras, J., Marquis, R., Brom, M., 2012b. Economic impacts of staffed and unassisted off-hour deliveries in New York City. Procedia – Soc. Behav. Sci.Holguín-Veras, J., Marquis, R., Campbell, S., Wojtowicz, J.M., Wang, X.C., Jaller, M.A., Hodge, S.D., Rothbard, S., Goevaers, R., 2013a. Fostering the use of

unassisted off-hour deliveries: operational and low-noise truck technologies. Transp. Res. Rec. 2379, 57–63.Holguín-Veras, J., Ozbay, K., Kornhauser, A.L., Brom, M., Iyer, S., Yushimito, W., Ukkusuri, S., Allen, B., Silas, M., 2011. Overall impacts of off-hour delivery

programs in the New York City metropolitan area. Transp. Res. Rec. 2238, 68–76.Holguín-Veras, J., Ozbay, K., Kornhauser, A.L., Shorris, A., Ukkusuri, S., 2010. Integrative freight demand management in the New York city metropolitan area

– Final Report. Retrieved Cooperative Agreement #DTOS59-07-H-0002, from <http://transp.rpi.edu/~usdotp/OHD_FINAL_REPORT.pdf>.Holguín-Veras, J., Pérez, N., Cruz, B., Polimeni, J., 2006a. Effectiveness of financial incentives to off peak deliveries to restaurants in Manhattan, New York.

Transport. Res. Record: J. Transport. Res. Board 1966, 51–59.Holguín-Veras, J., Sánchez-Díaz, I., 2016. Freight demand management and the potential of receiver-led consolidation programs. Transport. Res. Part A 84,

109–130. http://dx.doi.org/10.1016/j.tra.2015.06.013.Holguín-Veras, J., Silas, M.A., Polimeni, J., Cruz, B., 2007. An investigation on the effectiveness of joint receiver-carrier policies to increase truck traffic in the

off-peak hours: Part I: the behaviors of receivers. Netw. Spat. Econ. 7 (3), 277–295. http://dx.doi.org/10.1007/s11067-006-9002-7.Holguín-Veras, J., Silas, M.A., Polimeni, J., Cruz, B., 2008. An investigation on the effectiveness of joint receiver-carrier policies to increase truck traffic in the

off-peak hours: Part II: the behaviors of carriers. Netw. Spat. Econ. 8 (4), 327–354. http://dx.doi.org/10.1007/s11067-006-9011-6.Holguín-Veras, J., Wang, C., Browne, M., Hodge, S.D., Wojtowicz, J., 2014. The New York City off-hour delivery project: lessons for city logistics. Procedia-Soc.

Behav. Sci. 125, 36–48.Holguín-Veras, J., Wang, C., Sánchez-Díaz, I., Campbell, S., Hodge, S.D., Jaller, M., Wojtowicz, J., 2016. Fostering unassisted off-hour deliveries: the role of

incentives. Tranp. Res. Part A: Pol. Pract. (in press).Holguín-Veras, J., Wang, Q., Xu, N., Ozbay, K., Cetin, M., Polimeni, J., 2006b. The impacts of time of day pricing on the behavior of freight carriers in a

congested urban area: implications to road pricing. Transport. Res. Part A: Pol. Pract. 40 (9), 744–766.Holguín-Veras, J., Wojtowicz, J.M., Wang, C., Jaller, M., Ban, X.J., Aros-Vera, F., Campbell, S., Yang, X., Sanchez-Diaz, I., Amaya, J., González-Calderón, C.,

Marquis, R., Hodge, S.D., Maguire, T., Marsico, M., Zhang, S., Rothbard, S., Ozbay, K., Morgul, E.F., Iyer, S., Xie, K., Ozguven, E.E., 2013b. Integrative freightdemand management in the New York City metropolitan area: implementation phase, from <http://cite.rpi.edu/news/16-off-hour-delivery>.

Holguín Veras, J., Campbell, S., Kalahashti, L., Wang, C., 2016. Role and potential of a trusted vendor certification program to foster adoption of unassistedoff-hour deliveries. Transp. Rese. Part A: Pol. Pract. (in press).

IBGE, 2015. Instituto Brasileiro de Geografia e Estatística (Brazilian Institute of Geography and Statistics), from <http://www.ibge.gov.br/home> (Retrieved7/17/2016).

Jaller, M., Sánchez, S., Green, J., Fandiño, M., 2016. Quantifying the impacts of sustainable city logistics measures in the Mexico City Metropolitan Area.Transport. Res. Procedia 12, 613–626. http://dx.doi.org/10.1016/j.trpro.2016.02.015.

Journal of Commerce, 2009. New York delivers at night. J. Commer. 10 (35), 4A. September 7, 2009.Journal of Commerce, 2010, July 2, 2010. New York to expand off-peak truck program. J. Commer. (July 3, 2010).Lena, T.S., Ochieng, V., Carter, M., Holguín-Veras, J., Kinney, P.L., 2002. Elemental carbon and PM [sub 2.5] levels in an urban community heavily impacted by

truck traffic. Environ. Health Perspect. 110 (10), 1009–1015.Lepeule, J., Laden, F., Dockery, D., Schwartz, J., 2012. Chronic exposure to fine particles and mortality: an extended follow-up of the Harvard Six Cities study

from 1974 to 2009. Environ. Health Perspect. 120 (7), 965.Levy, J., Buonocore, J., Von Stackelberg, K., 2010. Evaluation of the public health impacts of traffic congestion: a health risk assessment. Environ. Health 9 (1),

1.Lin, J., Vallamsundar, S., 2013, 07/2013. Transportation conformity particulate matter hot-spot air quality modeling. Illinois Center for Transportation Series

No. 13-024 Retrieved FHWA-ICT-13-024, from <http://hdl.handle.net/2142/45750>.Manzoli, A., 2009. Análise das Emissões Veiculares em Trajetos Urbanos Curtos com Localização por GPS. Universidade de São Paulo, Doutor em Engenharia

Civil.McCormick, R.L., Graboski, M.S., Alleman, T.L., Yanowitz, J., 2000. Idle emissions from heavy-duty diesel and natural gas vehicles at high altitude. J. Air Waste

Manag. Assoc. 50 (11), 1992–1998. http://dx.doi.org/10.1080/10473289.2000.10464229.McCrea, B., 2016. From DC to Final Destination: Last Mile Dilemma. Supply Chain Management Review. MMH Editorial.McGeehan, P., 2016. Record Number of Tourists Visited New York City in 2015, and More Are Expected This Year, from <http://www.nytimes.com/2016/03/

09/nyregion/record-number-of-tourists-visited-new-york-city-in-2015-and-more-are-expected-this-year.html> (Retrieved 18/06/2016).Palmer, A., Piecyk, M., 2010. Time, cost and CO2 effects of rescheduling freight deliveries. Towards the Sustainable Supply Chain: Balancing the Needs of

Business, Economy and the Environment.Poudenx, P., 2008. The effect of transportation policies on energy consumption and greenhouse gas emission from urban passenger transportation.

Transport. Res. Part A: Pol. Pract. 42 (6), 901–909. http://dx.doi.org/10.1016/j.tra.2008.01.013.Rodríguez, R.A., Virguez, E.A., Rodríguez, P.A., Behrentz, E., 2016. Influence of driving patterns on vehicle emissions: a case study for Latin American cities.

Transport. Res. Part D: Transp. Environ. 43, 192–206. http://dx.doi.org/10.1016/j.trd.2015.12.008.


http://www.epa.gov/otaq/models/moves/documents/420b12001b.pdf

http://www.epa.gov/otaq/models/moves/documents/420b12001b.pdf

https://www.gpo.gov/fdsys/pkg/FR-2014-04-28/pdf/2014-06954.pdf

http://dx.doi.org/10.1111/risa.12612

http://www.rpi.edu/<ucode type=

http://www.rpi.edu/<ucode type=







http://onlinepubs.trb.org/onlinepubs/ncfrp/ncfrp_rpt_033.pdf

http://dx.doi.org/10.1007/s13676-013-0041-1

http://dx.doi.org/10.1007/s13676-013-0041-1

http://dx.doi.org/10.1016/j.ecotra.2015.04.002

http://dx.doi.org/10.1016/j.ecotra.2015.04.002



http://onlinepubs.trb.org/onlinepubs/nchrp/nchrp_rpt_739.pdf





http://transp.rpi.edu/<ucode type=



http://dx.doi.org/10.1016/j.tra.2015.06.013

http://dx.doi.org/10.1007/s11067-006-9002-7

http://dx.doi.org/10.1007/s11067-006-9011-6





http://cite.rpi.edu/news/16-off-hour-delivery

http://www.ibge.gov.br/home

http://dx.doi.org/10.1016/j.trpro.2016.02.015








http://hdl.handle.net/2142/45750



http://dx.doi.org/10.1080/10473289.2000.10464229

http://www.nytimes.com/2016/03/09/nyregion/record-number-of-tourists-visited-new-york-city-in-2015-and-more-are-expected-this-year.html

http://www.nytimes.com/2016/03/09/nyregion/record-number-of-tourists-visited-new-york-city-in-2015-and-more-are-expected-this-year.html




http://dx.doi.org/10.1016/j.tra.2008.01.013




Sathaye, N., Harley, R., Madanat, S., 2010. Unintended Environmental Impacts of Nighttime Freight Logistics Activities. Transport. Res. Part A: Pol. Pract. 44(8), 642–659.

SEADE, 2015. Fundação Sistema Estadual de Análise de Dados de São Paulo (State System of Data Analysis Foundation – São Paulo), from <http://www.seade.gov.br/>.

Segalou, E., Ambrosini, C., Routhier, J. des Transports, L., 2004. The environmental assessment of urban goods movement. Logistics systems for sustainablecities.

Silas, M.A., Holguín-Veras, J., 2009. Behavioral microsimulation formulation for analysis and design of off-hour delivery policies in urban areas. Transport.Res. Record: J. Transport. Res. Board 2097, 43–50. http://dx.doi.org/10.3141/2097-06.

Tomer, A., Kane, J., 2014. Mapping Freight: The Highly Concentrated Nature of Goods Trade in the United States, from <http://www.brookings.edu/research/reports2/2014/11/06-mapping-freight-tomer-kane?utm_campaign=Metropolitan+Policy+Program&utm_source=hs_email&utm_medium=email&utm_content=14841696&_hsenc=p2ANqtz–FGwrqD-S2D1W-Z4m-6ob9vWojaPg0ojevf76OfSk69yZHZ-prB6kPijXheAXhE59nI_CCr2uZw3TUIKkrXEZlAp-oPg&_hsmi=14841696>.

Transport Topics, 2010. Fleets Say They Discovered Time. Cost Bonanza Through New York Night-Delivery Experiment 1, 8.U.S. Census Bureau, 2016. U.S. and World Population, from <http://www.census.gov/population/international/data/idb/informationGateway.php>

(Retrieved 09-24-2016, 2016).US Bureau of Labor Statistics, 2016, 06/29/2016. New York Area Economic Summary, from <http://www.bls.gov/regions/new-york-new-jersey/summary/

blssummary_newyorkarea.pdf>.Vallamsundar, S., Lin, J., 2011. MOVES versus MOBILE: comparison of greenhouse gas and criterion pollutant emissions. Transp. Res. Rec. 2233, 27–35.Vallamsundar, S., Lin, J., 2013. Sensitivity Test Analysis of MOVES and AERMOD models. 92nd Transportation Research Board Annual Meeting. 500 Fifth

Street, NW, Washington, DC 20001 USA. paper #13-1590:19.Verlinde, S., Debauche, W., Heemeryck, A., Macharis, C., Van Hoeck, E., Witlox, F., 2010. Night-time delivery as a potential option in belgian urban

distribution: a stakeholder approach. WCTR Conference, Lisbon, Portugal.Verlinde, S., Macharis, C., Van Lier, T., Witlox, F., 2011. Which stakeholders benefit from rescheduling more freight deliveries to the off-peak hours? Results

of a pilot study in the retail industry. Transp. Res. Day, 85–96.Wall Street Journal, 2010. Congestion Fight Runs Into Night. Retrieved July 2, 2010, from <http://online.wsj.com/article/

SB10001424052748704334604575339292960610492.html>.Wood, R.T., 1970. Measuring Freight in the Tri-State Region. OECD, The Urban Movement of Goods, Paris, pp. 61–82.Wood, R.T., Leighton, R., 1969. Truck freight in the tri-state region. Traffic Quart. 23 (3), 323–340.Yannis, G., Golias, J., Antoniou, C., 2006. Effects of urban delivery restrictions on traffic movements. Transport. Plan. Technol. 29 (4), 295–311.Zanni, A.M., Bristow, A.L., 2010. Emissions of CO2 from road freight transport in London: trends and policies for long run reductions. Energy Pol. 38, 1774–

1786. http://dx.doi.org/10.1016/j.enpol.2009.11.053.Zárate, E., Belalcazar, L.C., Clappier, A., Manzi, V., Van den Bergh, H., 2007. Air quality modelling over Bogota, Colombia: combined techniques to estimate

and evaluate emission inventories. Atmos. Environ. 41 (29). http://dx.doi.org/10.1016/j.atmosenv.2007.03.011.Zhang, K., Batterman, S., Dion, F., 2011. Vehicle emissions in congestion: comparison of work zone, rush hour and free-flow conditions. Atmos. Environ. 45

(11), 1929–1939. http://dx.doi.org/10.1016/j.atmosenv.2011.01.030.




http://www.seade.gov.br/

http://www.seade.gov.br/

http://dx.doi.org/10.3141/2097-06

http://www.brookings.edu/research/reports2/2014/11/06-mapping-freight-tomer-kane?utm_campaign=Metropolitan+Policy+Program%26utm_source=hs_email%26utm_medium=email%26utm_content=14841696%26_hsenc=p2ANqtz--FGwrqD-S2D1W-Z4m-6ob9vWojaPg0ojevf76OfSk69yZHZ-prB6kPijXheAXhE59nI_CCr2uZw3TUIKkrXEZlAp-oPg%26_hsmi=14841696





http://www.census.gov/population/international/data/idb/informationGateway.php

http://www.bls.gov/regions/new-york-new-jersey/summary/blssummary_newyorkarea.pdf

http://www.bls.gov/regions/new-york-new-jersey/summary/blssummary_newyorkarea.pdf




http://online.wsj.com/article/SB10001424052748704334604575339292960610492.html

http://online.wsj.com/article/SB10001424052748704334604575339292960610492.html




http://dx.doi.org/10.1016/j.enpol.2009.11.053

http://dx.doi.org/10.1016/j.atmosenv.2007.03.011

http://dx.doi.org/10.1016/j.atmosenv.2011.01.030


Documents

Transportation Research Part D...Direct impacts of off-hour deliveries on urban freight emissions José Holguín-Verasa,⇑, Trilce Encarnacióna, Carlos A. González-Calderóna, James