Content uploaded by Kumar Venkat
Author content
All content in this area was uploaded by Kumar Venkat on Sep 15, 2019
Content may be subject to copyright.
211
J.I. Boye and Y. Arcand (eds.), Green Technologies in Food Production and Processing,
Food Engineering Series, DOI 10.1007/978-1-4614-1587-9_9,
© Springer Science+Business Media, LLC 2012
9.1 Introduction
Transportation is the largest end-use contributor toward global warming in the
United States and many other developed countries. The U.S. Department of Energy
(DOE 2009 ) calculates that CO 2 emissions from transportation surpassed two bil-
lion metric tons in 2007. Yet a survey by Golicic et al. ( 2010 ) fi nds that fewer than
10% of Fortune 500 companies have addressed the environmental impacts of trans-
portation, and even fewer are actively implementing improvements, despite the fact
that such initiatives would also tend to reduce fuel usage and costs in the long run.
Transportation has a signifi cant impact within the food and beverage sector
because food is often shipped long distances and not infrequently via air. Heller and
Keoleian ( 2000 ) estimate that diesel fuel use accounts for 25% of the total energy
consumed within the U.S. food system. Pirog et al. ( 2001 ) report that nearly half of
all fruit sold in the United States is imported, and that produce grown in North
America travels an average of 2,000 km from source to point of sale.
Although the impact of transportation is important, full life cycle analyses indi-
cate that for most foods transportation does not have the largest environmental
impact. Some analysts, such as Weber and Matthews ( 2008 ) , estimate that given the
W. Wakeland (*)
Portland State University , Portland , OR , USA
e-mail: wakeland@pdx.edu
S. Cholette
San Francisco State University , San Francisco , CA , USA
e-mail: cholette@sfsu.edu
K. Venkat
CleanMetrics Corporation , Portland , OR , USA
e-mail: kvenkat@cleanmetrics.com
Chapter 9
Food transportation issues and reducing
carbon footprint
Wayne Wakeland , Susan Cholette , and Kumar Venkat
212 W. Wakeland et al.
typical household food basket, aggregate transportation accounts for just 11% of
total carbon emissions associated with food production. We show in Sect. 9.5 that
freight transport accounts for just 6% of overall emissions in the U.S. food sector,
but its life cycle impact is greater in the case of plant-based foods that have rela-
tively low production emissions. Therefore, it is still worthwhile to consider improv-
ing the food distribution system. There are often many options for delivering food
to consumers, and these supply chain confi gurations can result in vastly differing
energy and emissions profi les. In this chapter, we provide the background and tools
for analyzing the energy intensity and resultant emissions of a food distribution
system, evaluating tradeoffs and identifying opportunities for signifi cant improve-
ment. Note that we use the terms “carbon emissions” and “greenhouse gas emis-
sions” interchangeably in this chapter, implying in both cases that all signifi cant
greenhouse gases emitted in a process are counted and reported as a single “carbon
dioxide equivalent” fi gure.
9.2 Supply chain basics
Before we can further investigate transportation impacts, we must fi rst introduce the
concept of the supply chain: the sequenced network of facilities and activities that
support the production and delivery of a good or service. Given the obvious impor-
tance of the supply chain, this fi eld is rife with terminology and buzzwords, many
of which are synonymous. For instance, supply chains are sometimes referred to as
“demand chains” or “value chains.” A supply chain starts with basic suppliers and
extends all the way to consumers via stages. These stages may include such facili-
ties as suppliers, factories, warehouses and other storage facilities, distribution cen-
ters, and retail outlets.
Figure 9.1 shows a sample supply chain, where the arrows denote the fl ow of a
product toward the consumer. This fi gure depicts both inbound logistics (the deliv-
ery of raw materials and packaging to the manufacturer) as well as outbound logis-
tics (the transportation and storage of the fi nished good to the end consumer). This
chapter focuses on outbound logistics, colloquially known as “gate-to-kitchen” and
“farm-to-fork” in the food and beverage industry. The emissions associated with
outbound logistics vary by origin and type of food. Weber and Matthews ( 2008 )
estimate that food transportation may account for 50% of total carbon emissions for
many fruits and vegetables, but less than 10% for red meat products. Although
Fig. 9.1 A simple supply chain
213
9 Food transportation issues and reducing carbon footprint
inbound logistics can require substantial energy use, it is considered part of the
production process and is discussed in earlier chapters.
Although the interrelationships between supply chain stages may be quite com-
plex, all supply chains have one aspect in common—they end with a consumer.
Supply chains for different products may be interlinked; one supply chain’s end
consumer may represent an intermediate node for another supply chain. Examples
include a fi rm that buys components and assembles them into consumer items, and
a soft drink producer that buys cylinders of compressed CO
2 to carbonate its
products.
Much supply chain complexity results from the fact that few supply chains are
completely controlled by one fi rm or vertically integrated. For example, producers
and retailers are not typically owned by the same fi rm. Companies may outsource
supply chain activities, especially transport and storage activities, which are handled
more effectively by third party logistic (3PL) providers. Outside fi rms that form a
part of a company’s supply chain are channel partners. These partnerships require
collaboration across organizations. We defi ne supply chain management (SCM) as
the coordination of business functions within an organization and between the orga-
nization and its channel partners. SCM strives to provide goods and services that
fulfi ll customer demand responsively, effi ciently, and sustainably.
SCM includes such functions as demand forecasting, purchasing (also known as
sourcing), customer relationship management (CRM), and logistics. Logistics con-
cerns the movement and storage of goods, services, and information. It is an umbrella
term for such important functions as transportation, inventory management, pack-
aging, and returns/reverse logistics. Some terminology will be helpful to understand
who is doing what. The shipper initiates the movement of the product forward into
the supply chain, the carrier is the party that does the actual moving of the product,
and the consignee receives the product.
9.2.1 Transport modes
Within the developed world there are four basic transport modes for shipping large
quantities of packaged products: water, rail, truck, and air. Trucking dominates,
comprising more than 75% of the total U.S. freight transit bill. Trucking variables
include truck type, ownership model (such as 3PL or company-owned fl eet), and
loading option (less-than-truckload or full-truckload). The dominant transport mode
has shifted over time. The fi rst transport revolution occurred when inland water
transport replaced animal caravans. In the mid-1800s railroads displaced inland
water as the dominant form of cargo transport, and in the mid-1900s trucking dis-
placed railroads. Air cargo is a more recent and growing transport mode popular for
short life cycle products such as fl owers and luxury foods. The U.S. DOE ( 2009 )
estimates that air transport accounts for 9% of U.S. transportation fuel usage.
Interestingly, water transport has started to make a comeback. In the United Kingdom
(UK), for example, Tesco is relying on inland waterway barges for transporting
214 W. Wakeland et al.
more of their beverage products. Short sea shipping, using ocean-going vessels for
delivering cargo domestically, is popular in Europe and also holds promise for
replacing many truck deliveries in the United States.
To compare transport modes with regard to energy usage and resultant emis-
sions, we defi ne a ton-km as the movement of 1 metric ton of cargo over 1 km.
Table 9.1 shows that these modes have very different energy and emissions profi les.
Caveats abound regarding the accuracy of these fi gures, but clearly air freighting is
much more energy and emission intensive compared to other modes, especially
water and rail. Of course, water and rail transport modes are contingent upon the
availability of navigable water and established railroad tracks. An additional consid-
eration is the potential need for supply chain responsiveness: air freight may be the
only viable option for long-distance transport when customer orders require imme-
diate fulfi llment.
9.2.2 Intermodal transport
Before we choose one mode over another, we should consider intermodal transport.
Defi ned as using more than one transportation mode to move a shipment between
two points, an intermodal route might involve shipping cargo by water, then by rail,
then by truck. Intermodal transport became practical with the advent of container-
ization, where products stay in the same container throughout their entire journey.
Containerization was made possible through global standardization of container
size and features, which dramatically reduced intermodal transfer times and signifi -
cantly increased cost effi ciency. From a sustainability viewpoint, the advantage of
intermodal transport is that we can utilize more effi cient modes for major transport
corridors, and then shift to trucks for transport to remote destinations. Shippers can
also use a 3PL provider to oversee the entire shipping process. One disadvantage of
intermodal transport is its inherent complexity of coordination and the information
technology support required to address that complexity. Another issue is the move-
ment and repositioning of empty containers.
Table 9.1 Energy and emissions per ton-km
MegaJoules per ton-km kg CO
2 e per ton-km
International water-container 0.2 0.14
Inland water 0.3 0.21
Rail a 0.3 0.18 a
Truck b 2.7 1.8
Air c 10 6.8
Note that utilization and backhaul rates will affect all fi gures
a May depend on whether diesel or electric power is used
b Depends on size and type of truck, power source
c Includes effects from radiative forcing
Source: Based on data from Weber and Matthews (
2008 )
215
9 Food transportation issues and reducing carbon footprint
9.2.3 Utilization and backhaul
Many carbon analyzers base calculations on only transport mode and shipping dis-
tance. In our analysis, we will take into account additional factors, including vehicle
utilization (how full the vehicle is) and backhaul (whether or not the vehicle carries
freight on its return journey). Although fully laden vehicles use more fuel than
nearly empty ones, most of the energy expended during a trip is used to move the
vehicle and not its cargo. Underutilized vehicles waste energy, as do vehicles that
return empty. Also, weight and volume limits must be respected, and all but the
lightest and bulkiest cargo loads tend to “weigh out” rather than “cube out.”
It can be diffi cult to determine utilization fractions and backhaul percentages, as
these are likely to vary with each trip. Such information is even more challenging to
obtain when transportation functions have been outsourced. However, some assump-
tions can be made. For example, vehicles chartered by 3PL providers are likely to
have higher utilization fractions because they often carry cargo from multiple com-
panies. Third party logistic providers are also likely to have higher backhaul rates,
because they have more opportunities for obtaining return freight owing to their
broader customer base.
9.2.4 Warehousing
Logistics involves not only the movement of goods, but their storage. Unless a prod-
uct is custom ordered by an onsite client, it is likely that the product will enter stor-
age at some point in its journey to the consumer. Such storage can occur at any
supply chain stage: at the producer, distributor warehouse, and/or retailer stock-
room. Intermediate supply chain stages range from pure storage centers to dedicated
cross-dock facilities, in which cargo from upstream supplier trucks/railcars is trans-
ferred directly to outbound trucks/railcars destined for downstream stages. In addi-
tion to storage, warehouses can provide additional services: pick and pack
(repackaging palletized products to smaller quantities destined for either retailers or
end consumers), customs clearance, or even house product-fi nishing functions such
as customizing goods to the local marketplace.
9.2.5 Packaging
Packaging decisions are inherently linked to the supply chain. Goods are frequently
shipped in bulk and broken into consumer-sized quantities at a warehouse or other
facility, and individual commodities are sometimes bundled into larger end-items,
such as multipacks, and palletized. Packaging materials (pallets, boxes, totes, slip-
sheets, etc.) for both fi nished goods and intermediate support functions may be
216 W. Wakeland et al.
designed to be recyclable, compostable, or reusable. Non-landfi lled packaging is
highly desirable, but creates other challenges, such as the impact of reusable pack-
aging in the reverse supply chain.
Packaging can often be reengineered to reduce package weight or bulk, which
can translate into savings in raw materials, landfi ll impacts, and transport/storage
energy use; but extra costs may be incurred elsewhere. For example, Safeway is
working with Kimberly-Clark to pilot palletless deliveries of paper products.
Although this may allow trucks to be packed with more products, labor costs are
likely to increase in receiving, and pallets would still not have been totally elimi-
nated from the supply chain because pallets would continue to be used at local
warehouses.
9.3 What makes food supply chains special?
We have shown that supply chains can be long and complex. Food supply chains are
some of the most diffi cult to manage as they must often address time constraints to
avoid spoilage, as well as concerns about contamination, high weight-to-value
ratios, fragility, unique packaging requirements, and the potential impact of food
being wasted rather than consumed. We show here how these considerations affect
outbound logistics.
One challenge relates to food production being inherently dependent on nature.
Not only is the cultivation of many foods restricted geographically, but also tempo-
rally. Fruits, vegetables, and grains typically have fi xed growing cycles with short
and specifi c annual harvest periods. However, North American and European demand
for many of these items is year round. There are three options for supplying fresh
produce that is out of season locally: sourcing from distant growing areas, using
long-term storage, or cultivating in a protected environment such as a greenhouse.
Importing produce often results in lower overall emissions than harvesting and
storing local produce for several months, as Hospido et al. ( 2009 ) and Milà i Canals
et al. ( 2007 ) show for lettuce and apples, respectively. Indeed, energy needed for
long-term cold storage can dominate a product’s overall emissions profi le. Carlsson-
Kanyama ( 1998 ) shows, for example, that storage accounts for 60% of the carbon
emissions associated with carrots. Higher emissions can result not only from the
energy needed for climate control, but also from the inherent yield losses that occur
during storage. Protected cultivation is even more energy intensive. Carlsson-
Kanyama et al. ( 2003 ) show that tomatoes produced locally in Swedish greenhouses
require ten times the energy as fi eld-grown tomatoes imported from Southern Europe.
Thus, long-distance supply chains, even though they are energy intensive, may yield
the lowest overall footprint for providing out-of-season product to consumers.
A second challenge is related to situations where similar food commodities are
produced locally as well as imported from distant locations, the emissions intensity
of the production methods must be considered in any comparison of overall supply
chain emissions. For example, Saunders and Barber ( 2007 ) fi nd that milk solids
217
9 Food transportation issues and reducing carbon footprint
produced locally in the United Kingdom generate 34% more emissions than the
same product imported from New Zealand, even with transport included. This result
refl ects the more energy-intensive dairy production system in the United Kingdom.
A third challenge is that highly perishable foods require special handling to avoid
yield loss and potential health issues. These foods often require cooling, refrigera-
tion, or freezing during transport and/or storage. It may also be necessary to control
other conditions, such as humidity, exposure to air, or contact with other items.
These requirements increase energy usage and emissions.
A fourth challenge is that the location of facilities within a food supply chain can
also affect emissions. For example, Sim et al. (
2007 ) fi nd that overall carbon emis-
sions can be signifi cantly reduced by locating processing and storage facilities in
countries where more electricity is generated from renewable fuels or cleaner
energy.
Fifth, when time is of the essence, as in the transport of highly perishable pro-
duce such as berries, air freight may be the only viable transport option. Air freight-
ing may also be necessary in regions such as Africa where no other viable alternative
exists for transporting produce to market. As previously shown, air freighting is
highly energy intensive. Scholz et al. ( 2009 ) report that fresh salmon air freighted
from overseas has about twice the environmental impact as frozen salmon trans-
ported by container ships over the same distance. The difference owing to transport
modes is far more signifi cant in this case than production choices such as wild ver-
sus farmed or organic versus conventional.
A sixth challenge is that safe food storage not only requires climate controls, but
also a high degree of sanitation. In most developed countries, warehouses must be
built and maintained to stringent guidelines to be certifi ed as “food grade.” In the
United States, wood pallets may not be reused and may soon be phased out as
unsanitary.
The process of packaging food is yet another challenge. Twede et al. ( 2000 )
emphasize that packaging beverage products is a high-speed automated process
involving expensive equipment. Such capital investment and the need for a con-
trolled environment favors centralizing packaging at the point of production, even if
it might be more energy effi cient to ship product in bulk. Food and beverage prod-
ucts typically require extensive packaging, which adds both weight and volume to
the product. Additional energy and materials are required to create the packaging
and transport it to the production site. Point ( 2008 ) performs a life cycle assessment
of the Nova Scotia wine industry and fi nds that the largest contribution to emissions
is owing to the production and transport of wine bottles.
9.4 Measuring transportation-related carbon emissions
This section presents the basics of performing a carbon audit and concludes with
some examples from practice. Although other gases such as methane and nitrous
oxide may contribute to global warming, aggregate greenhouse gas measures are
218 W. Wakeland et al.
typically reported in CO
2 equivalents (CO
2 e), which is kgs of CO
2 emitted per kg of
product. Carbon dioxide dominates, comprising 95% of total greenhouse gas emis-
sions by volume (World Resources Institute 2004 ) . Some points of confusion in
carbon measures exist. For instance, U.S. documents report tons emitted, where
most of the world measures in the SI (metric) units and reports in tons, as will we.
Emissions are colloquially called “carbon emissions,” which can lead to confusion
as some older studies only weigh the carbon component of the gas, which is 30% of
the total mass of CO
2 . It is also now standard practice to report all signifi cant green-
house gas emissions as a single carbon emissions fi gure. The scope of the analysis
depends on the purpose of the study. Scope 1 includes only direct emissions, whereas
Scope 2 also includes indirect emissions from any consumption of purchased elec-
tricity, heat, or steam. Scope 3 is the broadest, including all other indirect emissions,
such as the extraction and production of purchased materials and fuels, all out-
sourced activities, and waste disposal. Scope 3 can also include the substantial
impact of radiative forcing from the contrails in tallying airplane emissions.
Scope must be carefully considered because incomplete framing (inappropriate
scope) may lead to incorrect conclusions. For example, food miles are defi ned as the
distance between the production source and the retail store, or “farm-to-fork.” This
metric has received substantial attention in the popular press and has been adopted
by the business community. For example, the UK supermarket chain Tesco now
provides food mile information. However, there is often no consideration of the
energy used to transport supplies to the farm or the energy used for processing
or storage. For example, Saunders et al. ( 2006 ) estimate that grass-fed lamb from
New Zealand produces lower emissions overall than locally raised lamb fattened in
a feedlot. Transport modes such as ocean and rail may actually be more effi cient on
a per-weight basis, even over long distances. Considering another dimension of
sustainability, some African and South American farmers derive their livelihood
from the service export markets. Tradeoffs between different facets of sustainability
(environment, economics, equity; or planet, profi ts, people) can mean that carbon-
based metrics may be misleading, especially in the context of incomplete framing.
Now that we understand supply chain basics and special logistical issues faced in
food distribution, it would seem we are ready to collect data and enter this data into
an analysis tool to derive the defi nitive answer to the question: How much carbon
does our supply chain emit? Before we start broadcasting results with conviction
and certainty, however, let us consider the following scenario. A person travels from
San Francisco to New York and desires to purchase carbon offsets for the round-trip
fl ight. The Internet has many free online carbon calculators, often with donation
links for offsetting one’s carbon footprint. Table
9.2 shows results from several cal-
culators that target the typical U.S. consumer and report results in tons (the unit
used in the United States). Even for a well-defi ned trip, the emissions reported and
the recommended amount of carbon offsets to purchase vary widely. The amounts
vary for both logical reasons (such as whether radiative forcing is included or not)
and for obscure reasons (such as JetBlue’s claim to being almost twice as carbon
effi cient as United Airlines). Note also that there are many other factors that may or
may not have been considered by these carbon calculators, such as plane age and
219
9 Food transportation issues and reducing carbon footprint
model, weather, utilization, and backhaul. Table 9.2 also shows that even when the
amount of emissions being offset is about the same, the suggested donation is dif-
ferent. Different carbon calculators make different assumptions about the price per
unit of offset.
So what lessons can be learned? First, it may not be realistic to expect highly
accurate estimates of carbon emissions. It may be more important to strive for con-
sistency and to avoid using different tools or techniques when comparing across
scenarios. Johnson ( 2008 ) recommends that carbon footprints be defi ned sensibly
and transparently because defi nitive standards have not yet emerged. It is also
important to consider who will use the analyses and for what purpose. When provid-
ing consumers with recommendations for assessing personal transportation foot-
prints, it may be appropriate to acknowledge the disparities between calculators and
to provide a range of recommended offsets. There is a risk of backlash when savvy
Internet users survey websites and discover how divergent the results can be. As an
example of consumer skepticism, Rosenthal ( 2009 ) reports that in Sweden, where
carbon labeling is starting to appear, the typical consumer reaction to carbon labels
is bemusement.
In addition to providing consumers with information, carbon audits can provide
useful insights for companies evaluating their operations. However, we believe that
the fi gures from carbon audits should be viewed as guidelines rather than as precise
and absolute truths. Given these caveats, we are now ready to consider some results
from actual carbon audits performed for companies.
Case study: wine delivery to consumers
To illustrate the impact that supply chain design and implementation can have on
carbon emissions, our fi rst case considers the delivery of wine to the consumer.
Energy usage associated with postproduction logistics is high for wine because the
standard consumer packaging (a 750 mL glass bottle) is fragile, heavy, and bulky.
Wine comprises 50% of the weight and less than 40% of the volume of a case of
12 bottles. Wine is also sensitive to temperature and must be stored in a controlled
Table 9.2 Divergent results of online carbon calculators—round trip from San Francisco to
New York
Tons CO 2 e Recommended offset Implied $ per ton
Carbonfund.org 0.93 $9.34 $10.04
Adding radiative forcing 2.52 $25.22 $10.01
Terrapass.com
Via JetBlue 1.462 $11.90 $8.14
Via Virgin 1.584
Via United Airlines 2.215 $11.73 $5.30
Sustainabletravelinternational.org 1.86 $47.31 $25.44
Nativeenergy.com 2.055 $42.00 $20.44
Bonneville Education Foundation
www.b-e-f.org 4.192 $56.00 $13.36
220 W. Wakeland et al.
climate for all but the shortest periods. The outbound logistics network, depicted in
Fig. 9.2 , depicts paths with differing numbers of intermediary stages and several
retail channels. The majority of wine sold in the United States is delivered within
the framework of the 3-tier system, with product fl owing from the winery to a dis-
tributor/wholesaler, then to a retailer before reaching the end consumer. However,
in most states, wine can also be sold directly to the end consumer, either through
tasting room sales, or through wine clubs and mailing lists where delivery is typi-
cally supported through a 3PL provider such as IBG or ShipCompliant, Inc. A few
states such as California and Oregon allow wineries to self-distribute directly to
both on-premise and off-premise retailers.
Cholette and Venkat ( 2009 ) use an online carbon calculator to model each of these
options and provide a stage-by-stage view of resulting emissions for shipping a half
case of wine to both local and cross-country consumers. The tool utilized, CargoScope
(Venkat 2008 ) , considers transportation distances, mode, temperature control, utili-
zation, and backhaul rates for each link within the supply chain. Figure 9.3 provides
a comparison of some of the different options investigated both for small and long-
distance consumers. Not surprisingly, cross-country transport by rail (scenario D4) is
more effi cient than trucking (D1), which in turn is better than air freight (D2). Direct-
to-consumer small package local delivery (L4) is the most effi cient, in part because
the overall transportation distance is minimized and the vehicles servicing the deliv-
ery area are highly utilized because of its compactness. Making a dedicated trip to the
winery in a typical gasoline-powered car (L5a) results in 80 times more emissions
than the least carbon-intensive method (L4). Most local commercial delivery con-
fi gurations result in lower emissions than their long-distance counterparts, but there
is a notable exception: long-distance delivery via rail (D4) is effectively equivalent to
the standard, local 3-tier distribution scenario (L1).
Figure 9.3 also shows that the most energy-intensive transit link is often the last
one–driving to the store. This is not surprising, because other studies, such as those
by Browne et al. ( 2005 ) and Van Hauwermeiren et al. ( 2007 ) , fi nd that this link can
Fig. 9.2 An outbound logistics network for a winery
221
9 Food transportation issues and reducing carbon footprint
be the most carbon intensive even in European countries where consumers are tra-
ditionally more energy conscious than their U.S. counterparts. Because it is the least
measurable and the most diffi cult to control, the retail-to-consumer link is typically
outside the system boundary of most analyses. However, it may be worth consider-
ing the retail-to-consumer link when options include home delivery or when it may
be possible to infl uence consumer shopping behavior. For example, if wine produc-
ers or retailers could provide incentives for consumers either to make larger pur-
chases less often or to visit the store by bicycle, foot, or public transport, then the
energy intensity of this last segment could be reduced. However, for supply chains
involving multiple stages, such incentives may be impractical or diffi cult to imple-
ment and monitor.
Cholette and Venkat ( 2009 ) also fi nd that no single supply chain confi guration is
ideal for all wineries. Large wineries that sell in volume to retailers, whereby a typi-
cal delivery would fi ll a reasonably effi cient midsized truck, could consider self-
distribution. For small wineries, where a typical delivery would fi ll a less-effi cient
light truck, 3-tier distribution would be more effi cient than self-distribution. For
retail store chains, the key to reducing carbon emissions can be to design their sup-
ply chain to maintain high vehicle utilization rates. For example, if a store chain
moves suffi cient volumes to be able to keep their fl eet of delivery trucks fully uti-
lized transporting goods to and from their distributors’ warehouses, the third tier of
the 3-tier distribution channel (a central warehouse for the retailer) may not be nec-
essary, thereby saving considerable cost and reducing carbon emissions.
Fig. 9.3 Carbon emissions for local and long distance delivery scenario (Based on Cholette and
Venkat
2009 )
222 W. Wakeland et al.
Case study: an online supply chain emissions calculator
Organically Grown Co. (OGC), an Oregon-based regional distributor of organically
produced fruits and vegetables, undertook a major initiative in 2008 to track the
transport of every product that reached one of their three distribution centers located
in northwestern United States. The company wanted to be responsive to customer
requests for information about the environmental impacts of their products and to
be a part of the sustainability discussion in the region.
About one third of the company’s products are sourced from regional growers;
the remainder are transported long distances from other states such as California
and Hawaii, as well as from South America and other distant locations such as New
Zealand. Transport modes include air, ocean, and land, as well as multimodal trans-
port. Many products require temperature control during transit and storage. Products
that have very short shelf lives often must be transported by air, whereas others can
be sent refrigerated by ocean or land.
A web-based greenhouse gas emissions analyzer developed by CleanMetrics
Corp. (2010) made use of OGC’s database of thousands of suppliers and products,
to compare the transport impacts of products sourced from different suppliers and
locations. Figure
9.4 shows a typical result from the analysis of an international
supply chain for apples, a typical crop that has a fi xed harvesting cycle, which
necessitates diverse sourcing in order to meet year-round demand. The destination
Fig. 9.4 Analysis of a supply chain for apples
223
9 Food transportation issues and reducing carbon footprint
is a warehouse located in Eugene, Oregon, and the product is sourced from a variety
of producers, ranging from growers that are local (Oregon) and regional (California
and Washington) to international (Argentina and New Zealand). Results from these
analyses provide OGC with visibility into product supply chains and information
needed to incorporate transportation-related carbon footprint as one component of
their food purchasing decisions.
Case study: an international supply chain for packaged fruit products
Sundia Corporation, based in Oakland, California, produces and distributes fresh
fruits cut and packaged in plastic cups. Once packaged, the products do not require
refrigeration until they are placed on supermarket shelves. This process allows the
company to procure tropical fruits in Asian countries such as Vietnam and Thailand
where they are grown, process the fruits close to the source location, and then ship the
packaged products by ocean to stores in New York and California at a relatively low
carbon cost. Figure 9.5 depicts the entire supply chain. A carbon footprint analysis of
the supply chain shows that only about 30% of total emissions could be attributed to
transport. The longest transport leg, the ocean segment from an Asian port to the
United States, spanning 14,000–20,000 km, was responsible for just half of the trans-
port-related emissions. The remaining 70% of total emissions came from growing,
processing, and packaging the product. This is one of many recent industry examples
where transport distances have not been good indicators of overall life cycle green-
house gas emissions.
Fig. 9.5 A packaged fruit product supply chain
224 W. Wakeland et al.
Case study: home delivery of groceries versus consumers driving to stores
Many national and regional grocery chains now offer convenient home delivery
services. An example of this is the service provided by New Seasons Market in the
Portland, Oregon metropolitan area. The company uses a fl eet of delivery vans with
the capacity to make up to ten deliveries on a single route. The vans are fueled with
a blend consisting of 20% biodiesel, and they have separate cargo areas for unrefrig-
erated, refrigerated, and frozen goods. Deliveries are made directly from each of the
company’s retail stores to customers within a certain distance from that store.
Customers typically place their grocery orders online and receive an estimated
delivery time. Store employees shop for each customer and then load the fi lled
shopping bags into a designated van. Delivery routes are calculated and mapped in
advance using mapping and routing software such as MapPoint.
A carbon footprint study conducted by CleanMetrics (
2010 ) for New Seasons
Market compared the emissions produced by the delivery vans with the emissions that
would have been generated had the customers driven from their homes to the nearest
store. Using delivery data that included street addresses and delivery routes, the study
calculated carbon emissions for both scenarios using actual driving distances and
found that the delivery vans were more effi cient by a factor of almost two. This result
suggests that there may be signifi cant potential for emissions reductions in the last leg
of most supply chains where the product is fi nally delivered to consumers.
9.5 Putting transport emissions in context
Although transportation generally does not have the largest environmental impact in
food supply chains, it can play a signifi cant role depending on the specifi c supply
chain and the modes of transport used. In this section, we put transportation-related
greenhouse gas emissions in context by considering the major life cycle phases of
food products in the U.S. food sector.
Based on food consumption and food waste data published by the United States
Department of Agriculture (USDA 2008 ) , we analyzed the typical life cycles of
major food categories consumed in the United States, including meats, dairy, eggs,
seafood, grains, nuts, vegetables, and fruits. This data set includes quantities of raw
and processed food products delivered to retail locations, as well as percentages of
food wasted at the retail and consumer levels. The analysis was conducted using the
CarbonScope analytical tool (Venkat 2009 ) . Typical cooking and waste disposal
processes were assumed for the various food categories, with a standard freight
transport distance of 1,500 miles by semi-trailer truck from the farm or processing
facility to a typical retail location. Packaging, transport from retail stores to con-
sumers’ homes, and home refrigeration were not included in this analysis.
The results showed that freight transportation accounts for just 6% of the total
carbon emissions (in millions of metric tons of CO
2 e) for the food categories ana-
lyzed, indicated by the short bar in Fig. 9.6 . Considering both animal-based and
225
9 Food transportation issues and reducing carbon footprint
plant-based foods, production and processing dominate with 81.6% of the emissions,
followed by cooking, with 8.3%. Only waste disposal produces fewer emissions than
transport, primarily because of the fact that more than 44% of the methane emissions
from landfi lls are typically fl ared or converted to useful energy.
If we examine only the food products derived from animals (Fig. 9.7 )—includ-
ing meat, seafood, dairy and eggs—transportation contributes even less: just slightly
more than 3% of total carbon emissions. For animal products, the high production-
related carbon emissions dwarf emissions from all other factors.
Fig. 9.6 Life cycle carbon emissions (millions of metric tones of CO
2 e) for major food categories
in the United States
Fig. 9.7 Life cycle carbon emissions (millions of metric tons of CO
2 e) for animal-based foods
226 W. Wakeland et al.
On the other hand, if we consider only the plant-based products (Fig. 9.8 )—includ-
ing grains, nuts, fruits and vegetables—transportation contributes more than 16% of
the life cycle emissions, because of the relatively low emissions from production.
The preceding results indicate that the degree to which carbon emissions can be
reduced by optimizing the distribution network depends on whether the food items
are plant-based or animal-based.
If we replace our assumption of long-distance road transport with other distances
and transport modes, the results will change signifi cantly. For example, if air trans-
port is used to deliver fresh imported foods from distant production locations, trans-
portation will be a major contributor to the total life cycle emissions, regardless of
production emissions. Ocean transport, on the other hand, generally produces low
transport emissions per unit of freight. For foods that are imported via ocean, the
road transport to and from the sending and receiving ports often generates emissions
comparable to the [much longer] ocean segment, and therefore the total transport
carbon emissions for ocean-related segments are likely to end up being comparable
to the emissions associated with domestic truck transport for long distances.
9.6 Interactions and trade-offs
A classic challenge for supply chain managers is to strike the right balance between
transportation and storage costs. Because costs and carbon emissions are correlated,
Venkat and Wakeland ( 2006 ) examined the transportation and storage-related emis-
sion characteristics of a food supply chain that was adapted from Simons and Mason
( 2002 ) . Figure 9.9 shows the carbon emissions per unit of fi nal product. Venkat and
Wakeland ( 2006 ) also determined the sensitivity of the total transportation and storage
Fig. 9.8 Life cycle carbon emissions (millions of metric tones of CO
2 e) for plant-based foods
227
9 Food transportation issues and reducing carbon footprint
emissions to the maximum distance between points in the supply chain. The total
emissions would change by about 20% per 1,000 km (excluding any cold storage that
might be required during transportation). Next, the impact of using a lean supply
chain philosophy that emphasized small and frequent deliveries was analyzed. Results
showed that for distances greater than 200 km, the increase in transportation-related
carbon emissions would be greater than the reduction in storage-related carbon
emissions.
Packaging design, production methods, and distribution are inherently interde-
pendent. For example, the wine distribution scenarios explored in Sect. 9.4 consider
wines bottled at the winery in standard 750 mL bottles. A case of 750 mL bottles is,
by weight and volume, only about 50% product. As some wine producers shift to
packaging in polyethylene terephthalate (PET) bottles, Tetra Pak, or Bag-in-Box
formats, they realize energy and emissions savings by avoiding the transport of
excess packaging. Recent carbon footprint analyses done for a packaging manufac-
turer show that from a carbon emissions standpoint, plastic is often a better material
Fig. 9.9 Food supply chain-related carbon emissions (From Venkat and Wakeland 2006 , adapted
from Simons and Mason
2002 )
228 W. Wakeland et al.
for making bottles than glass. For example, considering their full life cycle, a
360 mL PET bottle generates 41% fewer greenhouse gas emissions than a compa-
rable glass bottle (Constar
2010 ) . Full lifecycle includes the carbon emissions asso-
ciated with transporting fi lled bottles to retail locations. Other winemakers have
started to experiment with light-weighting (decreasing the amount of glass used in
a bottle), although this can necessitate more careful handling in transport and stor-
age. Furthermore, some wine companies have started to ship wine in bulk from
Australia and bottle it in the United Kingdom, closer to the consumer market. Bulk
shipping requires a fi rm to geographically distribute its operations and to be able to
address preservation and contamination concerns. Of course, all alternative packag-
ing formats are dependent on consumer acceptance.
9.7 Taking action
We consider what companies can do to cost-effectively reduce logistics-related car-
bon emissions without compromising quality and service levels. Sometimes a single
project can provide a huge improvement opportunity. Jackson Family Wines recently
consolidated its ten warehouses into a single energy-effi cient distribution center
(DC) with a rail spur connecting it to the Union Pacifi c Railroad. According to
Bradley ( 2010 ) , the new facility allowed the wine producer to stop shuttling inven-
tory between storage locations and enabled increased usage of rail to transport its
fi ve million cases produced annually. Locating the DC near a key supplier sup-
ported the company’s backhaul initiative as well: after transporting wine from the
production facility to the DC, vehicles were able to pick up bottles from their nearby
supplier on the return journey.
In other situations, a series of separate initiatives may be more appropriate than
one large project. According to their website, by 2012 the UK-based retailer Tesco
intends to reduce by half the emissions associated with delivering a case of goods.
They plan to reach this goal through a variety of logistics improvements for moving
goods, including switching to larger vehicles, partnering for backhaul opportunities,
relying on rail for transport between DCs, and using barges to a greater extent.
A unifying factor in these initiatives is that they start with an initial study to
benchmark the current system performance and to discover potential opportunities.
Sometimes the results can be quite surprising, as seen in the following study. Recent
research sponsored by the Oregon Transportation Research and Education
Consortium (Pullman et al. 2009 ) compared a variety of packaging, food waste, and
transportation scenarios for three food items: fresh or frozen chicken, raw potatoes,
and processed diced tomatoes. These items were selected based on surveys and
interviews of institutional food purchasers. Packaging alternatives consisted of
waxed cardboard box versus plastic bag for chicken, cardboard versus reusable
plastic container (RPC) for potatoes, and can versus plastic bag (in box) for toma-
toes. Food waste scenarios considered uncooked and cooked waste that was either
229
9 Food transportation issues and reducing carbon footprint
partially composted or 100% landfi lled. Transportation alternatives included local
versus long distance and fresh versus frozen.
For each scenario, full life cycle carbon assessments were done to determine the
embodied carbon ( the total greenhouse gas emissions generated by the product
life cycle within a system boundary of interest and reported in kg of CO
2 e).
The CarbonScope analytical tool (Venkat 2009 ) was employed to do the analy-
sis. The primary standard used for the product life cycle greenhouse gas emissions
calculations was PAS 2050 (BSI Group 2008 ) . PAS 2050 relies on the ISO 14040
series of standards (International Standards Organization 2009 ) and the 2006 IPCC
Guidelines for National Greenhouse Gas Inventories (Intergovernmental Panel on
Climate Change 2009 ) .
Pullman et al. ( 2009 ) conclude that food waste has a large adverse impact on the
environment (especially if landfi lled, as is often the case with cooked vegetables
and meats); an unexpected fi nding that can be easily acted upon. Even raw vegeta-
bles that are composted rather than consumed increase carbon impact (see Table 9.3 ).
Institutions may want to consider policies to encourage the use of packaging that
reduces food waste (both before and after cooking). Similarly, “buy local” efforts
for the three products evaluated earlier would make sense and would contribute to
carbon reduction. Transportation-related carbon impacts were the most signifi cant
for frozen meat items.
Overall, “food miles” do not matter as much as other considerations when deter-
mining the carbon impact of food production, consumption, and disposal (except
perhaps for fresh food that is air freighted). Minimizing food waste and composting
the unavoidable food waste could have a much larger benefi t than switching from a
distant supplier to a local supplier. Also, when analyzed carefully, one must conclude
that plastic packaging generally has a smaller environmental footprint than steel,
paper, or glass, because of its low volume (thinness) and light weight. Considerations
may, however, be required on other health impacts of plastic packaging (e.g., physi-
ological impacts on animals who consume plastics in landfi lls, etc.).
As we have seen from previous examples, items transported by highly utilized
larger trucks result in lower emissions per unit than when transported in smaller
Table 9.3 Example life cycle analysis of food waste, packaging, and distance
Transportation
Local (within 100 miles) Major distributor
Packaging Packaging
Processed tomatoes #10 can
or equivalent
A B A B
Can (recycled) Bag-in-box Can (recycled) Bag-in-box
Food waste
(disposed
before
or after
cooking)
None 1.56 1.39 1.61 1.44
50% Compost,
before cooking
2.22 1.89 2.38 2.00
50% Landfi ll,
after cooking
4.09 3.76 4.20 3.86
From Pullman et al. (
2009 )
230 W. Wakeland et al.
vehicles. Such operational realities have led some analysts to posit a natural ecology
of scale must exist. Schlich and Fleissner ( 2005 ) provide examples of how the inter-
national sourcing of juices and lamb is less energy intensive on a per-unit basis than
the local (German) equivalent. They calculate that energy and cost savings arise
from large-scale production and distribution, as well as from the comparative natu-
ral advantages some countries have in growing or raising food. For example, Brazil’s
climate is naturally more conducive to growing fruit than the climate in most of
Europe, and Brazilian juice production typically occurs on a much large scale than
European producers are able to support. Consequently, Schlich and Fleissner (
2005 )
found that the lower emissions owing to producing juice in Brazil rather than Europe
more than offset the transportation-related emissions associated with transporting
the juice from Brazil to Europe. Such analyses tend to be controversial, in part
because the answer is not what many people expect. Also, these results do not mean
that local production can never compete with long distance sourcing regarding
energy use and carbon emissions. The key is for local operations to fi nd ways to be
effi cient and to achieve the economies that tend to be associated with large scale
operations. Local production also provides other benefi ts that are not as easily incor-
porated into analyses, such as local employment, rural conservation, and consumer
trust in origin and local brands.
It is not necessarily the case that only giant conglomerates can be highly effi -
cient. It is possible for smaller fi rms to mimic the effi ciency of larger entities. For
example, farmers in European wine regions have often formed cooperatives for the
production and marketing of their products. Indeed, Schlich ( 2010 ) documents the
Hessische Bergstraße cooperative within the Rhine Valley, which is comprised of
more than 500 family-owned vineyards. Despite the small size of these individual
winemakers, the Hessische Bergstraße cooperative coordinates production and dis-
tribution in a way that allows these farmers to capture effi ciencies of scale normally
associated with a much larger producer.
When aggregation occurs even further downstream in the supply chain, producer
collectivism is not required. The following example explores how the logistics
needed to support California farmers’ markets could be improved, while preserving
the small scale and independent nature of both the farms and the markets. Such
improvements will not only yield fi nancial benefi ts, but will also improve the energy
intensity associated with bringing these products to consumers.
The proliferation of farmers’ markets in the last few years has been a boon to
consumers looking to purchase fresh, locally grown fruits and vegetables. Worthen
( 2010 ) documents that California now has more than 500 markets, which are
attended by nearly 3,000 farmers. However, this growth in the number of markets
has negatively impacted many farmers, who report having to make multiple trips to
visit more markets in order to sell the same amount of produce. Jog ( 2010 ) profi les
Schletewitz Farms, one such fruit producer located near Fresno, California, that
sells year round to nine greater Bay Area markets, each 200–350 km away. Such
distances are typical, as many of the certifi ed market suppliers are located in the
Central Valley, whereas most markets are found in population centers closer to the
coast, as seen in Fig. 9.10 , which depicts a representative subset of farmers’ markets
and the farms that service them. Likewise, a map of vendors found at the Jack
231
9 Food transportation issues and reducing carbon footprint
London Square Farmers Market in Oakland shows that farmers come from as far
north as Yuba City and as far south as San Diego, almost 800 km away (Jog 2010 ) .
On a per unit basis, supply chains relying on extensive use of small capacity vehi-
cles are likely to be less energy effi cient than their larger-scale counterparts. To illus-
trate this, we consider Sanger Farm, a regular vendor at the San Francisco Ferry
Market. We assume that they nearly fi ll (90%) a large pickup truck with produce and
that they are able to completely sell the produce at the market (a best-case scenario).
Using CargoScope, we calculate that the round-trip journey to transport the produce
from Sanger Farm to San Francisco results in .45 kg CO
2 e emitted per kg of produce.
We next consider a neighboring farm that may sell high volumes of produce
through a supermarket such as Safeway. In this supply chain scenario, a midsized
commercial truck (which has about ten times the capacity of a pickup truck) would
Fig. 9.10 Subset of farmers’ markets, farms and candidate DCs (From Jog 2010 ) . The Northern
DC belongs to Safeway. The other DC represents a plausible consolidation point for Central Valley
Farmers
232 W. Wakeland et al.
transport produce from the farm to the Safeway DC in Tracy, where it could be
placed in cold storage for a period of time. The produce would then be transported
by a similar truck to a Safeway store in San Francisco, where it could take up to a
week to be sold. Considering typical supermarket volumes and dwell times, we
assume 90% utilization rates for both trucking links, and determine that .19 kg of
CO
2 e is emitted for the transport and storage of 1 kg of produce from farm to retail
gate. The farmers’ market produce may be fresher and have other ecological bene-
fi ts, but its carbon footprint is more than twice that of the comparable supermarket
produce. This difference would be even greater if the dwell time at the DC were
reduced for the supermarket-bound produce, because refrigerated storage contrib-
utes signifi cantly to emissions. Figure 9.11 compares the transport and storage-
related carbon emissions for the two preceding distribution scenarios and a third
scenario discussed next.
We fi nally consider modifying the farmers’ market supply chain to include a
consolidation DC that is near an interstate highway, but still relatively close to many
Central Valley farms (such as in the vicinity of the small town of Cantua Creek).
Rather than individually driving their produce across the state to farmers’ markets
in the Bay Area, participating farmers would deliver produce to the DC, where it
would be stored for a period of a few days and then consolidated into midsized com-
mercial trucks, each destined for a particular farmers’ market. Market stalls could
be staffed by employees of the DC or individuals hired at the local markets Assuming
the same high utilization rates would hold, Fig. 9.11 shows that this newly designed
supply chain would be much less carbon emissions intensive, approaching the rela-
tively low levels of emissions provided by the supermarket supply chain.
Jog ( 2010 ) explores the potential for a farmers’ market consolidation DC at a
regional level. Although it would be impractical to gather historical sales data for
every farmer at every market, Jog creates a simplifi ed approximation by modeling
the underlying network, which includes 135 Bay Area markets and more than 1,000
farms that sell produce at these markets. Jog’s analysis treats the farms within a
given zip code as a single larger farm that is equivalent in volume. The result is
slightly more than 200 equivalent farms. Using this approximation tends to slightly
Fig. 9.11 Emissions from farm to retail for different sales channels (calculations performed using
CargoScope)
233
9 Food transportation issues and reducing carbon footprint
underestimate the total travel distances. Considering one type of undifferentiated
produce, assuming uniform production rates for each of the equivalent farms, and
classifying each market into one of three sizes, Jog models this network as a trans-
portation problem. The optimization problem was to fi nd the minimum total travel
distance that could meet the market demand.
Without a consolidation DC, nearly 400 trucks would be needed to supply the
markets over the course of a week, with an average round trip of slightly more than
200 km and a total travel distance of 84,000 km. It should be noted that this optimal
solution greatly understates actual transit, as farmers typically travel more than
300 km per trip and visit several markets per week. In this minimized solution,
farmers would visit only two or three markets per week, and markets would have at
most a few vendors. Of course, such a lack of diversity would be counter to the mis-
sion of a farmers’ market.
Next, Jog adds the Cantua Creek consolidation DC to the model and fi nds that
the optimum in this case occurs when nearly 60% of the produce is routed through
the DC, which reduces the overall distance traveled by 20%. Furthermore, market
diversity would be supported, because trucks arriving to the market from the con-
solidation DC would carry produce from multiple farmers, a benefi t for small-scale
farmers who produce specialized products. Such farmers would be able to reach
more consumers than is currently possible, because it would be economically fea-
sible to ship fractional truckloads of their goods to different markets by consolidat-
ing them with goods from other small-scale farmers.
Of course, the reality of funding and implementing such a consolidation DC, sup-
porting transport from the DC to the markets, and arranging for staffi ng would be
much more complex than building models and calculating potential benefi ts. Perhaps
the DC could be created by the farmers as a cooperative, with collaborative staffi ng
and support, as well as partnerships with organizations such as the California
Federation of Certifi ed Farmers’ Markets. Alternatively, the DC could be created as a
joint venture funded by a major reseller. Another option might be to allow produce to
be sold at the DC or the market on a consignment basis. At present, California farm-
ers’ markets allow only direct producers. The concepts described above would tend to
reduce consumer contact with growers, which has been one of the primary benefi ts of
farmers’ markets. However, the current trend toward more and more farmers’ markets,
without corresponding improvements to the underlying support structures, will not be
sustainable in the long run for most small family farms and could ultimately deter
participation of just the sort of vendors these markets were created to showcase.
9.8 Conclusion
The transportation-related carbon footprint varies from a few percent to more than
half of the total carbon footprint associated with food production, distribution, and
storage. Supply chains are complex and varied, and food supply chains are especially
challenging because of seasonality, freshness, spoilage, and sanitary considerations.
Measuring transportation-related carbon footprint involves careful choice of the
234 W. Wakeland et al.
scope of the analysis, and there is much uncertainty in the results. Caution is warranted
regarding the absolute numbers from carbon assessments, so it may be best to focus
primarily on relative comparisons.
The winemaker case study showed that a local 3PL approach had the lowest car-
bon footprint, and that the highest carbon footprint resulted from consumers driving
to the winery. This does not mean, however, that local 3PL would be the best solution
for all wineries. The case study involving organic fruit and vegetable supply showed
how food carbon data can be provided to consumers in order to support their food
purchasing choices, and the case study regarding packaged fruit indicated that trans-
portation distances are not always a good indicator of total carbon footprint. Another
case study indicated that home delivery can cut the transportation-related carbon
footprint almost in half compared to consumers driving to the store individually.
Supply chain planners must carefully consider the trade-off between transporta-
tion-related energy cost & carbon footprint and storage-related energy cost & carbon
footprint. Also, the frequent small deliveries called for by lean manufacturing prac-
tices, although optimizing effi ciency within a facility, can increase overall carbon
footprint. Packaging is another important consideration, and the use of plastics
rather than glass tends to lower carbon footprint. Benefi ts for the environment and
health are further accrued by plastic recycling.
To reduce carbon footprint, suppliers are consolidating their operations, increas-
ing their use of rail and water transit, and increasing transport effi ciency by fi lling
trucks and considering backhaul opportunities. Food waste is another potentially
signifi cant contributor to carbon emissions, which could potentially be reduced via
alternative packaging options. Our research also indicates that food-miles, a metric
that many consider to be of primary importance, do not actually correlate very well
with overall carbon footprint. Finally, although farmers’ markets have many desir-
able attributes, they unfortunately tend to have a much higher carbon footprint than
conventional food distribution. This differential could be lessened considerably by
using consolidation DCs close to the farms.
Whether comprising a large or small share of a product’s total emissions, trans-
portation is an unavoidable step in the supply chain for nearly every food product.
However, the economic and environmental impacts of food transportation can be
moderated. We reiterate that although food transportation decisions can sometimes
be considered separately from other issues, this is not always appropriate. Food
transportation and storage involves trade-offs that necessitate taking an overall sys-
tem perspective.
References
Bradley, P. 2010. An inside look at Jackson Family Wines’ new eco-friendly warehouse. http://
www.dcvelocity.com/articles/20100222inside_an_eco_friendly_dc/ . Accessed 15 May 2010.
Browne, M., C. Rizet, S. Anderson, J. Allen, and B. Keita. 2005. Life cycle assessment in the sup-
ply chain: A review and case study. Transport Reviews 25(6): 761–782.
235
9 Food transportation issues and reducing carbon footprint
BSI Group. 2008. PAS 2050:2008—Specifi cation for the assessment of the lifecycle greenhouse
gas emissions of goods and services.
http://www.bsigroup.com/en/Standards-and-Publications/
Industry-Sectors/Energy/PAS-2050/ . Accessed 18 Mar 2009.
Carlsson-Kanyama, A. 1998. Climate change and dietary choices—How can emissions of green-
house gases from food consumption be reduced? Food Policy 23(3–4): 277–293.
Carlsson-Kanyama, A., M.P. Ekström, and H. Shanahan. 2003. Food and life cycle energy inputs:
Consequences of diet and ways to increase effi ciency. Ecological Economics 44(2–3):
293–307.
Cholette, S., and K. Venkat. 2009. The energy and carbon intensity of wine distribution: A study
of logistical options for delivering wine to consumers. Journal of Cleaner Production 17(16):
1401–1413.
CleanMetrics. 2010. http://www.cleanmetrics.com . Accessed 18 Sep 2010.
Constar. 2010. Sustainability analysis for PET food jars.
http://www.constar.net/pdf/Convert-It_
Analysis_Example.pdf . Accessed 24 June 2010.
DOE. 2009. U.S. CO 2 emissions from energy sources 2008 fl ash estimate. http://www.eia.doe.
gov/ . Accessed 20 Jan 2010.
Golicic, S., C. Boerstler, and L. Ellram. 2010. Greening transportation in the supply chain. MIT
Sloan Management Review 51(2): 47–55.
Heller, M., and G. Keoleian. 2000. Life cycle-based sustainability indicators for assessment of the
U.S. food system [CSS00-04] . Ann Arbor: Center for Sustainable Systems, School of Natural
Resources and Environment.
Hospido, A., L. Milà i Canals, S. McLauren, et al. 2009. The role of seasonality in lettuce con-
sumption: A case study of environmental and social aspects. International Journal of Life
Cycle Assessment 14: 381–391.
Intergovernmental Panel on Climate Change. 2009. 2006 IPCC guidelines for national greenhouse
gas inventories.
http://www.ipcc-nggip.iges.or.jp/public/2006gl/index.html . Accessed 18 Mar
2009.
International Standards Organization. 2009. ISO 14040: 2006 Environmental management–life
cycle assessment–Principles and framework (2).
http://www.iso.org/iso/iso_catalogue/cata-
logue_ tc/catalogue_detail.htm?csnumber = 37456 . Accessed 18 Mar 2009.
Jog, S. 2010 Optimizing the logistics of farmer’s markets: A carbon footprint study. Master’s the-
sis, San Francisco State University, California.
Johnson, E. 2008. Disagreement over carbon footprints: A comparison of electric and LPG fork-
lifts. Energy Policy 36(4): 1569–1573.
Milà i Canals, L., S.J. Cowell, S. Sim, and L. Basson. 2007. Comparing domestic versus imported
apples: A focus on energy use. Environmental Science and Pollution Research 14(5):
338–344.
Pirog, R., T. Van Pelt, K. Ensayan, and E. Cook. 2001. Food, fuel and freeways: An Iowa perspec-
tive on how far food travels, fuel usage, and greenhouse gas emissions . Ames: Leopold Center
for Sustainable Agriculture.
Point, E. 2008. Life cycle environmental impacts of wine production and consumption in Nova
Scotia, Canada. Master’s thesis, Dalhousie University, Halifax.
Pullman, M., R. Fenske, and W. Wakeland. 2009. Food delivery footprint: Addressing transporta-
tion, packaging, and waste in the food supply chain. OTREC Research Report, RR-08-154,
Oregon Transportation Research and Education Consortium, March 2009.
Rosenthal, E. 2009. To cut global warming, Swedes study their plates. The New York Times , 23 Oct
2009.
Saunders, C., and A. Barber. 2007. Comparative energy and greenhouse gas emissions of New
Zealand’s and the United Kingdom’s dairy industry . Christchurch: Lincoln University. AERU
Research Report No. 297.
Saunders, C., A. Barber, and G. Taylor. 2006. Food miles—Comparative energy/emissions perfor-
mance of New Zealand’s agriculture industry . Christchurch: Lincoln University. AERU
Research Report No. 285.
236 W. Wakeland et al.
Schlich, E. 2010. From vineyard to point of sale: Allocation of energy use and CO
2 emission to
entire supply chains of wine. In 4th annual meeting of American association of wine econo-
mists (AAWE) . Davis, CA, June 2010.
Schlich, E., and U. Fleissner. 2005. The ecology of scale: Assessment of regional energy turnover
and comparison with global food. International Journal of Life Cycle Assessment 10(3):
219–223.
Scholz, A., U. Sonesson, and P. Tyedmers. 2009. Catch of the freezer. The New York Times , 8 Dec
2009.
Sim, S., M. Barry, R. Clift, and S.J. Cowell. 2007. The relative importance of transport in deter-
mining an appropriate sustainability strategy for food sourcing. International Journal of Life
Cycle Assessment 12(6): 422–431.
Simons, D., and R. Mason. 2002. Environmental and transport supply chain evaluation with sus-
tainable value stream mapping. In Proceedings of the 7th logistics research network confer-
ence , Birmingham, UK.
Twede, D., R. Clarke, and J. Tait. 2000. Packaging postponement: A global packaging strategy.
Packaging Technology and Science 13(1): 105–115.
United States. Department of Agriculture. 2008. Loss-adjusted food availability: Spreadsheets.
http://www.ers.usda.gov/Data/foodconsumption/FoodGuideSpreadsheets.htm . Accessed 10
Sep 2010.
Van Hauwermeiren, A., H. Coene, G. Engelen, and E. Mathijs. 2007. Energy lifecycle inputs in
food systems: A comparison of local versus mainstream cases. Journal of Environmental Policy
Plan 9(1): 31–51.
Venkat, K. 2008. CargoScope web software tool.
http://www.cleanmetrics.com/html/cargoscope.
htm . Accessed 18 Mar 2009.
Venkat, K. 2009. CarbonScope web software tool.
http://www.cleanmetrics.com/html/carbon-
scope.htm . Accessed 18 Mar 2009.
Venkat, K., and W. Wakeland. 2006. Is lean necessarily green? In 50th anniversary conference of
the international society for the systems sciences , Sonoma, CA, July 2006.
Weber, C.L., and H.S. Matthews. 2008. Food-miles and the relative climate impacts of food choices
in the United States. Environmental Science and Technology 42(10): 3508–3513.
World Resources Institute. 2004. Greenhouse gas protocol, a corporate accounting and reporting
standard.
http://www.ghgprotocol.org/fi les/ghgp/public/ghg-protocol-revised.pdf . Accessed
20 June 2009.
Worthen, B. 2010. Farmers’ markets see risk from growth. Wall Street Journal , 11 Feb 2010.
http://
online.wsj.com/article/SB10001424052748704533204575048000839268376.html . Accessed
10 Mar 2010.
