Case Study: Multiple Objective Analysis of Intermodal

Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics Koji Kitazume, Candidate for Master of Environmental Management and Master of Business Administration Dr. Jay S. Golden, Advisor

December 2012

Master’s project submitted in partial fulfillment of the requirements for the Master of Environmental Management degree in the Nicholas School of the Environment of Duke University

Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics

i

Abstract

Nowadays, many logistics managers confront tradeoffs among keeping costs low, delivering

goods on time and reducing carbon footprint. In shipping finished goods from a manufacturing

plant in Asia to a distribution center in the eastern United States, how should a logistics manager

define and choose his preferred route and modes of transportation, taking into account the

potentially conflicting priorities?

This study explored a case of REI, an outdoor apparel brand/retailer, facing such a decision-

making question regarding its inbound logistics from the Port of Shanghai to its distribution center

in Bedford, Pennsylvania and approached it as a multiple objective problem. 15 possible intermodal

freight transportation routes with different attributes in terms of shipping costs, transit time and

greenhouse gas emissions were identified and associated data were collected. The preferred route

was derived by employing a simple additive model of preferences, using a pricing out method to

assess tradeoff weights and computing the overall utility of each alternative.

This framework quantified and visualized how the logistics manager’s choice is affected by

his preferences and the tradeoffs he is willing to make, thereby demonstrating its potential as a

practical aid for decision-making at the intersection of business and the environment. Accuracy of

the model used in this study could be improved by addressing uncertain data and omitted scope.

Furthermore, a versatile platform loaded and maintained with accurate and consistent data on

shipping costs, transit time and GHG emissions, covering multipoint-to-multipoint intermodal

freight transportation routes, could benefit shippers widely by enabling informed decision-making

to enhance their business and environmental performance.


ii

Executive Summary

This study applied a multiple objective analysis approach for decision-making on the

selection of an intermodal freight transportation route for REI’s inbound logistics. 15 possible

intermodal freight transportation routes were identified to ship one full 40-foot standard container

load of a pant product from the Port of Shanghai to the distribution center in Bedford, Pennsylvania.

On all routes, the freight is transported from the Port of Shanghai to a US marine port via ocean. The

routes can be broadly-divided into four groups based on the location of their landing port. In

particular, Routes 1 through 3 can be grouped as the Pacific Northwest group, Routes 4 through 10

the California group, Routes 11 through 13 the South Atlantic group, and Routes 14 and 15 the Mid

Atlantic group. All groups but the Mid Atlantic use rail from the landing port to one of the

intermodal rail terminals within 360 miles from the Bedford distribution center. The remaining

segments of the routes use truck. With such variation in transportation modes and distance, the 15

alternatives present different attributes in terms of shipping costs, transit time and GHG emissions.

Table ES-1 summarizes the attributes for each route.

Route Costs

(US$)

Transit Time

(d:hh:mm)

GHG Emissions (kg CO2)

1 7,902 23:04:08 2,110 2 7,949 23:10:43 2,154 3 8,029 22:23:33 2,199 4 8,149 24:20:56 2,250 5 8,196 25:03:31 2,294 6 8,276 24:16:21 2,338 7 8,547 19:01:28 2,281 8 7,782 19:18:17 2,086 9 7,862 18:22:26 2,165

10 7,903 13:13:06 2,171 11 7,337 28:09:04 2,521 12 6,616 28:20:15 2,430 13 6,492 29:14:22 2,429 14 5,999 32:04:14 2,412 15 6,573 41:14:26 2,339

Table ES-1: Routes and Modes of Transportation for Analysis


iii

The objectives of this decision-making question on inbound transportation route selection

are to keep the three attributes of costs, transit time and emissions all low. Overall utility of each

route was computed using the following simple additive model of preferences:

For this purpose, outcomes of attributes described in Table ES-1 were converted to utilities

on a scale of 0 to 1 proportionately, with the best outcome of the attribute under consideration

being a 1 and the worst being a 0. Table ES-2 lists the converted utilities.

Route Costs

Transit Time GHG Emissions

1 0.25 0.66 0.95 2 0.23 0.65 0.84 3 0.20 0.66 0.74 4 0.16 0.60 0.62 5 0.14 0.59 0.52 6 0.11 0.60 0.42 7 0.00 0.80 0.55 8 0.30 0.78 1.00 9 0.27 0.81 0.82

10 0.25 1.00 0.81 11 0.47 0.47 0.00 12 0.76 0.45 0.21 13 0.81 0.43 0.21 14 1.00 0.34 0.25 15 0.77 0.00 0.42

Table ES-2: Utility Scores of Attributes

Assessment of tradeoff weights used the pricing out method. Table ES-3 presents the weight

assignments derived with the given assumptions.

Costs Transit Time GHG Emissions

Weight Table ES-3: Weights of Attributes

Overall utility of each route was computed by substituting the converted utilities and

derived weight assignments into the simple additive model of preferences equation. Table ES-4

presents the results.


iv

Route Costs Transit Time GHG Emissions

Overall Utility Overall Rank

1 0.14 0.15 0.20 0.49 8 2 0.13 0.14 0.18 0.45 9 3 0.12 0.15 0.15 0.42 10 4 0.09 0.13 0.13 0.35 12 5 0.08 0.13 0.11 0.32 13 6 0.06 0.13 0.09 0.28 15 7 0.00 0.18 0.12 0.29 14 8 0.17 0.17 0.21 0.55 4 9 0.15 0.18 0.17 0.50 7

10 0.14 0.22 0.17 0.53 5 11 0.27 0.11 0.00 0.37 11 12 0.43 0.10 0.04 0.58 3 13 0.46 0.10 0.04 0.60 2 14 0.57 0.07 0.05 0.70 1 15 0.44 0.00 0.09 0.53 6

Table ES-4: Weighted and Overall Utility Scores and Ranking

Route 14 scored the highest overall utility. Therefore, this is logically the best route to ship

one full 40 foot standard container load of the pant product from the Port of Shanghai to the

Bedford DC, based on the given assumptions.

Sensitivity analysis of the tradeoff weight assessment was performed by increasing the

weight of one attribute while holding all other variables constant. Figure ES-1 illustrates how the

overall utility scores vary as the weight on the emissions attribute is increased. The preferred

alternative shifts from Route 14 to Route 8.


v

Figure ES-1: Sensitivity Analysis by Increasing Weight of Emissions Attribute

The same was done on the transit time attribute. Figure ES-2 illustrates the results. The

preferred alternative shifts from Route 14 to Route 10 as the weight on transit time is increased.

Figure ES-2: Sensitivity Analysis by Increasing Weight of Time Attribute

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0.0 0.2 0.4 0.6 0.8 1.0

Ove

rall

uti

lity

Weight on Emissions

Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.0 0.2 0.4 0.6 0.8 1.0

Ove

rall

uti

lity

Weight on Time

Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15


vi

Table of Contents

Abstract ......................................................................................................................................................................................... i Executive Summary ................................................................................................................................................................ ii List of Tables ............................................................................................................................................................................. vi List of Figures .......................................................................................................................................................................... vii Glossary of Key Terms and Acronyms ......................................................................................................................... viii 1. Introduction ..................................................................................................................................................................... 1

1.1. Objective .................................................................................................................................................................. 1 2. Materials and Methods ................................................................................................................................................ 2

2.1. Preconditions ......................................................................................................................................................... 2 2.2. Stage One ................................................................................................................................................................. 3 2.3. Stage Two ................................................................................................................................................................ 4 2.4. Stage Three ............................................................................................................................................................. 5

3. Findings and Results ..................................................................................................................................................... 6 3.1. Freight Transportation in the Context of Global Warming ................................................................. 6 3.2. Stage One ................................................................................................................................................................. 8

3.2.1. Intermodal Transportation .................................................................................................................... 8 3.2.2. Ports ................................................................................................................................................................. 9 3.2.3. Ocean Freight ............................................................................................................................................. 12 3.2.4. Rail Freight .................................................................................................................................................. 14 3.2.5. Road Freight ............................................................................................................................................... 20 3.2.6. Routes and Modes of Transportation for Data Collection ....................................................... 21

3.3. Stage Two .............................................................................................................................................................. 23 3.3.1. Volume and Mass ...................................................................................................................................... 24 3.3.2. Distance ........................................................................................................................................................ 25 3.3.3. Emission Factors ....................................................................................................................................... 25 3.3.4. Costs ............................................................................................................................................................... 26 3.3.5. Transit Time ............................................................................................................................................... 28 3.3.6. Greenhouse Gas Emissions ................................................................................................................... 29 3.3.7. Routes and Modes of Transportation for Analysis ..................................................................... 29

3.4. Stage Three ........................................................................................................................................................... 30 3.4.1. Multiple Objective Decision-Making ................................................................................................. 30 3.4.2. Sensitivity Analysis .................................................................................................................................. 35 3.4.3. Other Types of Containers .................................................................................................................... 38

4. Discussion of Findings and Future Research ................................................................................................... 38 5. Conclusion ....................................................................................................................................................................... 41 References ................................................................................................................................................................................ 43 Attachments ............................................................................................................................................................................. 47

List of Tables

Table ES-1: Routes and Modes of Transportation for Analysis ............................................................................ ii Table ES-2: Utility Scores of Attributes ......................................................................................................................... iii Table ES-3: Weights of Attributes .................................................................................................................................... iii Table ES-4: Weighted and Overall Utility Scores and Ranking ............................................................................ iv Table 3-1: North America’s Top 25 Container Ports............................................................................................... 11 Table 3-2: World’s Top 20 Ocean Container Carriers ............................................................................................ 12


vii

Table 3-3: Maersk Transpacific Services ..................................................................................................................... 13 Table 3-4: Terminal Locations and Distance to Bedford Distribution Center .............................................. 20 Table 3-5: Routes and Modes of Transportation for Data Collection .............................................................. 23 Table 3-6: Distance per Route by Mode of Transportation .................................................................................. 25 Table 3-7: Emission Factors by Mode of Transportation...................................................................................... 26 Table 3-8: Total Costs per Route by Mode of Transportation ............................................................................. 27 Table 3-9: Transit Times per Route by Mode of Transportation ....................................................................... 28 Table 3-10: Total Greenhouse Gas Emissions per Route by Mode of Transportation .............................. 29 Table 3-11: Routes and Modes of Transportation for Analysis .......................................................................... 30 Table 3-12: Ranking of Attributes .................................................................................................................................. 31 Table 3-13: Range of Attributes....................................................................................................................................... 32 Table 3-14: Utility Scores of Attributes ........................................................................................................................ 33 Table 3-15: Weighted and Overall Utility Scores and Ranking........................................................................... 35 Table 3-16: Comparison of Results by Container Type ......................................................................................... 38

List of Figures

Figure ES-1: Sensitivity Analysis by Increasing Weight of Emissions Attribute ............................................ v Figure ES-2: Sensitivity Analysis by Increasing Weight of Time Attribute ...................................................... v Figure 3-1: GHG Emission Factor by Mode of Freight Transportation.............................................................. 7 Figure 3-2: US Top 25 Container Ports ......................................................................................................................... 10 Figure 3-3: BNSF System Map .......................................................................................................................................... 15 Figure 3-4: UP Systems Map ............................................................................................................................................. 16 Figure 3-5: CSXT System Map........................................................................................................................................... 17 Figure 3-6: NS System Map ................................................................................................................................................ 18 Figure 3-7: CN System Map ............................................................................................................................................... 19 Figure 3-8: CPR System Map ............................................................................................................................................. 20 Figure 3-9: Map of Terminal Locations and Bedford Distribution Center ..................................................... 21 Figure 3-10: Ranking of Attributes................................................................................................................................. 32 Figure 3-11: Sensitivity Analysis by Increasing Weight of Emissions Attribute ......................................... 36 Figure 3-12: Sensitivity Analysis by Increasing Weight of Time Attribute ................................................... 36 Figure 3-13: Sensitivity Analysis by Decreasing Cost per Mile of Rail Transportation ............................ 37


viii

Glossary of Key Terms and Acronyms

BNSF: BNSF Railway.

BSR: Business for Social Responsibility.

Carrier: A firm that transports goods or people via land, sea or air (Thomas Publishing Company, 2012).

Class I Railroad: A Class I railroad in the United States, or a Class I railway (also Class I rail carrier) in Canada, is one of the largest freight railroads, as classified based on operating revenue. The exact revenues required to be in each class have varied through the years, and they are now continuously adjusted for inflation. The threshold for a Class I Railroad in 2006 was $346.8 million (Canadian National Railway Company, 2012).

Class I Railway: See Class I Railroad.

CN: Canadian National Railway.

CO2: Carbon Dioxide.

CO2e: Carbon Dioxide-Equivalent.

CPR: Canadian Pacific Railway.

CSXT: CSX Transportation.

DC: Distribution Center – The warehouse facility which holds inventory from manufacturing pending distribution to the appropriate stores (Thomas Publishing Company, 2012).

Drayage: The service offered by a motor carrier for pick-up and delivery of ocean containers or rail containers (Thomas Publishing Company, 2012).

EPA: Environmental Protection Agency.

FEC: Florida East Coast Railway.

GHG: Greenhouse Gas.

GHG Protocol: The Greenhouse Gas Protocol Initiative.

ICTF: Intermodal Container Transfer Facility.

Inbound Logistics: The management of materials from suppliers and vendors into production processes or storage facilities (Thomas Publishing Company, 2012).

Intermodal Transportation: Transporting freight by using two or more transportation modes, such as by truck and rail or truck and oceangoing vessel (Thomas Publishing Company, 2012).

kg: Kilogram.


ix

Maersk: Maersk Line – the global containerized division of the A.P. Moller – Maersk Group (Maersk Line).

mi: Mile.

Near-Dock: The ship-to-rail intermodal container transfer configuration that extends from the marine terminal and customs area to a nearby outside facility requiring drayage (Ashar & Swigart, 2007).

NGO: Non-Governmental Organization.

NS: Norfolk Southern Railway.

On-Dock: The ship-to-rail intermodal container transfer configuration that concludes within the marine terminal and customs area requiring no or minimal drayage (Ashar & Swigart, 2007).

Outbound Logistics: The process related to the movement and storage of products from the end of the production line to the end user (Thomas Publishing Company, 2012).

O-D Pair: Origin-Destination Pair.

Pallet: The platform which cartons are stacked on and then used for shipment or movement as a group. Pallets may be made of wood or composite materials (Thomas Publishing Company, 2012).

REI: Recreational Equipment, Inc.

Shipper: The party that tenders goods for transportation (Thomas Publishing Company, 2012).

Terminal: The end of a railroad or other transport route, or a station at such a point (Oxford University Press, 2012).

TEU: Twenty-Foot Equivalent Unit.

Transit Time: The total time that elapses between a shipment’s pickup and delivery (Thomas Publishing Company, 2012).

UP: Union Pacific Railroad.

US: United States.

WRI: World Resources Institute.

WTP: Willingness to Pay.


1

1. Introduction

You are the inbound logistics manager of an American apparel brand. Specifically, you are in

charge of transporting finished goods from factories in Asia to your distribution centers in the

United States. You have been striving to achieve cost optimization targets while also meeting lead-

time expectations imposed by your merchandizing colleagues. On top of that, your company

recently decided to introduce greenhouse gas emissions reduction goals, and you will soon be

tasked to manage the carbon footprint of your inbound logistics as well.

You have two distribution centers in the US; one in the western region that covers the retail

stores in that half of the country, and another in the eastern region that covers those in the other

half. You are particularly interested in the inbound transportation to the eastern distribution center,

where you have the option to land your shipment from Asia on either the Pacific coast or the

Atlantic coast. How would you go about choosing the optimal inbound logistics, taking into account

the potentially conflicting priorities?

1.1. Objective

This study aims to provide a practical framework for addressing such a multiple objective

question of selecting the best intermodal freight transportation route for inbound logistics,

considering shipping costs, transit time and GHG emissions. The study is motivated by two real-

world challenges: First, while many freight carriers and NGOs offer a “carbon calculator” for moving

goods through transportation networks, there is not yet a de facto tool that allows end-to-end

inventorying of GHG emissions for routes across multiple carriers or different modes of

transportation. Second, as illustrated above, logistics managers are virtually never rewarded for

simply reducing their carbon footprint; rather, they are typically confronted with tradeoffs among

keeping costs low, delivering goods on time and cutting GHG emissions, which altogether could

make the decision-making of choosing the optimal inbound transportation route a challenge.


2

As such, this study first identifies the financial, operational and environmental impacts of

inbound logistics for a specific origin-destination pair (i.e., transporting finished goods from the

factory in Asia to the distribution center in the US) via various routes and modes of transportation,

with a particular focus on intermodal freight transportation. Next, it attempts to demonstrate a

framework for decision-making involving such multiple objectives with tradeoffs. The hope is that

this study serves as a reference material for inbound logistics managers in their daily operations.

2. Materials and Methods

This study consists of three stages. The first stage will focus on identifying routes that are

available and suitable for analysis, through gaining a broad understanding of international

intermodal freight transportation. The second stage is spent on collecting cost, transit time and

GHG emissions data of the routes identified for detailed analysis. The third stage will be devoted to

modeling and analysis.

2.1. Preconditions

The apparel brand illustrated in the opening example is based on Recreational Equipment,

Inc., who agreed to support this study by sharing internal information and data. The outdoor

apparel brand/retailer has approximately $1.7 billion annual sales, 120 retail stores across the US,

and two distribution centers, one in the state of Washington covering the western region, and

another in Bedford, Pennsylvania covering the eastern region. This study will concentrate on REI’s

inbound logistics to the Bedford DC, for a specific men’s pant product which was sourced from a

contract manufacturer with its factory located in Nanjing, China near Shanghai for the 2011 season.

The systems boundary of this study is therefore set at the Shanghai Yangshan Deepwater

Port in China as the origin and REI’s eastern US distribution center in Bedford, Pennsylvania as the

final destination. The actual inbound transportation for the 2011 season pant product flowed from


3

the Port of Shanghai via ocean to the Port of Baltimore on the Atlantic coast, and from there via

road to the Bedford DC. Hence, this route will be used as the baseline where necessary.

Furthermore, the functional unit for the purpose of this study is assumed to be one full 40-foot

standard container load, or two twenty-foot equivalent units, of the pant product.

In order to keep this study focused and manageable, the following elements of the supply

chain are excluded from the scope:

Upstream supply chain beyond the origin port, due to infeasibility of data collection

Downstream supply chain beyond distribution centers (i.e., distribution center to retail

stores and customers) due to different characteristics of logistics

Inbound logistics to the western US distribution center, due to the high likelihood of

already achieving optimal state in terms of the objectives considered

Air freight, due to REI’s current practice of using air only for irregular, expedited

inbound shipments

2.2. Stage One

The goal of this stage is to identify intermodal freight transportation routes between the

Port of Shanghai and the Bedford DC that are available and suitable for analysis, through gaining a

broad understanding of international intermodal freight transportation. Information on intermodal

freight transportation in general, as well as on marine ports, ocean freight transportation, rail

freight transportation, and road freight transportation will be collected primarily through desktop

research, and supplemented by insights from practitioners engaged in transportation and logistics.

Types of information sources can be generally categorized as follows:

Government organizations

Environmental NGOs

Academia


4

Freight carriers and trade organizations

Ocean container ports and trade organizations

Logistics service providers

Shippers

Based on the information collected and understanding gained, the deliverable of this first

stage will be a narrowed-down list of inbound freight transportation modes and routes for data

collection and analysis in the subsequent stages.

2.3. Stage Two

The goal of this stage is to complete the list of inbound freight transportation routes

prepared in the previous stage by adding data for the objectives that will be considered in the

analysis in the following stage. The objectives in particular are total shipping costs, total transit

time and total GHG emissions per route. In order to calculate shipping costs, sub data such as

volume and mass of freight, distance of route, and price schedules will be collected. Similarly, in

order to calculate GHG emissions, sub data such as volume and mass of freight, distance of route,

and emission factors will be collected. Moreover, data will be required per transportation mode

that makes up each route.

Where possible, attempts will be made to collect primary data from the data owner. For

example, volume and mass of freight will be acquired from REI, and emission factors will be

acquired from the carrier, and so forth. In cases which primary data is not available, secondary data

will be obtained utilizing publicly available databases and tools. When neither primary nor

secondary data is available, proxy data and assumptions will be used.

Based on the data collected, the deliverable of this second stage will be a list of inbound

freight transportation routes with attributes in terms of shipping costs, transit time and greenhouse

emissions, for modeling and analysis in the following stage.


5

2.4. Stage Three

The goal of this stage is to select the best intermodal route for inbound transportation from

the list created in the previous section based on the data collected, through applying a multiple

objective problem approach. A Microsoft Excel-based simple additive model of preferences will be

created to compute the overall utility of each route and to perform sensitivity analysis.

The simple additive model equation is:

, where

utility of route on attribute ,



weight assignment to attribute ,



Overall utility assigned to route .

Outcomes of attributes in the list from the previous section will be converted to utility on a

scale of 0 to 1 proportionately, with the best outcome of the attribute under consideration being a 1

and the worst being a 0. Assessment of tradeoff weights will use the pricing out method. Weight

assignments for the three attributes will be derived by setting the cost attribute as the numeraire

and determining the willingness to pay to go from the worst time to best time performance, and the

WTP to go from the worst emissions to best emissions performance. By substituting the converted

utilities and derived weight assignments into the simple additive model of preferences equation,

the overall utility for each route can be computed, thus revealing the best intermodal route for

inbound transportation.

Sensitivity analysis will be performed on the weight assignments. This will visualize how

the results are affected by preferences and the tradeoffs to be made.


6

3. Findings and Results

In the first part of this section, the significance of freight transportation and inbound

logistics within the context of global warming is examined. In the remainder of the section, findings

and results are outlined according to the three stages of this study. Again, the first stage focused on

gaining an understanding of international intermodal transportation and identifying inbound

transportation modes and routes. The second stage was collecting input data to complete the list of

inbound transportation modes and routes with data on costs, transit time and emissions. The third

stage was modeling and analysis of the financial, operational and environmental impacts for

decision-making on the inbound logistics in question.

3.1. Freight Transportation in the Context of Global Warming

In the US in 2003, freight transportation sources accounted for approximately 438

teragrams of carbon dioxide-equivalents of GHG emissions, or 24.7 percent of GHG emissions from

all transportation sources and 6.3 percent of total GHG emissions (ICF Consultung, 2005). A similar

trend can be seen on the global level: A 2009 report estimates freight transportation accounts for

approximately 2,500 megatonnes of CO2e, or 5 percent of annual worldwide GHG emissions (World

Economic Forum, 2009). On both levels, road freight was the greatest contributor accounting for

77.8 percent of GHG emissions from freight transportation in the US and 63.8 percent of GHG

emissions from freight transportation worldwide.

Road freight being the biggest GHG emitter within the freight transportation sector does not

necessarily conclude it is the least environmentally-efficient; but it is the second least efficient

mode of transportation only after air freight. Figure 3-1 shows a comparison of emission factors by

freight transportation mode in terms of kilograms of CO2 emissions per ton-mile of freight moved,

adopted as default values for US vehicles in the GHG Protocol tool for mobile combustion (World

Resources Institute, 2012).


7

Figure 3-1: GHG Emission Factor by Mode of Freight Transportation (World Resources Institute, 2012)

In contrast to the data illustrated above, some studies and statistics point out that ocean

freight transportation is actually more environmentally-efficient than rail freight. For example,

ocean freight transportation is said to emit less than two-thirds of GHG per weight-distance

compared with rail (Dizikes, 2010), or ocean transportation is 32 to 55 percent more efficient than

rail at typical operating conditions (Herbert Engineering Corporation, 2011). This could be true

depending on which sets of data are used to derive an aggregated average. This study attempts to

address such issues by employing route- and mode-specific emission factors for analysis in the

following sections.

Finally, freight transportation and inbound logistics could be a particularly interesting area

to examine for apparel companies hoping to reduce their carbon footprint. In the case of Nike in

2009, inbound logistics accounted for 23 percent of the athletic apparel and footwear giant’s total

GHG footprint and was the second largest impact area only after manufacturing (Nike, Inc., 2010).

1.527

0.297

0.048

0.025

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8

Air

Road

Ocean

Rail

kg CO2/ton-mile


8

3.2. Stage One

3.2.1. Intermodal Transportation

Gerhardt Muller (1999) defined intermodal freight transportation as “the concept of

transporting freight using more than one mode of travel in such a way that all parts of the

transportation process are effectively connected and coordinated, safe, environmentally sound, and

offering flexibility.” In practical terms, intermodal freight transportation allows moving goods over

multiple modes of transportation, in particular ocean, rail and road, without the handling of the

actual freight itself at the point of interchange.

In ocean freight, intermodal transportation is synonymous with transporting containerized

freight by container ships, as opposed to other types of freight transportation such as shipping oil

and chemical by tankers or ore and grain by bulk carriers. Generally, shipping containers, also

called intermodal containers or ISO containers, are 20 or 40 feet in length and 8 feet 6 inches in

height, hence container capacity is commonly expressed in twenty-foot equivalent units, or TEUs.

Although less common, 45-foot containers are also used in intermodal freight. 45-foot containers

are typically 9 feet 6 inches in height and this variant is called high cube. 40-foot containers come in

both standard and high cube variants. While these variations create a range in container volumes,

40-foot standard, 40-foot high cube and 45-foot high cube containers are all considered 2 TEU.

Rail freight adds more complexity; rail intermodal could be transporting containers on well

cars capable of double-stacking containers, containers on a flatcar or trailers on a flatcar. Further, in

the US, domestic containers are typically 48 or 53 feet long. These factors, among others,

necessitate a distinction between international and domestic intermodal freight transportation on

the rail carriers’ part. Union Pacific Railroad, a US Class I railroad, defines international intermodal

as “intact containerized US rail shipments involving an immediately prior or subsequent ocean


9

movement” (Union Pacific Railroad Company). As such, not all of the rail carriers’ intermodal

equipment and facilities are capable of handling international intermodal freight.

With regards to the state of intermodal transportation in the US, intermodal freight on rail

increased from 3 million containers and trailers in 1980 to 11.9 million units in 2011. Over the

same period, the share of containers rose from 42 percent to 85.6 percent (Association of American

Railroads, 2012). Double-stacking of containers was first introduced in the US in 1984, and by 2004,

accounted for approximately 70 percent of intermodal freight transportation on rail (Pacer

International, Inc., 2004).

3.2.2. Ports

Locations of the US top 25 container ports in 2009 in terms container traffic in TEU per year

are illustrated in Figure 3-2 (US Department of Transportation, 2011). The Port of New York/New

Jersey includes the Port of Newark. The Port of Norfolk is part of the Port of Virginia, which is

sometimes referred to as Hampton Roads. The Port of Los Angeles and the Port of Long Beach are

located next to each other much like the Port of New York/New Jersey, and sometimes treated as a

combined single port.


10

Figure 3-2: US Top 25 Container Ports (US Department of Transportation, 2011)

Next, Table 3-1 lists North America’s top 25 container ports in 2010 in terms of container

traffic in TEU per year (American Association of Port Authorities). Combining the traffic of the Port

of Los Angeles and the Port of Long Beach would increase the number up to more than 14 million

TEU per year, nearly three times that of the Port of New York/New Jersey, the immediate follower.


11

Rank Port Coast Country Container Traffic

(TEU/year) 1 Los Angeles Pacific Coast United States 7,831,902

2 Long Beach Pacific Coast United States 6,263,499

3 New York/New Jersey Atlantic Coast United States 5,292,025

4 Savannah Atlantic Coast United States 2,825,179

5 Metro Port Vancouver Pacific Coast Canada 2,514,309

6 Oakland Pacific Coast United States 2,330,214

7 Seattle Pacific Coast United States 2,133,548

8 Hampton Roads Atlantic Coast United States 1,895,017

9 Houston Gulf Coast United States 1,812,268

10 San Juan Atlantic Coast United States 1,525,532

11 Manzanillo Pacific Coast Mexico 1,509,378

12 Tacoma Pacific Coast United States 1,455,466

13 Charleston Atlantic Coast United States 1,364,504

14 Montreal Atlantic Coast Canada 1,331,351

15 Honolulu Pacific Coast United States 968,326

16 Jacksonville Atlantic Coast United States 857,374

17 Miami Atlantic Coast United States 847,249

18 Lazaro Cardenas Pacific Coast Mexico 796,011

19 Port Everglades Atlantic Coast United States 793,227

20 Veracruz Gulf Coast Mexico 677,596

21 Baltimore Atlantic Coast United States 610,922

22 Altamira Gulf Coast Mexico 488,013

23 Anchorage Pacific Coast United States 445,814

24 Halifax Atlantic Coast Canada 435,461

25 New Orleans Gulf Coast United States 427,518 Table 3-1: North America’s Top 25 Container Ports (American Association of Port Authorities)

Based on the findings on container ports in the US and North America, this study tentatively

narrowed down the routes for analysis to those that transit through a port in Table 3-1 and is

located on either the Pacific or Atlantic coast of contiguous US, with the exception of Metro Port

Vancouver. Routes for analysis are further examined in conjunction with the availability of ocean

and rail carriers at the ports.


12

3.2.3. Ocean Freight

The world’s top 20 ocean container carriers in terms of operated capacity in TEU as of

January 2011 are presented in Table 3-2 (Cap-Marine Assurances & Réassurances SAS, 2011).

Maersk and Mediterranean Shipping Company are the two big players, with more than a 600,000

TEU capacity advantage ahead of CMA CGM Group, the number three.

Rank Carrier Operated Capacity (TEU)

Operated Capacity (Number of ships)

1 APM-Maersk 2,147,831 578 2 Mediterranean Shipping Company 1,863,449 450 3 CMA CGM Group 1,209,530 400 4 Evergreen Line 603,766 158 5 Hapag-Lloyd 596,774 136 6 APL 584,780 146 7 CSAV Group 579,296 155 8 COSCO Container Lines 544,857 139 9 Hanjin Shipping 476,955 104

10 China Shipping Container Lines 457,162 140 11 MOL Logistics 399,337 97 12 NYK Line 386,838 98 13 Hamburg Süd Group 370,851 116 14 OOCL 353,523 79 15 K Line 328,327 78 16 Zim Integrated Shipping Services 322,735 94 17 Yang Ming Marine Transport Corporation 322,091 79 18 Hyundai Merchant Marine 286,875 55 19 Pacific International Lines 263,558 142 20 United Arab Shipping Company 216,799 55

Table 3-2: World’s Top 20 Ocean Container Carriers (Cap-Marine Assurances & Réassurances SAS, 2011)

For the ocean freight segment, this study focused on Maersk, because the company’s Web

site provided richer information compared with that of other carriers, and further since

representatives from Maersk were willing to support this study by sharing data.

Maersk operates transpacific services from Shanghai to ports on both the Pacific and

Atlantic coasts of North America. Table 3-3 lists the Shanghai-originating services on transpacific

trade lanes as of October 2012 (Maersk Line).


13

Service Destination Port Destination Terminal Transit Days TP2 Long Beach, CA Total Terminals International/Pier T 15 TP3 Newark, NJ APM Terminal 32 TP3 Norfolk, VA APM Terminal 35 TP3 Savannah, GA Garden City Terminal 37 TP7 Miami, FL South Florida Container Terminal 25 TP7 Savannah, GA Garden City Terminal 26 TP7 Charleston, SC Wando Welch Terminal 28 TP8 Long Beach, CA Total Terminals International/Pier T 13 TP8 Oakland, CA International Container Terminal 17 TP9 Seattle, WA Terminal 18 13 TP9 Vancouver, Canada Deltaport Terminal 15

Table 3-3: Maersk Transpacific Services (Maersk Line)

Based on the findings on ocean freight carriers, this study narrowed down the routes for

analysis to those that go through the above landing ports called by Maersk. All of the above landing

ports were included in the tentative list from the screening conducted in the previous section. Ports

that were screened out from the tentative list were Tacoma, Jacksonville and Everglades.

Maersk was also a suitable ocean carrier to work with on this study, for its approaches to

environmental management and intermodal freight transportation. First, the carrier has been a

proponent of “slow steaming,” an operational strategy to slow vessels down which doesn’t require

adoption of new technology. According to Maersk, slow steaming helped the carrier reduce CO2

emissions per container by 12.5 percent over 2007 to 2009. A study by Cariou (2011) also

estimates that slow steaming led to an 11 percent CO2 emissions reduction in international shipping

over 2008 to 2010. Slow steaming not only cuts fuel consumption and thereby fuel costs and GHG

emissions, but also improves schedule reliability because it creates flexibility for ships to adjust

speed to meet delivery times (Maersk Line, 2010). In addition, in March 2012, Maersk introduced a

seamless intermodal freight service between major ports in Asia and Chicago, Dallas, Houston,

Memphis, and Northwest Ohio in the US via the Port of Los Angeles, in partnership with BNSF

Railway Company, a US Class I railroad. The collaboration between the ocean and rail carriers

allows faster transit and 95 percent on-time delivery (Maersk Line, 2012).


14

3.2.4. Rail Freight

There are five Class I railroads in the US providing freight transportation services, four of

which concern this study, namely BNSF Railway, Union Pacific Railroad, CSX Transportation, and

Norfolk Southern Railway. In addition, there are two Canadian Class I railways serving the

Vancouver area, namely Canadian National Railway and Canadian Pacific Railway.

3.2.4.1. BNSF Railway

BNSF mainly serves the western US region. Figure 3-3 presents the BNSF system map. The

Class I railroad has 25 intermodal facilities with international capability across its territory (BNSF

Railway Company, 2012). Of the landing ports which Maersk calls, BNSF has on-dock capability in

Seattle and Long Beach, and near-dock capability in Oakland. In Seattle, the Seattle International

Gateway – BNSF’s intermodal facility – is located just a half a mile from the port. At Long Beach,

BNSF’s Los Angeles Hobart intermodal facility is located approximately 20 miles from the port but

connected by the Alameda Corridor, a dedicated cargo rail expressway operated jointly by BNSF

and UP. In Oakland, although the Oakland International Gateway – also BNSF’s intermodal facility –

is a near-dock facility, it is located less than a mile away from the terminal.

Among the rail carriers that lack coverage in the eastern US region, BNSF is the only one

that publishes regular schedules for interline international intermodal services, implying a steel

wheel-based seamless interchange with eastern US railroads. From the Pacific coast to the Bedford

DC area, such routes are from the BNSF Los Angeles Hobart facility to CSXT facilities in Cleveland,

Ohio and Northwest Ohio, Ohio, and to NS facilities in Columbus, Ohio and Harrisburg, Pennsylvania.

International intermodal interchanges for other O-D pairs are also available, although they seem to

require individual arrangements.


15

Figure 3-3: BNSF System Map (Wikipedia, 2009)

3.2.4.2. Union Pacific Railroad

UP mainly serves the western US region and competes with BNSF. Figure 3-4 is a

representation of the UP system map. The Class I railroad has 27 intermodal facilities with

international capability across its territory (Union Pacific Railroad Company, 2012). Of the landing

ports which Maersk calls, UP has on-dock capability in Seattle, and near-dock capability in Oakland

and Long Beach. UP’s intermodal facilities are located just a mile from the port in Seattle and on-

port in Oakland. At Long Beach, the Intermodal Container Transfer Facility – UP’s intermodal

facility serving both the Port of Los Angeles and the Port of Long Beach – is located 5 miles from

both ports, however, containers are drayed from the ports to the ICTF by truck.

In order to deliver international intermodal freight from the Pacific coast facilities to the

Bedford DC area, UP interchanges with CSXT and NS via the Chicago facilities, although such

interchanges seem to require individual arrangements. Within the UP system, international

intermodal service is available from Seattle to the Chicago Global II, Chicago Global III and Chicago


16

Global IV intermodal facilities; from Oakland to the Chicago Global IV intermodal facility; and from

the Long Beach ICTF to the Chicago Global III and Chicago Global IV intermodal facilities.

Figure 3-4: UP Systems Map (Wikipedia, 2009)

3.2.4.3. CSX Transportation

CSXT mainly serves the eastern US region. Figure 3-5 presents the CSXT system map. The

Class I railroad has 43 intermodal facilities with international capability across its territory (CSX

Transportation, Inc., 2012). Of the landing ports which Maersk calls, CSXT has near-dock capability

in Charleston, Savannah and Miami, as well as on-dock capability in Savannah. The near-dock

intermodal facilities are located approximately 14 miles from the port in Charleston, 7 miles from

the port in Savannah and 14 miles from the port in Miami, respectively, and containers are drayed

from the port to the intermodal facility by truck at all three locations. The Miami facility belongs to

and is operated by Florida East Coast Railway, an exclusive railroad for ports in South Florida which

interchanges with CSXT in Jacksonville, Florida. As of October 2012, a project to connect FEC to an

on-port rail facility at the Port of Miami is underway (Florida East Coast Railway, 2011).


17

In order to reach the Bedford DC area from South Atlantic ports, the CSXT intermodal

facility with international capability that allows the shortest dispatch via road is the one in

Baltimore, Maryland. Service is available to the Baltimore facility from the three near-dock facilities

in Charleston, Savannah and Miami, however, not from the on-dock facility in Savannah.

As previously discussed, CSXT provides seamless international intermodal interchange

from the BNSF Los Angeles Hobart facility to CSXT facilities in Cleveland and Northwest Ohio. CSXT

also provides steel wheel-based international intermodal interchange from BNSF and UP facilities

in Seattle, Oakland, Los Angeles, and Long Beach to CSXT facilities in Cleveland, Columbus and

Northwest Ohio via Chicago, although such interchanges seem to require individual arrangements.

Figure 3-5: CSXT System Map (Wikipedia, 2009)

3.2.4.4. Norfolk Southern Railway

NS mainly serves the eastern US region and competes with CSXT. Figure 3-6 is a

representation of the NS system map. The Class I railroad has 51 intermodal facilities with

international capability across its territory (Norfolk Southern Corp., 2012). Of the landing ports

which Maersk calls, NS also has near-dock capability in Charleston, Savannah and Miami, and on-


18

dock capability in Savannah. However, no international intermodal service is available from these

four facilities to NS facilities in Maryland or Pennsylvania, which provide similar or shorter access

compared with the CSXT Baltimore facility to the Bedford DC.

As previously discussed, NS provides seamless international intermodal interchange from

the BNSF Los Angeles Hobart facility to NS facilities in Columbus and Harrisburg. Information on

availability and schedules of other international intermodal interchanges from western US

railroads to the Bedford DC area could not be obtained.

Figure 3-6: NS System Map (Wikipedia, 2009)

3.2.4.5. Canadian National Railway

CN mainly serves Canada and parts of US. Figure 3-7 presents the CN system map. Of the

landing ports which Maersk calls, CN has on-dock capability in Vancouver (Canadian National

Railway Company, 2012). The CN Vancouver intermodal facility is located in Surrey, British

Columbia, approximately 28 miles from the port. International intermodal service is available from

Vancouver to the CN Chicago facility in Harvey, Illinois. An agreement with CSXT to provide steel

wheel-based interchange in Chicago was announced in April 2012, however, information on start of


19

service and schedules could not be obtained. Prior to this agreement, CN and CSXT exchanged

container traffic in Chicago only by truck (Canadian National Railway Company, 2012).

Figure 3-7: CN System Map (Wikipedia, 2009)

3.2.4.6. Canadian Pacific Railway

CPR mainly serves Canada and parts of US and competes with CN. Figure 3-8 is a

representation of the CPR system map. Of the landing ports which Maersk calls, CPR has on-dock

capability in Vancouver (Canadian Pacific, 2012). The CPR Vancouver intermodal facility is located

in Pitt Meadows, British Columbia, approximately 35 miles from the port. International intermodal

service is available from Vancouver to the CPR Chicago facility in Bensenville, Illinois, however,

information on availability and schedules of interchanges to eastern US railroads could not be

obtained.


20

Figure 3-8: CPR System Map (Wikipedia, 2009)

3.2.5. Road Freight

The destination terminals from where the international intermodal freight is dispatched via

road to the Bedford DC have been identified from findings in the previous sections. Table 3-4

summarizes the terminals, their location and distance to the Bedford DC.

Terminal Location Road Miles NS Harrisburg 3500 Industrial Rd., Harrisburg, PA 17110 104 CSXT Baltimore 4801 Keith Ave., Baltimore, MD, 21224 148 Port of Baltimore 2700 Broening Hwy., Baltimore, MD 21224 149 CSXT Cleveland 601 E 152nd St., Cleveland, OH 44110 239 Port of Newark 5080 McLester St., Elizabeth, NJ 07207 262 CSXT Columbus 2351 Westbelt Dr., Columbus, OH 43228 285 NS Columbus-Rickenbacker 3329 Thoroughbred Dr., Lockbourne, OH 43217 289 CSXT Northwest Ohio 17000 Deshler Rd., North Baltimore, OH 45872 357

Table 3-4: Terminal Locations and Distance to Bedford Distribution Center


21

Figure 3-9 is a map of the selected terminals relative to the Bedford DC.

Figure 3-9: Map of Terminal Locations and Bedford Distribution Center

3.2.6. Routes and Modes of Transportation for Data Collection

Based on the findings from the previous sections, 14 combinations of alternative routes and

modes of inbound transportation in addition to the baseline were defined for data collection. The

routes can be broadly-divided into four groups based on the location of their landing port. In

particular, Routes 1 through 3 can be grouped as the Pacific Northwest group, Routes 4 through 10

the California group, Routes 11 through 13 the South Atlantic group, and Routes 14 and 15 the Mid

Atlantic group. All groups but the Mid Atlantic use rail from the landing port to one of the

intermodal rail terminals within 360 miles from the Bedford DC. The remaining segments of the

routes use truck. The 15 routes are listed in Table 3-5.


22

Origin Destination Mode Carrier Route 1 Port of Shanghai Port of Seattle Ocean Maersk BNSF Seattle BNSF Chicago Rail BNSF BNSF Chicago CSXT Cleveland Rail CSXT CSXT Cleveland Bedford DC Road N/A Route 2 Port of Shanghai Port of Seattle Ocean Maersk BNSF Seattle BNSF Chicago Rail BNSF BNSF Chicago CSXT Columbus Rail CSXT CSXT Columbus Bedford DC Road N/A Route 3 Port of Shanghai Port of Seattle Ocean Maersk BNSF Seattle BNSF Chicago Rail BNSF BNSF Chicago CSXT Northwest Ohio Rail CSXT CSXT Northwest Ohio Bedford DC Road N/A Route 4 Port of Shanghai Port of Oakland Ocean Maersk BNSF Oakland BNSF Chicago Rail BNSF BNSF Chicago CSXT Cleveland Rail CSXT CSXT Cleveland Bedford DC Road N/A Route 5 Port of Shanghai Port of Oakland Ocean Maersk BNSF Oakland BNSF Chicago Rail BNSF BNSF Chicago CSXT Columbus Rail CSXT CSXT Columbus Bedford DC Road N/A Route 6 Port of Shanghai Port of Oakland Ocean Maersk BNSF Oakland BNSF Chicago Rail BNSF BNSF Chicago CSXT Northwest Ohio Rail CSXT CSXT Northwest Ohio Bedford DC Road N/A Route 7 Port of Shanghai Port of Long Beach Ocean Maersk BNSF Long Beach BNSF Los Angeles Hobart Rail BNSF BNSF Los Angeles Hobart NS Harrisburg Rail BNSF-NS NS Harrisburg Bedford DC Road N/A Route 8 Port of Shanghai Port of Long Beach Ocean Maersk BNSF Long Beach BNSF Los Angeles Hobart Rail BNSF BNSF Los Angeles Hobart CSXT Cleveland Rail BNSF-CSXT CSXT Cleveland Bedford DC Road N/A Route 9 Port of Shanghai Port of Long Beach Ocean Maersk BNSF Long Beach BNSF Los Angeles Hobart Rail BNSF BNSF Los Angeles Hobart NS Columbus-Rickenbacker Rail BNSF-NS NS Columbus-Rickenbacker Bedford DC Road N/A


23

Route 10 Port of Shanghai Port of Long Beach Ocean Maersk BNSF Long Beach BNSF Los Angeles Hobart Rail BNSF BNSF Los Angeles Hobart CSXT Northwest Ohio Rail BNSF-CSXT CSXT Northwest Ohio Bedford DC Road N/A Route 11 Port of Shanghai Port of Miami Ocean Maersk Port of Miami FEC Miami Road N/A FEC Miami CSXT Baltimore Rail FEC-CSXT CSXT Baltimore Bedford DC Road N/A Route 12 Port of Shanghai Port of Savannah Ocean Maersk Port of Savannah CSXT Savannah Road N/A CSXT Savannah CSXT Baltimore Rail CSXT CSXT Baltimore Bedford DC Road N/A Route 13 Port of Shanghai Port of Charleston Ocean Maersk Port of Charleston CSXT Charleston Road N/A CSXT Charleston CSXT Baltimore Rail CSXT CSXT Baltimore Bedford DC Road N/A Route 14 Port of Shanghai Port of Newark Ocean Maersk Port of Newark Bedford DC Road N/A Route 15 (baseline) Port of Shanghai Port of Baltimore Ocean Maersk Port of Baltimore Bedford DC Road N/A

Table 3-5: Routes and Modes of Transportation for Data Collection

Routes that have been eliminated are:

UP routes, due to uncertain availability of interchange to eastern US railroads, and

overlap with BNSF routes

Port of Vancouver routes, due to uncertain availability of interchange from CN and CPR

to eastern US railroads

Port of Norfolk route, due to long ocean transit time and long road distance to Bedford

3.3. Stage Two

This stage focused on information and data collection to expand the list of routes and modes

of transportation from the previous stage for detailed analysis. The following data were collected:

Volume and mass


24

Distance

Costs

Transit time

Emission factors

Due to unavailability of sufficient data, the following components of inbound transportation

were excluded from this study:

Activities at ocean ports, such as loading, unloading and customs clearing

Activities at rail yards and terminals, such as switching, loading and unloading

3.3.1. Volume and Mass

Based on information and data provided by REI, the pant product is packed in various

quantities into cartons, loose-loaded without palletizing onto containers and transported together

with other products to the US over multiple shipments. The product weighs 13 ounces and 40 pairs

can be packed into a carton at maximum. Cartons are 0.60 meters long, 0.40 meters wide and 0.35

meters tall, and weigh 1.1 kilograms.

Internal dimensions of a typical 40-foot standard container are 12.03 meters in length, 2.35

meters in width and 2.39 meters in height. Tare weight of the container is 3,700 kilograms (Maersk

Line). For the purpose of this study, it is assumed that such a 40-foot standard container is loaded

to maximum capacity with cartons of only the pant product. It is also assumed that no pallets are

used. Based on an optimal stowage pattern of cartons on the container floor, 110 cartons can be laid

out per tier and stacked up six tiers. The payload weight, consisting of 660 cartons each packed

with 40 products, is 10,456 kilograms. Adding the tare weight of the container, the gross weight

becomes 14,156 kilograms, or 15.6 tons. Therefore, the functional unit of this study is one 40-foot

standard container, or 2 TEU, or 15.6 tons.


25

3.3.2. Distance

Ocean distance data was acquired primarily from Maersk. Where additional data was

required, materials from the National Oceanic and Atmospheric Administration (US Department of

Commerce, 2009) and the PortWorld Web site (Petromedia Ltd., 2012) were used.

Rail and road distance data were collected using PC*MILER, a routing, mileage and mapping

software for the transportation and logistics industry developed by ALK Technologies. In particular,

the PC*MILER Rail software hosts data for various modes of rail transport including intermodal

freight, and is adopted by Class I railroads such as BNSF, UP and CSXT for price calculation tools

based on rail distances. Trial versions of PC*MILER Version 26 and PC*MILER Rail Version 18 were

used for road and rail, respectively.

Distances in terms of miles by transportation mode per route are summarized in Table 3-6.

Route Ocean Rail Road Total Distance

1 5,851 2,734 239 8,825 2 5,851 2,756 285 8,893 3 5,851 2,706 357 8,915 4 6,971 2,846 239 10,056 5 6,971 2,868 285 10,124 6 6,971 2,818 357 10,146 7 6,537 3,308 104 9,949 8 6,537 2,555 239 9,332 9 6,537 2,596 289 9,423

10 6,537 2,524 357 9,418 11 11,210 1,162 161 12,533 12 11,676 647 154 12,477 13 11,793 545 162 12,500 14 12,221 0 262 12,483 15 12,338 0 149 12,487

Table 3-6: Distance per Route by Mode of Transportation

3.3.3. Emission Factors

Ocean emission factors were acquired from Maersk which are service-specific, based mostly

on data from 2011 and partially 2010. While emission factors for freight transportation are

generally expressed in emissions per weight-distance units (e.g., kilogram CO2 per tonne-kilometer),


26

those for ocean container transportation are typically expressed in emissions per volume-distance

units (e.g., kilogram CO2 per TEU-kilometer), and Maersk follows this practice.

Compared with WRI’s GHG Protocol tool for mobile combustion previously discussed,

Maersk’s emission factors, when converted to weight-distance unit, are much smaller than the

default value for ocean freight set in the tool, in fact, smaller than that of rail freight as some suggest.

Some factors that affect fuel efficiency and thus carbon emissions are speed, load, vessel and engine

type and age, and capacity utilization (Herbert Engineering Corporation, 2011). Given that Maersk

has been practicing slow steaming, it could be contributing to the variance between the emission

factors of Maersk and the GHG Protocol tool, together with the data aggregation issues previously

discussed. In addition, the inexactness of TEU could also be a source of variance when converting

volume-distance emission factors to weight-distance emission factors, which could be another

cause of the different opinions regarding carbon efficiencies of ocean and rail transportation.

Emission factors for rail and road were obtained from the US EPA SmartWay Carrier Data

(US Environmental Protection Agency, 2012). For rail, intermodal-specific emission factors by

railroad were available. For road, average of 103 carriers’ data for dray and average of 13 carriers’

data for truck intermodal, respectively, were adopted for the purpose of this study.

Emission factors by mode of transportation are summarized in Table 3-7.

Ocean

(kg CO2/TEU-mile)

Rail

(kg CO2/ton-mile)

Road Intermodal

(kg CO2/ton-mile)

Road Dray

(kg CO2/ton-mile) 0.0829-0.0898 0.0200-0.0242 0.0531 0.0866

Table 3-7: Emission Factors by Mode of Transportation

3.3.4. Costs

Ocean freight rates as of November 2012 were collected from the rate search tool on

Maersk’s Web site (Maersk Line). Maersk’s rates are charged per container and different rates

apply for different types of container. Road freight costs were taken from the PC*MILER tool.

Default settings of PC*MILER were used, of which the major components are fuel, toll and driver


27

costs. It was assumed that these costs are charged per container and that the same costs apply

regardless of container type.

Rail intermodal prices seem to generally consist of two components, a station-to-station

price and a fuel surcharge. However, none of the rail carriers but CN publishes their intermodal

prices. For the purpose of this study, CN intermodal prices were taken as a proxy and applied to the

15 routes for analysis. For the station-to-station component, prices of 50 sample CN O-D pairs

(Canadian National Railway Company, 2012) were converted to an average price per mile of $1.12.

CN’s fuel surcharge is 17.37 percent of the station-to-station price, as of November 2012 (Canadian

National Railway Company, 2005). CN prices are per container and the same price applies for 20-

foot, 40-foot and 45-foot containers. For the Alameda Corridor segment of the Port of Long Beach

routes, the published use fee of $21.60 per TEU was used (Alameda Corridor Transportation

Authority, 2012).

Costs in US dollars for one 40-foot standard container by transportation mode are

summarized in Table 3-8.

Route Ocean Rail Road Total Costs 1 3,907 3,594 401 7,902 2 3,907 3,623 418 7,949 3 3,907 3,558 564 8,029 4 4,007 3,741 401 8,149 5 4,007 3,770 418 8,196 6 4,007 3,705 564 8,276 7 4,007 4,363 177 8,547 8 4,007 3,374 401 7,782 9 4,007 3,428 428 7,862

10 4,007 3,332 564 7,903 11 5,557 1,528 252 7,337 12 5,525 851 240 6,616 13 5,525 717 250 6,492 14 5,585 0 414 5,999 15 6,341 0 232 6,573

Table 3-8: Total Costs per Route by Mode of Transportation


28

3.3.5. Transit Time

Transit times for ocean and rail freight were acquired from schedules published on carriers’

Web sites (Maersk Line) (BNSF Railway Company, 2012) (CSX Transportation, Inc., 2012). When

two or more transit times were quoted for a single service they were averaged. Concerning rail, one

additional day of transit time was added to routes that have non-seamless interchange, namely

Routes 1 through 3, the Port of Seattle routes, and Routes 4 through 6, the Port of Oakland routes.

An assumption of 45 minutes was made for the Alameda Corridor segment of Routes 7 through 10,

namely the Port of Long Beach routes, based on the average speed of trains of 35 to 40 miles per

hour over the 21.6-mile segment (Net Resources International). Transit times for road were

obtained from PC*MILER.

Transit times by transportation mode in terms of days, hours and minutes are presented in

Table 3-9 in d:hh:mm format.

Route Ocean Rail Road Total Transit Time

1 13:12:00 9:12:20 0:03:48 23:04:08 2 13:12:00 9:18:12 0:04:31 23:10:43 3 13:12:00 9:05:42 0:05:51 22:23:33 4 16:00:00 8:17:08 0:03:48 24:20:56 5 16:00:00 8:23:00 0:04:31 25:03:31 6 16:00:00 8:10:30 0:05:51 24:16:21 7 12:00:00 6:23:45 0:01:43 19:01:28 8 12:00:00 7:14:29 0:03:48 19:18:17 9 12:00:00 6:17:45 0:04:41 18:22:26

10 12:00:00 1:07:15 0:05:51 13:13:06 11 25:00:00 3:06:21 0:02:43 28:09:04 12 27:12:00 1:05:42 0:02:33 28:20:15 13 28:12:00 0:23:42 0:02:40 29:14:22 14 32:00:00 0:00:00 0:04:14 32:04:14 15 41:12:00 0:00:00 0:02:26 41:14:26

Table 3-9: Transit Times per Route by Mode of Transportation


29

3.3.6. Greenhouse Gas Emissions

GHG emissions were calculated by summing the products of volume or weight by distance

by emission factor per transportation mode. Total emissions in terms of kg CO2 by transportation

mode are summarized in Table 3-10.

Route Ocean Rail Road Total Emissions

1 1,051 861 198 2,110 2 1,051 868 236 2,154 3 1,051 852 296 2,199 4 1,156 896 198 2,250 5 1,156 903 236 2,294 6 1,156 887 296 2,338 7 1,084 1,111 86 2,281 8 1,084 804 198 2,086 9 1,084 842 239 2,165

10 1,084 792 296 2,171 11 2,013 367 141 2,521 12 2,097 202 131 2,430 13 2,118 170 141 2,429 14 2,195 0 217 2,412 15 2,216 0 124 2,339

Table 3-10: Total Greenhouse Gas Emissions per Route by Mode of Transportation

3.3.7. Routes and Modes of Transportation for Analysis

With the data collected, the list of routes and modes of transportation for analysis was

completed. The list is provided in Table 3-11.


30

Route Costs

(US$)

Transit Time

(d:hh:mm)


1 7,902 23:04:08 2,110 2 7,949 23:10:43 2,154 3 8,029 22:23:33 2,199 4 8,149 24:20:56 2,250 5 8,196 25:03:31 2,294 6 8,276 24:16:21 2,338 7 8,547 19:01:28 2,281 8 7,782 19:18:17 2,086 9 7,862 18:22:26 2,165

10 7,903 13:13:06 2,171 11 7,337 28:09:04 2,521 12 6,616 28:20:15 2,430 13 6,492 29:14:22 2,429 14 5,999 32:04:14 2,412 15 6,573 41:14:26 2,339

Table 3-11: Routes and Modes of Transportation for Analysis

3.4. Stage Three

The major purpose of this study was to provide a practical framework for decision-making

in an environment with multiple objectives involving tradeoffs. Given the 15 alternatives to choose

from, how should the logistics manager decide on which route to ship his pant product from

Shanghai to the Bedford DC? The following sections attempt to address this question.

3.4.1. Multiple Objective Decision-Making

The objectives of this decision-making question on inbound transportation route selection

are to keep the three attributes of costs, transit time and emissions all low. The study applied a

simple additive model of preferences to address this multiple objective problem. As previously

discussed, the goal was to compute the overall utility of each route using the following equation:

Procedures to reach this computation are presented in this section step-by-step.


31

3.4.1.1. Ranking and Dominance Check

First, outcomes of each attribute were converted to rankings as listed in Table 3-12.


1 8 6 2 2 10 7 3 3 11 5 6 4 12 9 7 5 13 10 9 6 14 8 10 7 15 3 8 8 6 4 1 9 7 2 4

10 9 1 5 11 5 11 15 12 4 12 14 13 2 13 13 14 1 14 12 15 3 15 11

Table 3-12: Ranking of Attributes

Figure 3-10 illustrates the attribute rankings in a chart format. It can be observed that

Route 8 dominates Routes 1 through 6. At this point, the three Port of Seattle routes and the three

Port of Oakland routes can be eliminated from further consideration, since the Long Beach to CSXT

Cleveland route is superior on all three attributes. However, dominance does not hold for the

remaining nine alternatives. It is visually and theoretically confirmed that there are tradeoffs

among the three objectives.


32

Figure 3-10: Ranking of Attributes

3.4.1.2. Trade-Off Weight Assessment

Next, a range of best and worst levels were defined for each attribute. The levels could

simply be set at the best and worst outcomes of the alternatives under consideration, or chosen

otherwise. For the purpose of this study, the former was employed as described in Table 3-13.

Costs

(USD)

Transit Time

(d:hh:mm)


Best 5,999 13:13:06 2,086 Worst 8,547 41:14:26 2,521

Table 3-13: Range of Attributes

Using the defined ranges, outcomes of attributes described in Table 3-11 were converted to

utilities on a scale of 0 to 1 proportionately, with the best outcome of the attribute under

consideration being a 1 and the worst being a 0. The converted utilities are listed in Table 3-14.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Cost Time Emissions

Ran

k

Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15


33


1 0.25 0.66 0.95 2 0.23 0.65 0.84 3 0.20 0.66 0.74 4 0.16 0.60 0.62 5 0.14 0.59 0.52 6 0.11 0.60 0.42 7 0.00 0.80 0.55 8 0.30 0.78 1.00 9 0.27 0.81 0.82

10 0.25 1.00 0.81 11 0.47 0.47 0.00 12 0.76 0.45 0.21 13 0.81 0.43 0.21 14 1.00 0.34 0.25 15 0.77 0.00 0.42

Table 3-14: Utility Scores of Attributes

The next step was to assess the tradeoff weights using the pricing out method. First, the

logistics manager needs to determine the maximum cost he is willing to pay if he were to improve

his transit time from worst performance to best performance. For the purpose of this study, it was

arbitrarily assumed that he is willing to pay the average cost of all the outcomes under

consideration which is $7,574. Conceptually, this is saying that if the cost of the expensive, fastest

route is at this level, the logistics manager would be indifferent between that route and the

inexpensive, slowest route. In mathematical terms, it is expressed as:

( ) ( ) ( ) ( )

The WTP of $7,574 is converted to a utility score of 0.38 using the cost utility curve from the

attribute range. By substituting utilities, the above equation becomes:

And by solving for , the equation becomes:


34

The same is done for the emissions attribute. For the purpose of this study, it was arbitrarily

assumed that the logistics manager is willing to pay the second highest cost of all the outcomes

under consideration which is $8,276, to improve his emissions performance from worst to best.

( ) ( ) ( ) ( )

Utility for WTP of $8,276 is 0.11 from the cost utility curve.

And by solving for , the equation becomes:

Finally, the sum of the three attribute weights must equal 1. Therefore:

The weights for the three attributes can be derived:

⁄

( )( )

( )( )

With the weights derived per above, the overall utility for each route can be computed.

Results are presented in Table 3-15.


35


Overall Utility Overall Rank

1 0.14 0.15 0.20 0.49 8 2 0.13 0.14 0.18 0.45 9 3 0.12 0.15 0.15 0.42 10 4 0.09 0.13 0.13 0.35 12 5 0.08 0.13 0.11 0.32 13 6 0.06 0.13 0.09 0.28 15 7 0.00 0.18 0.12 0.29 14 8 0.17 0.17 0.21 0.55 4 9 0.15 0.18 0.17 0.50 7

10 0.14 0.22 0.17 0.53 5 11 0.27 0.11 0.00 0.37 11 12 0.43 0.10 0.04 0.58 3 13 0.46 0.10 0.04 0.60 2 14 0.57 0.07 0.05 0.70 1 15 0.44 0.00 0.09 0.53 6

Table 3-15: Weighted and Overall Utility Scores and Ranking

Route 14, the Port of Newark route, scored the highest overall utility. Therefore, this is

logically the best route to ship one full 40-foot standard container load of the pant product from the

Port of Shanghai to the Bedford DC, based on the ranges of the attributes that were defined and the

amounts of WTP that were determined.

3.4.2. Sensitivity Analysis

Two types of sensitivity analysis were performed. First, how the best route is affected by

varying the weights of attributes was tested. Next, the assumption made for the uncertain station-

to-station price for rail was examined.

3.4.2.1. Sensitivity Analysis of Weighting Assessment

Sensitivity analysis of the weighting assessment was performed by increasing the weight of

one attribute while holding all other variables constant. Figure 3-11 illustrates how the overall

utility scores vary as the weight on the emissions attribute is increased from the base case weight of

0.209. It is observed that Route 8, the Long Beach to CSXT Cleveland route, becomes the preferred

alternative as the weight on emissions is increased.


36

Figure 3-11: Sensitivity Analysis by Increasing Weight of Emissions Attribute

The same was done on the transit time attribute. Route 10, the Long Beach to CSXT

Northwest Ohio route becomes the preferred alternative as the weight on transit time is increased

from the base case weight of 0.223. Figure 3-12 illustrates the results.

Figure 3-12: Sensitivity Analysis by Increasing Weight of Time Attribute

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0.0 0.2 0.4 0.6 0.8 1.0

Ove

rall

uti

lity

Weight on Emissions

Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.0 0.2 0.4 0.6 0.8 1.0

Ove

rall

uti

lity

Weight on Time

Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15


37

3.4.2.2. Sensitivity Analysis of Assumptions

The most uncertain data used in this study was the station-to-station price for rail. As

previously discussed, since none of BNSF, CSXT and NS publishes intermodal prices, those of CN, a

Canadian rail carrier, were averaged and used as a proxy. Sensitivity analysis of the cost per mile of

rail was performed by decreasing the assumption of $1.12 per mile while holding all other variables

constant. The dominance check conducted in Section 3.4.1.1 is not applicable (i.e., all 15 alternatives

need to be included) for this analysis, since varying this assumption will alter the original outcomes

and potentially the rankings of the Cost attribute.

Figure 3-13 illustrates how the overall utility scores vary as the cost per mile of rail is

decreased. It is observed that Route 8, the Long Beach to CSXT Cleveland route, could become the

best route as the cost per mile of rail is decreased.

Figure 3-13: Sensitivity Analysis by Decreasing Cost per Mile of Rail Transportation

Holding the weight assignments derived in the base case constant, the indifference point

between Route 14 and Route 8 is calculated to be at approximately $0.91 per mile. In other words,

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40

Ove

rall

uti

lity

Rail point-to-point price (US$/mi)

Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15


38

in the base case, the best route determined by this analysis is not impacted by the inaccuracy of the

rail cost as long as the true cost per mile of rail is not less than $0.91.

3.4.3. Other Types of Containers

In addition to the functional unit of this study of one 40-foot standard container, analyses

were performed for one 20-foot standard container, one 40-foot high cube container and one 45-

foot high cube container. Comparison of the results along with volume and mass of the respective

container types are provided in Table 3-16. The same assumptions for the price out assessment

were made for all types of containers.

20’ standard 40’ standard 40’ high cube 45’ high cube Number of containers 1 1 1 1 Volume (TEU) 1 2 2 2 Gross weight (ton) 7.9 15.6 17.7 20.2 Best route Route 14 Route 14 Route 14 Route 14 Potentially best route when weight on Emissions attribute is increased

Route 8 Route 8 Route 8 Route 14

Potentially best route when weight on Time attribute is increased

Route 10 Route 10 Route 10 Route 10

Potentially best route when Cost per mile of rail assumption is decreased

Route 8 Route 8 Route 8 Route 8 then Route 1 then Route 7

Table 3-16: Comparison of Results by Container Type

While the best route turned out to be the same for all container types under the

assumptions made, sensitivity analysis for 45-foot high cube container produced different results

compared with other types of containers.

4. Discussion of Findings and Future Research

To ship one full 40-foot standard container load of the pant product from the Port of

Shanghai to the Bedford DC, the ranking and dominance check first revealed that Routes 1 through

6, the Port of Seattle routes and the Port of Oakland routes, can be eliminated from further


39

consideration since they are dominated by Route 8, the Port of Long Beach to Cleveland route. Next,

the simple additive model of preferences concluded that Route 14, the Port of Newark route is the

best route, given the ranges of attributes defined and the WTP determined from pricing out.

Further, the sensitivity analysis on weight assignments presented that Route 8 could become the

preferred alternative as the weight on emissions is increased, and that Route 10, the Port of Long

Beach to Northwest Ohio route, could become the preferred alternative as the weight on transit

time is increased. Finally, the sensitivity analysis on cost per mile of rail indicated that given the

weight assignments of the base case, the result of Route 14 being the best route holds unless the

true cost per mile of rail is less than $0.91 versus the assumption of $1.12. In case the true cost per

mile of rail is less than $0.91, Route 8 becomes the best route.

Results of analysis for a particular container type cannot be generally applied across

different types of containers. This is due primarily to the different denominators of emission factors

and price schedules among different modes of transportation. For example, because emission

factors for ocean freight transportation are provided in per TEU-distance unit, emissions for the

ocean segment of a single route is the same between a 40-foot standard container and a 45-foot

high cube container, which are both 2 TEU. However, because emission factors for rail freight

transportation are provided in per weight-distance unit, emissions for the rail segment of the same

route are different between 40-foot standard and 45-foot high cube containers, which in this study,

the gross weight were assumed to be 15.6 tons and 20.2 tons, respectively. These could cause the

overall utility of a route with higher mix of rail to degrade relative to routes with lower mix of or no

rail as the gross weight of freight increases. Therefore, multiple objective analysis and tradeoff

weight assessment should be conducted based on the specific type of container used.

While accurate data for rail costs and a rigorous approach for determining the WTP would

make this analysis theoretically more robust, one of the objectives of this study was to provide the

logistics manager with a practical aid for decision-making on the selection of his preferred inbound


40

transportation route. Based on this analysis the logistics manager is already in a better position to

make an informed decision by gaining an understanding of how his choice is impacted by his

preferences, the tradeoffs he is willing to make and uncertain data. In any case, decision-making

should not rely solely on data or models, and in that respect, the model used in this study should

serve as an aid for decision-making rather than an automatic route selection tool.

Nevertheless, there are shortcomings of this model that can be improved. In terms of

accuracy, in addition to the uncertain data for rail costs, the exclusion of activities at marine ports

and rail facilities is another uncertainty that could potentially alter the outcomes when addressed.

In terms of scale, the model was built to analyze only one specific O-D pair of inbound

transportation; adding data for multiple O-D pairs would add more use to the model. In so doing,

adding overseas origin points to the model is relatively simple. However, increasing the number of

domestic destinations will likely require considerable efforts due to the complexity of interline

interchanging associated with rail transportation. Furthermore, enhancing the scope to include

outbound logistics would be another research area of its own, due to the different characteristics

and needs compared with inbound logistics.

Such improvements to the model or a development of a more versatile, generic platform

could be areas of interest for future research. Whichever direction that might take, one common

key to success will be cross-sector collaboration. By nature, intermodal freight transportation

involves many different players, and typically, there is lack of transparency at some point along the

supply chain, especially at points of interchange. The BSR Clean Cargo Working Group has been

working to bring participants primarily from ocean freight transportation together, and the US EPA

SmartWay program similarly to bring participants from domestic road and rail freight

transportation together. Although the main focus of both organizations is to address environmental

impacts, they could be potential partners for pursuing future research effectively and efficiently.


41

5. Conclusion

Logistics managers confront tradeoffs among keeping costs low, moving goods promptly

and reducing carbon footprint. Multiple objective analysis can be one practical framework to help

decision-making on problems involving such tradeoffs at the intersection of business and the

environment. This study attempted to demonstrate such an application by addressing a real-world

problem REI faces regarding the choice of intermodal freight transportation routes for its inbound

logistics. The application of the framework revealed, quantitatively and visually, how the logistics

manager’s choice is affected by his preferences and the tradeoffs he is willing to make.

The methodologies employed for this study were fairly straight forward; however,

collection of input data presented a challenge. In particular, there is limited information and data

publicly available regarding the rail segment of international intermodal freight transportation,

especially when interline interchange is involved. This could be due to the fact that evolution is

currently ongoing in this field. Carriers are investing in equipment and infrastructure and exploring

partnerships to make end-to-end freight transportation more seamless and efficient for shippers.

As evidenced by the launch of the seamless intermodal freight service by Maersk in partnership

with BNSF, standardized one-stop-shopping and one-stop-billing services could become the norm

as this evolution unfolds. Such trends should also drive improvements in data consistency,

transparency and accessibility for shippers.

Nevertheless, the current limitation in availability of data especially for the rail segment of

international intermodal freight transportation could hinder informed decision-making by shippers

to optimize shipping costs, transit time and carbon footprint of their inbound logistics. On the flip

side, it could also mean missed business opportunity for the rail carriers; given the assumptions,

the result of the base case of this study suggests the preferred intermodal route from the Port of

Shanghai to the Bedford DC is to not go via rail. What if the use of true costs in the model altered


42

this outcome? Furthermore, as indicated by the example of Nike, the impact of inbound logistics

within the shipper’s corporate value chain could be one not to be neglected.

A versatile platform loaded and maintained with accurate and consistent cost, transit time

and emissions data, covering multipoint-to-multipoint international intermodal freight

transportation routes, could be a valuable aid for decision-making by shippers to enhance business

and environmental performance. Cross-sector forums on freight transportation such as the BSR

Clean Cargo Working Group and the US EPA SmartWay program could be leveraged to gauge

interest in and facilitate development of such a platform. The Duke Center for Sustainability and

Commerce is uniquely positioned to drive such a pursuit with its expertise in corporate

sustainability, network with players from industry, government and NGOs, and resources offered

by the greater Duke community.


43

References

Alameda Corridor Transportation Authority. (2012, January 1). 2012 Charge per TEU (Twenty Foot Equivalent Unit). Retrieved from Alameda Corridor Transportation Authority Web site: http://www.acta.org/gen/charge_per_teu_2012.pdf

American Association of Port Authorities. (n.d.). North American Container Port Traffic (1990-2010). Retrieved from American Association of Port Authorities Web site: http://www.aapa-ports.org/Industry/content.cfm?ItemNumber=900#Statistics

Ashar, A., & Swigart, S. (2007). A Comparative Analysis of Intermodal Ship-to-Rail Connections at Louisiana Deep Water Ports. Retrieved from Louisiana Department of Transportation and Development Web site: http://www.dotd.la.gov/intermodal/marineandrail/documents/A_Comparative_Analysis_of_Intermodal_Ship_to_Rail_Connections_at_Louisiana_Deep_Water_Ports.pdf

Association of American Railroads. (2012, July). Rail Intermodal Keeps America Moving. Retrieved from Association of American Railroads Web site: http://www.aar.org/~/media/aar/Background-Papers/Rail-Intermodal.ashx

BNSF Railway Company. (2012). BNSF Railway Carbon Estimator. Retrieved from BNSF Railway Company Web site: http://domino.bnsf.com/website/carbon.nsf/ce?open

BNSF Railway Company. (2012, October 19). Intermodal Schedules. Retrieved from BNSF Railway Company Web site: http://www.bnsf.com/customers/pdf/international_P.pdf

BSR. (2009, October 19). Clean Cargo Working Group Intermodal Tool Training. BSR. (2011, March). How Brands Can Improve Supply Chain Sustainability Through the Clean Cargo

Working Group. Retrieved from BSR Web site: http://www.bsr.org/reports/CCWG_Report_Mar_2011_FINAL.pdf

Canadian National Railway Company. (2005, April 1). Fuel Surcharge. Retrieved from Canadian National Railway Company Web site: http://www.cn.ca/en/shipping-prices-tariffs-fuel-surcharge.htm

Canadian National Railway Company. (2012). CN Intermodal Terminals List. Retrieved from Canadian National Railway Company Web site: http://www.cn.ca/en/shipping-how-intermodal-terminals.htm

Canadian National Railway Company. (2012, April 2). CN offers new steel-wheel-interchange service with CSXT in Chicago, giving west coast imports efficient access to Ohio Valley markets. Retrieved from Canadian National Railway Company Web site: http://www.cn.ca/en/media-news-interchange-service-csx-20120402.htm

Canadian National Railway Company. (2012). CN T 008383-BQ INTERMODAL/Containers & Trailers/All Customers. Retrieved from Canadian National Railway Company Web site: http://ecprod.cn.ca/ebusiness/eDistribution/english/public/PriceDocumentDisplay

Canadian National Railway Company. (2012). Glossary of Terms. Retrieved from Canadian National Railway Company Web site: http://www.cn.ca/en/corporate-citizenship-safety-glossary.htm

Canadian Pacific. (2012). Facility List. Retrieved from Canadian Pacific Web site: http://www.cpr.ca/en/our-network-and-facilities/Pages/FacilityList.aspx?FacilityType=Intermodal&ProvinceState=British Columbia (BC)

Cap-Marine Assurances & Réassurances SAS. (2011). Shipping and Shipbuilding Markets 2011 Annual Review. Retrieved from Barry Rogliano Salles Web site: http://www.brs-paris.com/annual/annual_histo/annual_review_2011-a.pdf


44

Cariou, P. (2011). Is slow steaming a sustainable means of reducing CO2 emissions from container shipping? Transportation Research Part D: Transport and Environment, 260-264.

CSX Transportation, Inc. (2012). Carbon Calculator v2. Retrieved from CSX Transportation, Inc. Web site: http://www.csx.com/index.cfm/customers/tools/carbon-calculator-v2/

CSX Transportation, Inc. (2012, August 17). Customers. Retrieved from CSX Intermodal Web site: http://www.csxi.com/share/csxicustomer/main/docs/International_Service_Matrix-REF21399-REF24494.xls

CSX Transportation, Inc. (2012). Intermodal Schedules. Retrieved from CSX Transportation, Inc. Web site: http://shipcsx.com/public/ec.shipcsxpublic/Main?module=public.ischedule

Dizikes, P. (2010, November 8). The 6-percent solution. Retrieved from MIT news Web site: http://web.mit.edu/newsoffice/2010/corporate-greenhouse-gas-1108.html

Environmental Defense Fund. (2012, February 17). Smart Moves: Creative Supply Chain Strategies Are Cutting Costs and Emissions. Retrieved from Environmental Defense Fund Web site: http://business.edf.org/sites/business.edf.org/files/smartmoves_07_screen.pdf

Florida East Coast Railway. (2011). About. Retrieved from Florida East Coast Railway Web site: http://fecrwy.com/about

Herbert Engineering Corporation. (2011, June). Carbon Footprint Study for the Asia to North America Intermodal Trade. Retrieved from Port of Seattle Web site: http://www.portseattle.org/cargo/green-gateway/documents/carbon_footprint_study_20110610.pdf

ICF Consultung. (2005, April). Assessing the Effects of Freight Movement on Air Quality at the National and Regional Level. Retrieved from US Federal Highway Administration Web site: http://www.fhwa.dot.gov/environment/air_quality/publications/effects_of_freight_movement/

Maersk Line. (2010, August 23). Slow steaming is here to stay. Retrieved from Maersk Line Web site: http://www.maerskline.com/link/?page=news&path=/news/story_page/10/slow_steaming

Maersk Line. (2012, March 5). Maersk Line Brings Flagship Service Powered by BNSF Railway to North America. Retrieved from Maersk Line North America Web site: http://www.maersk-nam-marketing.com/advisories/2012/03/maersk-line-brings-flagship-service-powered-by-bnsf-railway-to-north-america/

Maersk Line. (n.d.). About Us. Retrieved from Maersk Line Web site: http://www.maerskline.com/link/?page=brochure&path=/about_us

Maersk Line. (n.d.). Containers. Retrieved from Maersk Line Web site: http://www.maerskline.com/link/?page=brochure&path=/our_services/containers

Maersk Line. (n.d.). Rates. Retrieved from Maersk Line Web site: https://www.maerskline.com/appmanager/maerskline/public?_nfpb=true&_nfls=false&_pageLabel=page_rate_search

Maersk Line. (n.d.). Reducing greenhouse gas emissions. Retrieved from Maersk Line Web site: http://www.maerskline.com/link/?page=brochure&path=/about_us/environment/reducing_gas_emissions

Maersk Line. (n.d.). Schedules. Retrieved from Maersk Line Web site: http://www.maerskline.com/appmanager/maerskline/public?_nfpb=true&_nfls=false&_pageLabel=page_schedules_location

Maersk Line. (n.d.). Transpacific. Retrieved from Maersk Line Web site: http://www.maerskline.com/link/?page=brochure&path=/routemaps/newnetwork/transpacific

Muller, G. (1999). Intermodal Freight Transportation. Eno Transportation Foundation, Inc.


45

Net Resources International. (n.d.). Alameda Corridor Freight Line, United States of America. Retrieved from railway-technology.com Web site: http://www.railway-technology.com/projects/alameda/

Nike, Inc. (2010). Corporate Responsibility Report FY07-FY09. Retrieved from Nike, Inc. Web site: http://www.nikebiz.com/crreport/pdf/

Norfolk Southern Corp. (2012, March 1). International Service Matrix. Retrieved from Norfolk Southern Intermodal Web site: http://www.nscorp.com/nscintermodal/Intermodal/Business/Business_Groups/International/Service_Matrix/international.xls

Oxford University Press. (2012). Dictionary. Retrieved from Oxford Dictionaries Web site: http://oxforddictionaries.com/?region=us

Pacer International, Inc. (2004, April 27). Investor Relations. Retrieved from Pacer International, Inc. Web site: http://phx.corporate-ir.net/phoenix.zhtml?c=107717&p=irol-newsArticle&ID=568736&highlight=

Petromedia Ltd. (2012). PortWorld Distance - Ship Voyage Distance Calculator. Retrieved from PortWorld Web site: http://www.portworld.com/map/

Thomas Publishing Company. (2012). Glossary of Supply Chain Terms. Retrieved from Inbound Logistics Web site: http://www.inboundlogistics.com/cms/logistics-glossary/

Union Pacific Railroad Company. (2012, September 17). Intermodal. Retrieved from Union Pacific Railroad Company Web site: http://www.uprr.com/customers/intermodal/attachments/intl_matrix.pdf

Union Pacific Railroad Company. (n.d.). International Intermodal. Retrieved March 2012, from Union Pacific Railroad Company Web site: https://www.uprr.com/customers/intermodal/intl/index.shtml

US Department of Commerce. (2009). Distances Between United States Ports. Retrieved from National Oceanic and Atmospheric Administration Web site: www.nauticalcharts.noaa.gov/nsd/distances-ports/distances.pdf

US Department of Transportation. (2011, January). America’s Container Ports: Linking Markets at Home and Abroad. Retrieved from Research and Innovative Technology Administration Web site: http://www.bts.gov/publications/americas_container_ports/2011/

US Environmental Protection Agency. (2008, May). Climate Leaders Greenhouse Gas Inventory Protocol Core Model Guidance: Optional Emissions from Commuting, Business Travel and Product Transport. Retrieved from US Environmental Protection Agency Web site: http://www.epa.gov/climateleadership/documents/resources/commute_travel_product.pdf

US Environmental Protection Agency. (2008, May). Climate Leaders Greenhouse Gas Inventory Protocol Core Module Guidance: Direct Emissions from Mobile Combustion Sources. Retrieved from US Environmental Protection Agency Web site: http://www.epa.gov/climateleadership/documents/resources/mobilesource_guidance.pdf

US Environmental Protection Agency. (2012, November 5). SmartWay Transport Partnership Freight Shippers. Retrieved from US Environmental Protection Agency Web site: http://www.epa.gov/smartway/documents/partnership/trucks/public-bin-export.xlsx

Wikipedia. (2009, February 23). File:BNSF Railway system map.svg. Retrieved from Wikipedia Web site: http://en.wikipedia.org/w/index.php?title=File:BNSF_Railway_system_map.svg&page=1

Wikipedia. (2009, June 7). File:Canadian National System Map.PNG. Retrieved from Wikipedia Web site: http://en.wikipedia.org/wiki/File:Canadian_National_System_Map.PNG

Wikipedia. (2009, June 9). File:Canadian Pacific System Railmap.PNG. Retrieved from Wikipedia Web site: http://en.wikipedia.org/wiki/File:Canadian_Pacific_System_Railmap.PNG


46

Wikipedia. (2009, February 23). File:CSX Transportation system map.svg. Retrieved from Wikipedia Web site: http://en.wikipedia.org/w/index.php?title=File:CSX_Transportation_system_map.svg&page=1

Wikipedia. (2009, February 23). File:Norfolk Southern Railway system map.svg. Retrieved from Wikipedia Web site: http://en.wikipedia.org/w/index.php?title=File:Norfolk_Southern_Railway_system_map.svg&page=1

Wikipedia. (2009, March 18). File:Union Pacific Railroad system map.svg. Retrieved from Wikipedia Web site: http://en.wikipedia.org/w/index.php?title=File:Union_Pacific_Railroad_system_map.svg&page=1

World Economic Forum. (2009). Supply Chain Decarbonization: The Role of Logistics and Transportation in Reducing Supply Chain Carbon Emissions. Retrieved from World Economic Forum Web site: https://members.weforum.org/pdf/ip/SupplyChainDecarbonization.pdf

World Resources Institute. (2009, June). GHG Protocol tool for mobile combustion. Version 2.0. Retrieved from Greenhouse Gas Protocol Web site: http://www.ghgprotocol.org/files/ghgp/tools/WRI_Transport_Tool.xls

World Resources Institute. (2012, August). Emission Factors from Cross-Sector Tools. Version 1.3. Retrieved from Greenhouse Gas Protocol Web site: http://www.ghgprotocol.org/files/ghgp/tools/Emission-Factors-from-Cross-Sector-Tools-(August-2012).xlsx


47

Attachments

Volume and Mass


48

Distance

Emission Factors

Cost: 20-Foot Standard Container


49

Cost: 40-Foot Standard Container

Cost: 40-Foot High Cube Container

Cost: 45-Foot High Cube Container


50

Time

Emissions: 20-Foot Standard Container

Emissions: 40-Foot Standard Container


51

Emissions: 40-Foot High Cube Container

Emissions: 45-Foot High Cube Container


52

Multiple Objective Analysis Results: 20-Foot Standard Container


53


54

Multiple Objective Analysis Results: 40-Foot Standard Container


55


56

Multiple Objective Analysis Results: 40-Foot High Cube Container


57


58

Multiple Objective Analysis Results: 45-Foot High Cube Container


59


60

Sensitivity Analysis: 20-Foot Standard Container

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.0 0.2 0.4 0.6 0.8 1.0

Ove

rall

uti

lity

Weight on Emissions

Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.0 0.2 0.4 0.6 0.8 1.0

Ove

rall

uti

lity

Weight on Time

Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40

Ove

rall

uti

lity


Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15


61

Sensitivity Analysis: 40-Foot Standard Container

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0.0 0.2 0.4 0.6 0.8 1.0

Ove

rall

uti

lity

Weight on Emissions

Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.0 0.2 0.4 0.6 0.8 1.0

Ove

rall

uti

lity

Weight on Time

Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40

Ove

rall

uti

lity


Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15


62

Sensitivity Analysis: 40-Foot High Cube Container

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0.0 0.2 0.4 0.6 0.8 1.0

Ove

rall

uti

lity

Weight on Emissions

Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.0 0.2 0.4 0.6 0.8 1.0

Ove

rall

uti

lity

Weight on Time

Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15

(0.2)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40

Ove

rall

uti

lity


Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15


63

Sensitivity Analysis: 45-Foot High Cube Container

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0.0 0.2 0.4 0.6 0.8 1.0

Ove

rall

uti

lity

Weight on Emissions

Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.0 0.2 0.4 0.6 0.8 1.0

Ove

rall

uti

lity

Weight on Time

Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15

(0.5)

0.0

0.5

1.0

1.5

2.0

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40

Ove

rall

uti

lity


Route 1

Route 2

Route 3

Route 4

Route 5

Route 6

Route 7

Route 8

Route 9

Route 10

Route 11

Route 12

Route 13

Route 14

Route 15

Documents

Case Study: Multiple Objective Analysis of Intermodal