Life Cycle Assessment of a Novel Closed-Containment Salmon

Article pubs.acs.org/est

Life Cycle Assessment of a Novel Closed-Containment Salmon Aquaculture Technology Keegan P. McGrath,*,† Nathan L. Pelletier, and Peter H. Tyedmers† †

School for Resource and Environmental Studies, Dalhousie University. 6100 University Ave, Suite 5010, PO Box 15000, Halifax, Nova Scotia B3H 4R2, Canada S Supporting Information *

ABSTRACT: In salmonid aquaculture, a variety of technologies have been deployed that attempt to limit a range of environmental impacts associated with net-pen culture. One such technology employs a floating, solid-walled enclosure as the primary culture environment, providing greater potential control over negative interactions with surroundings waters while limiting energy use required for water circulation, thermoregulation and supplemental oxygen provision. Here, we utilize life cycle assessment to model contributions to a suite of global-scale resource depletion and environmental concerns (including global warming potential, acidification potential, marine eutrophication potential, cumulative energy use, and biotic resource use) of such a technology deployed commercially to rear Chinook salmon in coastal British Columbia, Canada. Results indicate that at full grow-out, feed provisioning and on-site energy use dominate contributions across four of five impact categories assessed. For example, per tonne of salmon harvested, feed contributed approximately 72% to global warming potential, 72% to acidification potential, and accounted for 100% of biotic resource use. However, for both feed and on-site energy use, impacts are heavily influenced by specific sources of inputs; therefore efforts to improve the environmental performance of this technology should focus on reducing these in favor of less impactful alternatives.

1. INTRODUCTION Global cultured salmonid production has increased rapidly over recent decades, from 299 000 tonnes in 1990 to 1 900 000 tonnes in 2010.1 With its emergence as an ever more important industry, attention has increasingly focused on understanding and mitigating its resource use and environmental impacts.2,3 Particular concerns vary with production setting and culture technology but include nutrient loading and alteration of local receiving environments through the release of excess nutrients and other chemicals;4,5 the amplification and transmission of disease to nearby wildlife populations;6 and reduced fitness of wild conspecifics due to interactions and interbreeding with escapees.7,8 In addition to localized impacts, carnivorous species such as salmon are reared on nutrient-dense feedstuffs derived from a wide range of plant and animal sources.9−11 The provision of feeds along with the range of other operational and capital inputs contribute to a variety of environmental challenges (e.g., greenhouse gas emissions, acidification, overfishing) that span local, regional, and global scales.10,12−14 To mitigate the environmental impacts of salmon culture, considerable effort and research has been invested in the potential to move away from the industry dominant net-pen culture system toward “closed containment”15,16 or land-based recirculating aquaculture systems (RAS).17,18 These technologies are often considered environmentally superior as they theoretically allow managers to reduce, redirect, or eliminate © XXXX American Chemical Society

emissions of nutrients into receiving environments and negative interactions with wild species.17 Moreover, they allow managers a measure of control over environmental parameters such as influent water temperature, quality, oxygen content, and flow rate, which can be optimized for growth.19 Unfortunately, some of the attributes which improve the environmental performance of closed-containment and RAS may exacerbate other environmental issues (e.g., in particular those related to increased energy use).20 In addition, any change in the culture environment of any aquaculture system (RAS, net-pen or otherwise) that decreases overall efficiency of production (e.g., increased susceptibility to disease, higher mortalities, lower feed conversion etc..) represent potential changes in environmental performance. The result is that these intensive culture technologies may reduce local-scale environmental impacts (e.g., by reducing emissions of eutrophying chemicals)21 while exacerbating global-scale environmental challenges (e.g., climate change).20,22 Given the growing importance of salmon aquaculture and the environmental challenges it poses, it is essential that whenever possible future development simultaneously lower resource intensity and environmental impacts at all scales. To assess whether emerging Received: October 20, 2014 Revised: April 5, 2015 Accepted: April 6, 2015

A

DOI: 10.1021/es5051138 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 1. Life cycle flowchart for the material and energy inputs/outputs associated with the production of salmon in the solid-wall aquaculture system.

formulations; one, a high fish meal/fish oil feed used in the SWAS during production, the other, a feed representative of industry average salmon feed inputs used in western Canada in 2012. Lastly, we have undertaken an uncertainty analysis of life cycle impacts, an area of analysis that is often overlooked in LCA research.30

aquaculture systems are achieving this goal it is critical to understand the implications of these systems by employing comprehensive assessment techniques. Here we have assessed the life cycle impacts associated with salmonid production (Chinook salmon, Onchorhynchus tshawytscha) in a floating tank, flow-through, solid-walled aquaculture system (SWAS) developed for commercial application by AgriMarine Industries, Inc. (AgriMarine), using life cycle assessment (LCA) (see Supporting Information (SI) for a description of the SWAS). Resulting information is valuable for assessing the environmental performance of this technology and was used to identify environmental hotspots and explore strategies for improving environmental performance. The SWAS culture technology has attracted interest as it represents an intermediate between net-pens and RAS, potentially providing some of the local environmental benefits of RAS along with reduced global-scale impacts. These potential benefits may be achieved in three ways. First, with more control over influent, the producer can draw from deep waters that will presumably have fewer parasites and pathogens (e.g., salmon louse, Lepeophtheirus salmonis, which are thought to reside predominantly in the middle to upper water column23,24) and better manage the condition of the growing environment, theoretically improving fish health. Second, by partially isolating salmon from the local environment with a solid barrier there is potential for wastes to be captured for subsequent handling and the transmission of some pests and diseases to be mitigated. Lastly, by situating in the aquatic environment, the need to pump influent water significantly above its static level is reduced, as is the need to seasonally heat or cool influent, presumably reducing overall energy demand of the system. LCA has been used extensively over the past decade to explore environmental consequences of several species and culture technologies used in finfish aquaculture3,25−27 with several studies focusing specifically on salmon.10,20,28 This research has highlighted the importance of feed use and feed ingredient choice as drivers of environmental impacts.10,26,28 Also important is the recognition of the potentially large impacts that can be associated with energy carriers (e.g., electricity and fuel), particularly for “closed” systems such as RAS.20,27,29 Despite the breadth of LCA research undertaken on production of salmonids to-date, an analysis of the potential benefits and limitations of a solid-walled floating system (such as the SWAS) has not previously been reported. Moreover, to our knowledge this is the first published LCA of Chinook salmon. We also report two new feed

2.0. MATERIAL AND METHODS 2.1. Life Cycle Assessment. LCA is a biophysical assessment technique which is used to compile an inventory of the energy and material inputs/outputs associated with the life cycle of a good or service,31 and to quantify how these contribute to a defined suite of resource use and environmental impacts. It is typically described as a four step process that includes: definition of goal and scope, life cycle inventory (LCI), life cycle impact assessment (LCIA) and interpretation.32 The goal and scope of this study was to assess the life cycle environmental impacts associated with production of one live-weight tonne of salmon in the SWAS from cradle to farm-gate. This includes upstream and background processes required for production until salmon are harvested. It does not include postharvest processing, packaging, retail, use, or end-of life stages. 2.2. Life Cycle Inventory. The SWAS is a floating aquaculture system with the internal environment separated from the external by a solid wall. Influent water was pumped into the tank from depth and supplemented with oxygen. Salmon were grown from eggs for approximately one year at a hatchery. Upon smoltification they were transferred into the SWAS for grow-out. Here they were fed a compound feed (composed of 60% fish meal, 20% fish oil, and 20% wheat) until harvest. A grid connected waste capture system was used to intercept solid and semisolid material from the effluent, removing an estimated 5% of solids. Captured wastes were dewatered, pumped onto land, and composted. Inventory models were constructed using foreground operational data that encompassed all major subsystems required to produce salmon from cradle to farm-gate (Figure 1). Major subsystems were delineated based on relevant operational practices including provision of feed; production of juveniles (smolts); production and assemblage of SWAS infrastructure; on-site energy use; transportation of materials and; on-site emission of nutrients. To characterize the SWAS production cycle, operational inputs were elicited from AgriMarine staff; data/reports supplied by the company and; on-site inventory audits. Data on B


Article


Table 1. Contribution Analysis of Life Cycle Impacts Associated with the Production of One Live-Weight Tonne of Chinook Salmon in a Solid-Wall Aquaculture System (SWAS) for the Actual Production Cycle (APC) and Intended Production Cycle (IPC)

APC on-site emissions feed production production of juveniles (smolts) transportation infrastructure on-site energy use total IPC on-site emissions feed production production of juveniles (smolts) transportation infrastructure on-site energy use total

GWP (kg CO2 eq)

(%)

AP (kg SO2 eq)

(%)

MEP (kg N eq)

(%)

CEU (MJ)

(%)

BRU (Mg C)

(%)

0 2318 429 116 486 524 3874

0.0 59.8 11.1 3.0 12.5 13.5 100.0

0.0 14.60 7.22 0.59 2.09 0.99 25.5

0.0 57.2 28.3 2.3 8.2 3.9 100.0

67.10 1.40 0.66 0.03 0.09 0.05 69.30

96.8 2.0 1.0 0.0 0.1 0.1 100.0

0 36 324 5669 2009 9040 38 558 91 600

0.0 39.7 6.2 2.2 9.9 42.1 100.0

0 1429 0 0 0 0 1429

0 100 0 0 0 0 100

0 2180 189 106 214 337 3025

0.0 72.1 6.2 3.5 7.1 11.1 100.0

0.00 13.70 3.17 0.48 0.92 0.67 18.90

0.0 72.3 16.7 2.5 4.9 3.5 100.0

54.60 1.30 0.29 0.02 0.04 0.04 56.2

97.0 2.3 0.5 0.0 0.1 0.1 100.0

0 34 156 2491 1830 3972 24 616 67 064

0.0 50.9 3.7 2.7 5.9 36.7 100.0

0 1344 0 0 0 0 1344

0 100 0 0 0 0 100

alternative aquaculture technologies.20 While BRU is a somewhat novel impact category, it has been widely adopted in seafood system LCAs as it is highly sensitive to differences in trophic levels of biological inputs to production systems.3 Contributions to GWP (expressed as kg CO2-eq), AP (kg SO2-eq), and MEP (kg N-eq) were quantified using the Recipe Midpoint version 1.07 impact assessment method.37 Contributions to CEU (MJ eq) were quantified using Cumulative Energy Demand version 1.05;33 and BRU (Mg C) were quantified following previous practice in recent seafood LCAs.10,26,29 All of the calculations to conduct the impact assessment were facilitated by the use of SimaPro 7.1 from PRé consultants.33 2.4. Sensitivity, Uncertainty, and Scenario Analysis. Given the potential importance of uncertainty in LCA modeling in general, the novel nature of the SWAS, and limited real-world operational experience from which typical operational conditions could be modeled, effort was made to assess the effect of key input parameters on results. Additionally, since the IPC was modeled using known relationships, operations data, and expert opinion, there is the added uncertainty associated with the model itself and how well it approximates reality. Uncertainty was addressed in this study using sensitivity analysis while Monte Carlo simulation was used to characterize the cumulative effect of sources of uncertainty on model outputs (see SI). Sensitivity analysis was used to evaluate the effect of using alternative values for key inputs or methodological approaches including (1) the ratio of total feed used to grow one tonne of salmon at harvest, known generally as the economic feed conversion ratio (eFCR); (2) yields of fish meal and oil from wild-caught species (percentage of fish meal/oil obtained from the reduction of 1 kg of fish); (3) trophic level of fish species used in feed production; (4) life-expectancy for tank infrastructure; and (5) allocating impacts among coproducts by mass rather than nutritional energy. Monte Carlo simulation was used to assess the cumulative impact of operational foreground data as well as background data on some fisheries related inputs (a total of 16 variables using 10 000 runs, see Table S10 in SI). In addition to the impact of uncertainty and modeling choices, it is useful to assess how the study context could alter potential impacts of the SWAS. This was achieved using scenario analyses to evaluate the potential implications of (1) using a feed

background systems (e.g., resource extraction, transportation) were taken from peer-reviewed data sets (e.g., EcoInvent database v.2.2) accessible through SimaPro33 and adapted to the current context. Detailed descriptions of inventory data appear in SI. Unfortunately, as the SWAS is an emerging technology, operational data were only available from a single tank over most of a production cycle. The operational horizon from which data were available was truncated by a storm that compromised structural integrity of the tank after 13.5 months, resulting in the early harvest of salmon. This storm was abnormally powerful, with wind speeds over 130km/h, and exerted forces greater than what the SWAS was designed to withstand. Consequently, the data compiled do not represent the intended production of salmon but were used to model the actual production cycle (APC) based on the reality of production as it occurred. A model of the intended production cycle (IPC) was constructed in parallel based on APC data but modified as appropriate after consultation with industry experts and reviewing peer-reviewed literature to reflect more typical or expected salmonid growth to harvest parameters (see SI for detailed description of both APC and IPC). The purpose of modeling the IPC is to approximate the life cycle impacts of the SWAS as would be expected if there had not been a major storm event that caused system failure. Given the multioutput nature of many salmon feed subsystems, there was a need to address “coproduct allocation”. Coproduct allocation arises when there are two or more coproducts of a given process that are utilized somewhere in the technosphere.30,34 Here, allocation of inputs and resulting environmental burdens among coproducts was modeled in proportion to the nutritional energy embodied in each, following the rationale offered by Pelletier & Tyedmers.35 2.3. Life Cycle Impact Assessment. A problem oriented (midpoint) approach was used to evaluate the performance of the SWAS. Five impact categories were chosen: global warming potential (GWP), acidification potential (AP), marine eutrophication potential (MEP), cumulative energy use (CEU), and biotic resource use (BRU). These impact categories represent issues of concern for intensive aquaculture systems that have been highlighted by previous works,3,29,36 particularly GWP, AP, and CEU, which have been found to be a large concern in other C


Article


Table 2. Sensitivity and Scenario Analyses for Life Cycle Impacts Per Tonne of Live-Weight Salmon Produced in the SWAS, Including Monte Carlo Simulation and the Relative Change (%) from the Intended Production Cycle (IPC)

IPC sensitivity analysis mass allocation among feed inputs eFCR 1.1:1 eFCR 1.5:1 high oil/meal yielda low oil/meal yieldb low trophic levelc 10 year life expectancy IPC scenario analysis industry average Feed electricity Canada electricity NS Monte Carlo simulation mean median standard deviation

GWP (kg CO2 eq)

(%)

AP (kg SO2 eq)

(%)

3144 2573 3238 2707 3344 3025 3119

+3.9 −15 +7.1 −12 +11 +0 +3.1

20.1 16.1 20.3 17.4 20.5 18.9 19.3

+6.1 −15 +6.9 −8.1 +8.1 +0 +1.7

3167 4505 8245

+4.7 +49 +173

106.9 23.8 43.7

3080 3030 550

+464 +26 +131

19.2 18.8 1.98

MEP (kg N eq)

(%)

CEU (MJ)

(%)

BRU (Mg C)

(%)

56.3 38.2 64.7 56.2 56.3 56.2 56.3

+0.1 −32 +15 −0.2 +0.2 +0 +0

68 783 59 943 70 420 62 030 72 117 67 065 69 095

+2.6 −10.6 +5 −7.5 +7.5 +0 +3

1157 1078 1470 1114 1575 215 1344

−14 −20 +9.3 −17 +17 −84 1

68.2 56.4 57.1

+21 +0.3 +1.5

46 636 74 595 92 083

−31 +11 +37

15 1344 1344

−99 +0 +0

56.3 56.2 0.242

67 840 67 532 7598

711 578 453

a Yield of fish oil (5%) and fish meal (19%) from wild fisheries increased to 6% and 23% respectively. bYield of fish oil and fish meal from wild fisheries decreased to 4% and 17% respectively. cTrophic level for Pacific hake (4.4) and Pacific herring (3.2) decreased to 3.6 and 2.8 respectively.

(0.1%). Production of smolts was the second most important contributor to AP (28.3%) and fourth most important for GWP (11.1%). Infrastructure was a minor contributor to GWP (12.5%), AP (8.2%), and CEU (9.9%). Transportation plays only a minor role in all impact categories, the largest of which is GWP (3.0%) (Table 1). For all impact categories reported, the relative contribution of feed and transportation subsystems increased in the IPC compared to the APC while impacts associated with infrastructure, smolt production, and on-site energy use decreased. This occurred because feed provision and transport increased proportionally with increased production mass in the IPC. In contrast, total energy inputs increased but at a lower rate than fish production, resulting in a reduction of energy use relative to the functional unit. This was due to relatively static operational requirements associated with pumping and oxygenation of influent water, even as fish biomass increased. Total inputs associated with infrastructure and smolt production remain unchanged between the APC and the IPC, thus their relative contribution to impacts decreased in proportion to the increased production, reducing their relative importance. BRU is exclusively influenced by the provision of feed in both APC and IPC models, thus there is no change in relative contribution between the two. Comparison of total life cycle impacts from the IPC and APC revealed that for all impact categories reported, IPC resulted in lower impacts per tonne of live-weight salmon, with reductions ranging from 6 to 26% (Table 1). This result is not surprising given that feed production is the single largest driver of life cycle impact and the improved eFCR used to model the IPC represents approximately a 6% reduction in feed use per tonne of salmon harvested (SI Table S1). Consequently, the BRU impact category associated with IPC is reduced a comparable amount as it is driven exclusively by feed production. 3.3. Sensitivity, Uncertainty, and Scenario Analyses. Sensitivity analysis demonstrated the importance of values used in determining life cycle impacts, particularly those related to provision of feed. For example, when inputs to feed coproducts were allocated on the basis of mass, reductions of 13.9% to BRU

formulation representative of the Canadian industry average (SI Table S2); and (2) drawing electricity from alternative electrical grid mixes. Electrical grid mixes representative of the provinces of British Columbia (BC), Nova Scotia (NS), and the Canadian average were used. These are interesting because of the different sources from which they generate electricity; in NS the primary source of electricity generation is coal-fired combustion, in BC it is primarily hydro-power, while in Canada as a whole there is a diversity of electrical generating methods including hydro, coal, biomass, petroleum, natural gas and nuclear (SI Table S8). Sensitivity and scenario analyses along with the Monte Carlo simulation were applied only to the IPC as this was considered most relevant to the future applications of the SWAS, as it theoretically provides the best representation of a complete production cycle. See SI for a detailed explanation of choices.

3.0. RESULTS 3.1. Life Cycle Inventory. The SWAS was stocked with an estimated 56 108 Chinook salmon (Oncorhynchus tshawytscha) smolts, with an average mass of 35 g. A total of 43 366 fish were harvested with an average mass of 1.73 kg. The eFCR for salmon was 1.5:1 and 1.459:1 during the hatchery and grow-out phases, respectively. The mortality rate throughout grow-out was 17.8% (excluding mortalities at hatchery and escapes that occurred after the system was compromised by the storm). No antibiotics or antifouling paints were used during grow-out. A description of data collection activities and additional inventory data including a summary of key LCI data (Table S1) and detailed feed formulations (Table S2) are provided in SI. 3.2. Life Cycle Impact of Solid Wall Aquaculture System. Results of the contribution analysis indicate that for the APC, feed production was the most important driver of GWP, AP, and BRU, with major contributions to CEU (39.7%) (Table 1). The energy use subsystem, associated with on-site use of electricity and fossil fuels, was the primary contributor to CEU (42.1%) and second largest contributor to GWP (13.5%). Contributions to MEP were dominated by the on-site emission of nutrients (96.8%). Other contributors to MEP included feed production (2.0%), smolt production (1.0%), and infrastructure D


Article


reinforce previous findings regarding the importance of feed when addressing environmental impacts of aquaculture systems.10,20,26,27 With the exception of MEP, provision of feed contributed more than 39% for the APC and 50% for the IPC to all impact categories reported (Table 1). However, the low contribution of feed to MEP is somewhat misleading; on-site emissions (that contribute the majority of MEP impacts) are a direct result of the provision of feed to fish and subsequent emissions as uneaten losses or excreted wastes. From this perspective, “on-site emissions” and feed-related impacts are intimately linked and the provision of feed can be considered an important factor contributing to MEP.26 This distinction is important because it shows that reductions in feed use will actually reduce MEP. This is demonstrated in the sensitivity analysis of eFCR which shows that when eFCR is improved to 1.1:1 (from 1.37:1) modeled MEP is reduced by approximately 32% (Table 2). Changing the ingredients used in production of feed can substantially alter impacts of aquaculture. Fish meal and oil in particular are important contributors to the overall impact of fish feeds.9,13,22 Thus, reducing fish meal/oil in feed or replacing it with less impactful marine-derived substitutes can result in significant improvements in performance.9 For example, sourcing fish meal/oil from species with high yields (see high/ low yield scenarios Table 2), low fuel use in their capture,39 or low trophic levels can all result in lower impacts. Another option includes using agricultural products (including crop and/or animal-derived ingredients) in the place of fishery-derived ingredients.9,10 While this holds promise as a potential avenue to reduce impacts, research to date has yet to find consistent improvement in any impact category other than BRU.13,14 Indeed, many low-impact marine ingredients outperform other high-impact crop and livestock ingredients, indicating that it is perhaps more important to focus on the impacts of specific ingredients rather than whether they are of marine or agricultural origin. Important characteristics of those ingredients include yield, nitrogen fertilizer application rates, fuel use required for production and the geographic area of production.14 At a coarse scale, the impact of changing ingredients is illustrated in the industry average feed scenario that involved substantial reductions, relative to the actual diet fed, in the marine-derived fraction and a consequent increase in both crop and poultry-derived inputs (SI Table S2). In sum, this formulation increased GWP, AP, and MEP; while CEU and BRU were reduced. Noteworthy is the large increase in AP and the large decrease in BRU associated with increased inclusion of poultry byproducts and agricultural products, respectively. The poultry products increase AP because of large nitrogen emissions associated with management of poultry excrement while BRU is reduced because of the low trophic level of species used and the high inclusion rate of agricultural crop-based inputs. This result reinforces the need to address environmental impacts of feeds by ingredients, while simultaneously demonstrating trade-offs. The role of on-site energy use in contributing to the environmental impacts of SWAS is highly dependent on the primary source from which electricity is generated. Impacts of electricity use are not especially large in the current application of SWAS because of the high proportion (∼90%) of hydroelectricity in the BC grid. Scenarios using electrical grids representative of average Canadian and NS primary energy inputs, however, dramatically increase contribution of electricity to GWP, AP, and CEU. In the NS electrical grid scenario (nearly 60% coal-fired) electricity overtakes feed provision as the

and increases in GWP, AP, MEP, and CEU from 0.1 to 6.1% resulted (Table 2). While the production of feed does not contribute substantially to MEP, uneaten feed and wastes excreted by the salmon do. Thus, lowering the eFCR to 1.1:1 resulted in a 32% reduction in MEP impacts (together with reductions of between 11 and 15% across other impact categories) while increasing eFCR to 1.5:1 increased MEP impacts by 15% (and all other impacts between 5 and 7%). Increasing fish oil and fish meal yield rates from 5% to 6% and 20% to 23% respectively reduced impacts at farm-gate by 0.2− 17.1% while lowering yields a comparable increment increased impacts a similar amount. Reducing the trophic level of both Pacific herring (Clupea pallasii) and Pacific hake (Merluccius productus), the species from which byproduct fish meal were derived, by the standard error provided in FishBase (± 0.8 and 0.4 respectively)38 reduced the overall BRU by 84.0% but did not affect other impact categories. Halving the life-expectancy of tank infrastructure, from 20 years to 10 years, contributed to only minor increases in GWP, AP, and CEU (1.7−3.1%) (Table 2). Investigating the impact of differing scenarios on life cycle impact produced dramatic differences (Table 2). Replacing the actual feed used (20% wheat, 20% fish oil, 60% fish meal) with an industry-wide composite resulted in a minor increase in GWP (+4.7%); a moderate reduction in CEU (−31%); major reductions in BRU (−99%); a moderate increase in MEP (+21%); and a substantial increase in AP (+464%). Reductions in CEU and BRU are reflective of the lower fishery-derived ingredients in this feed. The massive increase in AP results from emissions of ammonia associated with rearing the poultry required to provide poultry byproducts and byproduct meals that make up ∼28% of the industry average diet (SI Table S2). Providing grid supplied electricity to the SWAS from energy mixes representative of average Canadian and Nova Scotia grids in 2012 resulted in moderate to substantial increases in GWP, AP, and CEU. This results from a larger proportion (13% for Canada as a whole and 82% for Nova Scotia) of generation powered by fossil fuel combustion in these electrical mixes. The effect of multiple sources of uncertainty on life cycle impact of the IPC for the SWAS was evaluated using Monte Carlo simulation using 10 000 runs (Table 2 and SI Table S11). The distributions around mean values in all impact categories showed considerable variation with coefficient of variation (CV) ranging from 0.43 to 63.7%. Comparing results with those of the IPC indicate that average GWP, AP, and CEU impacts increased while MEP and BRU decreased when all 16 uncertainty parameters are simultaneously modeled. Encouragingly, these results are not significantly different from the results of the IPC. Comparing results of individual scenario analyses with that of the Monte Carlo simulation, we see that each of the scenarios does result in some statistically significant change given the uncertainties modeled in the simulation. For the average Canadian feed scenario this includes increased AP and MEP, with reduced CEU and BRU. For both electricity grid mix scenarios the GWP and AP are significantly increased. CEU however, is only significantly increased in the electricity NS scenario, while the electricity Canada scenario falls within the standard deviation.

4.0. DISCUSSION 4.1. Growing Salmon in a Solid Wall Aquaculture System. Employing LCA to analyze the SWAS allows environmental hotspots to be identified and aids in finding ways to reduce environmental impacts. Results of the LCIA E


Article


reduced.43 This has the added benefit of improving eFCR, thus likely contributing to improved environmental performance on other impact categories influenced by provision of feed. When choosing feed ingredients or implementing management changes to reduce impacts highlighted by LCA research care must be taken for two reasons. First, changing ingredients may influence the performance of fish feed; if lower-impact feed formulations result in decreased growth of the cultured fish, then gains in environmental performance of the feed can be offset by increased eFCR. Second and perhaps more importantly, there is potential that choosing ingredients that perform well according to the impact categories being assessed will result in problemshifting whereby other problems not accounted for by the methodology are aggravated. For example, agriculture is often considered one of the most environmentally damaging activities globally,44,45 yet LCA methodology is currently unable to adequately address many related concerns such as its impacts on biodiversity.30 Likewise, depletion of fish stocks and the consumption of otherwise edible fisheries products are critically important social and environmental issues associated with the production of fish meals and fish oils46 that are not meaningfully addressed in any widely accepted LCA impact categories.47 For these reasons, decision-making based solely on LCA results could aggravate issues (environmental, social, political, etc.) for which the tool is not well suited or capable of assessing and it should only be used as a decision supporting tool.10,48,49 4.2. Actual Production Cycle (APC) Versus Intended Production Cycle (IPC). Ideally, in an LCA characterizing the typical production of a studied system, data from multiple growth cycles would be used to obtain average operational inputs/ outputs. Unfortunately the SWAS only operated 13.5 months; therefore, a model was constructed to approximate a complete production cycle (the IPC). Results indicate that for impact categories assessed, the IPC performed better than the APC, demonstrating the importance of increased production and the effect of losses (in this case due to escapes) as a means of improving life cycle environmental impacts per functional unit of production. It is hoped that the IPC provides a reasonable representation of typical production in the SWAS should it continue to operate in the future. However, there are two primary assumptions underlying this model: (1) the failure of the SWAS caused by the storm was a chance event that is not likely to reoccur and (2) the data collected for the APC are representative of future operations. Both assumptions are potentially false but the second in particular is problematic. This is because due to the complex nature of aquaculture operations there is typically a learning curve associated with management, especially for a novel system such as the SWAS. It is expected that as managers gain experience, issues such as prevalence of disease, feed losses, and animal stress will be reduced. This in turn is likely to be reflected in improved system parameters such as eFCR and mortality rate. Second, there are many factors which can affect production that are difficult to predict (e.g., weather, disease, technical malfunctions, etc.). The upshot of these is that they contribute to the already uncertain nature of results (see section 2.4). For this reason the results found here represents only a snapshot in time of an emerging technology rather than the average performance of a mature technology. Uncertainty is an area of active research in LCA;30 it has currently been applied in only a limited number of aquaculture LCAs.40,50 To characterize the cumulative uncertainty of the IPC, Monte Carlo simulation was performed. Results revealed

dominant contributor to GWP and AP. This mirrors similar results of other researchers who have found electricity generation by coal, oil, and natural gas contribute more to many impact categories than do sources such as hydro, wind, and solar among others.20,40 While the electrical grid to which aquaculture systems are connected is not typically something producers can control, it can be considered during site selection. Additionally, for producers in areas with high impact sources of electricity, the reduction of electricity use or installation of renewable energy technologies offers potential for environmental improvement. The life cycle environmental impacts attributable to the production of smolts are typically very low in salmon production.10,20 Here smolts contributed as much as 28.3% and 16.7% to reported impact categories for the APC and IPC respectively, an unusually large amount. This occurs because the hatchery providing smolts was not connected to the grid and all electricity used in their production was produced on-site using a diesel generator (SI Table S1). Infrastructure required to construct the SWAS contributed as much as 11.5% and 6.4% of overall impacts in the APC and IPC respectively. In salmon aquaculture, infrastructure typically does not contribute a large amount to life cycle impacts as the inputs/ outputs associated with it are distributed over the lifetime production of the system, rather than for a particular production cycle. This is demonstrated in the scenario where the lifeexpectancy of tank infrastructure is reduced from 20 to 10 years, increasing the impacts associated with infrastructure by 50% (which, at farm-gate resulted in an overall increase from 0 to 3.1%, Table 2). Contribution analysis indicates that the largest contribution from transportation was to GWP: 3.0% and 3.5% for APC and IPC respectively (Table 1). This result is broadly consistent with agricultural production in general, wherein greenhouse gas emissions are dominated by the production phase of the life cycle.41 In fact, distance of travel is often a poor indicator of overall environmental impacts because of its narrow focus that ignores most of the supply chain associated with food production.42 On-site emissions were of negligible concern with the exception of MEP where they contribute to over 98% of impacts. With moderate improvements to current waste-capture technology it is estimated that a solid waste-capture rate of 50−75% can be achieved (personal communication, Mr. Todd Adamson, April 2013). Due to the importance of on-site emissions in MEP, this would represent a significant improvement over the current system and warrants further investigation. Should this prove feasible, it might also be possible to use captured waste as fertilizer, potentially replacing the production of small amounts of conventional fertilizer. In this case, the environmental benefit associated with this substitution would need to be accounted for. However, it must also be noted that if improved waste-capture technology were employed it would result in increased energy use to operate the waste-capture system (personal communication, Mr. Todd Adamson, April 2013), increasing impacts of onsite energy use. A significant challenge in the reduction of MEP is the ability to capture soluble nutrients in effluent. While the current wastecapture technology is unable to do this, there are several other techniques in various stages of development which can (e.g., nitrification or denitrification reactors, constructed wetlands, algal systems).18 Alternatively, a strategy for reducing MEP without waste-capture involves optimizing nutritional management, whereby nutrient uptake is increased and emissions F


Article


technology (e.g., reduction in local ecological impacts such as transmission of parasites and disease to wild conspecifics) are not readily investigated through LCA methodology. For these reasons, it is currently impossible to draw definitive conclusions regarding the broader sustainability of the SWAS relative to other systems. Our research suggests reducing the impacts associated with provision of feed and energy use through increased efficiency and sourcing lower impact inputs are areas of importance for the SWAS. Nutritional management strategies can also be targeted to improve feed conversion and reduce the loss of undigested nutrients. This has the added benefit of reducing MEP associated with nutrients lost in fish wastes. Another method for reducing MEP is to improve waste-capture technology, especially the development of methods for removing soluble nutrients from effluent. However, as with many technologies, it is possible that such improvements result in environmental trade-offs that require careful assessment to determine optimal application. Without such robust analyses (e.g., LCA) it is probable that overly optimistic, and potentially unjustified impressions of their ability to mitigate environmental concerns will go unchallenged.

the effect of multiple sources of uncertainty on the life cycle impact of the SWAS (SI Table S11). The coefficient of variation for GWP, AP, and CEU were all above 10%, while it was below 1% and above 60% for MEP and BRU respectively. Due to this uncertainty and other sources of uncertainty not accounted for, results obtained here should be interpreted with caution. This is particularly the case with BRU which, based on the parameters modeled here, exhibits very large uncertainty. This is largely attributable to the uncertainty in trophic level of Pacific herring and Pacific hake used in the fish meal, as it exerts a dramatic influence on BRU as calculated. The Monte Carlo simulation used here is a useful tool for characterizing the range and effect of uncertainty based on known parameters but it was limited in two important ways. First, it only characterized uncertainty based on known variables, thus parameters which have not been assigned uncertainty were not accounted for. While effort was made to parametrize variables which were likely to contribute to uncertainty in important ways (e.g., foreground processes and fisheries related data), many variables were ignored (See SI, Monte Carlo simulation). Individually these are not likely to meaningfully change the results but cumulatively it is possible that they could have a significant impact. This is exemplified by the small range of uncertainty found for MEP. Given the considerable research demonstrating uncertainty regarding the amount, fate, and impact of eutrophying emissions in agriculture,51,52 it is clear that this result is a reflection of modeling decisions rather than the reality of the situation. Second, the parameters modeled here do not account for all possible events that can occur. This is because it is very difficult to predict stochastic events. Most notably, this includes events that might partially or completely compromise the system, such as a disease outbreak that causes complete mortality. Indeed, the storm event that compromised the actual production of the SWAS is an example of these. Other possible events including drought, stock collapse, climate change, etc. could alter the impacts associated with background inputs to the system. A challenge for future research is to improve the application of uncertainty in LCA so that these issues are minimized and/or stochasticity is incorporated. 4.3. Comparison with Net-Pens and Other Aquaculture Technologies. Debate concerning the appropriate method for culturing salmonids is widespread.15,17,53 At issue, in large part, is the relative environmental sustainability of culture techniques coupled with the values different members of society place on various forms and extent of impact. Commercial scale projects are currently underway to explore the feasibility of alternative culture technologies such as RAS;54,55 the SWAS is a unique example of these. While many of these technologies are designed specifically to limit or avoid local ecological and environmental impacts, it is useful to simultaneously consider the global and regional environmental impacts of these projects as we have done for the SWAS. The results found here represent another piece of evidence that can inform the comparison of various technologies (see Table S12 in SI for a summary of LCA results for salmonid aquaculture systems). At first glance, these data indicate that in many respects the SWAS results in intermediate life cycle impacts between net-pens and other alternatives that have been modeled (e.g., RAS and raceways). However, the large uncertainty in results and limited data availability for the SWAS together with contextual differences in LCA research on other systems make it difficult to draw definitive conclusions on their relative environmental performance. Moreover, many of the impacts associated with aquaculture and potential benefits of this

■

ASSOCIATED CONTENT

S Supporting Information *

A detailed list of data regarding subsystems and input parameters is included in Supporting Information. Additionally, details on calculations and assumptions are provided. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Thank you to all those who provided data for this project including: Dr. Brad Hicks, Dr. Jason Mann, Dr. Scott Wallace, Rob Morley, and Todd Adamson. A special thanks to Rob Walker for his continued support of this project, even when there were major hiccups. Funding for this project was generously provided by Agrimarine Industries, Mitacs accelerate, NSERC, Tides Canada, the Gordon and Betty Moore Foundation, and the Vancouver Aquarium.

■

ABBREVIATIONS AP acidification potential APC actual production cycle BC British Columbia BRU biotic resource use CEU cumulative energy use CO2eq carbon dioxide equivalents GHG greenhouse gas GWP global warming potential hp horse power IPC intended production cycle ISO International Organization for Standardization G


Article

Environmental Science & Technology kg kWh LCA LCIA LCI MEP MJ N NS P PO4eq RAS SO2eq SWAS WC

■

intensive floating tank system (SIFTS). Aquacult. Eng. 2006, 35, 109− 121. (17) Tal, Y.; Schreier, H. J.; Sowers, K. R.; Stubblefield, J. D.; Place, A. R.; Zohar, Y. Environmentally sustainable land-based marine aquaculture. Aquaculture 2009, 286, 28−35. (18) Martins, C. I. M.; Eding, E. H.; Verdegem, M. C. J.; Heinsbroek, L. T. N.; Schneider, O.; Blancheton, J. P.; d’Orbcastel, E. R.; Verreth, J. a. J. New developments in recirculating aquaculture systems in Europe: A perspective on environmental sustainability. Aquacult. Eng. 2010, 43, 83−93. (19) Wolters, W.; Masters, A.; Vinci, B.; Summerfelt, S. Design, Loading, and Water Quality in Recirculating Systems for Atlantic Salmon (Salmo Salar) at the USDA ARS National Cold Water Marine Aquaculture Center (Franklin, Maine). Aquac. Eng. 2009, 41, 60−70. (20) Ayer, N.; Tyedmers, P. Assessing alternative aquaculture technologies: Life cycle assessment of salmonid culture systems in Canada. J. Clean. Prod. 2009, 17, 362−373. (21) Aubin, J.; Papatryphon, E.; van der Werf, H. M. G.; Petit, J.; Morvan, Y. M. Characterisation of the environmental impact of a turbot (Scophthalmus maximus) re-circulating production system using life cycle assessment. Aquaculture 2006, 261, 1259−1268. (22) Pelletier, N.; Tyedmers, P. Life cycle assessment of frozen tilapia fillets from indonesian lake-based and pond-based intensive aquaculture systems. J. Ind. Ecol. 2010, 14, 467−481. (23) Costello, M. Ecology of sea lice parasitic on farmed and wild fish. Trends Parasitol. 2006, 22, 475−483. (24) Heuch, P.; Parsons, A.; Boxaspen, K. Diel vertical migration: A possible host-finding mechanism in Salmon Louse (Lepeophtheirus salmonis) copepodids? Can. J. Fish. Aquat. Sci. 1995, 52, 681−689. (25) Mungkung, R.; Gheewala, S. H. Use of life cycle assessment (LCA) to compare the environmental impacts of aquaculture and agrifood products. In Comparative Assessment of the Environmental Costs of Aquaculture and Other Food Production Sectors: Methods for Meaningful Comparisons; Bartley, D., Brugere, C., Soto, D., Gerber, P., Harvey, B., Eds.; FAO: Rome, 2007; pp 87−96. (26) Jerbi, M. A.; Aubin, J.; Garnaoui, K.; Achour, L.; Kacem, A. Life cycle assessment (LCA) of two rearing techniques of sea bass (Dicentrarchus labrax). Aquacult. Eng. 2012, 46, 1−9. (27) D’Orbcastel, R. E.; Blancheton, J.-P.; Aubin, J. Towards environmentally sustainable aquaculture: Comparison between two trout farming systems using life cycle assessment. Aquacult. Eng. 2009, 40, 113−119. (28) Ellingsen, H.; Aanondsen, S. A. Environmental impacts of wild caught cod and farmed salmonA comparison with chicken. Int. J. Life Cycle Assess. 2006, 11, 60−65. (29) Aubin, J.; Papatryphon, E.; van der Werf, H. M. G.; Chatzifotis, S. Assessment of the environmental impact of carnivorous finfish production systems using life cycle assessment. J. Clean. Prod. 2009, 17, 354−361. (30) Finnveden, G.; Hauschild, M. Z.; Ekvall, T.; Guinée, J.; Heijungs, R.; Hellweg, S.; Koehler, A.; Pennington, D.; Suh, S. Recent developments in life cycle assessment. J. Environ. Manage. 2009, 91, 1−21. (31) Heijungs, R.; Huppes, G.; Klein, R.; de Koning, A.; van Oers, L.; Sleeswijk, A. W.; Suh, S.; de Haes, H.; de Bruijn, H.; van Duin, R. Handbook on Life Cycle Assessment: Operational Guide to the ISO Standards; Guinée, J. B., Ed.; Kluwer Academic Publishers: Dordrecht, Netherlands, 2002; p 704. (32) Environmental ManagementLife Cycle AssessmentPrinciples and Framework; International Organization for Standardization: Geneva, Switzerland. 2006; http://www.iso.org/iso/catalogue_ detail?csnumber=37456. (33) PRé Consultants. SimaPro. http://www.pre.nl/simapro/default. htm. (34) Ayer, N.; Tyedmers, P.; Pelletier, N.; Sonesson, U.; Scholz, A. Coproduct allocation in life cycle assessments of seafood production systems: Review of problems and strategies. Int. J. Life Cycle Assess. 2007, 12, 480−487.

kilogram kilowatt-hour life cycle assessment life cycle impact assessment life cycle inventory marine eutrophication potential megajoule nitrogen Nova Scotia phosphorus phosphate equivalents recirculating aquaculture system sulfur dioxide equivalents solid-wall aquaculture system waste-capture

REFERENCES

(1) The State of World Fisheries and Aquaculture; FAO: Rome, 2012; p 209. (2) De Silva, S. S. Aquaculture: A newly emergent food production sectorand perspectives of its impacts on biodiversity and conservation. Biodiversity Conserv. 2012, 21, 3187−3220. (3) Henriksson, P. J. G.; Guinée, J. B.; Kleijn, R.; de Snoo, G. R. Life cycle assessment of aquaculture systemsA review of methodologies. Int. J. Life Cycle Assess. 2011, 17, 304−313. (4) Olsen, L. M.; Holmer, M.; Olsen, Y. Perspectives of nutrient emission from fish. Aquaculture Coast. Waters 2008, 87. (5) Folke, C.; Kautsky, N.; Troell, M. The costs of eutrophication and salmon farming: Implications for policy. J. Environ. Manage. 1994, 40, 173−182. (6) Krkosek, M.; Lewis, M.; Morton, A.; Frazer, L. N.; Volpe, J. P. Epizootics of wild fish induced by farm fish. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 15506−15510. (7) McGinnity, P.; Prodöhl, P.; Ferguson, A.; Hynes, R.; Maoiléidigh, N. O.; Baker, N.; Cotter, D.; O’Hea, B.; Cooke, D.; Rogan, G.; et al. Fitness reduction and potential extinction of wild populations of Atlantic Salmon, Salmo salar, as a result of interactions with escaped farm salmon. Proc. Biol. Sci. 2003, 270, 2443−2450. (8) Leggatt, R. A.; O’Reilly, P. T.; Blanchfield, P. J.; Mckindsey, C. W.; Devlin, R. H. Pathway of Effects of Escaped Aquaculture Organisms or Their Reproductive Material on Natural Ecosystems in Canada; Fisheries and Oceans Canada. Scientific Advisory Secretariat Research Document, 2010; p 70. (9) Pelletier, N.; Tyedmers, P. Feeding farmed salmon: Is organic better? Aquaculture 2007, 272, 399−416. (10) Pelletier, N.; Tyedmers, P.; Sonesson, U.; Scholz, A.; Ziegler, F.; Flysjo, A.; Kruse, S.; Cancino, B.; Silverman, H. Not all salmon are created equal: Life cycle assessment (LCA) of global salmon farming systems. Environ. Sci. Technol. 2009, 43, 8730−8736. (11) Aubin, J.; van der Werf, H. M. G. Fish farming and the environment: A life cycle assessment approach. Cah Agric. 2009, 18, 220−226. (12) Tyedmers, P. Salmon and Sustainability: The Biophysical Cost of Producing Salmon through the Commercial Salmon Fishery and the Intensive Salmon Culture Industry; University of British Columbia, 2000; p 258. (13) Papatryphon, E.; Petit, J.; Kaushik, S. J.; van der Werf, H. M. G. Environmental impact assessment of salmonid feeds using life cycle assessment LCA. Ambio 2004, 33, 316−323. (14) Boissy, J.; Aubin, J.; Drissi, A.; van der Werf, H. M. G.; Bell, G. J.; Kaushik, S. J. Environmental impacts of plant-based salmonid diets at feed and farm scales. Aquaculture 2011, 321, 61−70. (15) Chadwick, M. P.; Parsons, G. J.; Sayavong, B. Evaluation of ClosedContainment Technologies for Saltwater Salmon Aquaculture; National Research Council: Ottawa, Canada, 2010; p 160. (16) Partridge, G. J.; Sarre, G. a.; Ginbey, B. M.; Kay, G. D.; Jenkins, G. I. Finfish production in a static, inland saline water body using a semiH


Article

Environmental Science & Technology (35) Pelletier, N.; Tyedmers, P. An ecological economic critique of the use of market information in life cycle assessment research. J. Ind. Ecol. 2011, 15, 342−354. (36) Pelletier, N.; Ayer, N. W.; Tyedmers, P. H.; Kruse, S. a.; Flysjo, A.; Robillard, G.; Ziegler, F.; Scholz, A. J.; Sonesson, U. Impact categories for life cycle assessment research of seafood production systems: Review and prospectus. Int. J. Life Cycle Assess. 2006, 12, 414−421. (37) Goedkoop, M.; Heijungs, R.; Huijbregts, M.; De Schryver, A.; Struijs, J.; van Zelm, R. ReCiPe 2008: A Life Cycle Impact Assessment Method Which Comprises Harmonized Category Indicators at the Midpoint and Endpoint Level; Ministry of Housing, Spatial Planning and Environment: Netherlands, 2009; p 126. (38) Froese, R.; Pauly, D. FishBase Website. www.fishbase.org. (39) Tyedmers, P. Fisheries and energy use. In Encyclopedia of Energy; Cleveland, C., Ed.; Elsevier: Amsterdam, 2004, 683−694. (40) Cao, L.; Diana, J. S.; Keoleian, G. a; Lai, Q. Life cycle assessment of chinese shrimp farming systems targeted for export and domestic sales. Environ. Sci. Technol. 2011, 45, 6531−6538. (41) Weber, C. L.; Matthews, H. S. Food-miles and the relative climate impacts of food choices in the United States. Environ. Sci. Technol. 2008, 42, 3508−3513. (42) Edwards-Jones, G.; Milà i Canals, L.; Hounsome, N.; Truninger, M.; Koerber, G.; Hounsome, B.; Cross, P.; York, E. H.; Hospido, A.; Plassmann, K.; et al. Testing the assertion that “local food is best”: The challenges of an evidence-based approach. Trends Food Sci. Technol. 2008, 19, 265−274. (43) Bureau, D. P.; Hua, K. Towards effective nutritional management of waste outputs in aquaculture, with particular reference to salmonid aquaculture operations. Aquacult. Res. 2010, 41, 777−792. (44) Tilman, D.; Balzer, C.; Hill, J.; Befort, B. L. Global food demand and the sustainable intensification of agriculture. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 20260−20264. (45) Foley, J. A.; Brauman, K. A.; Cassidy, E. A.; Gerber, J. S.; Johnston, M.; Mueller, N. D.; O’Connell, C.; Ray, D. K.; West, P. C.; Balzer, C.; et al. Solutions for a cultivated planet. Nature 2011, 478, 337−342. (46) Tacon, A.; Metian, M.; Turchini, G. M.; De Silva, S. S. Responsible aquaculture and trophic level implications to global fish supply. Rev. Fish. Sci. 2009, 18 (1), 94−105. (47) Samuel-Fitwi, B.; Wuertz, S.; Schroeder, J. P.; Schulz, C. Sustainability assessment tools to support aquaculture development. J. Clean. Prod. 2012, 32, 183−192. (48) Reap, J.; Roman, F.; Duncan, S.; Bras, B. A survey of unresolved problems in life cycle assessmentPart 1: Goal and scope and inventory analysis. Int. J. Life Cycle Assess. 2008, 13, 290−300. (49) Reap, J.; Roman, F.; Duncan, S.; Bras, B. A survey of unresolved problems in life cycle assessmentPart 2: Impact assessment and interpretation. Int. J. Life Cycle Assess. 2008, 13, 374−388. (50) Jonell, M.; Henriksson, P. J. G. Mangrove−shrimp farms in VietnamComparing organic and conventional systems using life cycle assessment. Aquaculture. 2014, DOI: 10.1016/j.aquaculture.2014.11.001. (51) Payraudeau, S.; van der Werf, H. M. G.; Vertès, F. Analysis of the uncertainty associated with the estimation of nitrogen losses from farming systems. Agric. Syst. 2007, 2 (94), 416−430. (52) Seppällä, J.; Knuuttila, S.; Silvo, K. Eutrophication of aquatic ecosystems: A new method for calculating the potential contributions of nitrogen and phosphorus. Int. J. Life Cycle Assess. 2004, 9 (2), 90−100. (53) Cao, L.; Diana, J. S.; Keoleian, G. Role of life cycle assessment in sustainable aquaculture. Rev. Aquacult. 2013, 5, 61−71. (54) Namgis First Nation. Namgis Closed Containment Salmon Farm. http://www.namgis.bc.ca/CCP/Pages/default.aspx. (55) Langsand Laks. Langsand Laks. http://www.langsandlaks.dk/. (56) IFIA. International Fertilizer Industry Association. http://www. fertilizer.org. (57) CE. Centre for Energy. http://www.centreforenergy.com. (58) IPCC. Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, 4th ed.; Solomon, S., Qin, D.,

Manning, M., Chen, Z., Marquis, M., Averyt, K. B., Tignor, M., Miller, H. L., Eds.; Cambridge University Press: New York, NY, 2007; p 996. (59) Stock Assessment Report on Pacific Herring in British Columbia in 2012; Fisheries and Oceans; Nanaimo, BC, 2012; p 17. (60) Stewart, I. J.; Forrest, R. E.; Grandin, C.; Hamel, O. S.; Hicks, A. C.; Martell, S.; Taylor, I. G. Status of the Pacific Hake (Whiting) Stock in U.S. and Canadian Waters in 2011; Joint US and Canadian Technical Working Group; Seatle, WA, 2011; p 34. (61) Cho, C. Y.; Kaushik, S. J. Nutritional energetics in fish: Energy and protein utilization in rainbow trout. World Rev. Nutr. Diet. 1990, 61, 132−172. (62) Islam, M. S. Nitrogen and phosphorus budget in coastal and marine cage aquaculture and impacts of effluent loading on ecosystem: Review and analysis towards model development. Mar. Pollut. Bull. 2005, 50, 48−61. (63) Tacon, A. G. J.; Metian, M. Fish matters: Importance of aquatic foods in human nutrition and global food supply. Rev. Fish. Sci. 2013, 21, 22−38. (64) Papatryphon, E.; Petit, J.; van der Werf, H. M. G.; Sadasivam, K. J.; Claver, K. Nutrient-balance modeling as a tool for environmental management in aquaculture: The case of trout farming in France. Environ. Manage. 2005, 35, 161−174. (65) Aubin, J.; Tocqueville, A.; Kaushik, S. J. Characterisation of waste output from flow-through trout farms in France: Comparison of nutrient mass-balance modelling and hydrological methods. Aquat. Living Resour. 2011, 24, 63−70. (66) Sauvant, D.; Perez, J.-M.; Tran, G. Tables of Composition and Nutritional Value of Feed Materials, 2nd ed.; Wageningen Academic Publishers: Amstelveen, Netherlands, 2004; p 304. (67) Dumas, A.; de Lange, C. F. M.; France, J.; Bureau, D. P. Quantitative description of body composition and rates of nutrient deposition in rainbow trout (Oncorhynchus mykiss). Aquaculture 2007, 273, 165−181. (68) Iwama, K.; Tautz, A. F. A simple growth model for salmonids in hatcheries. Can. J. Fish. Aquat. Sci. 1981, 38, 639−646. (69) Jobling, M. The thermal growth coefficient (TGC) model of fish growth: A cautionary note. Aquacult. Res. 2003, 34, 581−584. (70) Wilfart, A.; Prudhomme, J.; Blancheton, J.; Aubin, J. LCA and emergy accounting of aquaculture systems: Towards ecological intensification. J. Environ. Manage. 2013, 121, 96−109.

I


Life Cycle Assessment of a Novel Closed-Containment Salmon

Recommend Documents