Multi-product strategy to enhance the environmental profile of the canning industry towards circular economy

Multi-product strategy to enhance the environmental profile of the canning industry towards circular economy Antonio Cortés, Xavier Esteve-Llorens, Sara González-García, Maria Teresa Moreira, Gumersindo Feijoo Accepted manuscript How to cite: Science of the Total Environment, 791 (2021), 148249. https://doi.org/10.1016/j.scitotenv.2021.148249 Copyright information: © 2021 Elsevier B.V. This manuscript version is made available under the CC-BY-NC-ND 4.0 license (http://creativecommons.org/licenses/by-nc-nd/4.0/) 1 Multi-product strategy to enhance the environmental profile of the canning industry 1 towards circular economy 2 Antonio Cortés*, Xavier Esteve-Llorens, Sara González-García, Maria Teresa Moreira 3 and Gumersindo Feijoo 4 CRETUS. Department of Chemical Engineering, Universidade de Santiago de Compostela, 5 15705 Santiago de Compostela (Spain) 6 * Corresponding author 7 E-mail. antoniojose.cortes.m[email protected]; [email protected] 8 Abstract 9 The sustainable and continued production of enough food to feed the entire world´s population is 10 one of the main concerns in the food industry. Spain, and in particular Galicia, which is an 11 eminently fishing region characterised by the consumption of large quantities of fish, both fresh 12 and processed, must face the challenge of shifting its seafood productive fabric towards a circular 13 economy. To achieve this objective, the first task is to demonstrate that circular economy 14 principles allow to reduce the environmental impacts associated with seafood production. In this 15 sense, this study proposes the environmental evaluation of the skipjack tuna (Katsuwonus 16 pelamis) value chain within a canning industry located in Galicia through the LCA methodology 17 from an attributional perspective, including the valorisation processes for biowaste (edible and 18 inedible by-products). Results indicate that the main crucial subsystems of the value chain are 19 tuna fishing and the canning process, as it was expected considering other similar studies on 20 seafood products. Moreover, this specific case study demonstrates that the multi-product strategy 21 applied to the canning sector is environmentally viable. Thus, although the environmental impacts 22 of the entire system are increased by including further valorisation operations, the environmental 23 loads assigned to the main product (canned tuna) decrease compared to the one-product system 24 by assigning environmental burdens to other value-added products (tuna pâté, fishmeal, and fish 25 oil). 26 2 Keywords 27 Life Cycle Assessment; Canned tuna; Value chain; Valorisation; By-products 28 29 3 1. Introduction 30 As the world´s population has been expanded, the demand for food and energy has seen a rapid 31 increase. In fact, all projections indicate that at least a 70% increase in food production will be 32 needed to meet food demand by 2050 (FAO, 2012), which is expected based on increased yields 33 and productivity of crops, livestock and fisheries. Food production is recognized as a major 34 contributor to environmental impacts in both developed and developing countries (Nemecek et 35 al., 2016), amounting to around 13.7 billion metric tons of CO2 eq (Poore and Nemecek, 2018), 36 which represent 26% of global anthropogenic greenhouse gas (GHG) emissions (Parker et al., 37 2018). Delving deeper into the key drivers of this high environmental impact, in addition to the 38 intrinsic impacts of food production itself, other "avoidable" impacts play a major role, such as 39 the environmental burdens related to food packaging and distribution worldwide (Yokokawa et 40 al., 2018). In a global market, the consumption of some products presents a large impact when 41 considering the entire production chain from a life-cycle perspective (cradle-to-plate approach). 42 In particular, focusing on the fisheries sector, the situation in the oceans is agonizing, with fish 43 stocks being decimated over the years all over the world (Wilson et al., 2020). The state of the 44 oceans is becoming extremely worrying over time. In 2017, the maximum peak of overfished 45 marine stocks (34.2%) and a minimum of underfished stocks (6.2%) was reached, according to 46 the results published in FAO (2020). In parallel to the increasing rise of overfishing, aquaculture 47 continues to grow steadily, to the point that today fish produced in aquaculture facilities account 48 for 46% of total fish production (FAO, 2020). At this point, an intense debate has started to emerge 49 regarding the long-term sustainability of wild fisheries or whether aquaculture should be chosen 50 as the main fish source (Ruiz-Salmón et al., 2021). The valorisation of waste and discard fractions 51 for the production of fishmeal and fish oil to be used for the formulation of feed for farmed fish 52 also needs to be considered (Fréon et al., 2014c). There is growing evidence that the approach to 53 utilize such fractions: fish bones, viscera, heads and other less desirable parts as raw material for 54 the production of value-added products such as omega-3 acids and collagen, although these 55 alternatives are at a less developed stage for industrial implementation (Laso et al., 2018a). 56 4 In Spain, a country that has traditionally been an important fishing nation from the point of view 57 of catching, processing and consumption (Vázquez-Rowe et al., 2014), 922,564 tons of fish and 58 seafood were landed in 2018, making Spain the first country in the European Union, both in terms 59 of volume and value, with almost 2,150 million euros (European Commission, 2020a). Spain has 60 also developed an important seafood processing sector, especially smoked, processed and, above 61 all, canned seafood. Domestic canned tuna production leads EU production, accounting for 62 approximately 70% of the total volume (García-del-Hoyo et al., 2017). Specifically, the volume 63 of canned seafood production by Spanish companies reached 353,000 tons in 2018, being Galicia 64 (NW Spain) the leading region at national level, accounting for more than 85% of Spanish 65 production (EUMOFA, 2019). 66 Galicia´s canning tradition means that it is home to 7 of the 10 largest companies on the Spanish 67 canning sector, including the Top-5 (Ardán, 2018). Moreover, the presence of small and medium- 68 sized enterprises (SMEs) is predominant in the Galician canning sector, with a high percentage 69 of small companies (<50 employees), which represent 66% of the total, even highlighting that 70 22% of the total number of companies are very small with less than 10 employees (Ardán, 2018). 71 Numerous initiatives and projects for the development of circular economy strategies are already 72 underway in different companies. However, due to the aforementioned characteristics of the 73 Galician canning industry, it is difficult for these initiatives to permeate the market and in the 74 present context the Galician canning sector must face the challenge of the current paradigm shift 75 from a linear economy to a circular economy in which the main objective of companies must be 76 to maximize production through the valorisation of waste (Ciccullo et al., 2021). In this sense, 77 among the actions to be developed to achieve a complete integration of the circular economy 78 within the canning sector, the following stand out: (i) the use of processing techniques with a low 79 environmental impact; (ii) the reduction of the packaging residues; (iii) the valorisation of 80 wastewater flows; and (iv) the accomplishment of the objective of zero biological waste, 81 valorising all biowaste fractions to produce new value-added products. With this in mind, the 82 canning industry in general, and the Galician canning industry in particular, has enormous room 83 5 for improvement, since a large part of the fish is directly discarded (heads, viscera, bones, etc.), 84 which can open the door to the development of new products (García-Santiago et al., 2020). 85 Circular economy emerges as an opposite solution to the current linear system as a sustainable 86 system where economic growth is decoupled from resources use, through the reduction in the 87 consumption and the recirculation of raw materials (Korhonen et al., 2018). On the main points 88 of the circular economy is the reduction of waste generated throughout the value chain, valorising 89 them as raw materials for the generation of added-value products. However, increasing circularity 90 does not necessarily translate into a direct reduction of environmental impacts (Niero and Kalbar, 91 2019), which creates a dilemma for decision-makers when selecting adequate circular practices 92 and innovations (Rufí-Salís et al., 2021). The impacts or benefits generated by these circular 93 strategies are often measured through the use of circularity metrics (Corona et al., 2019). Life 94 Cycle Assessment (LCA) methodology, as it is based on the quantification of the inputs and 95 outputs of a system, becomes a good example of a circularity assessment tool to quantify and 96 evaluate the benefits or impacts of circular economy strategies. The application of LCA 97 methodology to determine the environmental impacts of fish catches, farming (aquaculture), and 98 processing started in the mid-2000s. A long list of LCA seafood studies on diverse pelagic species 99 such as horse mackerel (Vázquez-Rowe et al., 2010), Peruvian and Cantabrian anchovy (Fréon et 100 al., 2014b; Laso et al., 2018b), carp (Hornborg and Främberg, 2020) or Atlantic mackerel (Ramos 101 et al., 2011) have been reported. Demersal species such as hake (Avadí et al., 2018; Vázquez- 102 Rowe et al., 2011b), cod (Svanes et al., 2011; Ziegler et al., 2013) or octopus (Vázquez-Rowe et 103 al., 2012b), crustacean species such as prawns (Farmery et al., 2015; Medeiros et al., 2017), 104 lobster (Driscoll et al., 2015) or goose barnacle (Vázquez-Rowe et al., 2013a), and bivalve species 105 such as Atlantic scallop (Cortés et al., 2021) have been reported. Regarding aquaculture, different 106 studies on mussels (Iribarren et al., 2010b, 2010c; Lourguioui et al., 2017; Tamburini et al., 2020), 107 oysters (Tamburini et al., 2019), turbot (Iribarren et al., 2012) or salmon farming (Philis et al., 108 2021) can be highlighted. It is also important to mention that traditionally not only fishing or 109 farming activities have been evaluated, but also the production of different fish- and seafood- 110 6 based products such as fish sticks (Vázquez-Rowe et al., 2013b), fishmeal and fish oil (Fréon et 111 al., 2017) and canned products (Almeida et al., 2015; Avadí et al., 2015, 2014; Iribarren et al., 112 2010a; Laso et al., 2017; Vázquez-Rowe et al., 2014). Review articles on fishing (Avadí and 113 Fréon, 2013), aquaculture (Bohnes et al., 2019; Bohnes and Laurent, 2019; Philis et al., 2019) and 114 processing (Ruiz-Salmón et al., 2021; Vázquez-Rowe et al., 2012a) stages have also been 115 evaluated. 116 This study proposes the environmental evaluation of the skipjack tuna (Katsuwonus pelamis) 117 value chain within a canning industry located in Galicia through the LCA methodology. The 118 production line focuses on gourmet products, with high added value, basing its production on 119 local raw materials, with traditional manufacturing methods and using, as far as possible, certified 120 organic ingredients. Thus, the processing plant meets some of the circular economy principles: (i) 121 the fish is caught with traditional techniques in national fishing grounds; (ii) traditional techniques 122 such as cooking in seawater and air-drying are followed; (iii) the primary packaging is made of 123 aluminium, so it is 100% recyclable; (iv) high quality by-products (non-canned edible parts) are 124 used to produce other products; (v) the low-quality by-products are valorised in the form of 125 fishmeal that could be used for animal feed. In this way, the canning plant minimises the 126 consumption of raw materials, minimises transport and follows a multi-product strategy. The 127 main objectives of the study are to determine the environmental viability of this approach and to 128 lay the foundations for the way forward for other companies to position themselves in a highly 129 competitive market. The main novelty of this study lies in the fact that the LCA methodology has 130 been used to analyse the environmental impacts of the entire canned tuna value chain, and not 131 only those impacts assigned to the production of the main product. The inclusion of the 132 valorisation processes of residual organic fractions within the system boundaries makes it possible 133 to analyse the product from a broader point of view and opens the door to the identification and 134 evaluation of opportunities for environmental improvement. 135 2. Materials and methods 136 2.1. Defining the goal and scope. Impact assessment methodology. 137 7 Moving towards a veritable circular economy requires taking small steps to demonstrate the 138 viability of multi-product processes from an environmental point of view. In this sense, this study 139 aims to assess the environmental sustainability of the entire canned tuna value chain following 140 the Life Cycle Assessment methodology (ISO 14040; 14044) from a attributional perspective. 141 Although the main product is canned tuna, all stages of the value chain were included within the 142 system boundaries, including the manufacture of by-products and the valorisation of organic 143 waste. Thus, the main objective of this study is to determine from an environmental point of view 144 whether the production of multiple value-added products is more sustainable than single-product 145 approaches. 146 A cradle-to-gate approach was considered in the study, that is, considering the extraction of raw 147 materials to produce the required inputs and the manufacture of the products, but not the 148 consumption and final disposal stages. This perspective was assumed since the main objective of 149 the study is to recognize the environmental implications of the production of tuna-based products. 150 The main raw material is skipjack tuna, so fishing and transport to the canning plant, as well as 151 the production and transport of other ingredients and packaging materials, were included in the 152 system boundaries. The Functional Unit (FU) considered for assessment was 1 tonne of raw tuna 153 at processing plant gate since it seems consistent to select a feedstock-based FU as the plant is 154 characterized by its multi-product nature. The software SimaPro 9.0 (PRe-Consultants, 2017) was 155 used for the computational implementation of the inventories. The life cycle impact assessment 156 step was carried out using the ReCiPe 2016 v1.1 methodology in a hierarchist perspective at 157 midpoint level (Huijbregts et al., 2017). The environmental burdens were calculated in terms of 158 the following impact categories: Global Warming (GW), Stratospheric Ozone Depletion (SOD), 159 Terrestrial Acidification (TA), Freshwater Eutrophication (FE), Marine Eutrophication (ME), 160 Freshwater Ecotoxicity (FET), Marine Ecotoxicity (MET), Mineral Resources Scarcity (MRS) 161 and Fossil Resources Scarcity (FRS). 162 2.2. Description of the system under study. 163 8 The value chain associated with canned tuna was divided in 4 different subsystems, as depicted 164 in Figure 1. Subsystem 1 is related to the fishing and transportation of tuna as the main raw 165 material to supply the canning plant. Subsystems 2-4 are linked to the different activities and 166 operations that take place within the canning factory. It is important to note that the aim is to use 167 the residual fractions of the process to produce value-added products; however, the treatment of 168 non-recoverable waste and wastewater has been included in all subsystems. All liquid fractions 169 are directly sent to a municipal wastewater treatment plant located close to the site. On the other 170 hand, packaging waste is recycled as far as possible and landfilled or incinerated according to the 171 Spanish profile. 172 173 Figure 1. System boundaries for the environmental assessment of canned tuna value chain. 174 15 Rowe et al., 2011b), and bait fishing and processing (Vázquez-Rowe et al., 2014). As can be 305 observed in Figure 3.a, fuel production and consumption during fishing operations turned out to 306 be one of the main sources of environmental impact, accounting for more than 83% of the total 307 impact in the TA category, 78% in FRS and 65% in GW. This result was expected and is in line 308 with other fishing fleets revised, where direct and indirect fuel emissions were highlighted as the 309 most important carrier of GHG emissions (Avadí et al., 2018; Sandison et al., 2021; Vázquez- 310 Rowe et al., 2014; Villanueva-Rey et al., 2018), as well as other environmental impacts, such as 311 terrestrial acidification (Ziegler et al., 2016). In this article, a Fuel Use Intensity (FUI) of 548.9 312 L/tonne was obtained for pole-and-line tuna fishing. This result can be compared with different 313 results published in scientific literature; e.g. Miller et al. (2017) estimated the fuel consumption 314 of the Maldivian pole-and-line fleet among different shoal of tuna, varying from 200 L/tonne to 315 almost 600 L/tonne. 316 It is worth noting that special emphasis is placed on the fact that bait accounts for approximately 317 15-20% of the amount of fuel consumed in a fishing trip, which coincides with the results obtained 318 in this study, since the impacts associated with fishing and bait production reach 71.6% in ME 319 and do not fall below 16% in any impact category. Pole-and-line tuna fishing methods have not 320 been traditionally studied, so there are only a few studies that quantify the fuel consumption of 321 this type of fishing gear with different target species, however, tuna fishing by different fishing 322 gears has been extensively studied. In this context, Hospido and Tyedmers (2005) quantified the 323 fuel consumption of the Spanish tuna purse seine fishing fleet in the Indian (373 L/tonne), Atlantic 324 (442 L/tonne) and Pacific (442 L/tonne) oceans, setting the framework for the quantification of 325 FUI of other fisheries in the future. Moreover, Parker et al., (2015) achieved a significantly larger 326 sample of vessels to update this result, obtaining on average that the tuna purse seine fishery 327 consumes on average 365 L/tonne. 328 16 Specifically, the purse seine fishery for skipjack tuna in Atlantic waters requires a fuel 329 consumption of 445 L/tonne according to their results. On the other hand, Parker and Tyedmers 330 (2015) estimated a fuel consumption of 1,612 L/tonne for large pelagic (mainly tuna) fisheries 331 using longlines and other forms of pole-and-lines. From another, much more generic point of 332 view, Parker et al. (2018) calculated the global CO2 emissions linked to fuel combustion in fishing 333 vessels, estimating the FUI of pelagic fish (> 30 cm) at 430 L/tonne. In brief, most of the values 334 provided for tuna fishing, except for longlines, are within the same range, which can give the idea 335 that the values obtained in this study are close to reality. However, it is important to note that FUI 336 measurements from previous studies may vary considerably depending on both the measurement 337 method and the analysed fishing gear (Parker et al., 2015). 338 On the other side, the production and consumption of antifouling has been revealed as a 339 differential element in the ecotoxicity categories (74.5% in FET and 71% in MET) due to copper 340 and zinc emissions during the use stage, as demonstrated in previous literature (Avadí and Fréon, 341 2013). It is also noteworthy the 20% of total fishing impact in MRS category, due to the extraction 342 of bauxite to produce cooper needed for antifouling formulation. The bauxite ore is treated with 343 dilute sulphuric acid over a period of months, dissolving copper to form a weak solution of copper 344 sulphate, from which copper can recovered by electrolysis. Ice production contributes much less 345 to the environmental burdens of the system, its contribution remains almost constant in all 346 categories ranging from a minimum of 0.5% to a maximum of 2%, except in the FET category, 347 where it reaches a maximum of 7%, mainly due to electricity consumption and the high 348 dependence on coal in the Spanish electricity profile. Within the GW category, the presence of 349 R-410A stands out, which is a refrigerant gas widely used in refrigeration machines and whose 350 leakage influences the carbon footprint of the process, as has been shown in previous literature 351 (Vázquez-Rowe et al., 2011b). Finally, the environmental impact of paint consumption is 352 practically negligible, while the treatment of the waste (both solid and liquid) is only remarkable 353 in the ME category due to the treatment of bilge wastewater. 354 17 355 Figure 3. Relative contribution to environmental impacts associated with Subsystem 1. Tuna 356 fishery (a) and Subsystem 2. Canning process (b). 357 According to the results shown in Figure 3.b, the largest contribution to the environmental impact 358 in 6 of the 9 impact categories analysed comes from the production of aluminium primary 359 packaging, proving to be one of the most important elements in the production of canned seafood 360 18 (Almeida et al., 2014; Hospido et al., 2006). In fact, the production of aluminium cans is the main 361 contributor to the MRS category, where it reaches the highest contribution to any impact category 362 (95%), linked to the production of virgin aluminium, which requires a high consumption of 363 bauxite (almost 5 kg per kg of aluminium) (Laso et al., 2017). Heat production by natural gas (for 364 cooking and sterilisation) is also an important contributor to GW (36.2%) and FRS (43%), 365 although its contribution to the other categories is in all cases less than 5%. Wastewater treatment, 366 as in the fisheries subsystem (Figure 3.a) is only relevant in the ME category (50%) due to 367 nutrient emissions of the treated effluent, while in the other categories, it hardly exceeds 6%. The 368 production of secondary packaging (cardboard and plastic) presents a relative constant 369 contribution around 5-7%, reaching 12.2% in FE and dropping to 2.9% in MRS. Electricity 370 consumption is almost negligible, only relevant in the FE and TA categories (9.2% and 8.3%, 371 respectively), while in the rest of the categories it hardly exceeds 4%. Finally, the production of 372 the agricultural ingredients for the garnish and seasoning (onion, garlic, pepper, etc.) and the 373 transport of waste to specialised plants have a very low contribution in all impact categories, 374 always below 0.5%. The numerical results related to the FU can be found in the Table S.4 of the 375 supplementary material. 376 3.2. Effect of the allocation strategies on the environmental profile 377 This study has assessed the entire canned tuna value chain from an environmental point of view, 378 from the production of raw materials to the processing of products and co-products and the 379 treatment of waste, trying to move towards the target of 0 bio-waste. This study is an example of 380 system expansion to avoid the use of allocation strategies, as prioritised in the ISO standards. 381 However, when the objective of the study is to analyse the environmental impacts associated with 382 a particular product within the value chain, and full segregation of material and energy 383 consumption for each production line is not possible, the use of allocation factors to accurately 384 report environmental burdens seems unavoidable (Ayer et al., 2007). Traditionally, the selection 385 of allocation factors is one of the procedures that generates the least consensus among LCA 386 practitioners. In this regard, different authors have proposed different allocation methods for 387 19 different case studies. Within the seafood-specific case studies, for example, Thrane (2006) used 388 system expansion strategies to handle the co-product allocation in the LCA of different fish 389 products. On the other side, Ziegler et al. (2003) applied economic allocation factors to calculate 390 the environmental burdens of cod fillets, but also they applied mass allocation considering that 391 price fluctuations may condition the reliability of the results. On this basis and considering that 392 the main objective of the canning industry is the production of marketable products that generate 393 an economic income, the economic allocation was calculated for canned tuna (95.8%), tuna pâté 394 (3.3%), fishmeal (0.7%) and fish oil (0.2%). 395 When analysing the environmental profile of the main product (canned tuna), one different 396 approach from system expansion corresponds to the exclusion of subsystems SS3 and SS4 from 397 the system boundaries, including within the boundaries of SS2 the organic waste treatment 398 processes corresponding to the co-products (635 kg per 1,000 kg of raw tuna). However, this 399 approach would not be entirely realistic, as currently, the processing of fish co-products into 400 fishmeal and fish oil for feed formulation is a widely used option for the treatment of bio-waste. 401 In any case, Figure 4 shows the variation in the environmental profile of the original value chain 402 (Base scenario) compared to the scenario in which only tuna fishing and processing to produced 403 canned tuna is included, but the valorisation of organic waste from the processing is excluded 404 (Non-circular scenario). 405 406 0,8 0,85 0,9 0,95 1GW SOD TA FE MEFET MET MRS FRS Base scenario Non-circular scenario 20 Figure 4. Comparative environmental profile of the two alternative canned tuna production 407 scenarios. 408 It seems obvious that removing SS3 and SS4 with all associated material and energy consumption 409 from the system boundaries will reduce the environmental impact of the entire value chain. 410 Especially relevant is the reduction in the categories ME and FE, where the contribution of the 411 SS4 subsystem was higher due to the environmental burdens of wastewater treatment, as can be 412 seen in Figure 2. However, it is remarkable that the impact of the non-circular approach is almost 413 equal to that of the baseline scenario in the categories of MET, FET and TA. Even more 414 remarkable is that the impact on the SOD category is higher in the non-circular scenario, as the 415 emissions from bio-waste treatment are higher than the emissions from the inputs and outputs of 416 SS3 and SS4. 417 In order to relativise the impacts of the system towards the production of canned tuna, in the case 418 of the baseline scenario it is necessary to apply economic allocation factors, while for the non- 419 circular scenario, the entire environmental burden is allocated to canned tuna, as it is the main 420 output of the system. Considering that only 365 kg out of 1,000 kg are canned and that 90 g of 421 tuna and 30 g of additive are put into each can, the carbon footprint of a can of tuna was quantified 422 as 0.98 kg CO2 eq in the baseline scenario (considering the economic allocation of all co- 423 products), while in the non-circular scenario it reached 1.01 kg CO2 eq per can. To give a broader 424 value and not so focused on the carbon footprint, the total environmental impacts were calculated 425 in terms of the ReCiPe methodology endpoint, ranging between 0.050 and 0.052 pts per can in 426 the baseline scenario and in the non-circular scenario, respectively. 427 In this way, the comparison between the environmental profile of the system evaluated in Figure 428 1 (Base scenario) and this new scenario (Non-circular scenario) allow to answer the question: is 429 the application of multi-product strategies environmentally viable when the objective of the 430 assessment is to assess only one of the products, without considering the products avoided and 431 the environmental credits? In the specific case studied and taking into account that the inclusion 432 of SS3 and SS4 has not had much influence on the total impact of the system (less than 6% on 433 21 average), the distribution of the loads between the 4 different products has made it possible to 434 reduce the life cycle impact of canned tuna, while the environmental performance of the whole 435 value chain is hardly modified. It can be seen that a circular economy approach is feasible and 436 effective and these principles are in line with those mentioned by the European Commission in 437 the Circular Economy action plan (European Commission, 2020b). 438 3.3. Benchmarking with other canned seafood 439 When comparing with other results available in the peer-reviewed literature, it is important to 440 note that there are some issues that need to be addressed in detail in order to make realistic 441 comparisons (Avadí and Fréon, 2013; Vázquez-Rowe et al., 2012a): (i) The life cycle impact 442 methodology should be detailed since, although the categories of different methodologies may be 443 analogous, in many cases the units of measurement are different; (ii) The functional unit selected 444 for the analysis, as well as the edible content of each product, to adequately estimate the associated 445 environmental impact per unit and per quantity of product (e.g. 1 kg); and (iii) the environmental 446 indicator used for comparison (environmental footprint, normalised impact factor, etc.). Taking 447 all this information into account, Table 3 shows the detailed results of the comparison between 448 the carbon footprint of different canned seafood products. 449 450 22 Table 3. Carbon footprint values for other canned seafood products 451 Product Assessment method Product kg CO2/kg Reference Tuna ReCiPe 2016 90 g tuna 30 g additive 20 g can 5 g board 8.2 Present study Pilchard CML-IA Baseline 95 g pilchard 35 g olive oil 20 g can 7.5 (Almeida et al., 2015) Pilchard ReCiPe 2008 85 g pilchard 35 g olive oil 20 g can 25.2 (Vázquez-Rowe et al., 2014) Mussels IPCC 2016 129 g mussels 120 g sauce 81 g can 12.7 g board 17.5 (Iribarren et al., 2010a) Tuna ReCiPe 2008 n.d. 3.7 (Avadí et al., 2015) Peruvian anchovy ReCiPe 2008 n.d. 1.7 (Avadí et al., 2014) Cantabrian anchovy ESA with metrics from ICheme 2002 30 g anchovy 20 g olive oil 15 g can 5 g board 4.7 (Laso et al., 2017) 452 Is important to note that Table 3 compiles the results with a cradle-to-gate approach, considering 453 the impacts related to fishing, processing and packaging. Thus, in those studies that considered 454 the distribution and consumption stages (Almeida et al., 2015; Laso et al., 2017; Vázquez-Rowe 455 et al., 2014), the environmental burdens corresponding to these stages were not taken into account. 456 The carbon footprint of canned tuna presents an intermediate value, much lower than the values 457 presented by sardine in olive oil (Vázquez-Rowe et al., 2014) and mussels, mainly due to the 458 production of the primary packaging, as these two cases were packed in tinplate. However, this 459 value is very similar to the associated with the can of sardines with olive oil assessed in Almeida 460 et al. (2015), where similar packaging is used and similar techniques are followed. 461 Cantabrian anchovy is also packaged in aluminium and, despite having a much lower fish/package 462 ratio than that obtained in this study, they reported only 4.7 kg CO2/kg. This is because the 463 processing operations are very different, and the catch ratios of this fishery are much higher. In 464 this sense, the low results for Ecuadorian canned tuna (Avadí et al., 2015) and Peruvian anchovy 465 23 (Avadí et al., 2014) can be explained by lower fuel use in the Ecuadorian and Peruvian fisheries, 466 mainly due to a better catch per unit effort in relation to a higher abundance of the resource (Fréon 467 et al., 2014a). While it is true that in all cases it was concluded that both the fishing stage and the 468 production of primary packaging (tinplate or aluminium) are the main drivers of environmental 469 impacts and all improvement actions should focus on them. 470 4. Conclusions 471 This study demonstrates that, from a product approach, the inclusion of by-product valorisation 472 processes to address a multi-product strategy improves the environmental profile of the main 473 product. It is a clear example of a system expansion to avoid burden allocation between products 474 when the focus is on the assessment of the entire value chain. When the focus is on the assessment 475 of a single product, the allocation of environmental burdens seems unavoidable. 476 It has been shown that the fishing and primary processing stages are the most relevant sub-systems 477 within the environmental profile of the canned tuna value chain. The inventory of the fishing stage 478 showed, as previous studies on different fishing fleets, that the impacts of the fishing stage come 479 mainly from the production and consumption of diesel and antifouling. In this case, the 480 importance of the bait used for fishing also stands out, as it requires the fishing and processing of 481 sardine for use as bait. Primary packaging presented the highest environmental impact in the life 482 cycle impacts of canned tuna. Aluminium production, lamination and extrusion had the highest 483 impact in almost all impact categories, as expected for canned products. By-product valorisation 484 processes, both edible and inedible, have proven to have a low impact. 485 This system has allowed an approximation of the EU target towards a cradle-to-cradle system 486 approach, achieving the goal of zero biowaste. The results show the need to improve the 487 application of the circular economy in the primary sector, converting waste into raw materials for 488 the production of new products, minimising the consumption of material and energy resources. 489 In this sense, the application of multi-product strategies has been shown to improve the 490 environmental profile of canned products through the allocation of environmental burdens among 491 the new products; although further analysis from a sustainability point of view is required. Similar 492 24 studies need to be further applied to specific primary sub-sectors in the future to continue the path 493 towards a more sustainable and circular food system. 494 Acknowledgements 495 This research was supported by the EAPA_576/2018 NEPTUNUS project. S.G-G would like to 496 express her gratitude to the Spanish Ministry of Economy and Competitiveness for financial 497 support (Grant reference RYC-2014-14984). The authors belong to CRETUS and the Galician 498 Competitive Research Group GRC ED431C 2017/29 co-founded by Xunta de Galicia and 499 FEDER (EU). 500 References 501 Abdou, K., Gascuel, D., Aubin, J., Romdhane, M.S., Ben Rais Lasram, F., Le Loc’h, F., 2018. 502 Environmental life cycle assessment of seafood production: A case study of trawler 503 catches in Tunisia. Sci. Total Environ. 610–611, 298–307. 504 https://doi.org/10.1016/j.scitotenv.2017.08.067 505 Abdou, K., Le Loc’h, F., Gascuel, D., Romdhane, M.S., Aubin, J., Ben Rais Lasram, F., 2020. 506 Combining ecosystem indicators and life cycle assessment for environmental assessment 507 of demersal trawling in Tunisia. Int. J. Life Cycle Assess. 25, 105–119. 508 https://doi.org/10.1007/s11367-019-01651-5 509 Almeida, C., Vaz, S., Cabral, H., Ziegler, F., 2014. Environmental assessment of sardine 510 (Sardina pilchardus) purse seine fishery in Portugal with LCA methodology including 511 biological impact categories. Int. J. Life Cycle Assess. 19, 297–306. 512 https://doi.org/10.1007/s11367-013-0646-5 513 Almeida, C., Vaz, S., Ziegler, F., 2015. Environmental Life Cycle Assessment of a Canned 514 Sardine Product from Portugal. J. Ind. Ecol. 19, 607–617. 515 https://doi.org/10.1111/jiec.12219 516 Ardán, 2018. Informe económico y de competitividad 2018. Capítulo 11: El sector de la pesca 517 en Galicia. 518 Avadí, A., Adrien, R., Aramayo, V., Fréon, P., 2018. Environmental assessment of the Peruvian 519 industrial hake fishery with LCA. Int. J. Life Cycle Assess. 23, 1126–1140. 520 https://doi.org/10.1007/s11367-017-1364-1 521 Avadí, A., Bolaños, C., Sandoval, I., Ycaza, C., 2015. Life cycle assessment of Ecuadorian 522 processed tuna. Int. J. Life Cycle Assess. 20, 1415–1428. https://doi.org/10.1007/s11367- 523 015-0943-2 524 Avadí, A., Fréon, P., 2013. Life cycle assessment of fisheries: A review for fisheries scientists 525 and managers. Fish. Res. 143, 21–38. https://doi.org/10.1016/j.fishres.2013.01.006 526 Avadí, A., Fréon, P., Quispe, I., 2014. Environmental assessment of Peruvian anchoveta food 527 products: Is less refined better? Int. J. Life Cycle Assess. 19, 1276–1293. 528 https://doi.org/10.1007/s11367-014-0737-y 529 Ayer, N.W., Tyedmers, P.H., Pelletier, N.L., Sonesson, U., Scholz, A., 2007. Co-product 530 allocation in life cycle assessments of seafood production systems: Review of problems 531 and strategies. Int. J. Life Cycle Assess. 12, 480–487. 532 https://doi.org/10.1065/lca2006.11.284 533 Bohnes, F.A., Hauschild, M.Z., Schlundt, J., Laurent, A., 2019. Life cycle assessments of 534 aquaculture systems: a critical review of reported findings with recommendations for 535 policy and system development. Rev. Aquac. 11, 1061–1079. 536