Journal of Cleaner Production xxx (xxxx) xxx Contents lists available at ScienceDirect Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture Juan J. Cadillo-Benalcazar a, Mario Giampietro a, b, *, Sandra G.F. Bukkens a, Roger Strand c a Institut de Ci noma de Barcelona, 08193 Bellaterra (Barcelona), Spain encia i Tecnologia Ambientals, Universitat Auto b  Catalana de Recerca i Estudis Avançats (ICREA), Barcelona, Spain Institucio c Centre for the Study of the Sciences and the Humanities, University of Bergen, Bergen, Norway a r t i c l e i n f o a b s t r a c t Article history: The steady increase in production volume of salmon aquaculture has sharpened concerns about its Received 7 April 2019 sustainability. In particular the production of salmon feed is a reason for concern given its reliance on Received in revised form scarce natural resources, such as wild fish captures. Multi-scale integrated analysis is put forward as a 6 September 2019 tool to anticipate the environmental and socio-economic impacts of large-scale implementation of Accepted 6 November 2019 Available online xxx alternative salmon feeds, considering both plant and insect sources as potential replacements of fish meal and fish oil. The proposed accounting framework, based on relational analysis across hierarchical ^ as de Handling editor; Cecilia Maria Villas Bo levels, describes the patterns of required inputs using biophysical and economic variables. It also con- Almeida siders the inputs used by external systems for the production of imported feed, thus providing a coherent assessment of the sustainability of the production system in terms of feasibility, viability, and desirability. Keywords: The analytical tool-kit is illustrated in conceptual terms and then applied to the Norwegian salmon Salmon aquaculture aquaculture, both in diagnostic (describing the actual situation) and anticipatory mode (examining feed Alternative aquafeeds scenarios). Results are used in an exercise of quantitative story-telling to check the quality of the nar- Insects ratives currently shaping policy discussions on aquaculture. Quantitative story-telling is a heuristic MuSIASEM approach aimed at checking the robustness of knowledge claims in face of uncertainty. It is concluded Quantitative story-telling Norway that rearing insects in the salmon feed production chain enlarges the option space of feed sources by opening up the possibility of using locally-produced seaweed and organic waste, but also raises the level of uncertainty with regard to the possible insurgence of negative side effects. © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). 1. Introduction (FAO, 2018b). This upward trend is expected to continue, but at a reduced annual rate of about 3e5% (Olafsen et al., 2012). The combined increase in per capita fish consumption and hu- Sooner or later exponential growth on a finite planet is expected man population over the last four decades has led to an increase in to run into scaling problems (Bartlett, 2004; Steffen et al., 2015). overall fish consumption by 2.8 times (FAO, 2018b; The World Bank, Indeed, the environmental pressure generated by salmon aqua- 2018). This increase was possible due to an impressive growth in culture and the increase in price of fishmeal and fish oil are the volume of aquacultural production, from 4.7 million tons in increasingly challenging the sustainability of this economic sector 1980 to 80 million in 2016 (FAO, 2018a). This growth is expected to (Rana et al., 2009). This is a highly relevant issue for Norway continue in the future but at a lower rate. In Norway in particular, because of the perceived relevance of salmon aquaculture in the favorable environmental, technological and economic conditions country and the ambitions for further growth (DKNVS/NTVA, have led to an impressive increase in salmon production by 1200 2012). times over the last five decades (Hersoug, 2015), thereby making One of the main concerns of the producers in the aquaculture this country the first salmon producer and exporter in the world industry in Norway (and other countries alike) is how to guarantee a robust feed supply. Diversifying feed supply alternatives is considered a desirable strategy to deal with the increasing problem of wild fish stock reductions and market price variations (Rana ncia i Tecnologia Ambientals, Universitat * Corresponding author. Institut de Cie et al., 2009). Various solutions have been proposed, such as noma de Barcelona, 08193 Bellaterra (Barcelona), Spain. Auto E-mail address: mario.giampietro@uab.cat (M. Giampietro). replacing the components of marine origin (meal and oil) with https://doi.org/10.1016/j.jclepro.2019.119210 0959-6526/© 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210 2 J.J. Cadillo-Benalcazar et al. / Journal of Cleaner Production xxx (xxxx) xxx products of plant origin and/or products of insect origin (Rumpold production systems considered, section 3 presents the results, and and Schlüter, 2013; Shepherd et al., 2017). However, as Sprague section 4 the discussion and conclusions. et al. (2016) have pointed out, by replacing the marine compo- nents of the salmon feed with terrestrial products, the content of 2. Methodology omega-3 fatty acidsdnotably eicosapentaenoic acid [EPA; 20:5n-3] and docosahexaenoic acid [DHA; 22:6n-3]din salmon may be 2.1. Quantitative story-telling reduced. This could have an important negative impact on the in- dustry as salmon is recognized as a main source of omega-3 in the Quantitative Story-Telling (QST) is a novel approach to the use of human diet in some countries (Dickhoff, 2010). Neither humans nor quantitative information to inform policy. QST has the goal of salmon can synthesize omega-3 fatty acids and therefore must checking the quality of the pre-analytical choice of narratives used obtain these through their diet. In the case of farmed salmon, to justify a given policy. According to Felt (2007) there are three omega-3 fatty acids are mainly derived from the fish oil used in types of narratives relevant in science for governance: (i) justifi- feed. cation narratives e WHY we want to apply a policy; (ii) normative Small-scale experiments with insect (Belghit et al., 2018; Gasco narratives e WHAT should be achieved with the policy; and (iii) et al., 2018; Lock et al., 2016; Rumpold and Schlüter, 2013; St-Hilaire explanation narratives e HOW to achieve the expected results. et al., 2007) and plant feed alternatives (Bell et al., 2004; Boissy Most commonly, quantitative analysis is used in relation to expla- et al., 2011; Gasco et al., 2018; Shepherd et al., 2017) in aquacul- nation narratives to supply information about how to achieve a ture have shown promising results. In addition, the EU recently result. However, in this paper, the quantitative analysis is aimed at approved the use of processed animal protein derived from farmed checking the validity of the justification narrative (the “why”): Why insects in aquaculture, provided conditions for rearing insects are should Norway invest additional money and resources in increasing met (Commission Regulation (EU) 2017/893 of May 24, 2017). One the actual level of salmon production? of these conditions concerns limitations on the substrate on which QST has first been used in the EU-funded Horizon2020 project insects are fed. MAGIC (MAGIC, 2019) as an alternative to providing “scientific However, alternative feed solutions may have unexpected side evidence” to inform sustainability policies. Applications have been effects when implemented and scaled up to industrial production. provided in relation to agricultural development (Giampietro, In general, it is difficult to anticipate what can go wrong until it is 2018), desalination (Serrano-Tovar et al., 2019) and alternative en- too late and the technology has become entrenched. This is ergy sources (Renner and Giampietro, 2020). Rather than crunching commonly known as the Collingridge dilemma (Collingridge, 1980). numbers to identify an ostensibly optimal course of action or pre- Thus, what is dearly needed is an overall analysis of the sustain- dict future states of the system, QST uses quantification to check the ability of the Norwegian salmon industry that is capable of antici- robustness of the story-telling associated with a given policy pating the potential problems arising from the continuous increase (Matthews et al., 2017; Saltelli and Giampietro, 2017, 2016). in production volume and innovations in this sector. In this particular study, QST is applied, in combination with This study is part of the AQUAFLY project, which explores the multi-scale integrated analysis of societal and ecosystem meta- option of using insect feeds in salmon aquaculture, and contributes bolism (see next section), to prime a reflection on the future of to a package of methods from ethics and environmental science, salmon production in Norway. A core set of quantitative results is informed by theory from Science and Technology Studies, for generated by considering relevant characteristics of the system integration into a STEM research project in a way that contributes simultaneously across different scales and dimensions of analysis, to Responsible Research and Innovation (RRI). It aims to illustrate a thus recognizing the existence of the legitimate and non-equivalent novel approach to explore potential advantages and problems of perceptions of performance of different relevant actors. QST starts changes in the salmon production process that may occur outside of from a first appraisal of the relevance of information referring to a the production system itself. The novel approach consists in the use general analysis of various aspects of the system. Then the various of a semantically open accounting framework (MuSIASEM) to insights are used to prime a reflection on the existence of synergies analyze narratives and support an informed stakeholder discussion and trade-offs among policy concerns. through quantitative story-telling (QST). The specific objective of the article is to illustrate the novel 2.2. Multi-scale integrated analysis of societal and ecosystem approach for the case of Norwegian salmon aquaculture perfor- metabolism mance through: Multi-scale integrated analysis of societal and ecosystem i. an evaluation of the current pressure of the Norwegian pro- metabolism (MuSIASEM) is a relatively novel approach to explore duction system on the local environment and other (external) the complex nature of the interactions between socio-economic social-ecological systems (externalized pressure due to im- systems and ecological systems with regard to sustainability ported feeds and labour); (Giampietro et al., 2014, 2013, 2012). MuSIASEM adopts the concept ii. an assessment of the potential changes in the environmental of metabolism of social-ecological systems: the socioeconomic and performance in response to changes in the salmon feed supply ecological system are considered as two components of a larger system. complex (the social-ecological system) that interact across different hierarchical levels of organization (Giampietro et al., 2014). This The illustration of the methodology is based on the feed pro- interaction involves a network of exchanges of matter and energy duction alternatives studied in AQUAFLY. Of special interest to the that provides the required conditions for the social-ecological project was the option of incorporating farmed insect ingredients system to express and reproduce its functions. The expression of in the salmon feed mix, using under-used marine resources not these exchanges over a given set of structural and functional ele- suitable for direct human consumption (kelp) as substrate for ments is what is called ‘the metabolic pattern’ (Giampietro et al., rearing the insects. Kelp is rich in omega-3 fatty acids and a readily 2012). available local resource in Norway. In MuSIASEM, the representation of the metabolic pattern of the The paper is structured as follows: section 2 presents the social-ecological system relies on relational analysis (Louie, 2013, methodological approach and describes the alternative feed 2009; Rosen, 1985, 1958) and the flow-fund model of Georgescu- Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210 J.J. Cadillo-Benalcazar et al. / Journal of Cleaner Production xxx (xxxx) xxx 3 Roegen (1971). The first step of the approach involves the identi- to the sustainability of aquaculture. An integrated analysis of fication of the hierarchical structure of the social-ecological system aquaculture must provide sufficient diversity of data so as to allow a in question (e.g., Norwegian salmon aquaculture) to describe the characterizing of the performance of the investigated system ac- functional relations among its constituent metabolic elements. cording to the legitimate but diverging interests of the different Following, the specific metabolic pattern of each one of the relevant social actors involved. For example, the information relevant for a elements is characterized using the concept of processor. The pro- Norwegian coastal community will be different from that needed cessors establish a relation among quantitative information refer- by a general manager in the salmon producing industry or by the ring to different hierarchical levels and different dimensions of national environmental protection agency. In the same way, the analysis by consistently describing the profile of inputs and outputs data required by decision makers in salmon feed exporting coun- associated with the expression of any given function or process in a tries will be different from that required by the feed importing common format (data array). Indeed, any metabolic element of the countries. Hence the challenge here is how to handle the diversity social-ecological system is considered an open system in itself with of data referring to different levels of analysis (a specific aquacul- an expected ‘behavior’ linked to the overall metabolic pattern. This ture system, a specific community, the whole country, the global behavior is described in terms of: (i) consumption of a profile of level) and different dimensions of analysis (monetary versus bio- inputs; (ii) expression of a profile of useful output; (iii) generation physical flows). of unwanted products (emissions/waste) (Gonza lez-Lo  pez and According to the steps mentioned above, Fig. 1 identifies the Giampietro, 2017). metabolic elements and the boundaries that will allow to establish MuSIASEM is a semantically-open approach. In its pre-analytical the functional and structural relationships of the salmon aquacul- framing, a series of decisions must be made consulting the stake- ture system in Norway. Note that the inputs used by this system are holders. First, it must be decided which boundaries and which obtained either from the technosphere (controlled by humans) and/ metabolic elements are relevant for the definition of the hierar- or from the biosphere (determined by and affecting the activity of chical structure of functional relations. Second, formal categories ecosystems). Furthermore, required inputs are obtained either from (e.g., energy, water, labour requirements) must be selected for a within the chosen boundary of the system or from outside the meaningful characterizationdusing the concept of processorsd of system (imports). Inputs from outside imply externalization (or the metabolic system under study. Finally, in the last step, it must outsourcing) in terms of impacts on other (external) ecosystems be decided how the quantitative information generated by the (processes taking place in the biosphere outside the system bor- processors is organized into a coherent assessment. The latter step ders) and other socio-economic systems (processes taking place in depends on the choice of logic and the purpose of the analysis. the technosphere outside the system borders). The same is true for Thus, the rich information space produced by MuSIASEM may the outputs produced. This distinction allows a study of the level of result in non-equivalent assessments (biophysical feasibility, eco- openness of the system and opens the discussion to the ‘desir- nomic viability, technical viability) tailored to the specific ability’ of exporting environmental problems. (research) questions or interests of specific stakeholders or story- Indeed, one important question in regard to the sustainability of tellers. These pre-analytical choices will therefore determine the Norwegian salmon aquaculture is how to relate information on relevance of the resulting information for different story-tellers. salmon output (system size) and feed quality to required inputs from both local and external sources. The various pieces of infor- mation needed to define this relation are shown in Fig. 2 and 2.3. Tailoring MuSIASEM to aquaculture production systems include: MuSIASEM has been applied to the societal energy metabolism P  The total quantity of salmon (to be) produced ( ). and agricultural production, but to the best of our knowledge never Fig. 1. Schematic overview of the relations among functional metabolic elements in salmon aquaculture systems. Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210 4 J.J. Cadillo-Benalcazar et al. / Journal of Cleaner Production xxx (xxxx) xxx Fig. 2. Identification of five holons in the set of relations of the salmon aquaculture system: salmon production; salmon feed processing; plant ingredient production (‘plant’), marine ingredient production (‘fish’); insect ingredient production (‘insect’).  The processors of salmon production, salmon processing and 1e3 represent the mapping of the characteristics of specific lower- feed processing (referring to a higher level of analysis: level level processes of production of feed inputs (Plants, Fish and In- nþ1). This information is generally readily available as it is sect), holon 4 represents the functional and structural elements routinely used in production management; entailment to the production of the feed and holon 5 the elements  The mix of products (plant-based feed, fish-based feed and related to the direct production of salmon. insect-based feed) used to feed the salmon (and expected to In turn, the description of holons allows identifying the pro- meet salmon dietary requirements). This particular level of the cessors, which are the units that describe the metabolic elements of hierarchy of relations is designed the level n of our analysis. the system in terms of fund and flow. Where, funds are those that  The unitary processors for obtaining the various ingredients that maintain their characteristics throughout the duration of the make up the mix of products used for the production of the analytical representation. While, flows describe what the system salmon feed (refers to the lowest hierarchy: level n-1 and level does and change their identity over the duration of the analytical n-2) representation, namely, they enter but do not exit or vice versa (Georgescu-Roegen, 1971). In general terms, Fig. 3 shows the Regarding the formalization of the characterization, MuSIASEM interaction of these elements with the processor. According to the adopts from the theories of complexity and hierarchy (Ahl and characteristics of the salmon aquaculture system, any processor is Allen, 1996; Allen and Starr, 1982; Pattee, 1973; Salthe, 1985) the considered within the technosphere. However, the interacting el- notion of holon (Koestler, 1978, 1968). The usefulness of this notion ements may be implied to the technosphere or the biosphere. In is to emphasize the epistemological problems inherent in the this analysis, fund elements are: land (ha), sea area (ha), labour elusive relation between structural and functional elements in the (hours). While, flow elements are: fertilizers (t), blue water (m3), representation of complex systems organized across different hi- green water (m3), pesticides as active ingredient (t), energy (J) and erarchical levels (for more details see Giampietro, 2018; Giampietro money (euros). et al., 2006). In our analysis five holons are distinguished: holons Note that the relation between the profile of inputs/outputs and Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210 J.J. Cadillo-Benalcazar et al. / Journal of Cleaner Production xxx (xxxx) xxx 5 Fig. 3. The inputs and outputs of flows and funds described by processors. the resulting output can be described either with intensive vari- based on a sequential pathway (material entailment). In other ables (unitary processor: expressed per unit of output) or (ii) words, when a series of processors is operating in the same extensive variables (extensive processor: expressed in absolute sequential pathway of a given production process, its overall amounts) (Di Felice et al., 2019). This dual system of accounting is characteristics can be represented as just one single processor illustrated on the right side of Fig. 3. (Fig. 4, upper graph) by summing the various homologous inputs in Quantitative representations based on processors can be the corresponding data arrays (e.g., summing kg of nitrogen fer- aggregated within and across hierarchical levels of analysis (Fig. 4). tilizer to kg of nitrogen fertilizer, hours of labour to hours of labour, Within the same hierarchical level of analysis, the aggregation is etc.). Across two different levels of analysis, the aggregation follows Fig. 4. Methods of aggregation of information carried by processors: Sequential pathway or material entailment (upper graph) and parallel composition (lower graph). Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210 6 J.J. Cadillo-Benalcazar et al. / Journal of Cleaner Production xxx (xxxx) xxx the logic of ‘parallel composition’ (Fig. 4, lower graph). In this case, Quantitative representations based on relations over processors relevant data for establishing a bridge across levels are: (i) the can be tailored to describe a specific existent situation (diagnostic profile of inputs and outputs of the individual sequential pathways mode) or scenarios (anticipation mode). Scenarios may consist of making up the functional units (defined at lower levels) and (ii) the changes in the existing feed supply systems (e.g., change in relative relative contribution of these functional units in the overall salmon contribution of marine ingredients; change in technological co- feed mix. In this way, we can move one level up in the hierarchy. efficients within a feed supply system) or the introduction of new For example, in Fig. 2, the mapping of the structural elements feed supply systems (holon 3 in Fig. 2). (the specific individual production processes) onto the corre- sponding functional element (the supply system) is done with the parallel composition for holons 1 and 2, and the sequential 2.4. Assessment of external inputs for the current salmon pathway for holon 3. The processor for the salmon ‘feed supply production system system’ (holon 4) is calculated from the processors of the functional elements of holons 1e3 following the functional pathway. Finally, The information about the current feed mix (plant-based feed, an overall processor, assessing the overall profile of input re- fish-based feed and insect-based feed) is represented by the vector quirements and outputs for the salmon production system as a vA1 (Fig. 5). This vector defines the relative contribution (in %) of whole (holon 5, made up by ‘feed production’, ‘salmon production’ the different feed inputs per unit of salmon feed produced: [vA11 e and ‘salmon processing’) is obtained following the sequential plant-based feed; vA12 e fish-based feed; vA13 e insect-based feed; pathway. In this way, the overall requirement of internal and and vA14 e additives (Premix)]. external inputs of the salmon production system can be tracked. Moving to a lower level of analysis, subsystems can be defined Thus, given the network of relations established, it is possible to for each element of vA1. These subsystems are represented by other scale-up (establishing a causal relation) the effects of changes in the vectors that describe their relative composition (see Fig. 5): feed supply systems in relation to the desired amount of salmon production. Implementation of this scaling requires knowledge of  vB1 e is the vector describing the relative contribution of crop the technical characteristics of the specific systems of production types in the plant-feed supply system: [vB11 e cereals; vB12 e (the structural elements) adopted in the various supply systems. oil-bearing; vB13 e pulses] Fig. 5. The set of relations over the various pieces of information required to assess the external inputs of salmon production. Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210 J.J. Cadillo-Benalcazar et al. / Journal of Cleaner Production xxx (xxxx) xxx 7  vB2 e is the vector describing the relative contribution of fish 2.5. Anticipatory mode types in the fish-feed supply system: [vB21 e demersal fish; vB22 e pelagic fish] Due to the modular features that were described in the pre-  vB3 e is the vector describing the profile of the different per- vious section, MuSIASEM facilitates the generation of scenarios at centages of feed inputs in the insect-feed supply system. In Fig. 8 any level. For example, exploring the effects generated by: an this vector has only one typology of feed-input and its value is increase in salmon production, changes in the relative contribu- zero because this figure corresponds to the diagnostic analysis tion of the feed supply system, or both. In this case, the second and this component is not yet used in the elaboration of salmon option will be evaluated. It is assumed that the necessary amount feed. It will be used for the anticipatory mode. of feed of the different combinations is the same, although this really is not the case, because the quantity can vary according to At the lowest level in Fig. 5 we have the specification of the the nutritional contributions that each ingredient contributes (see unitary processors related to production and processing (level n-2) section 2.5.3). However, this analysis goes beyond the scope of the that make up each of the subsystems defined at level n-1. For objective of this article. Fig. 9 shows the actual mix employed example: the element vB11 (cereals) is defined by the unitary pro- (described by vector vA1), predominant use of marine products cessors related to the production and processing of maize and (vA2), a total reliance on plant-based ingredients (vA3), and two wheat: [vC11 e Maize; vC12 e Wheat]. In Figs. 6e8 vectors at level n- scenarios with varying amounts of insect ingredients (vA4 and 2 are shown in detail. vA5). These combinations reflect the objectives of the Aquafly Using this set of relations in diagnostic mode, the profile of project (for more details, see Tables A1, A2 and A3 in Supple- external inputs required for salmon production was assessed based mentary data). on the information about the characteristics of existing production processes. Fig. 6. Unitary processors of feed-ingredient supply at level (n-2) in the plant-based supply system. Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210 8 J.J. Cadillo-Benalcazar et al. / Journal of Cleaner Production xxx (xxxx) xxx Fig. 7. Unitary processors of feed-ingredient supply at level (n-2) in the fish-based supply system. 2.6. Data sources 2015). For processing, the inputs and outputs are from (Schmidt, 2007), the labour data was considered the same as for 2.6.1. Feed production maize processing. For agricultural production: ⁃ Sunflower was assumed to be produced in Ukraine. The data of land, fertilizer, energy and pesticide were estimated from ⁃ Faba beans and pea: It was assumed that both were produced in (Wilfart et al., 2016) and labour input was assumed to be 32.8 h/t France. The inputs required for the crop production were esti- (UAB, 2018). For processing, due to the lack of information mated from the following sources: land, pesticide, fertilizers and Denmark data was used. The inputs and outputs are from energy were estimated from (Wilfart et al., 2016) and labour from (Schmidt, 2007), the labour data was considered the same as for (UAB, 2018). For processing, due to the lack of information, wheat maize processing. processing data was used as suggested by (Hognes et al., 2014). ⁃ Soybean: production data from Brazil were used. The data of ⁃ Maize was assumed to be produced in the United States of land, energy and fertilizer were estimated from (Wilfart et al., America. The inputs required for the crop production were 2016) and labour from (Ortega, 2003). For processing, the in- estimated from the following sources: land, labour, pesticide puts and outputs are from (Schmidt, 2007), the labour data was and energy from (Pimentel and Pimentel, 2008) and fertilizer considered the same as for the maize processing. from (UAB, 2018). For processing, the blue water, energy and ⁃ Wheat: It was assumed that it was produced in Germany. The outputs were estimated from (Buratti et al., 2008; Galitsky et al., data of land, pesticide and fertilizer were obtained from (UAB, 2003). Due to lack of data on the labour required, an estimate 2018), blue and green water from (Mekonnen and Hoekstra, based on European data was used (European Comission, 2016; 2010), energy from (Visser et al., 2012) and pesticide from Logatcheva and Galen, 2015). (Marinussen et al., 2012). For the elaboration of salmon feed, ⁃ Rapeseed was assumed to be produced in Germany. The inputs two by products which come from different processes are used, required for the crop production were estimated from the dry gluten and wheat starch. Due to the lack of information, following sources: land, labour and fertilizer were obtained Belgium data was used. Moreover, it was assumed that both from (UAB, 2018); pesticide and energy from (Queiro s et al., processes had the same inputs (Buchspies and Kaltschmitt, Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210 J.J. Cadillo-Benalcazar et al. / Journal of Cleaner Production xxx (xxxx) xxx 9 Fig. 8. Unitary processors of feed-ingredient supply at level (n-2) in the insect-based supply system. 2016). The proportion of outputs for the wheat starch process et al., 2005; SSB, 2018a). The procedure is presented in the was estimated from (FAO, 2017). The labour data was considered supplement. For processing, the data are from (Winther et al., the same as for the maize processing. 2009), The procedure for calculating the labour was similar to ⁃ Unless otherwise indicated, the data of blue and green water that of fishing and the data used is from (FAO, 2013). used in the agricultural production was obtained from ⁃ Regarding the analysis of externalization, it was assumed that (Mekonnen and Hoekstra, 2010). Blue whitling, Herring, Sande el, Norwau pout, Sprat and Cap- eling were of local origin. For marine products: For insect products: ⁃ The energy required by the different species was estimated from (Hognes et al., 2014, 2011). The labour in the catching and ⁃ The data for kelp production was obtained from the following processing for anchovy was estimated from (TASA, 2010, 2008). sources: the energy and blue water from (Alvarado-Morales For the rest of the fish the labour was estimated from et al., 2013), yield from (Skjermo et al., 2014), labour was esti- (Directorate of fisheries, 2018; Norway Working Days, 2018; Salz mated from (Meland and Rebours, 2012; Skjermo et al., 2014). Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210 10 J.J. Cadillo-Benalcazar et al. / Journal of Cleaner Production xxx (xxxx) xxx Fig. 9. Hypothetical mixes of feed-products to feed the salmon. ⁃ The data for the production of larvae were estimated from throughout the years). Although this makes the analysis less robust, (Hexeberg R., 2016; Smetana et al., 2019). Given that these data it should be kept in mind that the objective of this application is to refer to small-scale experimental production systems, they were show the potential of the methodological approach and not the slightly adapted to reflect expected benchmarks for industrial- outcome per se. Tables A4-A9 of the supplementary data show the scale production systems. Regarding labour requirements, a processors developed for the analysis. value of 130 working hours per ton of insect oil was adopted. Note that we did not consider allocation among the co-products. Instead we included the requirements of the entire process given For economic variables: that MuSIASEM considers the processes included in the analysis as functional units being generated by other functional units, gener- ⁃ The price for fishmeal and fish oil corresponds to the price of ating larger functional units (Giampietro et al., 2012). Any type of imports in Norway at 2017. It was obtained from (FAO, 2018c). allocation dividing categories of flows across different functional Because the prices for the ingredients of vegetable origin were units is inconsistent with complex system analysis. In this sense, not found for 2017, an average of the 2014e2016 prices was MuSIASEM differs from the approach used in life cycle analysis used. The data correspond to the Norwegian import price and (LCA). Difficulties of allocation among co-products have been were obtained from (FAO, 2018). Due to the difficulty to estab- documented by Ayer et al. (2007). lish a price for the ingredients coming from insects, the esti- The feasibility and viability of feed supply alternatives were mations were only made for the diagnosis assessed on a year basis assuming current volume of Norwegian salmon production. This approach has limitations as in reality salmon producers continuously switch feed composition in 2.6.2. Salmon production (feed processing, salmon production and response to stock reductions and market price variations. Again, we processing) re-iterate the point that quantitative story-telling does not have the The inputs and outputs were obtained from (Winther et al., 2009), goal of representing exactly the existing situation in a rigorous and the labour for feed processing was estimated from (Skretting Norway, detailed way. Instead it provides the big picture of an issue in 2015), for salmon production from (SSB, 2018b, 2018a; 2017a) and for relation to several non-equivalent narratives in order to identify salmon processing from (FAO, 2013). The salmon feed composition potential bottlenecks and constraints associated with large-scale was calculated from (Hognes et al., 2014, 2011). implementation of investigated policies. 2.6.3. Data limitations 3. Results Despite a major effort in locating data for all the inputs and outputs considered for the selected reference year (2017), some 3.1. Assessment of Norwegian aquaculture at the national level data gaps remained. As a result, some data refer to different years. (diagnostic mode) In particular, coarse assumptions had to be made for the labour hours in the fishery sector (where statistics vary markedly Salient environmental and economic features of the current Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210 J.J. Cadillo-Benalcazar et al. / Journal of Cleaner Production xxx (xxxx) xxx 11 Table 1 An overview of the relation of salmon production in Norway to both its local environment and the external context. Type of variables Phases Variables Unit per year Norway Externalization Biophysics Production of feed ingredients Land 10^6 ha e 0.7 Labour 10^6 h 32 46 Energy 10^15 J 10 16 Fertilizer 10^6 t e 0.1 Pesticide (a.i) 10^3 t e 1.6 Blue water 10^6 m3 negl 27 Green water 10^9 m3 e 3.2 Salmon production & Labour 10^6 h 25 e processing Sea area 10^3 ha 7.4 e Energy 10^15 J 8.3 e Blue water 10^6 m3 5.1 e Economic Trade Money for the imports Billions V e 1 Money from the exports Billions V 6 e Supply Amount of salmon in terms of weight Salmon supply t 1.7  10^4 1.1  10^6 Amount of salmon in terms of food energy and nutrients Food energy from salmon 10^15 Calories negl 2.3 Fat from salmon 10^15 Calories negl 1.4 Protein from salmon 10^15 Calories negl 0.9 Norwegian salmon production system are summarized in Table 1 Norwegian industry to keep doing “more of the same”: keep (Further details are available in Table A10 of the supplementary growing by importing more inputs from elsewhere. This is also an data.). This Table illustrates the biophysical and economic in- important point in the quantitative story-telling. teractions between Norway and the rest of the world e i.e. the level In relation to food security, the analysis shows that the salmon of openness of the system. The majority of the biophysical flows industry generates a supply of high quality nutrients (notably and funds required, either directly or indirectly, come from outside omega 3 fatty acids and protein). However, most of the production Norway, with the exception of natural gas (mainly of local origin). is export oriented, reflecting that it is an activity oriented to the Regarding water consumption, the feed that uses more in- generation of economic benefits. gredients of vegetable origin will have a higher water consumption. Also, because the main ingredients are imported, the results sug- gest that there is a significant amount of water externalized. 3.2. Assessment of salmon feed scenarios at the national level Especially, water captured directly from the biosphere (green (Norway) water). Regarding the labour required for the production of salmon Table 2 shows the profile of inputs that would be required to (expressed as hours of ‘human activity’ per year), it is estimated maintain the actual production of salmon in Norway with the five that around 55% are of local origin (57 Mhr) and the remaining 45% salmon diets illustrated in Fig. 9 and the definition of supply sys- are of external origin. The organization of information through tems illustrated in Fig. 5. As noted earlier, results reflect a coarse scales and levels facilitates the identification of direct and indirect assessment of potential problems and not accurate predictions. human activity that is involved in this production system. For As shown in Table 2, labour requirements vary among the feed example, 25 Mhr and 79 Mhr are directly and indirectly linked to supply systems. The greatest labour demand is observed for the the salmon production. While it is true that this type of analysis marine and insect/marine diets, it being respectively 1.8 and 1.7 should include working hours linked to transport and service to times higher than for the standard diet. It should be noted though obtain more robust results, the usefulness of the MuSIASEM as a that the estimates of working hours for the production of insects useful tool for decision making is already appreciated. are based on processes at experimental scale. If production will be Although the economic analysis is limited, it gives an idea of the scaled to an industrial level these values may decrease. trade balance (the relationship between the profits from the export The lower demand for working hours in the plant-based supply and the money that leaves Norway as a result of the purchase of systems comes at the cost of a greater pressure on the landdthis imported products). Due to the economic return made possible by scenario requires 5 times more land than the marine-based sol- imports, from an economic perspective it is convenient for the utiondand a greater consumption of fertilizers, pesticides and water. Water demand in the plant-based supply system is 6 times Table 2 Comparison of biophysical requirements at the focal level in 5 different feed scenarios (to produce 1.5  10^6 t feed). Biophysical resources Standard Marine based Plant based Insect -marine Insect e plant Flow elements (year) Crops (10^6 t) 3.0 0.8 5.0 0.8 2.0 Marine products (10^6 t) 4.0 11 e 5.4 2.0 Larvae products (10^`6 t) e e e 1.0 1.0 Energy (10^15 J) 26 43 16 40 32 Fertilizer (10^6 t) 0.1 negl 0.2 negl negl Pesticide (10^3 t) 1.6 0.4 2.4 0.4 1.0 Seaweed as feed (10^6 t) e e e 2.0 2.0 Blue water (10^6 m3) 27 8.0 38 22 32 Green water (10^9 m3) 3.2 0.8 4.7 0.8 2.0 Fund elements Labour (10^6 h) 78 140 39 135 103 Land (10^6 ha) 0.7 0.2 1.0 0.2 0.4 Sea area e seaweed (10^3 ha) e e e 75.0 75.0 Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210 12 J.J. Cadillo-Benalcazar et al. / Journal of Cleaner Production xxx (xxxx) xxx Table 3 Energy carriers required in the various salmon feed scenarios (in PJ to produce 1.5  10^6 t feed). Requirements have been calculated from the processors detailed in Tables A4- A9 and the vectors described in Tables A1-A3 of the supplementary data. Energy carrier Standard Marine-based Plant-based Insect-marine Insect-plant Electricity 3.8 2.9 4.3 7.4 7.8 Fuel 14.5 31.6 3.7 20.6 12.0 Heat 8.3 9.0 8.0 12.0 12.0 Total 26.6 43.5 16.0 40.0 31.8 higher than in the marine-based scenario and 1.5 times higher than of substituting marine feed ingredients with other sources will the standard supply system. continue in the future. Since the 1980s the fish catch worldwide has Regarding the energy use, the marine-based and the mixed remained constant and is unlikely to increase in the future (FAO, marine-insects systems are the most demanding. The former is 2.7 2017). Moreover, it is expected that fisheries will continue experi- times higher than the plant-based system and 1.7 times higher than encing diminishing return: fishing vessels need to travel a greater the current diet. distance to find shoal of fish. This will further increase the already In addition to the total amount of energy required, another key high energy (fuel) requirements and CO2 emission. The concomi- aspect to consider is the type of energy carriers required (see tant increase in price of marine ingredients has already caused a Table 3). For example, the marine-based scenario requires a greater change in the composition of salmon feed. While in 1990 feeds amount of fuel for fishing vessels. On the other hand, the plant- were based predominantly (90%) on marine ingredients, at present, based scenario requires a greater amount of electricity for pro- feeds are composed of a mix of terrestrial and marine ingredients cessing. The production of insects is also an activity that demands a (around 30%) (Ytrestøyl et al., 2015). This trend is expected to lot of energy (but, again, note that the benchmarks are based on continue. production on an experimental scale), therefore according to their Given these considerations it is obvious that salmon production combination (marine or plants) they will increase the energy will have to continuously adapt to the emergence of new bio- requirement. Thus it is observed that the marine-insect scenario is physical limits and threats and eventually ‘become something else’. 1.6 larger than the standard diet and 2.7 times larger than the lower Scaling up the production of aquaculture translates into an value scenario (based on plants). increasing level of uncertainty about the insurgence of possible Scenarios that include insects significantly reduce land biophysical problems. This will also imply taking risks and forcing requirement (the insect-marine scenario and the insect plant negotiations among social actors to avoid legitimate but unavoid- compared to the standard are 3.5 and 1.8 times lower respectively) able conflicts among new winners and losers. To support this and therefore the pressure on the terrestrial ecosystem in terms of process it is relevant and timely to start a reflection about what water, fertilizers and pesticides. These results are in line with criteria should be used and why for deciding the future of the nchez-Muros et al., 2014; Smetana et al., 2016), who point out (Sa salmon industry. that the production of insect-based protein is more beneficial to the environment than the conventional ingredients. But as noted 4.1. The risk of damaging the reputation of Norwegian high-quality above, they have a greater demand for hours of labour and energy. salmon Also, the proposal to use kelp as a source of omega 3 in order to reduce the dependence on fish oil will create a pressure on this Replacing marine ingredients with terrestrial ingredients im- resource in the marine ecosystem. For example, for both scenarios plies a larger demand for land, water, fertilizers and pesticides. This including insects, approximately 75,000 ha in the sea would be result is in concordance with findings of Torrissen et al. (2011). The required. For comparison, in 2016 the area used in the Norway sea production of salmon is currently following the same path of for the cultivation of seaweed was 277 ha (Ste vant et al., 2017). On development as intensive beef production in feed-lots. This solu- the other hand, the production of seaweed integrated in aquacul- tion implies further increases in crop production for animal feeds ture through Integrated Multi-Trophic Aquaculture (IMTA) systems (either salmon or beef) and associated land use changes (e.g., has as its main cost human labour (Chapman et al., 2014). This is competition with crop cultivation for direct consumption; defor- another important point for the story-telling. estation leading to soil erosion and loss of biodiversity) and envi- ronmental impacts of fertilizers and pesticides in rural areas 4. Discussion (Ytrestøyl et al., 2015). The progressive association of salmon pro- duction with negative effects on both the environment and rural The scenario of insect feed considered in the AQUAFLY project communities in the developing world may generate a negative refers to an innovation of low technology readiness level (in be- imagine for the salmon industry. This is another source of tween ‘basic technology research’ and ‘research to prove feasibility’ uncertainty. on the TRL scale). For this reason, the quantitative assessment is Results show that from an environmental perspective, insect approximate and not sufficiently robust for use as ‘scientific evi- feed ingredients may represent an interesting alternative, causing dence’. All the same, when adopting the logic of quantitative story- less pressure on terrestrial land than plant feed ingredients. How- telling the results are useful to anticipate potential troubles of a ever, energy and labour requirements are currently higher for in- continued growth of the Norwegian salmon industry, both inside sect feed than for plant-based alternatives. Alternative insect feed and outside the national borders, and to critically examine the substrates (wastes instead of kelp) could reduce sea area re- justification narratives for a further increase in production: Can quirements. However, both insect and plant alternatives are further growth be obtained without increasing environmental currently relatively poor in omega-3 and hence potentially impacts and is further growth desirable? endanger the status of salmon as a high quality food (Sprague et al., An informed discussion about the future of salmon production 2016). has to start from the feasibility, viability and desirability of this The solution of adding a new element to the metabolic pathway activity (the quality checks of quantitative story-telling). In relation (substrate e insects e salmon) enlarges the options of feed sources. to biophysical constraints (feasibility) it is expected that the trend Indeed, while the set of possible feed sources for salmon is limited, Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210 J.J. Cadillo-Benalcazar et al. / Journal of Cleaner Production xxx (xxxx) xxx 13 feed sources for insects are in principle virtually unlimited and important is the goal of preserving the existing cultural identity? open the way to the recycling of organic waste and sewage (circular What is wrong with the idea of remaining at a given level of salmon economy) and modulating omega-3 content. However, this solution production and focus on greening of the salmon industry and may also lead to the emergence of new and unexpected problems, safeguarding adaptability in the case of perturbations to boundary such as bio-accumulation of toxic substances that could damage conditions? Who has decided that we always have to strive for the image of Norwegian salmon. In this case we are dealing with a increasing the productivity of existing economic activities? Does very serious source of uncertainty. the moral of Aesop’s fable about the dog losing its bone “in order to have more” apply here? 4.2. Is further growth in production desirable? To make the proposed analysis more robust for quantitative story-telling, it should be carried out in co-production with the A common justification narrative about the future of Norwegian people using the results. This is particularly important with regard salmon production sees an expansion of the industry into new to: i) selection of the indicators of performance in setting up the markets as an option for generating greater economic benefit and processors (variables included in the analysis); and ii) checking the direct and indirect jobs (ISFA, 2018). How large are these economic plausibility of the assumptions adopted in the definition of the benefits for Norway? Benefits for whom? What type of direct and various processors describing the production systems (Ci); and iii) indirect jobs will be created? For whom? checking the choice of the definition of the elements to be included The results of Table 1 suggest that the main destination of the in the two vectors vAi (the diet for the salmons) and vBi (the mix of salmon production in Norway is the export, obtaining from this supply systems used in the production of feed). trade around 6 billion euros in 2017 (SSB, 2017b). When its value is This work has focused on salmon feed. Other environmental compared with the Norwegian GDP (350 billion euros) it is problems with salmon aquaculture do exist, such as the effect of sea observed that this represents just the 1.7%. This suggests that it may lice on wild fish (Hersoug, 2015; Nilsen et al., 2018), that have put be more relevant to focus on local employment and preservation of pressure on the Norwegian government to reconsider its policies traditional activities and landscapes, rather than increasing the for expanding the salmon industry. revenues for the country in this way. However also in terms of job creation, the salmon industry does not significantly contribute to Acknowledgement employment at the level of the national economy (Bailey, 2014). In addition, it is reasonable to assume that additional growth of The authors gratefully acknowledge funding from the Research the salmon industry (by increasing installed operational capacity Council of Norway (Norges forskningsråd) under grant agreement and maintaining current productive efficiency) will increase the 238997/E40 (Insects as natural feed ingredients for sustainable impact on the environment in terms of a greater demand of bio- salmon farmingeAQUAFLY). The Institute of Environmental Science physical resources and a greater pressure on the ecological systems, and Technology (ICTA) has received financial support from the besides a progressive increase in the costs of the local production Spanish Ministry of Science, Innovation and Universities, through process associated with increased intensity of salmon farming. the “María de Maeztu” program for Units of Excellence (MDM- When this pressure will exceed the resilience of the ecological 2015-0552). This work reflects the authors’ view only; the funding systems, a slash-back on the salmon production system may occur. agencies are not responsible for any use that may be made of the In the same way, the results in Table 1 suggest that a further information it contains. increase in the production of salmon cannot be justified in relation to providing food security for Norway. The actual production of Appendix A. Supplementary data salmon is orders of magnitude larger than its internal consumption. As for global food security, salmon is a luxury food item and while Supplementary data to this article can be found online at its nutritional quality is excellent for preventing cardio-vascular https://doi.org/10.1016/j.jclepro.2019.119210. diseases in affluent countries, it is unlikely to contribute to the food security of the low-income classes in developing countries References most exposed to the risk of malnutrition due to a systemic problem of low disposable income. Ahl, V., Allen, T.F.H., 1996. Hierarchy Theory : a Vision, Vocabulary, and Episte- As explained earlier, our analysis does not pretend to identify mology. Columbia University Press, New York. “the best” pathway for the development of salmon aquaculture, but Allen, T.F.H., Starr, T.B., 1982. Hierarchy : Perspectives for Ecological Complexity. University of Chicago Press, Chicago. intends to support a reflection and discussion on the future of Alvarado-Morales, M., Boldrin, A., Karakashev, D.B., Holdt, S.L., Angelidaki, I., Norwegian salmon aquaculture through quantitative story-telling. Astrup, T., 2013. Life cycle assessment of biofuel production from brown Quantitative story-telling avoids the trap of ‘solving’ the seaweed in Nordic conditions. Bioresour. Technol. 129, 92e99. https://doi.org/ 10.1016/j.biortech.2012.11.029. complexity associated with our interaction with the external world Ayer, N.W., Tyedmers, P.H., Pelletier, N.L., Sonesson, U., Scholz, A., 2007. Co-product by simplification. Our analysis stimulates a reflection on the allocation in life cycle assessments of seafood production systems: review of following questions: Why does Norway want to further increase its problems and strategies. Int. J. Life Cycle Assess. 12, 480e487. https://doi.org/ 10.1065/lca2006.11.284. production if salmon is not produced for internal food security nor Bailey, J., 2014. Looking for sustainable solutions in salmon aquaculture. Etikk i for boosting the national economy? What is to be gained by whom, praksis - Nord. J. Appl. Ethics 8, 22e40. https://doi.org/10.5324/eip.v8i1.1801. and what is to be lost by whom? Is “more” necessarily “better”? The Bartlett, A.A., 2004. The Essential Exponential! for the Future of Our Planet. Center for Science, Mathematics & Computer Education, University of Nebraska, idea of incorporating insects in the food chain certainly has merit Nebraska, USA. because it enlarges the option space for technical innovations, but Belghit, I., Liland, N.S., Waagbø, R., Biancarosa, I., Pelusio, N., Li, Y., Krogdahl, Å., at the same time it opens a Pandora box of uncertainties. Techno- Lock, E.-J., 2018. Potential of insect-based diets for Atlantic salmon (Salmo salar). Aquaculture 491, 72e81. https://doi.org/10.1016/j.aquaculture.2018. logical innovations are always associated with justification narra- 03.016. tives: Why do we have to change? Decisions about changes require Bell, J.G., Henderson, R.J., Tocher, D.R., Sargent, J.R., 2004. Replacement of dietary discussions over the moral issues related to the ethical implications fish oil with increasing levels of linseed oil: modification of flesh fatty acid of proposed changes. Should we protect the existing values by compositions in atlantic salmon (Salmo salar) using a fish oil finishing diet. Lipids 39, 223e232. https://doi.org/10.1007/s11745-004-1223-5. limiting the options of changes to be considered? Or should we Boissy, J., Aubin, J., Drissi, A., van der Werf, H.M.G., Bell, G.J., Kaushik, S.J., 2011. update/replace existing values so as to expand the options? How Environmental impacts of plant-based salmonid diets at feed and farm scales. Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210 14 J.J. Cadillo-Benalcazar et al. / Journal of Cleaner Production xxx (xxxx) xxx Aquaculture 321, 61e70. https://doi.org/10.1016/j.aquaculture.2011.08.033. Nutr. 22, 1202e1213. https://doi.org/10.1111/anu.12343. Buchspies, B., Kaltschmitt, M., 2016. Life cycle assessment of bioethanol from wheat Logatcheva, K., Galen, M.A. van, 2015. Primary Food Processing; Cornerstone of and sugar beet discussing environmental impacts of multiple concepts of co- Plant-Based Food Production and the Bio-Economy in Europe (No. Report/LEI product processing in the context of the European Renewable Energy Direc- 2015-121). The Hague. tive. Biofuels 7, 141e153. https://doi.org/10.1080/17597269.2015.1122472. Louie, A.H., 2009. More than Life Itself: A Synthetic Continuation in Relational Buratti, C., Barbanera, M., Fantozzi, F., 2008. Environmental balance of bioethanol Biology. De Gruyter, Inc., Berlin/Boston. from corn grain: evaluation of different procedures of co-products allocation. Louie, A.H., 2013. The Reflection of Life. Springer, New York. In: 16th European Biomass Conference & Exhibition, 2-6 June 2008, MAGIC, 2019. Moving towards adaptive governance in complexity: informing nexus pp. 2621e2628. Valencia: Spain. security [WWW Document]. H2020 No. 689669. URL. https://cordis.europa.eu/ Chapman, A., Ste vant, P., Schipper, J., Krakas, O., 2014. Markedsvurdering for project/rcn/203266/factsheet/en. accessed 8.21.19. Bærekraftig Algedyrking I Integrert Multitrofisk Akvakultur Anlegg. Ålesund, Marinussen, M., Kernebeek, H. van, Broekema, R., Groen, E., Kool, A., Zeist, W.J. van, Norway. Dolman, M., Blonk, H., 2012. LCI Data for the Calculation Tool Feedprint for Collingridge, D., 1980. The Social Control of Technology. St. Martin’s Press, New Greenhouse Gas Emissions of Feed Production and Utilization. Gouda, the York. Netherlands. Di Felice, L.J., Ripa, M., Giampietro, M., 2019. An alternative to market-oriented Matthews, K.B., Blackstock, K.L., Rivington, M., Waylen, K., Miller, D.G., Wardell- energy models: nexus patterns across hierarchical levels. Energy Policy 126, Johnson, D., Kovacic, Z., Renner, A., Ripa, M., Giampietro, M., 2017. Delivering 431e443. https://doi.org/10.1016/j.enpol.2018.11.002. more than the “Sum of the Parts”: using Quantitative Storytelling to address the Dickhoff, D., 2010. Effects of farmed seafood consumption on trends in U.S. per challenges of conducting science for policy in the EU land, water and energy capita omega-3 fatty acid consumption. In: Aquaculture 2010. World Aquacul- nexus. In: 22nd International Congress on Modelling and Simulation. Hobart, ture Society, San Diego, CA, USA. Tasmania, Australia, pp. 15e21, 3 to 8 December 2017. Australia. Directorate of fisheries, 2018. Fishermen, fishing vessels and licenses [WWW Mekonnen, M.M., Hoekstra, A.Y., 2010. The green, blue and grey water footprint of Document]. URL. https://www.fiskeridir.no/English/Fisheries/Statistics/ crops and derived crop products. In: Value of Water Research Report Series Fishermen-fishing-vessels-and-licenses. accessed 10.17.18. No.47. (Delft, the Netherlands) (No. Volume 2: Appendices). DKNVS/NTVA, 2012. Verdiskaping Basert På Produktive Hav I 2050 (Oslo). Meland, M., Rebours, C., 2012. The Norwegian Seaweed Industry. Work Package European Comission, 2016. Euro area and EU working days to build Calendar 1&2. Bodø, Norway. Adjustment Regressor [WWW Document]. URL. https://ec.europa.eu/eurostat/ Nilsen, F., Ellingsen, I., Finstad, B., Helgesen, K.O., Karlsen, O., Sandvik, A.D., cros/content/euro-area-and-eu-working-days-build-calendar-adjustment- Saegrov, H., Ugedal, O., Vollset, K.W., Qviller, L., 2018. Assessment of Salmon regressor_en. accessed 10.22.18. Lice-Induced Wild Fish Mortality Per Production Area in 2018. Report from FAO, 2013. Fishery and aquaculture country profiles. The kingdom of Norway Expert Group for Assessment of Lice Impact. [WWW Document]. URL. http://www.fao.org/fishery/facp/NOR/en#Country Skretting Norway, 2015. Sustainability Report 2015. Norway. Sector-SectorSocioEcoContribution. accessed 10.18.18. Norway Working Days, 2018. Calculator working days [WWW Document]. URL. FAO, 2017. Technical Conversion Factors for Agricultural Commodities. https://norway.workingdays.org/#a20. accessed 10.17.18. FAO - Food and Agriculture Organization of the United Nations, 2018. Faostat Olafsen, T., Winther, U., Olsen, Y., Skjermo, J., 2012. Value Created from Productive [WWW Document]. URL. http://www.fao.org/faostat/en/#home. accessed Oceans in 2050, Report of a Working Group Appointed by the Royal Norwegian 12.19.18. Society of Sciences (DKNVS) and the Norwegian Academy of Technological FAO, 2018a. The State of World Fisheries and Aquaculture 2018. Meeting the sus- Sciences (NTVA) (Norway). tainable development goals, Rome, Italy. Ortega, E., 2003. A soja no brasil: modelos de produç~ ao, custos, lucros, external- FAO, 2018b. GLOBEFISH Highlights, GLOBEFISH Highlights. FAO, Rome, Italy. idades, sustentabilidade e políticas públicas (Brazil). FAO, 2018c. GLOBEFISH Monthly Trade Statistics. Pattee, H.H., 1973. Hierarchy Theory; the Challenge of Complex Systems. G. Braziller, Felt, U., 2007. Takimg European Knowledge Society Seriously. Report of the Expert New York. Group on Science and Governance to the Science. Economy and Society Pimentel, D., Pimentel, H.M., 2008. Food, Energy, and Society, third ed. ed. Fl. : CRC Directorate, Directorate-General for Research, European Commission, European Press/Taylor & Francis G., cop, Boca Raton. Commission, Belgium. Queiro s, J., Malça, J., Freire, F., 2015. Environmental life-cycle assessment of rape- Galitsky, C., Worrell, E., Ruth, M., 2003. Energy Efficiency Improvement and Cost seed produced in Central Europe: addressing alternative fertilization and Saving Opportunities for the Corn Wet Milling Industry. An ENERGY STAR® management practices. J. Clean. Prod. 99, 266e274. https://doi.org/10.1016/ Guide for Energy and Plant Managers, Berkeley, CA. j.jclepro.2015.03.016. Gasco, L., Gai, F., Maricchiolo, G., Genovese, L., Ragonese, S., Bottari, T., Caruso, G., Rana, K.J., Siriwardena, S., Hasan, M.R., 2009. Impact of Rising Feed Ingredient Prices 2018. Feeds for the Aquaculture Sector: Current Situation and Alternative on Aquafeeds and Aquaculture Production (No. 541). FAO Fisheries and Aqua- Sources. Springer, Cham, Switzerland. https://doi.org/10.1007/978-3-319- culture Technical Paper, Rome, Italy. 77941-6. Renner, A., Giampietro, M., 2020. Socio-technical discourses of European electricity Georgescu-Roegen, N., 1971. The Entropy Law and Economic Process. Harvard decarbonization: contesting narrative credibility and legitimacy with quanti- University Press, Cambridge, MA, USA. tative story-telling. Energy Res. Soc. Sci 59, 101279. https://doi.org/10.1016/ Giampietro, M., 2018. Anticipation in agriculture. In: Poli, R. (Ed.), Handbook of j.erss.2019.101279. Anticipation. Springer International Publishing, Cham, pp. 1e35. https://doi.org/ Rosen, R., 1958. A relational theory of biological systems. Bull. Math. Biophys. 20, 10.1007/978-3-319-31737-3_23-1. 245e260. https://doi.org/10.1007/BF02478302. Giampietro, M., Allen, T.F.H., Mayumi, K., 2006. The epistemological predicament Rosen, R., 1985. Organisms as causal systems which are not mechanisms: an essay associated with purposive quantitative analysis. Ecol. Complex. 3, 307e327. into the nature of complexity. In: Rosen, R. (Ed.), Theoretical Biology and https://doi.org/10.1016/j.ecocom.2007.02.005. Complexity. Three Essays on the Natural Philosophy of Complex Systems. Ac- Giampietro, M., Mayumi, K., Sorman, A.H., 2012. The Metabolic Pattern of Societies: ademic Press, pp. 165e203. https://doi.org/10.1016/C2013-0-11405-5. where Economists Fall Short. Routledge, cop., London ([etc.]). Rumpold, B.A., Schlüter, O.K., 2013. Potential and challenges of insects as an inno- Giampietro, M., Mayumi, K., Sorman, A.H., 2013. Energy Analysis for a Sustainable vative source for food and feed production. Innov. Food Sci. Emerg. Technol. 17, Future : Multi-Scale Integrated Analysis of Societal and Ecosystem Metabolism. 1e11. https://doi.org/10.1016/j.ifset.2012.11.005. Routledge. 2013, London; New York. Saltelli, A., Giampietro, M., 2016. The fallacy of evidence-based policy. In: Giampietro, M., Aspinall, R., Ramos-Martin, J., Bukkens, S.G.F., 2014. Resource Ac- Benessia, A., Funtowicz, S., Giampietro, M., Guimar~ ^ Ravetz, J.R., aes Pereira, A., counting for Sustainability Assessment. The Nexus between Energy, Food, Saltelli, A., Strand, R., Sluijs, J.P. van der (Eds.), Science on the Verge. Consortium Water and Land Use. Routledge, cop., London. for Science, Policy & Outcomes, pp. 31e69. Tempe, Ariz. Gonza lez-Lo  pez, R., Giampietro, M., 2017. Multi-scale integrated analysis of charcoal Saltelli, A., Giampietro, M., 2017. What is wrong with evidence based policy, and production in complex social-ecological systems. Front. Environ. Sci. 5, 54. how can it be improved? Futures 91, 62e71. https://doi.org/10.1016/ https://doi.org/10.3389/fenvs.2017.00054. J.FUTURES.2016.11.012. Hersoug, B., 2015. The greening of Norwegian salmon production. Maritain Stud. 14, Salthe, S.N., 1985. Evolving Hierarchical Systems : Their Structure and Representa- 16. https://doi.org/10.1186/s40152-015-0034-9. tion. Columbia University Press, New York. Hexeberg, R.,I., 2016. Life Cycle Assessment of Fish Feed Produced from the Black Salz, P., Buisman, E., Malvarosa, L., Daures, F., Leblond, E., Andersen, J.L., Soldier Fly (Hermetia Illucens). Norwegian University of Science and Andersen, K.N., Nielsen, R., Anton, S., Curtis, H., 2005. Calculation of Labour Technology. Including FTE (Full-time Equivalent) in Fisheries (No. Study No FISH/2005/14 Hognes, E.S., Ziegler, F., Sund, V., 2011. Carbon footprint and area use of farmed Final Report). Norwegian salmon. Project No.. 830276. Norway. https://doi.org/10.13140/RG.2. nchez-Muros, M.J., Barroso, F.G., Manzano-Agugliaro, F., 2014. Insect meal as Sa 1.4168.2808. renewable source of food for animal feeding: a review. J. Clean. Prod. 65, 16e27. Hognes, E.S., Nilsson, K., Sund, V., Ziegler, F., 2014. LCA of Norwegian Salmon Pro- https://doi.org/10.1016/j.jclepro.2013.11.068. duction 2012. SINTEF Fiskeri og havbruk, Norway. A26401. Schmidt, J.H., 2007. Life Cycle Assessment of Rapeseed Oil and Palm Oil: Ph.D. ISFA, 2018. Salmon Farming Sustaining Communities and Feeding the World Thesis, Part 3: Life Cycle Inventory of Rapeseed Oil and Palm Oil. Ph. D. Thesis. (Norway). Aalborg University, Denmark. Koestler, A., 1968. The Ghost in the Machine. The MacMillan Co., New York. Serrano-Tovar, T., Pen ~ ate Sua rez, B., Musicki, A., de la Fuente Bencomo, J.A., Koestler, A., 1978. Janus: A Summing up. Hutchinson & Co, USA. Cabello, V., Giampietro, M., 2019. Structuring an integrated water-energy-food Lock, E.R., Arsiwalla, T., Waagbø, R., 2016. Insect larvae meal as an alternative source nexus assessment of a local wind energy desalination system for irrigation. of nutrients in the diet of Atlantic salmon (Salmo salar) postsmolt. Aquacult. Sci. Total Environ. 689, 945e957. https://doi.org/10.1016/J.SCITOTENV.2019. Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210 J.J. Cadillo-Benalcazar et al. / Journal of Cleaner Production xxx (xxxx) xxx 15 06.422. Steffen, W., Broadgate, W., Deutsch, L., Gaffney, O., Ludwig, C., 2015. The trajectory of Shepherd, C.J., Monroig, O., Tocher, D.R., 2017. Future availability of raw materials the anthropocene: the great acceleration. Anthropol. Rev. 2, 81e98. https:// for salmon feeds and supply chain implications: the case of Scottish farmed doi.org/10.1177/2053019614564785. salmon. Aquaculture 467, 49e62. https://doi.org/10.1016/j.aquaculture.2016. vant, P., Rebours, C., Chapman, A., 2017. Seaweed aquaculture in Norway: recent Ste 08.021. industrial developments and future perspectives. Aquacult. Int. 25, 1373e1390. Skjermo, J., Aasen, I.M., Arff, J., Broch, O.J., Carvajal, A.K., Christie, H.C., Forbord, S., https://doi.org/10.1007/s10499-017-0120-7. Olsen, Y., Reitan, K.I., Rustad, T., Sandquist, J., Solbakken, R., Steinhovden, K., TASA, 2008. Reporte de sostenibilidad 2008 (Lima). Wittgens, B., Wolff, R., Handå, A., 2014. A New Norwegian Bioeconomy Based on TASA, 2010. Reporte de sostenibilidad 2010 (Lima). Cultivation and Processing of Seaweeds: Opportunities and R&D Needs (No. The World Bank, 2018. Population, total [WWW Document]. URL. https://data. SINTEF A25981-Unrestricted). Project No 6020913. Norway. worldbank.org/indicator/SP.POP.TOTL. accessed 12.3.18. Smetana, S., Palanisamy, M., Mathys, A., Heinz, V., 2016. Sustainability of insect use Torrissen, O., Olsen, R.E., Toresen, R., Hemre, G.I., Tacon, A.G.J., Asche, F., Hardy, R.W., for feed and food: life Cycle Assessment perspective. J. Clean. Prod. 137, Lall, S., 2011. Atlantic salmon (Salmo salar): the “super-chicken” of the sea? Rev. 741e751. https://doi.org/10.1016/j.jclepro.2016.07.148. Fish. Sci. 19, 257e278. https://doi.org/10.1080/10641262.2011.597890. Smetana, S., Schmitt, E., Mathys, A., 2019. Sustainable use of Hermetia illucens in- UAB, 2018. Database - magic project [WWW Document]. URL. https://magic-nexus. sect biomass for feed and food: attributional and consequential life cycle eu/. assessment. Resour. Conserv. Recycl. 144, 285e296. https://doi.org/10.1016/ Visser, C. de, Buisonje , F. de, Ellen, H., Stanghellini, C., Voort, M. van der, j.resconrec.2019.01.042. Gołaszewski, J., Brodzin  ski, Z., Myhan, R., Olba-Zie˛ ty, E., Stolarski, M., Baptista, F., Sprague, M., Dick, J.R., Tocher, D.R., 2016. Impact of sustainable feeds on omega-3 Leopoldo Silva, L., Murcho, D., Meyer-Aurich, A., Ziegler, T., Ahokas, J., long-chain fatty acid levels in farmed Atlantic salmon, 2006-2015. Sci. Rep. 6, Jokiniemi, T., Mikkola, H., Rajaniemi, M., Balafoutis, A., Briassoulis, D., 1e9. https://doi.org/10.1038/srep21892. Mistriotis, A., Panagakis, P., Papadakis, G., 2012. State of the Art on Energy Ef- SSB, 2017a. Growing numbers working in fish farming Norway [WWW Document]. ficiency in Agriculture. Country data on energy consumption in different URL. https://www.ssb.no/en/jord-skog-jakt-og-fiskeri/artikler-og-publikasjoner/ agroproduction sectors in the European countries (No. AGREE Project Deliver- growing-numbers-working-in-fish-farming. able 2.1). SSB, 2017b. Aquaculture [WWW Document]. URL. https://www.ssb.no/en/ Wilfart, A., Espagnol, S., Dauguet, S., Tailleur, A., Gac, A., Garcia-Launay, F., 2016. fiskeoppdrett. accessed 9.3.18. ECOALIM: a dataset of environmental impacts of feed ingredients used in SSB, 2018a. Quantity and value of catch from Norwegian vessels. Selected fish French animal production. PLoS One 11, 1e17. https://doi.org/10.1371/ species. Preliminary figures [WWW Document]. URL. https://www.ssb.no/en/ journal.pone.0167343. fiskeri. accessed 10.17.13. Winther, U., Ziegler, F., Skontorp, E., Emanuelsson, A., Sund, V., Ellignsen, H., 2009. SSB, 2018b. Aquaculture [WWW Document]. URL. https://www.ssb.no/en/jord- Carbon Footprint and Energy Use of Norwegian Seafood Products (No. Report skog-jakt-og-fiskeri/statistikker/fiskeoppdrett. accessed 10.22.18. SFH80 A096068) (Norway). St-Hilaire, S., Sheppard, C., Tomberlin, J.K., Irving, S., Newton, L., McGuire, M.A., Ytrestøyl, T., Aas, T.S., Åsgård, T., 2015. Utilisation of feed resources in production of Mosley, E.E., Hardy, R.W., Sealey, W., 2007. Fly prepupae as a feedstuff for Atlantic salmon (Salmo salar) in Norway. Aquaculture 448, 365e374. https:// rainbow trout, Oncorhynchus mykiss. J. World Aquac. Soc. 38, 59e67. https:// doi.org/10.1016/j.aquaculture.2015.06.023. doi.org/10.1111/j.1749-7345.2006.00073.x. Please cite this article as: Cadillo-Benalcazar, J.J et al., Multi-scale integrated evaluation of the sustainability of large-scale use of alternative feeds in salmon aquaculture, Journal of Cleaner Production, https://doi.org/10.1016/j.jclepro.2019.119210