The Impact of Alternative Packaging on the Life Cycle of Wine on Tap

33 Sustainability is a key objective of development policies within international organizations, and it is 34 also progressively gaining importance in the wine industry as a whole and, more specifically, in the 35 draught wine market. The competitive conditions of the wine sector and the evolution of consumption 36 styles have led to an increasing need for more accurate management strategies and analysis activities 37 to determine the performance of wineries. 38 This study aims to analyse both the environmental and the economic concerns of a commercial 39 development strategy implemented by an Italian winery that uses three packaging formats (glass 40 bottle, one-way PET keg, and reusable steel keg) in the sale of Falanghina PGI wine on three different 41 markets (domestic, Italy; regional, Germany; and international, USA). By assessing the 42 environmental and economic impact of the different formats on the three scenarios through LCA and 43 LCC analysis, it is revealed that the economic and environmental sustainability of packaging types 44 can vary significantly depending on the market destinations. In any case, the results show that PET, 45 and especially reusable materials such as steel, can lead to a marked reduction in impacts on the 46 market for tapped wine. 47 48 49


Introduction 65
Sustainability is undoubtedly a key objective of development policies within international 66 organizations. The European Union (EU), through its Europe 2020 Strategy, aims to promote smart, 67 sustainable, and inclusive growth. Sustainability has also gained importance in the wine sector and 68 has led to companies and consumers being more aware of this issue within the wine supply chain 69 [1,2]. 70 Consumer awareness in particular plays a central role in encouraging wine producers to pay close 71 attention not only to economic aspects but also to the environmental impact of wine at different stages 72 of its life cycle [3]. 73 About 258 million hl of wine were produced during the 2019 campaign, according to the International 74 Organisation of Vine and Wine (http://www.oiv.int). Moreover, competitive landscapes in the wine 75 sector and high fragmentation of consumer behaviour have led to the need for management planning 76 and tighter monitoring of costs. The precise estimation of the production cost of a litre of wine is an 77 essential basis for setting up the different processing steps and for developing appropriate marketing 78 strategies [4][5][6]. 79 The consumption of wine at entertainment venues such as restaurants and bars highlights the need to 80 find a compromise between environmental and economic costs, in order to ensure the consumer has 81 a pleasant, reasonably-priced and sustainably-valued consumption experience. To this end, the choice 82 of wine packaging can impact significantly on limiting environmental impacts and reducing costs. In 83 recent years, several alternative packaging options have been adopted in the beverage sector. In 84 addition to traditional glass bottles, wine is marketed to on-premise markets in large bag-in-box 85 containers, PET (polyethylene terephthalate) kegs, and steel kegs [7]. 86 Two distinct and contrasting aspects arise in the choice of packaging: on the one hand, retailers prefer 87 large-volume packaging due to its convenience; on the other, consumers prefer glass-bottled wine 88 due to environmental concerns about plastic pollution [8]. Another important aspect, as indicated by 89 several authors [9,10] is that consumers' purchasing decisions are influenced by the end-of-life of the 90 product rather than the environmental impacts in the production and transport phases. 91 With regard to on-premises markets, PET and steel kegs appear to be the most promising competitors 92 to glass. Both contain more volume for the same weight, and steel can be used multiple times, 93 theoretically with endless use. In addition, the distribution phase is more critical for glass, due to the 94 fragility of the material, which also has a major impact on secondary and tertiary packaging and on 95 the type of materials used (pallets, films, and carton boxes) [11]. 96 5 2. Methodology 129

Functional Unit and system boundaries 130
The volume of the beverage is typically chosen as the functional unit (FU) for LCA and LCC analyses 131 and, in particular, other studies that have focused on wine have defined their FU as 0.75 l or 1 l of 132 wine [27,28]. When analysing the consumption of wine on the premises, we chose a 125 ml glass as 133 FU, because that allowed us to compare Falanghina PGI wine marketed in different volume packages. 134 We considered the on-trade markets to carry out a cradle-to-grave environmental analysis, while the 135 economic analysis was conducted from cradle to wholesale. Cultivation of grapevine, winemaking, 136 packaging, transport, refrigeration and waste management were considered and the allocation method 137 by mass was used, considering that the wine yield of a unit mass of grapes is about 70%. 138 The decision to adopt two systems lies in the fact that there are limitations when estimating both the 139 transport costs from the wholesaler to the retailer and the product handling phase at the point of sale: 140 the wine storage and service phase by the retailer could not be calculated because of the high 141 variability due to the intrinsic characteristics of the shops, which results in very different costs (

Inventory analysis and impact assessment 148
The software tool SimaPro 8.5 (PRE Consultants, Amersfoort, The Netherlands) was used to perform 149 the LCA. 150 For vine growing, winemaking, packaging and transport, we obtained primary data from the winery; 151 for refrigeration and disposal, we used background data from the Ecoinvent v.3.7 database. 152 The environmental impacts of the three packaging techniques and the three markets were calculated 153 by adopting the IMPACT 2002+ method. 154

Grape cultivation 156
Grape cultivation was analysed from cradle to farm gate. We assumed that Falanghina PGI grapes 157 are grown with a conventional farming model. All input was provided by the farmers, and we 158 processed it considering the production cycle in the following phases: fertilisation, fungicide 159 treatments, pesticide treatments, pruning, inter-row management, irrigation, and harvesting. 160 It was assumed that the vineyard is in full production, and vineyard establishment and end-of-life 161 were excluded from the assessment as these stages represent minor impacts due to the long (and 162 uncertain) lifespan of the vineyard. 163

Winemaking 169
The vinification phase considers two steps: 170 Step 1 − Winemaking with all related operations (Table 2) 171 Step 2 − Filtration and finishing with addition of pre-packaging products (  Table 3 -Amount of all input in the second wine-making step (per 125 ml of wine). 177

Packaging 179
Three different wine packaging systems were considered: 0.75 l glass bottle, 20 l PET keg and 20 l 180 steel keg. 181 Both the glass bottle and the PET keg are one way, while the steel keg is recyclable; therefore, the 182 amount of steel per FU depends on the reference market scenario (Italy, Germany, or USA) and on 183 the lifetime of the kegs. The winery declared that the life cycle of steel kegs lasts about 10 years and 184 the number of roundtrips depends on the destination: 9 roundtrips/year for the Italian scenario, 5 185 roundtrips/year for the German scenario, and 2 roundtrips/year for the US scenario. 186 Considering the weight of the 20 l steel keg (6.4 kg, or 40 g FU -1 ), its lifespan and the number of 187 roundtrips, the right amount of steel FU -1 for each scenario is the following: 188  The wine wholesalers are located in Verona (for the Italian scenario), Frankfurt (for the German 205 scenario) and New York (for the US scenario) respectively, thus causing a different environmental 206 impact due to both the distance and the vehicles used during transport. The estimation of vehicle 207 emissions was carried out considering the average values of emissions from the use of Euro 4, Euro 208 5 and Euro 6 lorries [29], light commercial vehicles and, for the US scenario only, transoceanic ships. 209 -Italian and German scenarios: The distance between the winery and the wholesaler is 700 km 210 for the Italian scenario and 1,450 km for the German scenario; a lorry (32 tonnes) was 211 considered for the transport. The average distance from the wholesaler to the retailers was 212 assumed at 150 km for both scenarios, considering a light commercial vehicle. 213 -US scenario: For overland transport from the winery's headquarters to the port of Livorno 214 (Italy), 550 km were assumed with a 32-tonne truck. For transport from Italy to the wholesaler 215 located in the port of New York, a transoceanic ship with cooling was considered. From the 216 wholesaler to the retailer, a light commercial vehicle was considered, for an average distance 217 from wholesaler to retailer of 50 km. 218 219

Retailer refrigeration 220
During the refrigeration phase, the electricity consumption for the glass bottle scenario was assumed 221 to be 1.025E-03 kWh UF -1 in 12 hours of refrigeration (average time assumed before wine tapping). 222 For the PET keg and steel keg scenarios, the use of electricity is limited to the tapping phase (Table  223 7) and the refrigeration is managed using inert gases, leading to a refrigerant loss, which was also 224 considered. The leakage of R404A and its three components for UF are shown in Table 7; as also 225 reported by Amienyo and Azapagic [24], the Global Warming Power of R404A was estimated in 226 3.860 kg CO2 eq. kg -1 . 227

Scenario modelling 240
A key option for reducing the environmental impact of wine consumption is closely related to the 241 weight of packaging. In defining the alternative scenario, the potential reduction in life cycle 242 environmental impact was analysed by considering alternative packaging containers with a weight 243 reduction of 33% per glass bottle, PET keg and steel keg. 244 245

LCC methodology 246
Life cycle costs were evaluated according to [24,33] the methodological approach given in Hunkeler 247 [34] and Swarr [35] concerning the Conventional LCC calculation. 248 The following equation (Eq. 1) includes the phases and material useful to calculate the LCC of 125 249 ml of wine from field to wholesaler. 250 251 (1) 252

Where: 253
LCCw wine life cycle costs of 125 ml of wine 254

Cc costs of vine cultivation 255
Cp costs of packaging (glass bottle or PET keg or steel keg) 256 Cwpb costs of wine production and bottling 257 Ct costs of transport to wholesaler (Italy, Germany or United States) 258 259 All cost items are given per functional unit and reported in the unit of measure € 125ml -1 . 260 All costs for cultivation, packaging, wine production and bottling were collected directly from the 261 case study company. The LCC was conducted following an activity-based costing approach. In 262 addition, different cost separation criteria and cost centres were taken into account in order to 263 elaborate the balance sheet data set. 264 In analytical cost accounting, the most commonly used categories are direct and indirect costs [36]. 265 Therefore, the primary criterion for separating costs is based on the distinction between: 266 -Direct Costs, which are allocated directly to cost objects, based on an objective measurement of the 267 input consumed by the cost object; 268 -Indirect Costs, which are allocated or charged indirectly to the cost object because the amount of 269 the input consumed by the cost object in question has not been objectively measured. 270 The above categories have been broadly divided (as shown in Table 9) into direct and indirect costs.  The transport phase for the three scenarios is external to the company and was calculated through the 279 analysis of contracts with transport companies. 280 281

Results and discussion 282
As shown in Figure 2  As the cultivation and winemaking techniques are the same regardless of the type of packaging, their 296 environmental impact is equal for each scenario. In terms of resources used, the sum of their values 297 is the highest of all the phases considered (2.3 MJ primary energy FU -1 ), while in terms of GWP they 298 barely reach 0.06 kg CO2 eq. FU -1 . 299 In terms of resource consumption, the vineyard cultivation and winemaking phases remain among 300 the most impactful; however, in the case of scenarios involving the use of glass bottle packaging, the 301 latter becomes relevant in defining the overall impact. 302 The production steps that lead to a greater difference between the three scenarios considered are 303 packaging and transport for two reasons: 1) the amount and type of raw material used in packaging 304 and 2) the distance between cellar and retailer. The glass bottle is the most impactful packaging, 305 followed by the PET keg and the steel keg; the higher quantity of raw material used for the glass 306 bottles and the total weight of each batch led to a greater environmental impact. 307 Regarding the impact of transport, it is obviously linked to distance: the greater the distance, the 308 higher the environmental impact. Therefore, the "US scenario" has the highest values, followed by 309 the "German scenario" and the "Italian scenario".  Table 10 shows the results of the scenario analysis, highlighting the different impact of packaging 324 weight reduction in the market scenarios investigated. A significant change emerges with the use of 325 the glass bottle as the primary packaging container. In this case, the Climate Change indicator shows 326 a reduction in impact of as much as 1/3 for the commercial scenario on a domestic scale, clearly 327 evidencing the impact of this type of container on the product life cycle. Less sharp results were 328 obtained on a regional and international scale, but again there is evidence that a significant share of 329 the overall impact is attributable to the container. For the PET keg container, the reduction results, 330 although appreciable, are more limited, also considering the large volume transported per single unit. 331 The use of the steel keg shows no significant difference, considering the re-use of the container for 332 several trips. These results also express the relationship between packaging weight and distance 333 travelled to market, highlighting the strong environmental impact of packaging for short-marketed 334 products.   Climate change -30% -26% -15% -8% -7% -6% 0% 0% 0%

339
The cost analysis shows the high competitiveness of the steel keg format compared to the PET keg 340 and the glass bottle, due to the possibility to reuse the packaging and thus spread the purchase costs 341 over many trips. The least competitive scenario is where the glass bottle is used, mainly because of 342 the cost of buying glass. Considering the costs incurred by the winery to deliver the wine in the three 343 scenarios (Table 11), it can be seen that, on the domestic market, the most competitive format is the 344 steel keg; for the European destination, the choice of one of the two keg formats analysed does not 345 influence the total cost. For the US scenario, the most competitive format is the steel keg (-3.4% 346 compared to PET). 347 Looking at individual cost items, raw material (wine), category A is the item that alone accounts for 348 most of the costs in the PET and steel keg scenarios, while for the glass bottle, category B represents 349 the highest costs. 350 The packaging (category C) in steel keg accounts for 0.3% of production costs; this value rises to 351 3.5% for PET keg and 4.1% for the glass bottle. 352 Going into greater detail, the wine production and bottling phase (categories B and D) differ in the 353 use of the three types of packaging, due to manual labour in the bottling phase and in all the phases 354 prior to bottling, such as the cleaning of each container, the management of the bottling line, and the 355 subsequent activities of warehouse logistics. In particular, category B is higher for the glass bottle 356 because of the higher energy consumption of the various machines that constitute the line, compared 357 to the keg plant and the high incidence of the fixed costs of the plant. Conversely, category D is higher 358 for kegs because the incidence of the cost of personnel employed in the various operations is higher 359 than for other production lines. 360 Turning to transport costs, carriers define unit costs that depend on the kilometres travelled and the 361 type of material. PET packaging is the cheapest on all routes because it is the lightest in terms of 362 volume transported. Glass remains competitive on the domestic market, but not on the European and 363 US markets. For steel packaging, the return of the empty container is also considered in the costs 364 shown. 365 366 Packaging in the food industry has to consider various environmental and economic requirements in 369 addition to marketing, logistics, and production. As another study [37] points out, there are two 370 central elements to focus on when choosing the right packaging: the packaging material and the 371 packaging end-of-life. The packaging sector evolved initially because of the need to produce new 372 materials for technological reasons related to wine transport and preservation. Currently, the need to 373 find effective ways to reduce costs and environmental impact have led to new design paradigms [38]. 374 This study shows that wine steel and PET have comparable and significantly better economic 375 performance than glass packaging, with steel achieving the best environmental results. Similar 376 considerations were expressed by Brock and Williams [39] who found that glass and the recycled 377 glass bottle are still the most impactful packaging. Another study confirms the findings of this work 378 for beer [40], with glass containers appearing to be the most expensive compared to steel. Reusable 379 packaging systems therefore appear to be more competitive in the supply chain than single-use 380 packaging, as also demonstrated by Mahmoudi and Parviziomran [41]. 381 In these terms, it is difficult to find alternative solutions considering on the one hand the tradition of 382 using the glass bottle container and, on the other hand, the perception of the consumer. 383 Not all studies agree on the importance of wine packaging, but it seems that bottle design may play 384 an important role in some old-world markets that are more tied to tradition [42], but also in relation 385 to more innovative products, as for fruit wines that highlight the fundamental role of packaging in 386 defining the attractiveness of the product [43]. A recent study [44], indicates that Portuguese 387 consumers associate the heavier glass bottle with better quality and a higher price, while at the same 388 time expressing concerns about the presence of plastic in the packaging that may reduce recyclability 389 and reuse. 390 This condition is less evident for tap wine, but the cultural link with tradition can potentially influence 391 the choice. Nevertheless, the role of the consumer has been changing in recent years, and more and 392 more attention is being paid to environmental claims and to the communication of the role of limiting 393 impacts by wineries [45], which now consider their carbon neutrality and containment process as 394 development objectives in the medium and long term. 395 Moreover, in the last few years, experiments are being conducted to evaluate alternative packaging 396 such as bioplastic bottles, which would guarantee a reduced environmental footprint but would be 397 more expensive [46,47]. Compared to our case study, the use of PET kegs seems to be interesting 398 from an economic point of view; however, from a circular economy perspective and considering the 399 increasing awareness of consumers on the use of recyclable and reusable products compared to the 400 classical disposable ones, it seems inevitable for companies operating in the beverage sector to adopt 401 green strategies [48]. 402 In addition, when considering wine packaging, one must actually refer to three levels of packaging: 403 primary packaging, which includes the container intended for the end consumer and with the function 404 of protecting and advertising the product; secondary packaging, used to group bottles, for example in 405 cardboard boxes; tertiary packaging, such as containers used to combine groups of packages into 406 larger loads for transport [49]. The discourse, therefore, becomes broader and refers to many 407 materials, paper, cardboard, plastic, and wood in primis. These materials are also chosen by the 408 industry according to the form of distribution. 409 The transport of wine has emerged as one of the main causes of environmental impact both because 410 of direct emissions, mainly due to fuel consumption during logistics and product handling, but also 411 indirectly because it determines the choice of packaging materials, especially secondary and tertiary 412 packaging, and therefore requires more effective solutions. Other studies also confirm the results of 413 this research and emphasise the need to analyse the role of packaging in the agro-food system from a 414 holistic point of view considering its interaction with the logistics phase [50]. 415 Finally, focusing on the end-of-life results obtained by the different packaging systems, the glass 416 bottle generated the greatest environmental benefits, due to its efficient waste management system, 417 mainly based on recycling. However, its impact is greater than the other two systems, as reuse, in the 418 case of steel keg, seems to be a strong point for sustainability, as confirmed by other authors [51]. In 419 order to limit environmental impacts and costs, new packaging, such as bag-in-box and Tetra Pak 420 with integrated use of cardboard or paperboard layers, has entered the wine market in recent years, 421 with the dual aim of maximising the volume transported and containing costs, while at the same time 422 reducing environmental impact at the end of life. However, even these products are only partially 423 recyclable [52]. 424

Conclusions 425
This study lays the basis to support wineries, merchants, and retailers in their choice of wine 426 packaging, taking into account the different target markets. 427 This is the first study in the wine sector to consider the entire product life cycle, by assessing both 428 the dynamics and environmental impacts and costs with reference to all phases of the life cycle 429 (production, transformation, distribution, consumption, and end of life). In this way, it has been 430 possible to respond to a need of the industrial and logistics worlds that until now were not in a position 431 to highlight the cost and environmental impact hot spots of the various phases that characterise wine 432 consumption. We have been able to confirm that the glass bottle is still the most popular and 433 appreciated packaging among consumers, probably for sentimental reasons and links with tradition. 434 However, this container has obvious limits from the point of view of the circular economy, 435 considering the limited volume transported for the same weight of the container, compared to other 436 alternatives available on the market today. 437 Considering the above, companies are studying the possibility of using alternative packaging on the 438 on-premise market, given that the use of glass bottles requires skilled employees, high cost 439 technology, large space for storage and bottling equipment such as additional pack accessories: cork, 440 screwcap, or cardboard. In addition, as the scenario analysis also showed, the traditional packaging 441 consisting of the glass bottle makes a strong environmental contribution to the entire life cycle of the 442

wine. 443
In this respect, PET kegs prove to be particularly competitive, especially because of their limited 444 weight and considering that each keg carries the equivalent of more than 26 glass bottles; moreover, 445 wineries do not have to consider backhaul and handling charges and there is no need to store empty 446 containers. In addition to the obvious advantages for logistics and limited costs, the one-way use of 447 this container, coupled with not always guaranteed recyclability, introduces doubts about its use from 448 an environmental point of view. 449 The steel keg has interesting technological features, theoretically no end-of-life (unlimited use), and 450 cleaning, filling, and packaging technologies that are much easier to handle than the bottle crate, and 451 which are much less expensive. Furthermore, this container has a high material performance in terms 452 of wine shelf life and is also suitable for sparkling wines. 453 On the other hand, this packaging has return transport costs, administration (book-keeping) and 454 handling costs for the management of a keg, initial investment costs for the keg, and repair costs 455 (higher for long routes or constant circulation rate). Therefore, companies need a surplus of containers 456 throughout the year to manage seasonal fluctuations. Moving empty kegs over long routes increases 457 the environmental impact and transport costs, and the process of washing and sanitising kegs before 458 each use wastes water, energy, and chemicals. 459 Reusable packaging systems appear to be a viable alternative to replace single-use packaging in 460 supply chain systems. The decision-making processes of companies should therefore include an 461 analysis of the feasibility of using reusable packaging systems considering environmental and 462 economic factors. 463 The future of research could lie in new forms of packaging eco-design, using materials with low 464 environmental impact throughout the life cycle, aimed at improving container management in the 465 logistics system. Therefore, with a view to optimising the whole chain, both environmental and 466 economic factors should be considered organically through optimisation models applicable at cellar 467 level. Furthermore, for future studies, it will be useful to consider case studies related to larger volume 468 production, as the case examined refers to a production example of a medium-high range, low-volume 469 wine. Likewise, the research should also investigate other markets, including emerging ones. 470 Another aspect concerns the consumer's approach to wine from different containers, which often 471 favours glass. Consumer behaviour could be directed towards less impactful packaging with 472 appropriate information campaigns both on the quality aspects of wine − which does not vary in 473 containers made of different materials − and on the social commitment to reduce the impact of wine 474 on the climate. 475 Finally, a central role could be played by institutions at various levels, both central and local, which 476 could promote market-based schemes to reduce emissions based on taxes on environmental 477 externalities, to internalise society's costs for the use of impactful packaging, and to translate 478 environmental impacts into economic form. 479