The energy and carbon intensity of wine distribution A study of

Document Sample
The energy and carbon intensity of wine distribution A study of Powered By Docstoc
					Our reference: JCLP 1977                                                                                        P-authorquery v6

                                                AUTHOR QUERY FORM

                   Journal: JCLP                  Please e-mail or fax your responses and any corrections to:


                   Article Number: 1977           Fax: +31 2048 52789

Dear author,

During the preparation of your manuscript for typesetting, some questions may have arisen. These are listed below and
highlighted by flags in the margin of the proof. Please check your typeset proof carefully and mark any corrections in the
margin of the proof or compile them as a separate list. In case artwork needs revision, please consult

Electronic file usage
Sometimes we are unable to process the electronic file of your article and/or artwork. If this is the case, we have proceeded by:

,      Scanning (parts of) your article         ,        Rekeying (parts of) your article          ,       Scanning the artwork

Queries and/or remarks

,      Articles in Special Issues: Please ensure that the words ’this issue’ are added (in the list and text) to any references to
       other articles in this Special Issue.

 Uncited references: References that occur in the reference list but not in the text - please position each reference in the text
 or delete it from the list
 Missing references: References listed below were noted in the text but are missing from the reference list - please make the
 list complete or remove the references from the text.
 in article Query / remark                                                 Response
 Q1            Please check whether the edit made to the sentence
               ‘‘Typical research can be seen.’’ is ok.
 Q2            Please     check   ‘‘produce’’   in    the    sentence
               ‘‘CleanMetrics, the company which.’’.
Thank you for your assistance.
                                                                   ARTICLE IN PRESS                                             JCLP1977_proof Š 16 June 2009 Š 1/13

                                                                      Journal of Cleaner Production xxx (2009) 1–13

                                                                    Contents lists available at ScienceDirect

                                                             Journal of Cleaner Production
                                                    journal homepage:

 1                                                                                                                                                                         56
     The energy and carbon intensity of wine distribution: A study of logistical                                                                                           58
 4   options for delivering wine to consumers                                                                                                                              59

 5                                                                                                                                                                         60
                                a, *                          b                                                                                                            61
 6   Susan Cholette                , Kumar Venkat
 7                                                                                                                                                                         62
 8     Assistant Professor of Decision Sciences, San Francisco State University, College of Business, 1600 Holloway Avenue, San Francisco, CA 94132,United States          63
       President and Principal Engineer, CleanMetrics Corp, 4888 NW Bethany Blvd. Suite K5, #191 Portland, Oregon 97229, United States
 9                                                                                                                                                                         64

10                                                                                                                                                                         65
11                                                                                                                                                                         66
12   a r t i c l e i n f o                                   a b s t r a c t                                                                                               67
13                                                                                                                                                                         68
     Article history:                                        Logistics within the food and beverage sector are often energy-intensive, especially for the wine industry.
14                                                                                                                                                                         69
     Received 15 January 2009                                We consider how California wines may be routed to U.S. consumers near and far, basing scenarios and
15   Received in revised form                                                                                                                                              70
16   27 May 2009
                                                             supporting data on interviews and literature review. We use a web-based tool, CargoScope, to calculate
                                                             the energy and carbon emissions associated with each transportation link and storage echelon. We find
17   Accepted 31 May 2009                                                                                                                                                  72
     Available online xxx
                                                             that supply chain configurations can result in vastly different energy and emissions’ profiles, varying by
18                                                           up to a factor of 80, and discuss how these results could be incorporated into a winery’s overall             73
19                                                           sustainability strategy.                                                                                      74
20                                                                                                                                  Ó 2009 Published by Elsevier Ltd.      75
     Carbon emissions
21   Logistics                                                                                                                                                             76

22   Supply chain management                                                                                                                                               77
23   Wine industry                                                                                                                                                         78
24                                                                                                                                                                         79
25                                                                                                                                                                         80
26                                                                                                                                                                         81

27   1. Introduction                                                                             been published on the carbon intensity of basic outbound logistics.       82
28                                                                                                                ¨
                                                                                                 Seuring and Muller’s recent extensive survey [22] of peer-reviewed        83
29       In the past few years mainstream corporate interest in envi-                            articles shows few directly consider energy and emissions impact of       84
30   ronmental sustainability has blossomed, especially with regards to                          supply chains. The lack of guidance from the research community           85
31   reducing energy usage and carbon emissions. A key component of                              creates a relative vacuum that may inadvertently aid the promul-          86

32   such understanding is the ability to create a model to analyze the                          gation of potentially simplistic and misleading metrics. For              87
33   problem, quantify metrics for success and evaluate alternatives                             instance, some retailers are considering labeling products with           88
34   based on their effectiveness. We present an analysis of the carbon                          ‘‘food-miles,’’ defined as the distance that a product has traveled        89
35   and energy profiles of wine distribution, using a U.S. case study of                         from manufacture to point of sale. Even Tsoulfas and Pappis [23] in       90
36   logistical options for delivering wine to consumers, supported by                           their well-delineated decision model, frame their first principle for      91

37   a model developed in CargoScope. We show that different supply                              transportation as ‘‘minimizing distance covered.’’                        92
38   chain configurations vary dramatically in overall energy and                                     Yet different transport modes vary greatly in energy and emis-        93
39   emissions impact, and provide recommendations that wineries can                             sions’ profiles, and higher transportation emissions may offset            94
40   consider for improvement.                                                                   emissions produced elsewhere in the supply chain. For instance,           95
41       Despite recent media awareness to what is popularly known as                            Saunders and Barber [21] show that lamb raised in New Zealand             96
42   ‘‘carbon foot printing,’’ measuring the carbon intensity of the                             and shipped to the UK on ocean-going vessels is more carbon               97

43   supply chain has received comparatively scant research attention.                           efficient than lamb from British feed lots. Lebel and Lorek [14] point     98
44   Kleindorfer et al.’s comprehensive review [13] of the extant liter-                         to examples where localization may reduce emissions but result in          99
45   ature on sustainability in a respected operations management                                greater negative ecological or social effects. Even just considering      100
46   journal focuses on three topics: production and process develop-                            energy and emissions, other factors within a supply chain may             101
47   ment, waste minimization through lean operations, and re-                                   dominate pure distances. Delivery lot sizes have a profound effect        102
48   manufacturing through closed loop supply chains. While reverse                              on carbon emissions in the food and beverage sector; Venkat and           103
49   logistics has generated much recent excitement, fewer articles have                         Wakeland [27] show that the extra energy needed for transporting          104
50                                                                                               more partial loads may be less than that associated with stockpiling      105
51                                                                                               products in cold storage for greater durations, making lean opera-        106
52       * Corresponding author.                                                                 tions less attractive. Van Hauwermeiren et al. [25] demonstrate that      107
53         E-mail addresses:, (S. Cholette).          the organically grown food is not necessarily more carbon efficient        108
54                                                                                                                                                                         109
     0959-6526/$ – see front matter Ó 2009 Published by Elsevier Ltd.

         Please cite this article in press as: Cholette S, Venkat K, The energy and carbon intensity of wine distribution: A study of logistical options for..., J
         Clean Prod (2009), doi:10.1016/j.jclepro.2009.05.011
                                                                     ARTICLE IN PRESS                                           JCLP1977_proof Š 16 June 2009 Š 2/13

           2                                                  S. Cholette, K. Venkat / Journal of Cleaner Production xxx (2009) 1–13

110        than its conventional counterpart, as economies of scale may make                        2. Distributing wine in the United States                                   175
111        the latter less intensive to transport and dominate the net carbon                                                                                                   176
112        impact.                                                                                  2.1. An overview of the U.S. wine market                                    177
113            With supply chains that span long distances, the transportation                                                                                                  178
114        and storage of food products can be very energy-intensive. Trans-                            The logistics network for the U.S. wine market is complex, with         179
115        portation, namely diesel fuels from trucking, is estimated by Heller                     many echelons and options, as seen in Fig. 1. This complexity exists        180
116        and Keoleian [10] to account for 25% of the total energy consumed                        for historical and regulatory reasons. At the repeal of prohibition,        181
117        within the U.S. food system. We consider the wine industry, one of                       the 3-tier system was designed to prevent over-consumption by               182
118        the pioneering consumer goods sectors in respect to addressing                           requiring alcohol producers to sell to retailers via distributors, all of   183
119        environmental issues. Much of this sector’s efforts concern                              which must be separately owned entities. Cholette [4] emphasizes            184
120        sustainable growing practices or improving the process of wine-                          that although distributors in other industries can coordinate funds         185
121   Q1   making. Typical research can be seen in Marchettini et al.’s [15]                        and information while the actual products may be shipped directly           186
122        quantification of energy inputs, erosion factors, pesticides and                          from the manufacturer to the retailer, alcohol distributors are             187

123        fertilizers and Ruggieri et al.’s LCA study [20] investigates reducing                   legally bound to take physical possession of the stock. Additionally,       188
124        and reusing winemaking wastes.                                                           supermarkets and other chain stores with several outlets in                 189

125            We target another area, namely the logistical processes that                         a geographical area may consolidate merchandise at regional                 190
126        occur after wine has been packaged for consumer sale. While it is                        distribution centers before delivery to the store. Cholette [5]             191
127        often myopic to consider just a single area (logistics) and just                         reports that nearly half of wine in the U.S. is sold through such           192
128        a single impact (CO2 emissions), we feel this is justified for the                        retailers, effectively adding an additional echelon to the supply           193
129        following reasons. Most wineries have a fractional share of the                          chain.                                                                      194
130        overall consumer market, so a unilateral attempt by a winery to                              Although most U.S. produced wine is shipped to domestic                 195

131        redefine package formats or make other significant changes                                 consumers via the 3-tier system, Fig. 1 shows alternative routings          196
132        requiring acceptance by supply chain partners and, ultimately, the                       exist. Wineries can self-distribute in California, although this option     197
133        end consumer, would be difficult. Decisions made for supporting                           is typically not practical for smaller wineries. Wineries can apply to      198
134        this part of the product cycle are separable from the sourcing and                       sell wine directly to consumers in many states. The traditional             199
135        winemaking processes and also any post-consumer recycling/                               direct sales channel is for consumers to visit a tasting room at            200
136        recovery efforts. Energy usage associated with post-production                           a winery. Purchases can either be carted away by the consumer or            201
137        logistics is high for wine as the standard consumer packaging is                         shipped to the consumer’s home, via a small package carrier.                202
138        fragile, heavy and bulky. Wine itself comprises just half the weight                     Wineries may also support direct sales through a mailing list or            203
139        and under 40% of the volume of a case of twelve 750 ml glass                             a website, where customers select from the wines advertised and             204
140        bottles. Wine is also sensitive to temperature and must be stored in                     place orders from their home. Additionally, many wineries offer             205
141        a controlled climate for all but the shortest periods. In short,                         wine clubs, where members periodically receive deliveries of small          206

142        changes to a winery’s outbound supply chain can have a high                              allotments of pre-selected wines. Smaller wineries often utilize 3PL        207
143        impact and be implemented quickly without requiring major                                (third party logistics) providers to support these direct-to-               208
144        retooling of producers or extensive re-education of consumers.                           consumer sales’ programs.                                                   209
145        Over the longer term a winery may be able to reconsider all aspects                          In many states wine sold directly to consumers can either be            210
146        of production, marketing and logistics.                                                  picked up by the consumer or shipped to the consumer’s home.                211

147            Of the research reviewed, only 2 works consider the outbound                         However, direct-to-consumer delivery is illegal in some states. In          212
148                                                      ¨
           supply chain for wineries. Colman and Paster’s lifecycle study of                        such locales, wineries may be able to route customer orders                 213

149        wine [6] shows that outbound logistics may contribute to over                            through a certified wholesaler who in turn sends the wine to                 214
150        half of the total carbon emissions for many regions’ wines. Point                        a retailer close to the consumer. Although this is not an issue for         215
151        [19] performs a life cycle assessment for Nova Scotia wines and                          direct shipping within California, we consider this logistical option       216
152        assumes localized consumption, as Nova Scotia wines are not                                                                                                          217
153        widely distributed in other provinces or export markets. Point                                                                                                       218
154        [19] shows that post-production logistics, even given the short                                                                                                      219

155        distances of her study, are the second highest contributor to CO2                                                                                                    220
156        emissions, after the emissions associated with producing and                                                                                                         221
157        transporting bottles. Both of these works assign a single outbound                                                                                                   222
158        logistics routing to a winery. Our research attempts to help fill                                                                                                     223
159        this gap by examining the carbon intensity of several different                                                                                                      224

160        options that a winery may have for delivering products to                                                                                                            225
161        consumers.                                                                                                                                                           226
162            The remainder of this paper is organized as follows. We                                                                                                          227
163        provide an overview of the U.S. wine distribution system, dis-                                                                                                       228
164        cussing the available options to reach U.S. consumers. We                                                                                                            229
165        construct a representative network to model delivery of specialty                                                                                                    230
166        wines to end consumers both nearby and cross-country. We                                                                                                             231
167        introduce the software used to estimate the energy usage and                                                                                                         232
168        carbon emissions associated with these delivery scenarios. We                                                                                                        233
169        compare scenario results and show how different supply chain                                                                                                         234
170        configurations can impact emissions. We suggest how these                                                                                                             235
171        findings could be of use within a winery’s emissions reduction                                                                                                        236
172        program, as a component of an overall corporate social respon-                                                                                                       237
173        sibility (CSR) strategy. Lastly, we suggest directions for future                                                                                                    238
174        research.                                                                                                       Fig. 1. The supply chain for U.S. wineries.          239

               Please cite this article in press as: Cholette S, Venkat K, The energy and carbon intensity of wine distribution: A study of logistical options for..., J
               Clean Prod (2009), doi:10.1016/j.jclepro.2009.05.011
                                                                 ARTICLE IN PRESS                                              JCLP1977_proof Š 16 June 2009 Š 3/13

                                                          S. Cholette, K. Venkat / Journal of Cleaner Production xxx (2009) 1–13                                      3

240   for the following reason. Many wine retailers support customer                            of the regional customer warehouse (RCW) and select Richmond as           305
241   ordering of limited availability wines not normally stocked in store.                     a representational site since Cost Plus World MarketÔ, a retail           306
242   Wine ordered this way follows a similar path as direct-to-consumer                        chain noted for its wine sales, has a large facility here.                307
243   sales that must be routed through a distributor and retailer.                                 For direct shipping to this local market, we consider one of the      308
244                                                                                             major third party logistics (3PL) providers for California wine           309
245                                                                                             shipments, which is New Vine Logistics (NVL). New Vine’s fulfill-          310
      2.2. Mapping the logistical network for domestic distribution
246                                                                                             ment center is located in American Canyon. While New Vine                 311
247                                                                                             partners with several small package carriers, we select Federal           312
          In order to analyze the energy usage of the various supply chain
248                                                                                             Express, which has a sorting center and warehouse (FDX) in South          313
      options, we consider the case of a Sonoma winery that is
249                                                                                             San Francisco. Distances between points are calculated via GoogleÔ        314
      attempting to deliver specialty wine to consumers located in San
250                                                                                             maps to determine appropriate routes. We locate both the retail           315
      Francisco and in Manhattan. We pick these two regions as they are
251                                                                                             store (RS) and the consumer (CU) in San Francisco and assume that         316
      centers for wine consumption, especially of specialty wines, and
252                                                                                             the consumer is located 3.6 km from the store. This distance is           317

      allow for consideration of local and long distance supply chains. In
253                                                                                             below the national average of 10 km, as BAEF research [1] shows           318
      addition to the literature sources provided throughout the paper,
254                                                                                             that consumers in the Bay Area typically have to travel much              319
      the structure and data of our model are based on input from

255                                                                                             shorter distances. While many researchers, such as Hutchins and           320
      professionals representing every echelon, save for the distributor/
256                                                                                             Sutherland [11], terminate the supply chain at the retail outlet, we      321
      wholesaler tier, as summarized in Table 1, as well as from a carrier
257                                                                                             include transport to the end consumer for reasons that shall shortly      322
      and a 3PL provider.
258                                                                                             become apparent.                                                          323
          As Table 1 shows, we engaged in discussions with several
259                                                                                                 Servicing the metropolitan New York market requires consid-           324
      wineries. Our representative Sonoma winery, is based most closely
260                                                                                             ering a much larger geographical area and additional transport            325

      upon Cline Cellars, a medium-sized winery with a line of moder-
261                                                                                             modes. We add the following nodes: OAK, as Oakland houses the             326
      ately priced wines (approximately $10/bottle) in retail stores
262                                                                                             Bay Area’s pre-eminent cargo rail terminal, and SFO, as this airport      327
      nationwide as well as several higher-end wines ($25/bottleþ),
263                                                                                             services much of the region’s outbound air cargo. We include two          328
      many of which are primarily available thorough direct-to-
264                                                                                             hubs: rail companies often route East-bound trains through Chi-           329
      consumer channels. Although large firms with low-margin prod-
265                                                                                             cago (CHI), and Memphis (MEM) is the super hub through which              330
      ucts like the Wine Group can use alternate packaging formats such
266                                                                                             much of FedEx’s air cargo travels. Newark has both an airport and         331
      as bag-in-a-box and TetraPakÔ or even ship product in bulk for
267                                                                                             rail terminal (EWR). New Jersey has the sorting/distribution centers      332
      bottling closer to the retail market, these options are not currently
268                                                                                             for FedEx in Edison (SSE) and for Southern/Glazer’s in Monroe             333
      feasible for most smaller wineries or for those with more upscale
269                                                                                             Township (DC–NJ). The retail store in Manhattan is designated as          334
      wines. Wine is predominantly sold in 750 ml glass bottles, and
270                                                                                             RS–NY. The location of all Northern California and Metropolitan           335
      Twede et al. [24] emphasize that packaging beverage products is
271                                                                                             New York nodes are shown side by side to the same scale in Fig. 2.        336
      a high-speed automated process involving expensive equipment,

272                                                                                                                                                                       337
      favoring centralization. We can reasonably assume that most Cal-
273                                                                                                                                                                       338
      ifornia wineries bottle and warehouse products onsite, as Cline                           3. Solution methodology and model scenarios
274                                                                                                                                                                       339
      Cellars indeed does. Dividing the standard 12-bottle case of wine
275                                                                                                                                                                       340
      into 2 separate customer orders of six bottles each represents                                We first introduce the software utilized, presenting both the
276                                                                                                                                                                       341

      a typical order size.                                                                     mechanics and the interface. We then describe the options avail-
277                                                                                                                                                                       342
          We select representative locations for the logistical echelons for                    able to our representative Sonoma winery in fulfilling delivery of
278                                                                                                                                                                       343
      each of the two regional markets and code them with acronyms. For                         a half cases (6 bottles) of wine to a consumer located in San Fran-

279                                                                                                                                                                       344
      instance, Southern/Glazer’s, which distributes over 80% of the wine                       cisco. This construct mimics the business model of a wine club. We
280                                                                                                                                                                       345
      and spirits sold in the U.S. [29], has a large regional facility in Union                 next consider the order fulfillment options available for delivering
281                                                                                                                                                                       346
      City. Union City is thus chosen as the location for our representative                    that half case wine to a Manhattan consumer. Each of the scenarios
282                                                                                                                                                                       347
      distributor’s warehouse (DW). We also consider the optional layer                         depicts a different configuration for transporting wine from the
283                                                                                                                                                                       348
                                                                                                winery’s onsite warehouse to an end consumer.
284                                                                                                                                                                       349

285   Table 1                                                                                                                                                             350
286   Interviews by echelon and transportation partner.                                         3.1. CargoScope: introducing the software                                 351
287   Echelon                              Information provided                                                                                                           352
288   Wineries: Cline Cellars,             Direct shipment frequency and volumes,                   In order to be understood and usable by non-specialists, models       353
289     Hess Collection, LionHeart         rough percentages of sales supported each            must balance simplicity and usability with analytic power. Devel-         354

        Wines, Nicholson Ranch             by delivery options
290                                                                                             oped and maintained by CleanMetrics, CargoScope is a web-based            355
      Retailer: Cost Plus World            Location of stores and RDC, dwell times,
        Market                             storage and replenishment policies. Inbound          tool that allows users to build a supply chain network and define          356
292                                        and outbound transportation modes.                   the storage, transit and processing parameters for every echelon.         357
293   Distributor: None                    Location of warehouses available online at           While many websites support calculators for determining personal          358
294     interviewed, SWS/Glazer’s Parameters and                 ‘‘carbon footprints’’ there are fewer, if any, tools online that allow    359
        selected as representative         policies are assumed to be comparable to
                                           that of the retailer’s RDC
                                                                                                the user to configure a general supply chain. CargoScope was also          360
296   Carrier: FedEx                       Locations of nodes and routes, inbound and           selected for this study because it was free, and trial subscriptions      361
297                                        outbound transport modes utilized, rough             are available on request. CargoScope’s built-in parameters are            362
298                                        estimates of utilization and backhaul rates,         based on data from U.S. governmental [7–9] and international [30]         363
                                           dwell times at sorting center
299                                                                                             agencies, as well as academic studies [18]. Users can create, share       364
      3PL provider: New Vine               Inbound and outbound transportation
        Logistics                          modes, dwell times, estimates of backhaul            and revise their models and CargoScope will calculate and display         365
301                                        and utilization rates for inbound shipments.         the energy needs and equivalent amount of carbon emitted. While           366
302                                        Corroboration of winery-related data                 more detailed documentation on underlying software mechanics is           367
303                                        (shipping frequency and volumes).                    available from CleanMetric’s website [3] and Venkat [26], this            368
304   Individuals’ names have been withheld upon request.                                       section briefly presents the functionality relevant for our analysis.      369

       Please cite this article in press as: Cholette S, Venkat K, The energy and carbon intensity of wine distribution: A study of logistical options for..., J
       Clean Prod (2009), doi:10.1016/j.jclepro.2009.05.011
                                                                     ARTICLE IN PRESS                                           JCLP1977_proof Š 16 June 2009 Š 4/13

           4                                                  S. Cholette, K. Venkat / Journal of Cleaner Production xxx (2009) 1–13

370                                                                                                                                                                          435
371                                                                                                                                                                          436
372                                                                                                                                                                          437
373                                                                                                                                                                          438
374                                                                                                                                                                          439
375                                                                                                                                                                          440
376                                                                                                                                                                          441
377                                                                                                                                                                          442
378                                                                                                                                                                          443
379                                                                                                                                                                          444
380                                                                                                                                                                          445
381                                                                                                                                                                          446
382                                                                                                                                                                          447

383                                                                                                                                                                          448
384                                                                                                                                                                          449

385                                                                                                                                                                          450
386                                                                                                                                                                          451
387                                                                                                                                                                          452
388                                                                                                                                                                          453
389                                                                                                                                                                          454
390                                                                                                                                                                          455

391                                                                                                                                                                          456
392                                                                                                                                                                          457
393                                                                                                                                                                          458
394                                                                                                                                                                          459
395                                                                                                                                                                          460
396                                                                                                                                                                          461
397                                                                                                                                                                          462
398                                                                                                                                                                          463
399                                                                                                                                                                          464
400                                                                                                                                                                          465
401                                                                                                                                                                          466

402                                                                                                                                                                          467
403                                                                                                                                                                          468
404                                                                                                                                                                          469
405                                                                                                                                                                          470
406                                                                       Fig. 2. The location of supply chain nodes.                                                        471

407                                                                                                                                                                          472
408                                                                                                                                                                          473

409        The user-defined inputs, fixed parameters and output relevant to                           (100%) utilization, powered by electricity from the Pacific region.       474
410        our model are summarized in Table 2.                                                     Selecting ‘‘transport properties’’ in Fig. 3 instead would display       475
411            CargoScope can be used to model the production, storage and                          Fig. 5. Notice the user-specified parameters that define the link to       476
412        distribution of any discrete packaged good. Venkat [26] documents                        the downstream echelon; non-temperature controlled midsized              477
413        that SEAT, a prior version of CargoScope that was not web-enabled,                       trucks travel from NVL to FDX with high (100%) utilization but no        478
414        has been used to model the supply chains of diverse goods such as                        (0%) backhaul. It should be noted that carrying limits are calculated    479

415        automotive supplies, printers, dairy products, biscuits, and frozen                      both for weight and volume. As bottled wine is heavy, carrying           480
416   Q2   foods. CleanMetrics, the company which created and maintains                             capacity will be maxed out by weight instead of volume for all           481
417        CargoScope, has worked with clients to develop detailed models for                       commercial vehicles utilized in these scenarios.                         482
418        supply chains supporting the distribution of cleaning products, soy                          Using characteristics of road transport modes, distances,            483
419        milk, produce, and textiles. While many food products have been                          regional energy estimates for power generation, and other industry       484

420        analyzed, this is the first time that CargoScope has been used to                         data, CargoScope calculates the energy usage and carbon emissions        485
421        model the distribution of wine.                                                          associated with transport and storage for each echelon. While            486
422            A model is constructed in CargoScope by starting with the end                        CargoScope is a more general tool that allows energy and emissions       487
423        consumer as the first node and then adding nodes for each echelon                         from processing to be calculated, present scenarios consider no          488
424        in the supply chain. Fig. 3 illustrates a high level view of one the                     other energy usage beyond that associated with transportation or         489
425        scenarios studied, that of 3PL local fulfillment through New Vine                         temperature-controlled storage.                                          490
426        Logistics (NVL) via FedEx (FDX). Each node represents either                                 While it is possible to perform similar analyses with custom         491
427        a storage or processing echelon, and the inter-echelon connections                       spreadsheet models, we feel CargoScope is more intuitive for non-        492
428        represent transportation links, where the user specifies the                              specialists, with its visual, interactive interface. Users can quickly   493
429        distance, selects from a predefined list of transport modes, and sets                     configure a model with predefined menus listing types of transit           494
430        three key parameters: temperature control, utilization rate and                          options or regional power sources. The user can redefine key              495
431        backhaul rate. Fig. 3 shows that the user has opted for a closer view                    parameters, such as dwell times or distances traveled and select         496
432        of storage properties for NVL, one of the echelons. The user would                       from different menu options for quick comparative analyses. This         497
433        then be presented with Fig. 4, which shows that products reside 14                       makes CargoScope a useful tool for demonstration purposes and            498
434        days in a temperature-controlled (cooler) storage with very high                         initial evaluations of supply chain processes.                           499

               Please cite this article in press as: Cholette S, Venkat K, The energy and carbon intensity of wine distribution: A study of logistical options for..., J
               Clean Prod (2009), doi:10.1016/j.jclepro.2009.05.011
                                                                ARTICLE IN PRESS                                              JCLP1977_proof Š 16 June 2009 Š 5/13

                                                         S. Cholette, K. Venkat / Journal of Cleaner Production xxx (2009) 1–13                                                  5

500   Table 2                                                                                                                                                                        565
501   Inputs, parameters and outputs for CargoScope.                                                                                                                                 566
502   User-supplied inputs                                                                                     Parameters provided by CargoScope                                     567
503   Universal                               Product weight                                                   Transport                             Energy usage, per km            568
504   Inputs                                  Product volume                                                   Mode                                  CO2 emissions profile, per km    569
505                                           Overall supply chain configuration                                Parameters                            Carrying capacity, by volume    570
506                                                                                                                                                  Carrying capacity, by weight    571
      Transportation                          Distances between nodes
507                                                                                                                                                                                  572
      Inputs                                  Transport mode                                                   Storage                               Energy usage, per day
508                                           Level of temperature control                                     Parameters                            Emissions profile, per day       573
509                                           Utilization rate                                                                                                                       574
510                                           Backhaul rate                                                                                                                          575
                                                                                                               Outputs calculated by CargoScope
511                                                                                                                                                                                  576
      Storage                                 Dwell times                                                      Energy usage for each node and link
512                                                                                                                                                                                  577

      Inputs                                  Location and type of power used                                  CO2 emissions by node and link
513                                           Level of temperature control                                                                                                           578
514                                           Utilization rate                                                                                                                       579

515                                                                                                                                                                                  580
516                                                                                                                                                                                  581
517   3.2. Scenarios supporting local deliveries                                               back to the distributor’s warehouse utilizes the full capacity of the                 582
518                                                                                            truck. As the distances are relatively short, the vehicles used in all                583
519      We first consider how our representative Sonoma winery could                           local scenarios are assumed not need any temperature control to                       584
520   fulfill the orders of San Francisco consumers. Table 3 provides                           prevent wine spoilage.                                                                585

521   a summary breakdown of these methods. We describe a base case                                As the wine has been ordered by end consumers, we utilize                         586
522   scenario in detail and then indicate how alternative scenarios differ.                   a pull model. We assume wine spends a week at the distributor’s                       587
523   Scenario configurations and data were drawn from discussions with                         warehouse until another midsized truck is used to transfer wine                       588
524   operations’ managers at various echelons, as seen in Table 1. We                         from the distributor’s warehouse to the retailer store (RS), making                   589
525   provide justification for assumptions when data are unavailable.                          such deliveries every week. We assume that the wine remains in                        590
526                                                                                            temperature-controlled storage at the retailer for a week before the                  591
527   3.2.1. Standard scenario L1: 3-tier distribution                                         customer (CU) drives to the store and back in a gasoline powered                      592
528       The base scenario for local distribution (L1) is represented by the                  Honda Accord, at a fuel efficiency of 9.8 l per 100 km, for the sole                   593
529   3-tier system, as it is the predominant outbound logistical method;                      purpose of picking up the wine, thus utilizing only 24% of the car’s                  594
530   Cholette [4] shows that it supports 90% of all U.S. wine purchases.                      stated hauling capacity by weight. We also consider two scenario                      595
531   Midsized trucks are used to transport wine from the winery’s                             variants. In L1a the consumer reaches the retail store without a car                  596

532   warehouse (WW) to the distributor’s warehouse (DW). We assume                            and in L1b the consumer more effectively utilizes the car by fully                    597
533   that the rest of the truck’s capacity is utilized efficiently to transport                loading it with other purchases.                                                      598
534   other products from nearby wineries to the same destination.                                 Wine storage facilities should be cooled but not refrigerated,                    599
535   Indeed, for the delivery portion of a trip we assume that capacity is                    with 13  C the ideal temperature. The energy cost associated with                    600
536   utilized with 100% efficiency in all commercial vehicles for every                        warehousing wine is calculated by determining the area necessary                      601

537   scenario. Although our interviewees and other data sources could                         to store the wine and the duration of the stay. We assume that the                    602
538   not provide us with definitive backhauling and utilization rates, we                      warehouse is highly utilized and record energy use only for when                      603

539   can partially justify assuming high utilization rates by use of                          wine remains in storage and not after the wine has been moved to                      604
540   significant dwell times at all intermediate warehousing echelons.                         another echelon. We also cease considering energy usage associ-                       605
541   Unless stated otherwise, no backhauling is assumed to occur. For                         ated with storage after final delivery to the consumer has occurred.                   606
542   instance, the model considers that the truck is empty when it drives                     While some consumers may possess wine refrigerators, most store                       607
543   to the winery’s warehouse from the distributor, but that the trip                        wine at ambient house temperature.                                                    608
544                                                                                                                                                                                  609

545                                                                                                                                                                                  610
546                                                                                                                                                                                  611
547                                                                                                                                                                                  612
548                                                                                                                                                                                  613
549                                                                                                                                                                                  614

550                                                                                                                                                                                  615
551                                                                                                                                                                                  616
552                                                                                                                                                                                  617
553                                                                                                                                                                                  618
554                                                                                                                                                                                  619
555                                                                                                                                                                                  620
556                                                                                                                                                                                  621
557                                                                                                                                                                                  622
558                                                                                                                                                                                  623
559                                                                                                                                                                                  624
560                                                                                                                                                                                  625
561                                                                                                                                                                                  626
562                                                                                                                                                                                  627
563                                                                                                                                                                                  628
564                                                        Fig. 3. Graphical view of example supply chain in CargoScope.                                                             629

       Please cite this article in press as: Cholette S, Venkat K, The energy and carbon intensity of wine distribution: A study of logistical options for..., J
       Clean Prod (2009), doi:10.1016/j.jclepro.2009.05.011
                                                                ARTICLE IN PRESS                                           JCLP1977_proof Š 16 June 2009 Š 6/13

      6                                                  S. Cholette, K. Venkat / Journal of Cleaner Production xxx (2009) 1–13

630                                                                                                                                                                      695
631                                                                                                                                                                      696
632                                                                                                                                                                      697
633                                                                                                                                                                      698
634                                                                                                                                                                      699
635                                                                                                                                                                      700
636                                                                                                                                                                      701
637                                                                                                                                                                      702
638                                                                                                                                                                      703
639                                                                                                                                                                      704
640                                                                                                                                                                      705
641                                                                                                                                                                      706
642                                                                                                                                                                      707

643                                                                                                                                                                      708
644                                                                                                                                                                      709

645                                                                                                                                                                      710
646                                                                                                                                                                      711
647                                                                                                                                                                      712
648                                                             Fig. 4. Configuring a storage echelon in CargoScope.                                                      713
649                                                                                                                                                                      714
650                                                                                                                                                                      715

651   3.2.2. L2: 3-tier distribution via retailer warehouse                                    3.2.4. L4: fulfillment via 3PL                                             716
652       Given that many chain retailers make use of regional distribu-                           This scenario (L4) considers wine that is shipped to customers        717
653   tion centers, we modify the network to route the distributor’s                           through direct sales channels, via New Vine Logistics, a leading 3PL      718
654   midsized truck to the additional echelon of a regional centralized                       provider focused on wine industry clients. Midsized trucks from           719
655   warehouse (RCW) instead of the retail store (RS). We assume that                         New Vine Logistics (NVL) pickup wine from the winery’s ware-              720
656   the wine will stay in temperature-controlled storage at the RCW for                      house (WW) for transport back to NVL’s temperature-controlled             721
657   an additional week. This assumption also allows us to justify high                       warehouse. The small package carrier sends a midsized truck to            722
658   (100%) utilization rates for transit to the retail store. Otherwise this                 pickup wine from New Vine and bring it to the sorting center (FDX)        723
659   scenario (L2) is similar to the base scenario (L1).                                      in South San Francisco every 2 weeks. The sorting center is not           724
660                                                                                            climate controlled, but as packages reside only briefly, spoilage is       725
661   3.2.3. L3: winery self-distribution                                                      unlikely to occur. The wine is then sent by a light parcel truck to the   726

662       The difference from the base scenario (L1) is that the winery is                     end consumer in San Francisco. Carriers such as Federal Express           727
663   now permitted to engage in self-distribution. Although 3-tier                            have domain expertise in being efficient, and parcel trucks                728
664   distribution is the most common channel, some California wineries                        returning from customer drop offs will pickup outbound parcels            729
665   have filed the paperwork to obtain the legal right to bypass                              from urban drop points in the return trip to the sorting center.          730
666   distribution for direct sale to an instate retailer. The winery                          Therefore, both high utilization (100%) and that significant (50%)         731

667   provides or contracts for a truck to deliver wine directly to the retail                 backhauling are assumed to occur. This is the only transport link in      732
668   store (RS) from the winery’s warehouse (WW). As always, we                               any of the local scenarios to have a non-zero backhauling rate.           733

669   assume 100% utilization. As some wineries may not generate                                                                                                         734
670   sufficient order volumes to fill a midsized truck, with a 6250 kg of                       3.2.5. L5abc: consumer drives to winery                                   735
671   carrying capacity equivalent to 344 cases of wine, we additionally                           The final local scenarios also result from the direct sales channel,   736
672   consider utilizing a light truck with a vastly reduced capacity of                       but consider consumers who make dedicated trips to the winery to          737
673   a mere 600 kg, the equivalent of 33 cases of wine. Removing the                          take possession of wine orders. This supply chain option is the           738
674   distributor echelon results in one less week of storage costs and                        simplest and considers only the fuel used in the round trip. We           739

675   slightly decreases the total distance traveled in this scenario (L3).                    continue to employ the same car that consumes 8.9 l of gasoline per       740
676                                                                                                                                                                      741
677                                                                                                                                                                      742
678                                                                                                                                                                      743
679                                                                                                                                                                      744

680                                                                                                                                                                      745
681                                                                                                                                                                      746
682                                                                                                                                                                      747
683                                                                                                                                                                      748
684                                                                                                                                                                      749
685                                                                                                                                                                      750
686                                                                                                                                                                      751
687                                                                                                                                                                      752
688                                                                                                                                                                      753
689                                                                                                                                                                      754
690                                                                                                                                                                      755
691                                                                                                                                                                      756
692                                                                                                                                                                      757
693                                                                                                                                                                      758
694                                                           Fig. 5. Configuring an inter-echelon link in CargoScope.                                                    759

          Please cite this article in press as: Cholette S, Venkat K, The energy and carbon intensity of wine distribution: A study of logistical options for..., J
          Clean Prod (2009), doi:10.1016/j.jclepro.2009.05.011
                                                                     ARTICLE IN PRESS                                              JCLP1977_proof Š 16 June 2009 Š 7/13

                                                             S. Cholette, K. Venkat / Journal of Cleaner Production xxx (2009) 1–13                                                  7

760   Table 3                                                                                                                                                                            825
761   Summary of local scenarios’ inter-echelon links.                                                                                                                                   826
762               L1: 3-tier distribution       L2: 3-tier distribution          L3ab: self-distribution           L4: delivery via 3PL                 L5abc: consumer drives           827
763                                             with RDC                                                                                                                                 828
764   Echelon     WW > DW         112 km,       WW > DW             112 km,      WW > RS        72 km,             WW > NVL               29 km,        WW > CU        72 km, car (a),   829
765     1–2                       midsized                          midsized                    light (a) or                              midsized                     hybrid (b) or     830
766                               truck                             truck                       midsized (b)                              truck                        mid-pickup (c)    831
767                                                                                                                                                                                      832
      Echelon     DW > RS         60 km,        DW > RCW            48 km,       RS > CU        112 km,            NVL > FDX 75 km,
768     2–3                       midsized                          midsized                    midsized           midsized                                                              833
769                               truck                             truck                       truck              truck                                                                 834
770   Echelon     RS > CU         3.6 km,       RCW > RS            32 km,                                         FDX > CU               10 km,                                         835
        3–4                       car                               midsized                                                              light truck
771                                                                                                                                                                                      836
772                                                                                                                                                                                      837

      Echelon                                   RS > CU             3.6 km,
773     4–5                                                         car                                                                                                                  838
774                                                                                                                                                                                      839

775                                                                                                                                                                                      840
776   100 km, for scenario L5a. We also consider a scenario variant L5b,                           winery warehouse (WW) to the distributor warehouse (DW) is                            841
777   a variant where the car in question is a hybrid, averaging 4.3 l per                         filled with wine destined for both local and far markets. A heavy-                     842
778   100 km. Additionally, we consider a further extension to this                                duty diesel truck with cooling is used to make the cross-country                      843
779   scenario to model the consumer who may take the trip and                                     journey to the company’s distribution center in Monroe Township,                      844
780   consolidate several purchases, such as picking up wine club                                  New Jersey (DW-NJ). Because this is a long, expensive link, we                        845

781   purchases on behalf of neighbors and nearby friends, none of                                 assume that the distributor sets capacity and backhauling rates at                    846
782   whom need to drive any distance to receive their orders from this                            100%. Such efficiencies are possible as Southern/Glazer’s also                         847
783   generous driver. We thus assume the consumer in Scenario L5c                                 distributes European imported wine and Eastern produced spirits                       848
784   fully utilizes a midsized pickup truck, which has half the cargo                             to California retailers. The example retailer we consider, Whole                      849
785   space of a light commercial truck and holds 33 half cases of wine. It                        FoodsÔ, does not have any distribution warehouses in New Jersey                       850
786   should be noted that individuals or companies offering such                                  so we bypass the optional retailer warehouse echelon, with the                        851
787   a service for a fee would need special permits to avoid legal issues                         distributor sending wine to the Manhattan retailer via midsized                       852
788   associated with transporting and distributing alcohol.                                       truck. With the density of retail outlets and residential housing in                  853
789                                                                                                Manhattan, we assume that consumers need to travel at most                            854
790   3.3. Scenarios supporting long distance delivery                                             0.8 km (0.5 mile) to reach the store and that they take public transit                855
791                                                                                                or walk to the store. As a half case of wine is fragile, heavy and                    856

792       For the Manhattan consumer, the sheer distances change the                               awkward to carry by hand, our hypothetical consumer hails a cab                       857
793   scenarios under consideration. No rational consumer would make                               for their return trip, effectively resulting in a 100% backhaul rate.                 858
794   a dedicated cross-country drive for a wine purchase. Nor is winery                                                                                                                 859
795   self-distribution an option with interstate sales. However, a variety                        3.3.2. D2: long distance fulfillment via 3PL ground delivery                           860
796   of other network configurations exist. In addition to traditional 3-                              New York state has allowed direct-to-consumer sales from Cal-                     861

797   tier distribution, 3PLs such as New Vine Logistics, supported by                             ifornia since 2005. Wineries often offer remote consumers a choice                    862
798   carriers such as FedEx, offer a choice between air shipping and                              between ground and air delivery. This scenario (D2) considers                         863
      ground based delivery via truck. We also consider an intermodal                              ground delivery, supported by a service such as FedEx Ground, with

799                                                                                                                                                                                      864
800   transport option, utilizing rail for the cross-country link. The                             New Vine Logistics as the 3PL provider. The supply chain is identical                 865
801   scenarios are summarized in Table 4.                                                         to that of scenario L4 in Section 3.2.4, up to the point at which the                 866
802                                                                                                wine is ready to leave the NVL facility. As wine transported long                     867
803   3.3.1. D1: standard long distance scenario: 3-tier distribution                              distance by truck may be subject to spoilage, we assume that New                      868
804       The 3-tier distribution system is the prevalent method for                               Vine Logistics packs shipments in a proprietary multi-day temper-                     869

805   supporting longer distance wine supply chains within the U.S. We                             ature-regulating packaging, as documented by their partner’s                          870
806   continue to make use of the same distributor’s warehouse, as                                 website [28]. We account for the energy associated with this addi-                    871
807   Southern/Glazer’s is also the dominant player in the New York                                tional packaging by modeling all subsequent links as being cooled.                    872
808   market. The initial part of the supply change is identical to that of                        After the wine is transported to the FedEx center in South San                        873
809   scenario L1, described in Section 3.2.1; the midsized truck from the                         Francisco (FDX-CA), a heavy-duty diesel truck carries the wine cross-                 874

810                                                                                                                                                                                      875
      Table 4
811                                                                                                                                                                                      876
      Summary of long distance scenarios’ inter-echelon links.
812                                                                                                                                                                                      877
813               D1: 3-tier distribution                D2 3PL fulfillment via truck            D3 3PL via air                        D4 3PL via rail                                    878
814   Echelon     WW > DW-CA         112 km,             WW > NVL          29 km,               WW > NVL              29 km,          WW > NVL          29 km, midsized truck            879
        1–2                          midsized                              midsized truck                             midsized
815                                                                                                                                                                                      880
                                     truck                                                                            truck
816   Echelon     DW-CA >            4700 km,            NVL > FDX-CA      75 km,               NVL > FDX-CA          75 km,          NVL > FDX-CA      75 km, midsized truck, cooler    881
817     2–3       DW-NJ              heavy-duty                            midsized                                   midsized                                                           882
818                                  truck, cooler                         truck, cooler                              truck                                                              883
819   Echelon     DW-NJ >            74 km,              FDX-CA > ESS      4675 km,             FDX-CA > NWR,         4960 km,        FDX-CA > OAK      50 km, midsized truck, cooler    884
        3–4       RS-NY              midsized                              heavy-duty           via MEM               Airfreight
820                                                                                                                                                                                      885
                                     truck                                 truck, cooler
821   Echelon     RS-NY > CU         0.8 km, car         ESS > CU          53 km, light         EWR > CU              21 km, light    OAK > EWR,        5500 km, rail, cooler            886
822     4–5                                                                truck, cooler                              truck           via CHI                                            887
823   Echelon                                                                                                                         EWR > CU          21 km, light truck, cooler       888
824                                                                                                                                                                                      889

       Please cite this article in press as: Cholette S, Venkat K, The energy and carbon intensity of wine distribution: A study of logistical options for..., J
       Clean Prod (2009), doi:10.1016/j.jclepro.2009.05.011
                                                                    ARTICLE IN PRESS                                           JCLP1977_proof Š 16 June 2009 Š 8/13

      8                                                      S. Cholette, K. Venkat / Journal of Cleaner Production xxx (2009) 1–13

890   country to the FedEx sorting center in Edison (ESS). We justify a 100%                       associated with each echelon and link in the supply chain, where             955
891   backhaul rate on this link, as the same truck is used to ship packages                       each order is 6 bottles of wine.                                             956
892   west to California clients. Once the package reaches Edison, it is sent                                                                                                   957
893   by parcel truck to home of the Manhattan consumer.                                           4.1. L1: local 3-tier distribution results                                   958
894                                                                                                                                                                             959
895   3.3.3. D3: long distance fulfillment via 3PL airfreight                                           Table 5 lists the energy and emissions associated with each link         960
896       Most 3PL providers offer air shipment as well as delivery service                        and node that can be assigned to the half case of wine being routed          961
897   by truck. Many clients are willing to pay the higher price for                               through the supply chain in our base case scenario for local distri-         962
898   airfreight not only for faster delivery, but also because transporting                       bution (L1). In total, 31 MJ of energy are utilized in getting this order    963
899   wine cross-country on trucks without temperature controls can                                from the winery to the end consumer’s home, resulting in 2.18 kg of          964
900   spoil wines. This scenario (D3) replaces the long distance diesel                            CO2 being emitted. Transportation link emissions are presented in            965
901   truck link with a FedEx air cargo route, routing the plane from San                          top-down order, followed by storage echelons emissions, ordered              966
902   Francisco Airport, adjacent to FDX-CA, to Newark International                               top-down. Emissions associated with transportation from the                  967

903   Airport (NWR) via Memphis. CargoScope assigns both a 100%                                    winery to the retail store (0.46 kg of CO2 per half case) dominate           968
904   utilization and backhaul rate to airfreight. As this link is of                              those from storage (0.04 kg of CO2 per half case) by a factor of ten.        969

905   comparatively short duration, temperature-controlled packaging is                            While dwell times at the different echelons may vary from our                970
906   not necessary. The New Jersey FedEx facility, also very near the                             assumptions, these results suggest that dwell times have minimal             971
907   airport, is assumed to dispatch a Manhattan-bound parcel truck to                            impact on emissions and are of less concern for this analysis.               972
908   the consumer, assuming the same utilization rate (100%) and                                      The eye-catching result from Table 5 is that the most energy-            973
909   backhaul (50%) as its Bay Area counterpart.                                                  intensive transit link is the last one. Given our assumptions, driving       974
910                                                                                                to the retail store on dedicated trips accounts for over three fourths       975

911   3.3.4. D4: long distance fulfillment via 3PL utilizing rail                                   of the total supply chain emissions. This result may seem surprising         976
912       Although carriers like FedEx have both extensive ground and air                          with the short distance involved. However, per-case energy usage is          977
913   networks, they do not have the same presence in rail in part because                         much lower for freight vehicles, which tend to be more highly                978
914   of a lack of an open, national rail network. However, public pressure                        utilized than individual personal vehicles. Our assumed low utili-           979
915   and rising fuel costs may convince companies to increase rail usage.                         zation rate for consumer vehicles is echoed by a government study            980
916   We consider a scenario (D4) where the long distance link is via rail,                        [16] showing average Americans do not tend to engage in energy-              981
917   through a company such as CSX, one of the dominant rail carriers in                          saving behaviors, such as carpooling to work. Other studies in               982
918   the U.S. This scenario has the same configuration as that of scenario                         apparel [2] and food [25] also find that the retailer-to-consumer             983
919   D2, until it is time for the package to leave the FedEx facility in South                    link can be the most carbon intensive, even in European countries            984
920   San Francisco. At that point, a midsized truck is sent to the Oakland                        where consumers are traditionally more energy conscious than                 985
921   rail terminal (OAK), with 100% loading and 0% backhaul. The rail                             their U.S. counterparts. If San Francisco consumers walk or take             986

922   company would then route the shipment to the rail terminal (NWR)                             well-utilized public transit, only 0.50 kg of CO2 per half case in total     987
923   adjacent to Newark International Airport. CargoScope assigns both                            would be emitted. More realistically, if these consumers drive but           988
924   100% utilization and backhaul rate to all rail cargo. As with the air                        make additional purchases to fully utilize the car’s cargo space up to       989
925   shipping scenario (D3) we assume that the package does not dwell                             the specified weight limit, emissions drop to 0.90 kg of CO2 per half         990
926   for any measurable time at FedEx’s EWR facilities, but instead is sent                       case. We discuss implications of these findings in Section 5.                 991

927   on a Manhattan-bound parcel truck to the consumer’s home. Given                                                                                                           992
928   this journey takes several days on vehicles lacking temperature                              4.2. L2: 3-tier distribution via retailer warehouse results                  993

929   control, NVL would package the wine in the same temperature-                                                                                                              994
930   regulating packaging as featured in scenario D2.                                                Comparing Table 6 with Table 5 reveals that inclusion of                  995
931                                                                                                a regional centralized warehouse (RCW) increases overall energy              996
932   4. Model results                                                                             usage and emissions by 3%. However, it should be noted that our              997
933                                                                                                standard assumption is that outbound transit from the distributor            998
934      We present and interpret the results for each scenario and then                           results in 100% utilization. Use of this consolidation echelon would         999

935   perform a summary comparison across all scenarios, local and long                            result an overall efficiency gains if our distributor instead typically      1000
936   distance. Although we include figures for both energy usage and                               provides relatively small volumes to the client store or set of stores      1001
937   emissions, we focus on the latter. Transportation energy usage                               and routinely fills a midsized truck to 50% or less of capacity.             1002
938   dominates that associated with storage, and the emissions’ profiles                                                                                                       1003
939   of the various fuels consumed by different transport modes are                               4.3. L3ab: winery self-distribution results                                 1004

940   similar. Thus, total energy expended correlates closely with emis-                                                                                                       1005
941   sions. Results are presented in terms of per-order emissions                                     We now consider what happens when we bypass the distribution            1006
942                                                                                                tier. As a winery may not send large orders to the retailer, we             1007
943                                                                                                compare results from use of a light truck (L3a) to that of a midsized       1008
      Table 5
944                                                                                                truck (L3b), assuming 100% utilization of each vehicle. As can be seen      1009
      Energy and emissions by link and echelon, scenario L1.
945                                                                                                in Table 7, the choice of which truck to use has great impact. If           1010
946   Scenario L1: local 3-tier, standard           Distance/     Energy – MJ      Carbon –        a winery generates sufficient volume of sales, self-distribution with        1011
      scenario                                      time                           kg CO2
947                                                                                                highly utilized midsize trucks is more efficient (at 1.89 kg of CO2 per      1012
948   Transport     Midsize truck, diesel           112 km         4.05            0.3             half case) than the previously presented standard scenario (2.18 kg of      1013
      Transport     Midsize truck, diesel           60 km          2.17            0.16
949                                                                                                CO2 per half case). However, for smaller wineries that have insuffi-         1014
      Transport     HondaAccord, gasoline           3.6 km        24.22            1.68
950   Storage       None, electricity-US-Pacific     0 days         0               0               cient volume to fill a midsized truck, the use of a distributor would        1015
951   Storage       Cooler, electricity-US-Pacific   7 days         0.37            0.02            result in lower emissions than would self-distribution via a light          1016
952   Storage       Cooler, electricity-US-Pacific   8 days         0.37            0.02            truck or a highly underutilized midsized truck. As consumer’s driving       1017
      Storage       None, electricity-US-Pacific     0 days         0               0
953                                                                                                still dominates all of these local scenarios, the efficiency differences     1018
                                                    Total         31.19            2.18
954                                                                                                are even greater than the emissions’ totals would suggest. Fully            1019

          Please cite this article in press as: Cholette S, Venkat K, The energy and carbon intensity of wine distribution: A study of logistical options for..., J
          Clean Prod (2009), doi:10.1016/j.jclepro.2009.05.011
                                                                        ARTICLE IN PRESS                                            JCLP1977_proof Š 16 June 2009 Š 9/13

                                                               S. Cholette, K. Venkat / Journal of Cleaner Production xxx (2009) 1–13                                                   9

1020   Table 6                                                                                       Table 8                                                                                1085
1021   Energy and emissions by link and echelon, scenario L2.                                        Energy and emissions by link and echelon, scenario L4.                                 1086
1022   Scenario L2: local 3-tier, with retailer          Distance/      Energy –     Carbon – kg     Scenario L4: local 3PL delivery                          Distance/ Energy – Carbon –   1087
1023   warehouse                                         time           MJ           CO2                                                                      time      MJ       kg CO2     1088
1024   WW > DW       Transport   Midsize truck, diesel   112 km          4.05        0.3             WW > NVL     Transport   Midsize truck, diesel           29 km     1.05     0.08       1089
1025   DW > RCW      Transport   Midsize truck, diesel   60 km           1.74        0.13            NVL > FDX    Transport   Midsize truck, diesel           75 km     2.71     0.20       1090
1026   RCW > RS      Transport   Midsize truck, diesel   32 km           1.16        0.09            FDX > CU     Transport   Light truck, diesel             10 km     1.32     0.10       1091
       RS > CU       Transport   HondaAccord,            3.6 km         24.22        1.68            WW           Storage     Cooler, electricity-US-Pacific   0 days    0        0
1027                                                                                                                                                                                        1092
                                 gasoline                                                            NVL          Storage     Cooler, electricity-US-Pacific   14 days   0.75     0.04
1028   WW            Storage     Cooler, electricity-    0 days          0           0               FDX          Storage     None, electricity-US-Pacific     0 days    0        0          1093
1029                             US-Pacific                                                           Customer     Storage     None, electricity-US-Pacific     0 days    0        0          1094
1030   DW            Storage     Cooler, electricity-    7 days          0.37        0.02                                                                     Total     5.83     0.42       1095
1031                                                                                                                                                                                        1096
       RCW           Storage     Cooler, electricity-    7 days          0.37        0.02
1032                                                                                                                                                                                        1097

1033   RS            Storage     Cooler, electricity-    7 days          0.37        0.02            amount produced from distribution through the 3-tier system. Even                      1098
1034                             US-Pacific                                                           if the consumers were to utilize a hybrid car such as a Toyota Prius                   1099

1035   Customer      Storage     None, electricity       0 days          0           0                                                                                                      1100
                                                         Total          32.28        2.25
                                                                                                     (L5b), emissions would still total 14.5 kg of CO2 per half case, over six
1036                                                                                                 times of those associated with the 3-tier scenario (L1). Of course,                    1101
1037                                                                                                 many consumers may not undertake a round trip to the wine                              1102
1038   utilized light truck usage results in nearly 5 times the emissions                                                                                                                   1103
                                                                                                     country just to pick up a single wine shipment. Drivers may justify
       (0.94 kg of CO2 per half case) of those from a fully loaded midsized                                                                                                                 1104
                                                                                                     such trips by picking up additional wine orders from nearby
1040   truck (0.19 kg of CO2 per half case). These results are similar to                                                                                                                   1105
                                                                                                     wineries. An analogous situation would be that of a consumer col-

       findings by Van Hauwermeiren et al. [25] economies of scale from                                                                                                                      1106
                                                                                                     lecting additional orders for neighbors and nearby friends. Scenario
1042   consolidating transit dramatically impact emissions efficiency of the                                                                                                                 1107
                                                                                                     L5c thus represents an extreme version of the latter possibility. It
       overall supply chain.                                                                                                                                                                1108
                                                                                                     assumes the consumer fills a midsized pickup truck, representing
1044                                                                                                 a total of 33 half case orders. Emissions would drop to 1.43 kg of CO2                 1109
1045   4.4. L4: results for local fulfillment via 3PL                                                 per half case, but this efficiency holds only if none of the other                      1110
1046                                                                                                 consumers require a special car trip to the pickup truck owner’s                       1111
1047      Table 8 shows that the direct shipping option (L4) produces the                            home to get their orders. Note that even with this unrealistic                         1112
1048   lowest emissions of all: 0.42 kg of CO2 per half case, or 19% of the                          expectation, per-order emissions are still higher than those from the                  1113
1049   emissions associated with the standard 3-tier scenario (L1). Much                             3PL scenario, in part because large personal vehicles, even when                       1114
1050   of this improvement can be traced to eliminating driving to the                               fully loaded, are less efficient than well utilized commercial ones.                    1115
1051   store. End-customer delivery is comparatively fuel efficient as                                                                                                                       1116

1052   parcel trucks are assumed to have 100% utilization in delivery and                                                                                                                   1117
1053   employ some (50%) backhauling. If we consider removing                                                                                                                               1118
                                                                                                     4.6. D1: results for the standard long distance scenario
1054   consumer driving from the standard scenario, the direct shipping                                                                                                                     1119
                                                                                                     of 3-tier distribution
1055   scenario would result in only slightly lower (88%) emissions. The                                                                                                                    1120
1056   minor savings can be attributed to the more efficient routing and                                                                                                                     1121

                                                                                                         We now consider the results for cross-country orders. The base
1057   the services of the 3PL provider. For instance, if the winery had to                                                                                                                 1122
                                                                                                     long distance scenario of shipping a half case of wine via the 3-tier
1058   drive orders to a consolidation point or if FedEx had to directly send                                                                                                               1123
                                                                                                     distribution system results in 48.61 MJ of energy usage and 3.62 kg

1059   parcel trucks to the winery, emissions would likely increase.                                                                                                                        1124
                                                                                                     of CO2 emitted. These emissions are only 66% more than those from
1060                                                                                                                                                                                        1125
                                                                                                     local 3-tier shipping. This result can be explained by examining
1061   4.5. L5abc: results for the consumer driving to winery                                        Table 9. While the trip between the California and New Jersey
1062                                                                                                                                                                                        1127
                                                                                                     distribution centers contributes the most to emissions, this link is
1063      If a casual observer might be tempted to expect that eliminating                           relatively efficient, accounting for 78% of the emissions, but 96% of
1064   layers in a supply chain automatically increases energy efficiency,                                                                                                                   1129

                                                                                                     the distance covered. Additionally, the Manhattan consumer travels
1065   the following results would put this misconception to rest. Driving                           a shorter distance by car, resulting in the least amount of emissions
1066   a conventional car to the winery (L5a) results in the most emissions                          produced of all the scenario’s transit links. It can also be seen that
1067   being produced, 33.75 kg of CO2 per half case, over 15 times the                              Mid-Atlantic electricity results in more carbon emissions than
1068                                                                                                                                                                                        1133
                                                                                                     Pacific electricity, although emissions from cold storage have
1069                                                                                                                                                                                        1134
                                                                                                     minimal impact in our results.

1070   Table 7                                                                                                                                                                              1135
       Energy and emissions by link and echelon, scenarios L3a and L3b.
1071                                                                                                                                                                                        1136
1072   L3a: Winery self-distribution, via light truck          Distance/ Energy – Carbon –           Table 9                                                                                1137
                                                               time      MJ       kg CO2             Energy and emissions by link and echelon, scenario D1.
1073                                                                                                                                                                                        1138
1074   WW > RS     Transport   Light truck, diesel             72 km         12.68       0.94                                                                                               1139
                                                                                                     D1: Long distance: 3-tier distribution           Distance/time Energy – MJ Carbon –
       RS > CU     Transport   HondaAccord, gasoline           3.6 km        24.22       1.68
1075                                                                                                                                                                            kg CO2      1140
       WW          Storage     None, electricity-US-Pacific     0 days         0          0
1076   RS          Storage     Cooler, electricity-US-Pacific   7days          0.37       0.02        Transport   Midsize truck, diesel                112 km         4.05        0.3        1141
1077   CU          Storage     None, electricity               0 days         0          0           Transport   Heavy-duty truck, diesel, Cooler     4700 km       38.07        2.82       1142
1078                                                           Total         37.27       2.64        Transport   Midsize truck, diesel                74 km          2.68        0.20       1143
       L3b: Winery Self-Distribution, via midsized truck                                             Transport   HondaAccord, gasoline                0.8 km         2.69        0.19
1079                                                                                                                                                                                        1144
       WW > RS Transport Midsize truck, Diesel                 72 km          2.6        0.19        Storage     None, electricity-US-Pacific          0 days         0           0
1080                                                                                                 Storage     Cooler, electricity-US-Pacific        7 days         0.37        0.02       1145
       RS > CU   Transport HondaAccord, gasoline               3.6 km        24.22       1.68
1081   WW        Storage     None, electricity-US-Pacific       0 days         0          0           Storage     Cooler, electricity-US-MidAtlantic   7 days         0.37        0.05       1146
1082   RS        Storage     Cooler, electricity-US-Pacific     7days          0.37       0.02        Storage     Cooler, electricity-US-MidAtlantic   7 days         0.37 0.05              1147
1083   CU        Storage     None, electricity                 0 days         0          0           Storage     None, electricity-US-MidAtlantic     0 days         0           0          1148
                                                               Total         27.2        1.89                                                         Total         48.61        3.62
1084                                                                                                                                                                                        1149

        Please cite this article in press as: Cholette S, Venkat K, The energy and carbon intensity of wine distribution: A study of logistical options for..., J
        Clean Prod (2009), doi:10.1016/j.jclepro.2009.05.011
                                                                 ARTICLE IN PRESS                                          JCLP1977_proof Š 16 June 2009 Š 10/13

       10                                                 S. Cholette, K. Venkat / Journal of Cleaner Production xxx (2009) 1–13

1150   Table 10                                                                                                                                                             1215
1151   Energy and emissions by link and echelon, scenario D2.                                                                                                               1216
1152   D2: Long distance 3PL fulfillment via truck                                                              Distance/time             Energy –          Carbon-kg CO2    1217
1153                                                                                                                                     MJ                                 1218
1154   WW > NVL                     Transport              Midsize truck, diesel                               29 km                      1.05             0.08             1219
1155   NVL > FDX-CA                 Transport              Midsize truck, diesel, cooler                       75 km                      2.72             0.20             1220
1156   FDX-CA > ESS                 Transport              Heavy-duty truck, diesel, cooler                    4675 km                   37.87             2.80             1221
       ESS > CU                     Transport              Light truck, diesel, cooler                         53 km                      7                0.52
1157                                                                                                                                                                        1222
       WW                           Storage                None, electricity-US-Pacific                         0 days                     0                0
1158   NVL                          Storage                Cooler, electricity-US-Pacific                       14 days                    0.75             0.04             1223
1159   FDX-CA                       Storage                Cooler, electricity-US-Pacific                       0 days                     0                0                1224
1160   ESS                          Storage                Cooler, electricity-US-MidAtlantic                  0 days                     0                0                1225
       CU                           Storage                None, electricity-US-MidAtlantic                    0 days                     0                0
1161                                                                                                                                                                        1226
                                                                                                               Total                     49.39             3.64
1162                                                                                                                                                                        1227

1163                                                                                                                                                                        1228
1164   4.7. D2: results for long distance 3PL fulfillment via ground delivery                    would be even greater for shipments routed between points with              1229

1165                                                                                            better established rail infrastructure, such as sending cargo from          1230
1166       As can be seen from comparing Tables 9 and 10, negligible                            Los Angeles to Chicago. It should also be expected that a 3-tier            1231
1167   difference in overall emissions exists between this scenario and                         distribution plan utilizing rail would result in emissions efficiencies      1232
1168   that of 3-tier distribution. Most emissions occur on the cross-                          similar to those realized in this scenario.                                 1233
1169   country routing of the truck. Slight savings from this scenario’s                                                                                                    1234
1170   decreased electricity usage are offset by the fact that highly utilized                                                                                              1235

1171                                                                                            4.10. Comparison across all scenarios                                       1236
       light parcel trucks, even with some backhauling, are still less effi-
1172   cient than midsized trucks. Thus, the inbound Manhattan transit                                                                                                      1237
1173                                                                                               Our study considers many scenarios with a variety of transport           1238
       link results in more emissions even though FedEx’s staging center
1174                                                                                            modes, echelons and distances. One informative way to present               1239
       for receiving cross-country shipments is slightly closer to the city
1175                                                                                            results is to consider the emissions and energy totals from all the         1240
       than is Southern/Glazer’s distribution center. Likewise, parcel
1176                                                                                            scenarios and list them in increasing order of emissions generated.         1241
       delivery’s elimination of having the consumer drive to the retail
1177                                                                                            Table 13 illustrates that significant emissions difference exist. The        1242
       store has less impact when that trip to the retail store is much
1178                                                                                            least efficient scenario, driving to the winery in a typical gas             1243
       shorter and has effective backhauling.
1179                                                                                            powered car (L5a), results in 80 times the emissions that would             1244
1180                                                                                            occur if that local delivery were handled via our 3PL scenario (L4).        1245
       4.8. D3: results for long distance 3PL fulfillment via airfreight                         While most local supply chain configurations produce lower
1181                                                                                                                                                                        1246

1182                                                                                            emissions than their long distance counterparts, there are some             1247
          Opting for 3PL delivery via airfreight (D3) instead of trucking                       notable exceptions. In particular, long distance 3PL delivery via rail
1183                                                                                                                                                                        1248
       (D2) increases total emissions by over a factor of seven. Although                       (D4) makes the top half of the list and is effectively equivalent to the
1184                                                                                                                                                                        1249
       carriers can be presumed to maximize their air fleet’s utilization                        standard, local 3-tier distribution scenario (L1). Total emissions for
1185                                                                                                                                                                        1250
       and backhauling rates, flying the half case from San Francisco to                         3PL rail are 60% of those associated with trucking (D2), and only 8%
1186                                                                                                                                                                        1251

       Newark (via Memphis) results in over 25 kg of CO2 emitted. Table                         of those associated with airfreight (D3). Interestingly, the most
1187                                                                                                                                                                        1252
       11 shows that air transit is responsible for 98% of the scenario’s total                 emissions-intensive scenario of our study involves one with the
1188                                                                                                                                                                        1253
       emissions.                                                                               least amount distance traveled, that of the consumer driving to the

1189                                                                                                                                                                        1254
1190                                                                                            winery (L5a). In determining efficiency, the utilization of vehicles         1255
1191   4.9. D4: results for long distance 3PL fulfillment utilizing rail                         repeatedly dominates pure distance traveled.                                1256
1192                                                                                               Our results suggest that wineries should focus more on mini-             1257
1193      Were rail to become a viable option for 3PL providers, significant                     mizing the emissions from transportation instead of those from              1258
1194   emissions savings could be realized for long-haul land shipments.                        storage, which contribute very little, no doubt because cool, rather        1259

1195   Table 12 shows that total emissions for this scenario (D4) are 60% of                    than cold storage is required. Thus stockpiling larger inventory            1260
1196   those associated with 3PL trucking (D2). Although routing through                        buffers at echelons may be useful if it enables the intra-echelon           1261
1197   Chicago increases total distance traveled by over 800 km, the lower                      transit links to be more fully utilized. Our results are supported by       1262
1198   energy usage of rail results in about half as much emissions as the                      Van Hauwermeiren et al.’s [25] calculations that emissions from             1263
1199   cross-country trucking link. The outbound logistics for a South San                      transportation dominate those associated with storage and pro-              1264

1200   Francisco based carrier such as FedEx are more complicated, since                        cessing for most of the plant-derived foods they study. Supply              1265
1201   Oakland has the closest commercial rail terminal. These savings                          chains for foods that require more intensive cooling will have              1266
1202                                                                                                                                                                        1267
1203   Table 11                                                                                                                                                             1268
       Energy and emissions by link and echelon, scenario D3.
1204                                                                                                                                                                        1269
1205   D3: Long distance 3PL fulfillment via airfreight                                                      Distance/time             Energy – MJ         Carbon – kg CO2   1270
1206   WW > NVL                     Transport              Midsize truck, diesel                            29 km                       1.05               0.08             1271
1207   NVL > FDX-CA                 Transport              Midsize truck, diesel                            75 km                       2.71               0.20             1272
1208   FDX-CA > NWR                 Transport              Air-LongHaul, JetFuel                            4960 km                   362.59              25.64             1273
       NWR > CU                     Transport              Light truck, diesel                              21 km                       2.77               0.21
1209                                                                                                                                                                        1274
       WW                           Storage                None, electricity-US-Pacific                      0 days                      0                  0
1210   NVL                          Storage                Cooler, electricity-US-Pacific                    14 days                     0.75               0.04             1275
1211   FDX-CA                       Storage                None, electricity-US-Pacific                      0 days                      0                  0                1276
1212   NWR                          Storage                None, electricity-US-MidAtlantic                 0 days                      0                  0                1277
       CU                           Storage                None, electricity-US-MidAtlantic                 0 days                      0                  0
1213                                                                                                                                                                        1278
                                                                                                            Total                     369.88              26.17
1214                                                                                                                                                                        1279

        Please cite this article in press as: Cholette S, Venkat K, The energy and carbon intensity of wine distribution: A study of logistical options for..., J
        Clean Prod (2009), doi:10.1016/j.jclepro.2009.05.011
                                                                     ARTICLE IN PRESS                                            JCLP1977_proof Š 16 June 2009 Š 11/13

                                                            S. Cholette, K. Venkat / Journal of Cleaner Production xxx (2009) 1–13                                          11

1280   Table 12                                                                                   conversations with managers across the supply chain, we found                   1345
1281   Energy and emissions by link and echelon, scenario D4.                                     that very few were comfortable with estimating the utilization and              1346
1282   D4: long distance 3PL fulfillment via rail                                                  backhaul rates of inbound and/or outbound vehicles. As shown in                 1347
1283   WW > NVL       Transport   Midsize truck, diesel                29 km    1.05 0.08         Table 1, only the 3PL provider (New Vine Logistics) and the carrier             1348
1284   NVL > FDX-CA   Transport   Midsize truck, diesel, cooler        75 km    2.72 0.20         (FedEx) provide information about either utilization or backhaul                1349
1285   FDX-CA > OAK   Transport   Midsize truck, diesel, cooler        50 km    1.81 0.13         rates. While it seems safe to assume that negligible backhauling                1350
       OAK > NWK      Transport   Rail, diesel, cooler                 5500 km 20.62 1.53
1286                                                                                              occurs in most local transport links, our assumption of a 100%                  1351
       NWK > CU       Transport   Light truck, diesel, cooler          21 km    2.78 0.21
       WW             Storage     None, electricity-US-Pacific          0 days   0    0            utilization rate is, by definition, bound to be optimistic. To be                1352
1288   NVL            Storage     Cooler, electricity-US-Pacific        14 days  0.75 0.04         consistent, we assume the same high utilization rate holds for all              1353
1289   FDX-CA         Storage     Cooler, electricity-US-Pacific        0 days   0    0            commercial vehicles. Were utilization rates significantly lower, our             1354
1290   OAK            Storage     Cooler, electricity-US-Pacific        0 days   0    0            absolute per-unit emissions figures would increase. However, the                 1355
       NWK            Storage     Cooler, electricity-US-MidAtlantic   0 days   0    0
       CU             Storage     None, electricity-US-MidAtlantic     0 days   0    0
                                                                                                  relative ranking of the different scenarios would not be greatly                1356
1292                                                                                              affected, save for those that naturally lead to higher utilization for          1357

                                                                       Total   29.72 2.19
1293                                                                                              some links, such as scenarios that rely on retailer warehouses. As              1358
1294                                                                                              utilization rates drop, inserting a consolidation echelon would                 1359

1295   different results. For instance, Venkat and Wakeland’s [27] frozen                         likely improve overall supply chain efficiency and emissions’                    1360
1296   food system requires more energy to store products than to deliver                         profiles, even as mileage and lead-times increase.                               1361
1297   them and can be made more efficient by having less filled trucks                                 Table 1 also shows that we were unable to have a conversation               1362
1298   making more frequent deliveries, to reduce the overall amount of                           with a representative at the distributor tier. We thus assume that              1363
1299   inventory stockpiled at each echelon, and thus reduce usage of cold                        many of the characteristics of the retailer’s regional distribution             1364
1300   storage. Likewise, Van Hauwermeiren et al.’s [25] sample meat and                          center would apply to the distributor echelon. If, say, the distributor         1365

1301   dairy products result in comparatively more processing and storage                         contracts with a trucking company for different sizes of trucks for             1366
1302   emissions than transportation emissions.                                                   outbound distribution than we assume, our results would be                      1367
1303       The last 2 columns in Table 13 indicate the transport link in the                      compromised. Even more fundamental to our analysis is that Car-                 1368
1304   scenario that contributes the most to emissions. Not surprisingly the                      goScope requires certain assumptions that may not hold univer-                  1369
1305   cross-country transit link is responsible for the most emissions for                       sally, introducing some inflexibility into the modeling process. In              1370
1306   all the long-distance scenarios. For most local scenarios the step                         particular, CargoScope assumes consolidated transit modes (ocean,               1371
1307   linking the retailer to the consumer dominates. Even if the consumer                       rail and airfreight) have 100% utilization and 100% backhaul rates.             1372
1308   effectively loads a standard gas car to full utilization (L1b) nearly half                 Should significant inefficiencies exist for a particular case, Cargo-             1373
1309   the emissions result from this segment. Only by eliminating this link,                     Scope results would have to be adjusted manually.                               1374
1310   perhaps by having consumers walk, cycle or use efficient public                                 Finally, it should be noted that our analysis is based upon                 1375
1311   transit, would emissions approach those of the local 3PL scenario.                         a representative supply chain, but that no single winery’s supply               1376

1312                                                                                              chain is likely to conform exactly to our sample. Wineries in more              1377
1313   4.11. Caveats and limitations                                                              remote wine regions like Mendocino and Lake County will natu-                   1378
1314                                                                                              rally incur more transportation emissions than our Sonoma                       1379
1315       Researchers who have undertaken analyses similar to this one                           Winery, especially as these wineries may not be convenient to                   1380
1316   know that accurate and reliable data may not always be available                           delivery routes, such as the winery pickup service provided by New              1381

1317   for every input, requiring assumptions and estimates to be made. If                        Vine Logistics Retailers with RDCs more remotely located than Cost              1382
1318   these are inaccurate, results will be compromised. In our                                  Plus’s Richmond facility will likewise result in higher emissions. By           1383

1319                                                                                                                                                                              1384
1320                                                                                                                                                                              1385
1321                                                                                                                                                                              1386
       Table 13
1322                                                                                                                                                                              1387
       Ranked summary comparison of scenarios.
1323                                                                                                                                                                              1388
1324   Scenario                                              Local or distant       Energy – MJ             Emissions-kg CO2             Link with greatest   Link’s percent of   1389

                                                                                                                                         emissions            total emissions
1325                                                                                                                                                                              1390
       L4: local 3PL delivery                                Local                     5.83                   0.42                       NVL > FDX            48%
1326                                                                                                                                                                              1391
       L1a: local 3-tier, standard scenario, with            Local                     6.97                   0.50                       DW > RS              32%
1327     consumer using public transit or walking                                                                                                                                 1392
1328   L1b: local 3-tier, standard scenario, with            Local                    12.82                   0.91                       RS > CU              45%                 1393
1329     consumer fully loading the car                                                                                                                                           1394

1330   L5c: consolidation run. consumer                      Local                    20.67                   1.43                       n/a                                      1395
         utilizes 100% of CargoScope of
1331                                                                                                                                                                              1396
         midsized pickup
1332   L3b: winery self-distribution, via                    Local                    27.2                    1.89                       RS > CU              89%                 1397
1333     midsized truck                                                                                                                                                           1398
1334   L1: local 3-tier, standard                            Local                    31.19                   2.18                       RS > CU              77%                 1399
       D4: long distance 3PL fulfillment via rail             Distant                  29.72                   2.19                       OAK > NWK            70%
1335                                                                                                                                                                              1400
       L2: local 3-tier, with retailer warehouse             Local                    32.28                   2.25                       RS > CU              74%
1336   L3a: winery self-distribution, via light              Local                    37.27                   2.64                       RS > CU              64%                 1401
1337     truck                                                                                                                                                                    1402
1338   D1: 3-tier distribution, Long Distance                Distant                  48.61                   3.62                       DW-CA > DW-NJ        78%                 1403
1339   D2: long distance 3PL fulfillment via                  Distant                  49.39                   3.64                       FDX-CA > ESS         77%                 1404
1340                                                                                                                                                                              1405
       L5b: consumer drives a hybrid                         Local                  208.89                  14.47                        n/a
1341   D3: Long distance 3PL fulfillment via                  Distant                369.88                  26.17                        FDX-CA > NWR         98%                 1406
1342     airfreight                                                                                                                                                               1407
1343   L5a: consumer drives a regular car to                 Local                  487.41                  33.75                        n/a                                      1408
         the winery
1344                                                                                                                                                                              1409

        Please cite this article in press as: Cholette S, Venkat K, The energy and carbon intensity of wine distribution: A study of logistical options for..., J
        Clean Prod (2009), doi:10.1016/j.jclepro.2009.05.011
                                                              ARTICLE IN PRESS                                          JCLP1977_proof Š 16 June 2009 Š 12/13

       12                                              S. Cholette, K. Venkat / Journal of Cleaner Production xxx (2009) 1–13

1410   performing a detailed case study for a particular winery and its                      Lebel and Lorek [14] emphasize that a full life cycle assessment is               1475
1411   downstream network, we would be able to better estimate per-case                      often more appropriate than optimizing one single factor. In Point’s              1476
1412   emissions and address that winery’s specific concerns.                                 [19] detailed life cycle assessment of the Nova Scotia wine region,               1477
1413                                                                                         contributions to global warming are but one of the 8 environmental                1478
1414   5. Conclusions                                                                        factors examined. Even further, a fully complete CSR strategy for                 1479
1415                                                                                         a winery would encompass more than ecological concerns, as grape                  1480
1416       Wine is an image-focused, luxury product that generates strong                    harvesting typically makes heavy use of migrant laborers.                         1481
1417   emotional ties with consumers. Wineries are concerned with                                However, our isolated focus on logistics is justifiable because                1482
1418   attracting and retaining consumers and creating an image of                           most aspects of the distribution process are independent of the                   1483
1419   sustainability. With the recent and growing focus on reducing                         winery’s growing and operational processes. Thus, the delivery                    1484
1420   greenhouse gas emissions, wineries face increasing pressure to                        portion of supply chain can be evaluated separately by activities                 1485
1421   demonstrate their commitment to minimizing their ‘‘carbon foot-                       further upstream. We thus recommend wineries consider imple-                      1486
1422   print.’’ Many wineries are taking steps to reduce the energy usage                    menting this type of analysis as a part of their overall sustainability           1487

1423   associated with grape production and winery operations. For                           portfolio. Klassen and McLaughlin [12] show that companies often                  1488
1424   instance Cline Cellars, the winery that matches our model most                        benefit financially from improving their environmental perfor-                      1489

1425   closely, has solar panels on facility roofs to provide the majority of                mance, especially in industries that are already categorized as                   1490
1426   the winery’s power needs. Other wineries are actively attempting to                   environmentally friendly, as is the wine industry. Wineries could                 1491
1427   prevent soil erosion, reduce water usage and eliminate pesticide and                  reap rewards from well-considered efforts. Attempting to docu-                    1492
1428   herbicide usage. Some wineries even purchase credits to offset                        ment energy usage and carbon emissions using models such as                       1493
1429   carbon emissions, as reported by Penn [17]. We propose that eval-                     those presented here would be a positive first step.                               1494
1430   uating and redesigning the outbound supply chain will be consid-                          Speaking of first steps, we recognize that our model makes some                1495

1431   ered as additional tool, as wineries typically have many options for                  generalizations and assumptions that may not apply universally. We                1496
1432   downstream order fulfillment, and our results show that these                          plan next to undertake detailed case studies for specific wineries                 1497
1433   options can have very different energy and emissions’ profiles.                        and their logistical networks. Such studies would allow the partic-               1498
1434       Wineries should focus more on transit than storage, as the latter                 ipants to better understand their supply chains and their options for             1499
1435   contributes little to overall emissions. First, wineries can promote                  improving efficiency. Comparisons between participating wineries                   1500
1436   use of 3PLs for supporting direct-to-consumer sales, as this is very                  would provide a better understanding of the commonalities within                  1501
1437   efficient for local delivery and can be comparable to 3-tier distri-                   the wine industry and help us to better support generalizations                   1502
1438   bution for long distance fulfillment. For the latter, wineries should                  about obtainable emissions’ improvements.                                         1503
1439   encourage clients to select ground rather than airfreight delivery                        Additionally, our research to date assumes that supply chain                  1504
1440   and use 3PLs that provide temperature-controlled packaging to                         network decisions are made with existing products, facilities and                 1505
1441   guard against spoilage on these longer journeys. Although small                       equipment. We could extend our research to consider designing                     1506

1442   wineries are unlikely to have significant leverage with their supply                   a supply chain with equipment and placement of facilities selected to             1507
1443   chain partners, these wineries could favor supply chain partners                      minimize net energy usage and emissions. Evaluations of capital                   1508
1444   who use rail instead of trucks for long distance deliveries.                          investments for new or existing firms may explicitly address                       1509
1445       While it would be naive to advise wineries to discourage tasting                  sustainability issues in the future. These considerations would                   1510
1446   room visits, we recommend them to encourage wine club members                         become even more probable were the U.S. to adopt a cap and trade                  1511

1447   to receive additional purchases via package delivery services by                      emissions program similar to those found in the European Union. If                1512
1448   offering discounts on shipping. Another possibility would be for the                  so, emissions saved as a result of implementing a more efficient                   1513

1449   winery to coordinate round trip van transport from club members                       supply chain could then be credited to the winery. For instance, if the           1514
1450   from nearby cities for promotional winery events. Not only would                      winery were able to ship more wine through efficient third party                   1515
1451   such a service lessen the risk of inebriated drivers, but also it would               logistics providers in lieu of more energy-intensive delivery options             1516
1452   allow the winery to better approach the efficiencies realized by the                   or even redesign product packaging to be more compact and light-                  1517
1453   consolidation scenario L5c.                                                           weight, overall emissions reduction could be calculated and applied               1518
1454       The high carbon intensity associated with consumer driving is                     as a credit towards a winery’s emission budget. Modeling and                      1519

1455   troublesome from a policy perspective. This link is the least traceable               analyzing such strategies, supported by use of tools such as Cargo-               1520
1456   and also the one a winery has least control over. Through positive                    Scope will help in quantifying the costs and benefits of different                 1521
1457   informative campaigns, however, wineries could promote their                          supply chain options and will support management decisions.                       1522
1458   involvement in reducing carbon emissions and, at the same time,                           As a last word, we find that a winery can have an immediate and                1523
1459   nudge consumers to consider their own contributions. At the very                      effective impact on emissions, even within our present limited                    1524

1460   least, volume discounts would encourage consumers to purchase                         scope. Wineries should focus on finding opportunities to make                      1525
1461   more bottles at a time, leading to per-order emissions savings.                       transport use more efficient, rather than focusing only on pure                    1526
1462       Our results also show that no single supply chain configuration is                 distances. They can support more direct-to-consumer sales through                 1527
1463   ideal for all wineries. Larger wineries that sell sufficient quantities to             3PL providers and ask supply chain partners to support long                       1528
1464   California retailers, where a typical delivery would fill a midsized                   distance deliveries via rail rather than by truck and, most of all,               1529
1465   truck, should consider self-distribution. Otherwise, a winery should                  avoid airfreight. Likewise, wineries with sufficient volume can                    1530
1466   rely on 3-tier distribution rather than self-distributing smaller                     consider routing deliveries though fewer echelons. Lastly, when-                  1531
1467   volumes with light trucks or underutilizing midsized trucks. Simi-                    ever possible, wineries should encourage their customers to                       1532
1468   larly, if stores sell sufficient volumes to validate our assumption of                 consolidate purchases and otherwise minimize the highly emis-                     1533
1469   fully utilized delivery vehicles, there is little value in adding the                 sions-intensive last link in the supply chain.                                    1534
1470   echelon of the retailer warehouse.                                                                                                                                      1535
1471       As previously noted, the emissions associated with delivering                                                                                                       1536
1472   wine are a significant portion but still, only a portion of a winery’s                                                                                                   1537
1473   total carbon emissions. Likewise, carbon emissions are but one                         [1] Bay Area Economic Forum (BAEF). Supercenters and the transformation of the
1474   component of a company’s overall environmental performance, and                            Bay Area grocery industry: issues, trends, and impacts, Bay Area Economic    1539

        Please cite this article in press as: Cholette S, Venkat K, The energy and carbon intensity of wine distribution: A study of logistical options for..., J
        Clean Prod (2009), doi:10.1016/j.jclepro.2009.05.011
                                                                       ARTICLE IN PRESS                                              JCLP1977_proof Š 16 June 2009 Š 13/13

                                                                S. Cholette, K. Venkat / Journal of Cleaner Production xxx (2009) 1–13                                                13

1540          Forum Report Series. Available from:                         [17] Penn C. Parducci: the first winery in the U.S. to be carbon neutral. Wine Busi-   1571
              html#FREP; 2004 [accessed August 2007].                                                      ness Online. Available from
1541                                                                                                                                                                                        1572
        [2]                                                ¨
              Browne M, Rizet C, Anderson S, Allen J, Keıta B. Life cycle assessment in the                cfm?dataId¼47813; 2007 [accessed January 2008].
1542          supply chain: a review and case study. Transport Reviews 2005;25(6):761–82.             [18] Pirog R, Van Pelt T, Ensayan K, Cook E. Food, fuel and freeways: an Iowa         1573
1543    [3]   CleanMetrics. An energy and emissions analyzer for supply chains. Available from:            perspective on how far food travels, fuel usage, and greenhouse gas emissions.   1574
1544; 2007 [accessed August 2007].                      Ames, Iowa: Leopold Center for Sustainable Agriculture; 2001.                    1575
        [4]   Cholette S. A tale of two regions: a comparison of the French and Californian           [19] Point E. Life cycle environmental impacts of wine production and consump-
1545          wine sectors. International Journal of Wine Marketing 2004;15(2):24–48.                      tion in Nova Scotia, Master’s Thesis, Dalhousie University, Canada; 2008.        1576
1546    [5]   Cholette S. A novel problem for a vintage technique: matching wineries and                                               ´
                                                                                                      [20] Ruggieri L, Cadena E, Martınez-Blanco J, Gasol C, Rieradevall J, Gabarrell X,    1577
1547          distributors with mixed integer programming. Interfaces 2007;37(3):231–9.                    et al. Recovery of organic wastes in the Spanish wine industry: technical,       1578
        [6]                ¨
              Colman T, Paster P. Red, white and green – the cost of carbon in the global wine             economic and environmental analyses of the composting process. Journal of
1548          trade. American Association of Wine Economists. Available from:                    Cleaner Production 2009;. doi:10.1016/j.jclepro.2008.12.05.
1549; 2007 [accessed January 2008].                [21] Saunders C, Barber A. Comparative energy and greenhouse gas emissions of         1580
1550    [7]   Davis S, Diegel S. Transportation energy data book, 26e. Available from: www.                New Zealand’s and the United Kingdom’s dairy industry. Christchurch, New         1581
    ; 2007 [accessed January 2008].                             Zealand: Lincoln University; 2007 [No. 285].
1551    [8]   Energy Information Administration. Commercial buildings energy consump-                                    ¨
                                                                                                      [22] Seuring S, Muller M. From a literature review to a conceptual framework for
1552          tion survey. Available from:                     supply chain management. Journal of Cleaner Production 2008;16:1699–710.         1583

1553          warehouse.html; 1999 [accessed January 2008].                                           [23] Tsoulfas G, Pappis C. A model for supply chains environmental performance        1584
        [9]   Environmental Protection Agency. Emissions & generations resource inte-                      analysis and decision making. Journal of Cleaner Production 2008;16:1647–57.
1554                                                                                                                                                                                        1585

              grated database, v2.1. Available from:                  [24] Twede D, Clarke R, Tait J. Packaging postponement: a global packaging
1555          resources/egrde/index.html; 2007 [accessed January 2008].                                    strategy. Packaging Technology and Science 2000;13(1):105–15.                    1586
1556   [10]   Heller M, Keoleian G. Life cycle-based sustainability indicators for assessment         [25] Van Hauwermeiren A, Coene H, Engelen G, Mathijs E. Energy lifecycle inputs       1587
              of the U.S. food system. Ann Arbor, MI: Center for Sustainable Systems, School               in food systems: a comparison of local versus mainstream cases. Journal of
1557                                                                                                                                                                                        1588
              of Natural Resources and Environment; 2000 [CSS00-04].                                       Environmental Policy & Planning 2007;9(1):31–51.
1558   [11]   Hutchins M, Sutherland J. An exploration of measures of social sustainability           [26] Venkat K. Analyzing and optimizing the environmental performance of supply       1589
1559          to supply chain decisions. Journal of Cleaner Production 2008;16:1688–98.                    chains. In: Proceedings of the ACCEE Summer Study on Energy Efficiency in         1590
1560   [12]   Klassen R, McLaughlin C. The impact of environmental management on firm                       Industry. White Plains, New York: U.S. Department of Energy; 2007.               1591

              performance. Management Science 1996;42(8):1199–214.                                    [27] Venkat K, Wakeland W. Is lean necessarily green? In: Proceedings of the 50th
1561   [13]   Kleindorfer P, Singhal K, Van Wassenhove L. Sustainable operations manage-                   Annual Meeting of the ISSS. York, UK: International Society for the Systems      1592
1562          ment. Production and Operations Management 2005;14(4):482–92.                                Sciences; 2006.                                                                  1593
1563   [14]   Lebel L, Lorek S. Enabling sustainable production-consumption systems.                  [28] Wine Assure. Wine Assure test shipment: 7 day temperature profile. Available      1594
              Annual Review of Environmental Resources 2008;33:241–75.                                     from:; 2008 [accessed October 2008].
1564   [15]   Marchettini N, Panzieri M, Niccolucci V, Bastianoni S, Borsa S. Sustainability          [29] Wine Business Insider. Southern and Glazer’s form Southern/Glazer’s Distrib-
1565          indicators for environmental performance and sustainability assessment of                    utors of America. Available from:         1596
1566          the production of four fine Italian wines. International Journal of Sustainable               webarticle.cfm?dataId¼57915; 2008 [accessed October 2008].                       1597
              World Ecology 2003;10:275–82.                                                           [30] World Research Institute. Greenhouse gas protocol, a corporate accounting
1567   [16]   NHTS, 2001 Nationwide Household Transportation Survey. Available from                        and reporting standard. Available from:
1568; 2001 [accessed September 2007].                                               protocol-revised.pdf; 2004 [accessed June 2007].                                 1599
1569                                                                                                                                                                                        1600
1570                                                                                                                                                                                        1601
                                                                  E            CT

        Please cite this article in press as: Cholette S, Venkat K, The energy and carbon intensity of wine distribution: A study of logistical options for..., J
        Clean Prod (2009), doi:10.1016/j.jclepro.2009.05.011

Shared By: