The ecological limits of aquaculture: Comparative performance of salmon production systems

(1)

by Valerie Ethier

B.A., Princeton University, 2005

A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of

MASTER OF SCIENCE

in the School of Environmental Studies

 Valerie Ethier, 2013 University of Victoria

(2)

Supervisory Committee

The ecological limits of aquaculture: Comparative performance of salmon production systems

by Valerie Ethier

B.A., Princeton University, 2005

Supervisory Committee

John Volpe, School of Environmental Studies Supervisor

Jason Fisher, School of Environmental Studies Departmental Member

(3)

Abstract

Supervisory Committee

John Volpe, School of Environmental Studies Supervisor

Jason Fisher, School of Environmental Studies Departmental Member

Aquaculture is one of the fastest growing animal protein production industries and

accounted for 47% of the world’s food fish consumption in 2010. Aquaculture production is expected to increase to compensate for projected shortfalls in seafood supply by

capture fisheries. Current assessments and scenarios predicting the outcome of this increased production have limited scope and ability to distinguish alternative courses of action.

Using the Global Aquaculture Performance Index (GAPI) as a starting point, I have developed an ecologically comprehensive and quantitative farm level assessment. I selected salmon as the candidate to compare production scenarios due to being economically important, data rich and farmed in a diversity of production systems. In applying the farm-level assessment to conventional net-pen salmon production and four alternative systems, I determined the ecological impact per unit of production to be significantly different.

It is possible to produce a greater volume of fish for less ecological impact. While there are benefits and trade-offs in the alternative production systems, the results indicate that projected food fish demands can be met in a more sustainable manner. The Farm Level Aquaculture Performance Index (FLAPI) provides a quantitative, performance-based tool that accounts for all ecological impacts and the resulting assessments can be used to benchmark and guide future development of aquaculture.

(4)

List of Tables

Table 1.1 Ecological impacts of aquaculture as identified by seafood sustainability

initiatives ... 7

Table 1.2 Commonly used terms for discussing feed and energy efficiency in aquaculture ... 9

Table 2.1 Farm level data sample size by production system and group ... 38

Table 2.2 Summary of ANOVA statistics comparing FLAPI criteria and final scores by production system (net-pen, IMTA, solid-sided, flow-through, RAS) ... 42

Table 2.3 Summary of Tukey's HSD by production system ... 43

Table 2.4 Directional relationships results from Tukey's HSD by production systems ... 43

Table 2.5 Summary of ANOVA statistics comparing FLAPI criteria and final scores by group (open, closed, partial) ... 44

Table 2.6 Comparison of ANOVA p-values between production system and groups ... 44

Table 2.7 Summary of Tukey's HSD by group ... 45

Table 2.8 Summary of relationships results from Tukey's HSD comparing groups ... 45

Table 2.9 Comparison of Kruskal-Wallis and ANOVA p-values by production system and group ... 46

Table 3.1 Benefits/costs of moving BC (75,000 mT) and global (1,426,000 mT) net-pen farm salmon to alternative production systems (assuming a maximum reduction of 100%) ... 56

(6)

List of Figures

Figure 1.1 Top four Atlantic salmon aquaculture producing countries 1987-2010 (FAO 2011) ... 5 Figure 1.2 Spectrum of aquaculture production system by reliance on the surrounding ecosystem ... 17 Figure 2.1 Salmon production systems by ecosystem reliance ... 36 Figure 2.2 Production system FLAPI performance boxplots ... 40 Figure 2.3 Comparison of average FLAPI performance scores by production system .... 47 Figure 2.4 Comparison of FLAPI performance scores by group ... 48 Figure 2.5 Ecological performance in percent differences from net-pen production using FLAPI scores (by production system) ... 49 Figure 2.6 Ecological performance in percent differences from the Open grouping using FLAPI scores (by grouped production systems) ... 50 Figure 2.7 Simplified diagram outlining the characteristics of net-pen and RAS

production ... 55 Figure 3.1 Summary of performance differences between RAS and net-pen production in % and potential production gain/loss for the same ecological impact ... 60

(7)

Acknowledgments

I would like to thank the many people who have helped me through the completion of this thesis. My sincerest thanks are extended to my project supervisor, John Volpe, for his support and guidance. His ability to ask incisive and challenging questions has

undoubtedly strengthened my critical abilities and hopefully this thesis.

Thank you to my supervisory committee, Drs. Jason Fisher and Rosemary Ommer, for their valuable comments and suggestions. Their different research interests have helped improve my communication skills and the quality of my thesis.

I would like to thank Pew Charitable Trusts for the financial support for the GAPI project that began me on this research and continuing it to the farm level.

To Jen, Martina and Jenna, thank you for making the hours in a small office (or van or car) pass like nothing, for the data gathering (real and distracting) and for all the hard work making this project what it is. I know we’re stuck with the “GAPI Girls” moniker and I, for one, feel honoured.

Thanks to my excellent lab mates and ES cohort. You have inspired me with your work and made me believe that despite how impossible something looks it can be done. I am beyond grateful to my Mom and Dad, who both instilled a love of the outdoors, a sense of curiosity in marine life and knowledge that if I were to pursue this as a career it wouldn’t be for the money. Thank you for your support regardless of what I decided. Finally, much love and thanks go to Doug. Someone once told me that partners of graduate students get to experience all of the hard work and stress with none of the reward. Thank you, I wouldn’t have been able to finish this without your support,

patience, and encouragement. And for understanding that sometimes I just needed to run. “So long, and thanks for all the fish.” D. Adams

(8)

The challenge of increasing food fish demands

It is predicted that in 2013, for the first time in human history, more of the fish consumed globally will be from aquaculture than wild fishery stocks (FAO 2012). As our last major wild food source, fisheries are not expected to disappear, but are unable to meet the increasing demand for seafood despite historically being considered limitless (Tidwell & Allen 2001). Expanding human populations have resulted in an even greater growth in demand for fish and an increasingly intensified aquaculture industry (Delgado et al. 2003). Aquaculture has been touted as a solution to meet food-fish demands and protect wild fish stocks, but may become less sustainable as global production grows (Volpe et

al. 2010).

Global fishery production can be divided into capture fishery and aquaculture. Capture fishery production is that from wild fish stocks in the oceans, lakes and rivers and is “captured” or fished. Aquaculture is the farming of aquatic organisms including fish, molluscs, crustaceans and aquatic plants. Capture production has remained relatively stable over the past decade while global fishery production has continued to increase (FAO 2012). Total fishery production gains are due to increasing exploitation of wild fisheries (87.3% of the worlds fishery stocks were at least fully exploited as of 2009 (FAO 2012)) and a rapidly growing aquaculture industry. Aquaculture has been

expanding at an annual growth rate of 8.3% over the past forty years and accounted for nearly half of the world’s food fish (47%) in 2010 (FAO 2012).

Although the idea of aquaculture as a solution to the depletion of wild fisheries is relatively recent for the world community, fish farming is not a new concept. The practice has been documented as far back as 2000 B.C., beginning with pond culture of common carp (Cyprinus carpio) in China (Nash 2011). While the origins are not well documented, the beginning of the practice entailed catching and holding fish in

(9)

reservoirs. Aquaculture has been improved upon over the years by exploring different species and farming practices, completing the full life cycle in captivity and learning how to increase production efficiency.

As of 2010, approximately 600 species were being farmed in 190 countries: total aquaculture production of food fish was 60 million tonnes (FAO 2012). Aquaculture is expanding and expected to supply food fish for the foreseeable future, but the growth potential is somewhat unknown. Aquaculture production is fuelled by feed from reduction fisheries (i.e. those reduced into fishmeal and fish oil rather than directly consumed) and reliant upon the surrounding ecosystem’s ability to absorb organic, inorganic and biological wastes (Volpe et al. 2010). Both capture fisheries and aquaculture continue to diminish natural resources from marine ecosystems without a comprehensive understanding of how much production is sustainable.

Annual global fish consumption continues to grow and has increased from 12.6 kg per capita in the 1980’s to 18.6 kg per capita in 2010 (FAO 2012). Wild fisheries are expected to reach a maximum annual yield of approximately 80 million tonnes before 2020 (Brugère & Ridler 2004), falling well short of anticipated demands. Aquaculture is expected to compensate for the 71-117 million tonne deficit predicted by 2050 and as such sustainable solutions to meet this need must be investigated (Brugère & Ridler 2004).

While aquaculture production has averaged 8.3% growth over the past forty years, capture fishery production has remained at 1.2% for the same period (Natale et al. 2013). The ability of aquaculture to meet future demands typically reflects business-as-usual growth, which assumes similar production increases as historically achieved (Brugère & Ridler 2004). Future aquaculture estimates usually include only fishmeal and fish oil requirements and their supply through capture fishery production or scenarios with improved fish in-fish out ratios, meaning fewer marine resources are required per unit of production (Delgado et al. 2003). Estimates do not include other known ecological impacts associated with aquaculture production such as chemical use, escapes, nutrient effluent, energy use and disease. Narrowly focused aquaculture production forecasts

(10)

result in an inability to predict how production might be limited by amplification of other ecological impacts.

The aquaculture industry is built on a business model of profitability co-varying with the extent to which carrying capacity of the surrounding ecosystem is exceeded. Increased carrying capacity is accomplished through the addition of certain classes of inputs (i.e. feed and industrial energy) and removal or displacement of outputs and biological emissions (i.e. nutrient effluent, chemicals, pathogens and escapes). The ecological impact of fish farms varies by production species and system, scale and farming region. Future production of aquaculture is not limited by the carrying capacity of the ecosystems in which it is located, but by a combination of inputs, outputs and biological impacts. Although the absolute limit of aquaculture is a moving target and difficult to project, examining the effect of different production scenarios may provide some idea of whether current best practices and technology can produce enough fish to meet anticipated

demands. The definition of sustainability is similarly difficult to outline, but for the purposes of this research, a sustainable aquaculture industry is one that can maintain a given production level for the foreseeable future. The limit of aquaculture is a conceptual point at which benefits (or production) are maximised and costs (or impacts) are

minimized. Therefore, when assessing where this point lies it is important to include all ecological impacts that may limit production.

Research and reviews of the risks posed by intensive marine aquaculture have identified the same key ecological issues for over two and a half decades (Beveridge 1984, Folke & Kautsky 1992, Goldberg & Naylor 2005, Gowen & Bradbury 1987, Naylor et al. 1998, Schlag 2010, Silvert 1992, Tyedmers 2000, Wu 1995). Some of the impacts have been addressed to varying degrees in net-pen production through improved feed formulations, implementation of best management practices for chemicals, disease and escapes and increased energy efficiency. During the last two decades alternative production systems have shown increasing promise for improved ecological performance through the same methods as net-pen production with the addition of increased separation from the surrounding ecosystem (Pelletier et al. 2009, Volpe et al. 2010). These alternative systems may improve the ability of aquaculture to meet future food fish demands, but

(11)

their performance needs to be evaluated in an ecologically comprehensive and quantitative manner against each other and conventional net-pen aquaculture.

Ecological impacts can vary by production system, but also between species. Comparing a single species in many production systems increases the likelihood of differences being reflective of the production system. Although salmon aquaculture represents only about 2% of total global production (including fish, molluscs, crustaceans and plants) (FAO 2011), this data-rich species group is grown in a diversity of farm types and accounts for a high proportion of marine impact (Bartley et al. 2006, Volpe 2010). For example, salmon production in 2008 was roughly 2% of global aquaculture, but accounted for 13.7% of global fishmeal use and 36.6% of global fish oil use (FAO 2011, FAO 2012). Salmon production causes the introduction and spread of non-native species, pathogens, and increased marine pollution (Krkošek et al. 2005, Naylor et al. 2005, Wu 1995). Atlantic salmon is the dominant species in production due to their superior growth rates and health in captivity, but all salmon species in aquaculture have similar ecological requirements and impacts (Volpe et al. 2010).

Consequently, salmon are potentially an ideal candidate to examine the relative impact of alternative production systems. However, there is an absence of adequate methodology to evaluate the relative impacts of different systems. What is needed is a clear, quantitative and robust means of analyzing and contrasting the ecological impact of different

production systems at the farm level to incorporate into projection scenarios. Here I explore alternative farming methods and provide guidance as to how pursuing different paths may impact the future limit of aquaculture.

Salmon farming

Conventional salmon cage culture did not emerge until the 1960’s. Atlantic salmon aquaculture began in rural areas of Norway as a government subsidized activity to bolster communities affected by diminishing wild fisheries (Aarset 1998, Willoughby 1999 & Hjelt 2000 in Liu & Sumaila 2008). It began experimentally in Norway with the first trial completed in 1969 and was quickly followed by production in the UK (1970). Expansion was rapid in both countries as the market rewarded the initial smaller quantities with high

(12)

prices relative to production costs (Bjørndal 2002). These results motivated other

countries to begin production and soon after both Canada (1979) and Chile (1987) joined their ranks as the world’s top salmon producing countries (FAO 2011). In a period of 23 years, annual Atlantic salmon aquaculture production has grown from just under 70,000 tonnes, to a 2010 harvest of 1.43 million tonnes, more than a 2000% increase.

Despite the prodigious growth in the farmed salmon industry, relatively little has changed in the main production methodology since the early days. Most gains in profit have been due to improved technology to reduce man-hours required for production, increased efficiency of feeding routines, feed formulas and improved survival rates (Bjørndal et al. 2003). The current industry follows the same basic model as the early farms, but has been intensified and produces a larger quantity of fish at higher densities and lower production costs. Greater production has also resulted in a dramatic increase in spatial coverage. Figure 1.1 Top four Atlantic salmon aquaculture producing countries 1987-2010 (FAO 2011)

Increased production efficiency has led some to speculate on the potential of aquaculture to meet market demand for seafood protein while relieving pressure on wild fisheries (Sachs 2007). Although the economic cost of producing salmon has decreased over the years, the ecological cost may not have declined at the same rate. Net-pen farms are open

(13)

systems, which rely heavily upon services from ecosystems surrounding the farms as well as environments further away. In the immediate area, nutrient cycling and water

provision maintain water quality conditions suitable for salmon culture (e.g. dissolved oxygen, ammonia levels). As such, it is the surrounding environments that must assimilate biological and chemical wastes and cope with increased pathogen loads. External feed provided for salmon farms are sourced from global fish stocks, agricultural crops and livestock. Each source has its own associated problems, which include

overfishing, erosion, fertilizer and other chemical runoff and energy demands (Gerber et

al. 2007, Pauly 2002, Pimentel et al. 2005). Ecological costs not reflected in the price of

producing salmon are known as externalities, and are widespread and well documented in the agro-food industry (Clark 2005). Externalities of the salmon farming industry are unintended negative (or positive) side effects bourn by the environment rather than producers or consumers. If the total ecological and economic costs of production are not included in the price, the growth potential of aquaculture cannot be accurately estimated. Without assessing all factors it is not possible to determine the juncture where

externalities, or ecological impacts not included in the cost of production, grow to the point of affecting productive capacity.

Ecological impacts

Intensive salmon aquaculture has increased pressure on ecosystems; thus the emphasis in literature remains on environmental challenges (Beveridge 1984, Folke & Kautsky 1992, Goldberg & Naylor 2005, Gowen & Bradbury 1987, Naylor & Burke 2005, Naylor et al. 1998, Naylor et al. 2000, Schlag 2010, Silvert 1992, Tyedmers 2000). The main

ecological impacts include escapes (Naylor et al. 2005), organic and inorganic effluents (Burridge et al. 2008b, Wu 1995), reliance on wild fish for feed (Tacon & Metian 2008), and increased pathogen loads (Krkošek et al. 2005). The cost of producing salmon would increase if the industry had to compensate for all impacts of provisioning and

assimilation. Given that externalities are considered unquantifiable (or very difficult to quantify) in dollar terms, one alternative option is to explore production systems that reduce or eliminate externalities.

(14)

Table 1.1 Ecological impacts of aquaculture as identified by seafood sustainability initiatives

Inputs

Outputs

Biological

Feed Parasiticides Pathogens

Ecological energy Antibiotics Escapes Industrial energy Biocides

Source of Stock Effluents

The vast majority of Atlantic salmon production (>99%) occurs in open-net pens as opposed to solid-sided pens or land based facilities; they have been found to be the most cost-effective method to raise most mid-to-high value fish (Halwart et al. 2007,

McCowan 2012). The exchange of water through the nets replenishes oxygen and

removes wastes, externalizing these production costs to the surrounding ecosystem. Other inputs are necessary in the production process, including feed and industrial energy. As open systems, any other production discharges are released directly into the environment they are located, as are any biological contaminants such as escapes or pathogens. Inputs

Certain classes of input (i.e. feed, energy and broodstock) are employed to increase fish growth and production beyond natural carrying capacity. Several of these are related to the external feed, which is a very important component of production that allows for concentrated and rapid growth of species with high protein requirements. Feed inputs often account for a high proportion of overall cost and energy demand (Vinci 2012). Supplemental feeding from capture fisheries results in ecological energy being removed from the source ecosystem and transferred into aquaculture production and its supporting environment (Volpe et al. 2010). The capture and processing of feed along with the running of the production systems also require the use of industrial energy (Ayer & Tyedmers 2009). Concentrated farming of a species is made possible through feed inputs, but wild fisheries may be relied upon as a source of stock as well. Depending on the domestication status of broodstock and seed supply, stocking of aquaculture systems can further reduce wild fish populations (Lovatelli & Holthus 2008).

(15)

Feed

Ecological-impact literature challenges the notion that industrial-scale aquaculture operates as a substitute for capture fisheries, thus relieving pressure on overexploited stocks (Goldberg & Naylor 2005, Naylor & Burke 2005). Through an examination of raw energy flow data, aquaculture production consumes fewer marine resources per unit of a given species than does commercial fishing (Welch et al. 2010). However, given the proportion of marine ingredients in feed inputs, aquaculture adds to capture production rather than replacing it. Aquaculture feed production drives a transition in capture fisheries towards lower trophic fish for indirect use (Pauly et al. 2002). Energy is lost with each successive conversion up trophic levels, and as such, the industrial farming of high trophic-level species (i.e. carnivorous finfish) is inefficient by its very nature. Rather than humans directly consuming three to four thousand tonnes of forage fish, they are processed into feed for aquaculture, which yield one thousand tonnes of farmed fish. The high protein and fat required to maximize growth requires feed inputs often in the form of fishmeal and oil from wild marine sources (Deutsch et al. 2007, Naylor et al. 2000, Tacon & Metian 2008). It is important to note that feed efficiency commonly refers to the feed conversion ratio (FCR): the amount of feed in to the amount of product out (Table 1.2). The FCR ratio does not accurately describe the actual volume being input to the system, as feed is usually a compound pellet requiring significantly more than 1 kg of wild fish, livestock or crops to produce 1 kg of feed. Using yield rates for fishmeal (22.5%) and fish oil (5%), it is possible to calculate a transfer coefficient for feed that gives a more accurate account of wild fish in to aquaculture product out, also known as fish in-fish out (FIFO) (Tacon & Metian 2008). With an average FIFO for fish oil of 2.8, it is clear that Atlantic salmon aquaculture is a net consumer of marine biomass not a net provider (Skretting 2011).

Compound feeds are produced as pellets using rendered marine, crop and livestock ingredients. Fishmeal and fish oil in compound feeds are derived from capture fisheries such as anchoveta, herring, mackerels and sardines. Seven of the ten largest fisheries (by weight) are reduction fisheries and a decline in mean trophic level of capture fishery

(16)

landings has been observed over the past forty years (FAO 2008, Pauly 2002). Although some reduction fisheries have shown minor improvements in management and stock status, none are currently managed using an ecosystem-based framework and others ignore precautionary advice in setting targets or are not strictly enforced (SFP 2012). For example, total allowable catch for anchoveta in Chile is set 60% above the scientifically advised levels and Atlantic menhaden fishing mortality is more than three times the target (SFP 2012). In 2006 it is estimated that 23.8 million tonnes of small pelagic fish were consumed for aquaculture feed production (Tacon & Metian 2009). Aquaculture using compound feed yielded 19.3 million tonnes of food fish, a global average conversion rate of 1.23. Instead of alleviating pressure on wild populations, farming carnivorous species like Atlantic salmon simply shifts and may amplify that pressure.

Table 1.2 Commonly used terms for discussing feed and energy efficiency in aquaculture

Term Definition

FCR (feed conversion ratio) Ratio between the dry weight of feed fed and the weight of yield gain (e.g. FCR = 2.8 means that 2.8 kg of feed is needed to produce 1 kg of fish live weight)

eFCR (economic feed conversion ratio) Takes into account all the feed used, including losses through wastage and fish mortalities

FIFO (fish in fish out) &

FFDR (forage fish dependency ratio)

Expresses how many kg of wild fish are required to produce one kg of farmed fish, requires fishmeal and fish oil yield and percentage in feed

Ecological energy

Energy in biological systems begins as solar energy, which is captured by photosynthesis and converted into biologically available forms. Consumption of ecological energy by aquaculture systems can be measured by converting biological consumption to net primary productivity (NPP) equivalents (Haberl et al. 2004). NPP describes the net transfer of atmospheric carbon into plants through photosynthesis and can examine the amount of energy redirected through aquaculture feed consumption that would otherwise be utilized in ecosystems. Energetic transfer up through the food chain is inefficient (as per the first law of thermodynamics) and with each successive step, only approximately

(17)

10% of energy contained in a food item is converted into the predator biomass (Chapman & Reiss 1999). Examining NPP allows a more direct approximation of how much energy is appropriated through aquaculture production (Pauly & Christensen 1995). Measuring ecological energy provides an accurate picture of how much energy is actually required for production by converting both the feed and resulting farm biomass to a consistent resource (units of carbon) that can be directly compared. Marine sources are only one aspect of feed inputs and contributions of livestock and plant ingredients can be examined in the same manner using NPP and provide a total primary productivity footprint for aquaculture production. NPP can also be expressed as a footprint of land and ocean required to produce the aquaculture feed.

Industrial energy

In addition to ecological energy, aquaculture often has industrial energy demands.

Industrial energy is that provided by resources such as fossil fuels or hydroelectricity. For conventional open net-pen systems, feed production accounts for up to 94% of industrial energy use (Ayer & Tyedmers 2009, Pelletier et al. 2009). This percentage includes the capture of marine inputs, growth and harvest of agriculture ingredients and the rendering and processing of both. Technology and transportation on the farm also require industrial energy. The type of aquaculture system, sourcing and production of feed inputs need to be considered to account for all industrial energy imbedded in farmed salmon.

Source of stock

While feed accounts for a considerable portion of the ecological impact from inputs, some aquaculture systems with undomesticated species require further appropriation of marine resources in the form of seed stock. These systems, termed capture-based aquaculture (CBA), rely on the capture of wild fish at various life stages to stock their facilities (Lovatelli & Holthus 2008, Ottolenghi et al. 2004). CBA occur in circumstances where capture is more cost-effective than hatcheries, breeding in captivity is not possible or broodstock are being supplemented from the wild (Lovatelli & Holthus 2008). If older life stages are being caught, the impact of seed/broodstock fisheries can be the same (on size and age structure of a population) as for a capture fishery. The impact can be much

(18)

greater than capture fisheries if large quantities of juveniles are being harvested. Capture of seed stock appropriates biological potential, which would otherwise contribute to the reproductive capacity of a stock. In addition to disrupting the structure of a population, harvest of some species involves ecosystem-damaging activities, such as the application of cyanide to reefs to capture grouper juveniles (Ottolenghi et al. 2004).

Discharges

Highly toxic chemical compounds are not used in the supplying of salmon broodstock, but typical salmon production facilities are open systems meaning that anything added to the net pens eventually becomes available to the surrounding ecosystem. Three classes of chemicals used in salmon aquaculture are of particular ecological concern: parasiticides, antibiotics and biocides. Toxicity, persistence and number of organisms affected differ by chemical and will vary the degree of ecosystem impact (Burridge et al. 2008b). While short-range toxicity (24 to 96 hours) of these chemicals is fairly well studied through laboratory lethality research, there is a lack of comprehensive knowledge of their potential long-term use and cumulative effects in the marine environment.

Parasiticides

Research on chemicals to treat parasites suggests that in addition to acute toxicity there are chronic and sub-lethal neurotoxicological effects that influence survival and spawning for aquatic crustaceans and other marine invertebrates (Bright & Dionne 2005, Burridge

et al. 2008a). Parasite resistance to treatment is also a potential problem, requiring

increased dosages or application of new chemicals (Jones et al. 1997). While management regulations, such as those reducing stocking densities, decrease the likelihood of parasite occurrence, the conditions of food production on a commercial scale are often ideal for outbreaks (Collins & Wall 2004). Multiple applications per grow out cycle (or the time from placing salmon in net pens to harvest) are common and environmental fate of parasiticides is often suspended and settled particles. Not only should the site-specific amount and toxicity of parasiticides be considered, but also the cumulative load of parasiticides both in space and time must be accounted for (Bright &

(19)

Dionne 2005). Examining each treatment in isolation ignores the potential for interactive chemical effects and increased chemical accumulation due to long degradation periods.

Antibiotics

As with most conventional animal protein production systems, the application of antibiotics is standard practice. Antibiotics are used mostly as treatment, but occasional prophylactic, or preventative, use occurs (Burridge et al. 2008b, Redshaw 1995).

Antibiotics are provided through medicated feed as well as immersive baths, using a tarp to surround the fish where they are treated for 1-2 hours at which point both fish and treated water are released. Unfortunately the dominant treatment methods often result in effects seen beyond farms including toxicity to non-target organisms, persistence in sediment and the water column and antibiotic resistance (Cabello 2006, Christensen et al. 2006). Non-linear effects, or those great than anticipated by the additive impact, of certain antibiotic combinations have been demonstrated experimentally (Christensen et

al. 2006) and establish the need for further research before applying novel chemical

mixtures. As the use of antibiotics in commercial animal production has intensified, so has the occurrence of antibiotic-resistant bacteria (Schwarz & Chaslus-Dancla 2001). Antibiotic resistance is of particular concern due to the more than thirty different antibiotics considered critically or highly important in both human and veterinary medicine that are used in aquaculture (FAO, WHO & OIE 2008).

Biocides

In addition to treatments applied to the production species, antifoulant paint is used to inhibit the growth of marine organisms on netting. The most common active biocidal ingredient in these paints is copper (Burridge et al. 2008b), although other more toxic compounds such as tributyltins are applied in countries where regulation allows (Karlsson

et al. 2006, Ytreberg et al. 2010). Antifoulants are applied to nets to prevent biofouling

that decreases the durability of nets and water flow, leading to lower oxygen exchange. The copper in these paints can leach into the surrounding ecosystem, or be added through wastewater after net cleaning occurs. Although it is a naturally occurring element,

(20)

bivalves (Kullman et al. 2007), polychaetes, branchiopods (Mayor et al. 2008), microbial communities (Webster et al. 2001), algae and copepods (Karlsson et al. 2006). Copper is also very persistent and remains bioavailable even when bound in sediment (Chou et al. 2002). Due to toxic and sub-lethal effects in some marine organisms, continual or punctuated releases of copper into the marine environment have the potential to

drastically alter community structure around aquaculture sites (Burridge et al. 2008b).

Effluent

Not all discharges from aquaculture are inorganic. Feed inputs result in a large nutrient supplement from uneaten food and wastes; mass balance models of salmon and trout farms have estimated that 28% of the phosphorous and 50% of the nitrogen in feed winds up as dissolved nutrients (Mente et al. 2006). The impacts of these wastes have been well documented and include the creation of anaerobic conditions beneath net pens and

localized eutrophication (Wu 1995) and benthic communities that are altered by the significant increase of deposit feeders (Hargrave et al. 1997). Effects of added nutrients from aquaculture have been observed at up to 1000 meters away from production sites, and possibly further as this was the maximum distance sampled (Sarà et al. 2006). The extent of the effect from this organic effluent will depend largely on hydrodynamics and character of an ecosystem (whether it is oligotrophic or eutrophic) (Ackefors & Ennell 1990), but the environment will be altered from its original state.

Biological

The outcomes of waste are well known, but the effects of biological releases are greatly influenced by local conditions. Many production systems are vulnerable to escape events of varying magnitudes and the effect of escapes will vary depending on life-history characteristics of the farmed species and the receiving environment. Certain escape events yield greater impact due to differences in competition for food and habitat,

likelihood of predatory behaviour or modification of habitat through feeding, foraging or settlement. The lack of impermeable barrier between farms and the surrounding

ecosystem that makes escapes possible also creates risk for pathogens, diseases and parasites to move between farmed and wild populations.

(21)

Pathogens

The use of veterinary medicine for both prevention and treatment cannot avert all

pathogen outbreaks. Effects from an epidemic of infectious salmon anemia (ISA) in Chile that first surfaced in 2007 are still being felt four years later as production has decreased from nearly 400,000 tonnes in 2006 to an estimated 100,000 tonnes in 2010 (Asche et al. 2009). Increased wild stock mortality is linked to pathogen outbreaks on farms (Krkošek

et al. 2007). Transmission of diseases and parasites from farmed to wild fish is

unavoidable when they are sympatric (Johnson & Jensen 1991, Krkošek et al. 2005). Although it is unlikely for salmon aquaculture to amplify pathogen loads in molluscs, crustaceans or other invertebrates, susceptibility has been shown to increase with organism relatedness. Certain pathogens, such as infectious salmonid anemia (ISA) are more limited to salmon, while others such as furunculosis can infect many species of marine and freshwater finfish (McCarthy 1975). Aquaculture systems may act as sites of intensive breeding and acute release for pathogens, amplifying those that naturally occur (Johnson & Jensen 1991, Krkošek et al. 2007) in addition to being responsible for the introduction of novel diseases and pathogens (Asche et al. 2009).

Escapes

Transmission of pathogens from farms may take place when wild fish swim past production sites, or when farmed fish escape. Not only do they present disease risks to wild fish, but also both native and non-native farmed fish present a variety of risks to ecosystems. Introduction of exotic species through escape can alter the ecosystem structure in unpredictable ways, competing with wild fish for mates, habitat and food (Naylor et al. 2005). If escapees are able to produce successful, viable offspring, the repercussions of these escapes may persist, or increase if these fish out-compete native species and establish feral or even hybrid stocks (if interbreeding occurs). Even if farmed salmon are not establishing viable populations, aquaculture may act as a continual source of individuals and decrease the chance for escapee eradication. These problems can happen with the escape of farmed fish native to the region as well, with additional risk of them breeding with native populations and leading to decreased fitness (Araki et al. 2007). Atlantic salmon (Salmo salar) show a reduced lifetime success of hybrids 27% to

(22)

89% of their wild counterparts, and very low second-generation embryo survival (McGinnity et al. 2003). Genetic variability, the currency of adaptive capacity, can be irreversibly lost through this addition of a more genetically homogenous group of farmed fish (Naylor et al. 2005).

Production systems

Ecological impacts have mainly been researched and associated with open net-pen salmon farms, but are known issues for most production systems (Ayer & Tyedmers 2009, Volpe et al. 2010). Although open net pens are the primary production system for Atlantic salmon, alternative production systems have been explored. Operations can differ in efficiency, extent of control over production, reliance on the surrounding ecosystem for habitat use and water quality management, feed inputs and use of technology. Production parameters can exist in many combinations and may result in greatly differing environmental performance.

Open net-pen systems

Over 99% of current Atlantic salmon aquaculture is farmed using open net-pen or cage systems (McCowan 2012), and is thus identified as conventional production. Open net pens or cages are anchored in place and enclosed. These systems allow organic and inorganic wastes to pass freely into the surrounding ecosystem. Although farmers supply feed for the fish, water quality and habitat are provided by the environment and keep the cost for production and technology use low. Stocking density is between 20-25 kg/m3 to maintain fish health and optimal growth conditions (Turnbull et al. 2005).

Integrated multi-trophic aquaculture (IMTA)

One approach researchers and producers have explored to address the addition of organic waste through fish farming is integrated multi-trophic aquaculture (IMTA) (Troell et al. 2009). The salmon portion of this production system uses open net pens. Salmon

represent the fed component of the system and are paired with extractive species that take in dissolved inorganic nutrients (such as seaweeds) and those which consume particulate organic matter (such as shellfish). Seaweeds and shellfish also provide additional

(23)

harvestable biomass. Aside from the addition of extractive species, IMTA remains very similar to conventional open net-pen salmon production.

Solid-sided

Creating barriers between farmed populations and the surrounding ecosystem is one way to increase control over production. Floating solid-sided systems are located in the same offshore areas as net pens and IMTA, however, water and wastes do not pass freely between the farm and surrounding ecosystem. Water is pumped from near the ocean floor into the system and supplemented with oxygen. Before the water is pumped back into the environment, organic wastes can be filtered and removed. Slightly higher stocking densities than conventional systems are possible without serious health implications. Solid sides also act as a barrier preventing biological and predator interactions, but are still vulnerable to extreme weather that may damage infrastructure and permit escapes.

Flow-through

Increasing the separation between farm and surrounding environment beyond solid-sided systems means moving production onto land. Flow-though production systems represent a spectrum of farms based on levels of control and water usage. In a traditional flow-though system water passes once through the farm, resulting in high influent and effluent levels. This type of flow-through is more likely in areas where suitable water is abundant and can be gravity fed. Technology can be added for filtration and oxygenation, to reduce water demand and decrease nutrients in effluent. Added technology also increases the level of control producers have over their farm environments. A connection still exists between flow-through farms and the surrounding ecosystem by way of their effluent and what it may contain (i.e. chemicals, nutrients, escapes and pathogens).

Recirculating aquaculture system (RAS)

Water reuse occurs in land based production systems anywhere from 0 to 99%. Systems reusing between 85-99% of their water are referred to as recirculating aquaculture systems (RAS). In addition to removing solids and adding oxygen, further water quality treatment is necessary to allow for increased water reuse. RAS uses biofiltration to

(24)

remove ammonia nitrogen (a toxic metabolic waste of fish). RAS producers have maximum control over culture conditions and can remove most, if not all, connection to the environment immediately outside the farm. Both flow-through and RAS production systems are able to accommodate greater stocking densities (up to 100 kg/m3) without the same fish health problems as net pens and IMTA (Vinci 2012).

Spectrum of ecosystem reliance

Aquaculture production attempts to increase efficiency of the farmed organisms, growing biomass more quickly than it would naturally. The previously described salmon

production systems vary in a number of ways, but one way to describe their differences is by where they lie along a spectrum of ecosystem reliance. The ecosystems surrounding the farms are used to supply certain requirements for production and depending on the production system, this can be nearly full reliance to very little.

Figure 1.2 Spectrum of aquaculture production system by reliance on the surrounding ecosystem

At one end of the spectrum there is extensive aquaculture (completely reliant on the surrounding ecosystem) and at the other end is intensive, closed-containment aquaculture (with little reliance on the immediately surrounding ecosystem). A number of production

(25)

systems exist in between the two, with an intermediate amount of reliance on the ecosystem surrounding the farms.

Extensive aquaculture operations are energetically porous by nature, obtaining energy in the form of feed, brood and ecosystem services and losing it through natural mortality, organic waste and predation. These systems allow for a flow of energy between production and the surrounding system that includes natural processes rather than recreating them. Near-exclusive reliance on the surrounding ecosystem limits harvest biomass of extensive production systems (per unit of area) due to the fact that energy is diverted from production of the target species to other ecological processes.

Neither extensive nor semi-extensive aquaculture are used to farm salmon, as these production systems are not financially viable for producing homogenous global

commodities. Moving away from complete reliance on surrounding ecosystems, net pens are the first production systems along the spectrum used to raise salmon. Net pens are open to the surrounding ecosystems and make use of the optimal rearing conditions they maintain. However, unlike extensive aquaculture, open net pens are supplied with external feed among other inputs to increase production, increasing carrying capacity beyond that inherent in the surrounding ecosystem.

On the far right of the spectrum are intensive closed-containment systems, which remove production almost entirely from the surrounding ecosystem. Intensive closed-containment uses technology to create the ideal growth environment for farmed species and to remove solid wastes and treat water for reuse. The inputs and discharges to the edge of the tank can be very similar to those of open net-pen production, but in intensive

closed-containment production it is possible to prevent discharges from entering the ecosystem. Preventing the outputs further diminish dependence on the surrounding ecosystem. The two ends of the ecosystem reliance spectrum (extensive production and RAS) are often credited as having less environmental impact than intensive net-pen aquaculture. However, there are trade-offs with each of these systems. The resources required for extensive aquaculture (such as appropriate wetlands/ponds and feed sources) are not abundantly available, especially close to large population centres. RAS and flow-through

(26)

systems have been criticized for several reasons as well, often due to the amount of industrial energy used being considered unjustified for the amount of food produced (Ayer & Tyedmers 2009).

Each different type of production system represents a range of values rather than a point on the spectrum, allowing for some overlap of ecosystem reliance. The total ecological impact of these systems includes not only the impacts on the immediately surrounding ecosystem, but also those generated away from the farm. While the trade-offs in production systems have often been described qualitatively, it is necessary to examine what the costs and benefits mean in quantitative ecological terms.

The ecological limit of aquaculture

Models of future food fish supply predict that dependence on aquaculture will continue to increase (Brugère & Ridler 2004, Delgado et al. 2003). The potential ecological

implications of greater aquaculture production have generally been examined in terms of feed sources required, especially from marine sources (Merino et al. 2010, Merino et al. 2012, Tacon et al. 2011). Although fishmeal and fish oil availability may be a limiting factor in the growth and potential development of aquaculture, it is not the only

ecological impact that may constrain what aquaculture can produce.

The ecological limit of aquaculture is not an absolute point, but rather a concept that sustainable aquaculture (which can continue for the foreseeable future) will be able to produce a finite amount based on ecological limits. Production may be limited by any of the ecological impacts (i.e. feed, chemical use, effluent, energy use, escapes, pathogens and source of stock) and will differ by species, production region and production system. Questions regarding the limits of aquaculture may be more theoretical in nature and providing data to contrast different ecological limits based on species and production system can help guide understanding of relative costs and benefits. For example, the ecological performance for producing one kilogram of salmon in an open net-pen might be compared against that for one kilogram of milkfish produced in an intensive pond. These profiles can then be scaled up to predict the cumulative differences in ecological

(27)

performance for pursuing each option. Here I limit the scope to salmon in different production systems. I ask:

1. Is the ecological impact per unit of production the same for all production systems? 2. Can a greater amount of farmed salmon be produced for the same level of impact in different production systems?

H0 1. Ecological performance does not differ significantly between the production systems.

H0 2. The quantity of farmed salmon will not differ significantly among the various production systems for the same level of impact.

(28)

Chapter 2 Ecological impact of aquaculture at the farm-level

Introduction

To inform decision makers about how future development of the aquaculture industry should be guided, there need to be robust ecological characterizations of different production systems. Many assessment tools focus in detail on one or two impact areas, such as effluent or raw material use, or are too generalized like in life-cycle analysis (i.e. can be used in for impacts in any industry). These assessments do not provide an

adequate ecological performance profile by which to compare systems, as many will perform well in one area of impact and not in others. For example, conventional net pens typically do not have high industrial energy demands compared to RAS. However, the opposite performance might be seen with respect to effluent. All aspects of production and their impacts must be included to create a fair platform for comparison.

Many tools exist to assess the various ecological impacts of aquaculture (Aubin et al. 2009, Ayer & Tyedmers 2009, Bunting 2002, Emerson et al. 2010, MBA 2012, Monfreda

et al. 2004, Pelletier et al. 2009, Roth et al. 2000, Singh et al. 2009, Tlusty 2011, Volpe et al. 2010, WWF 2007). Indicators to measure different components of sustainability

have gained momentum as the best manner in which to do this as they enable robust comparisons of complex ecological processes and services in more simplified terms. The aim of ecological indicators is to maximize explanatory power with the fewest number of variables. The reduced complexity, but high descriptive ability of indicators is useful for management decisions and policy creation. Certain initiatives are aquaculture oriented, whereas others are constructed for wider application and thus are less specific.

Global aquaculture performance index

Of the various initiatives seeking to quantify impacts, sustainability indices show the greatest promise for transforming vague concepts into meaningful, measurable values. Sustainability indices do so using a suite of specially developed ecological indicators to assess the performance of various activities, outlining the important components of all impacts in simplified terms. Rather than including every detail available for each impact,

(29)

the main, driving variables are compiled to benchmark, compare and project ecological performance. The Global Aquaculture Performance Index (GAPI; Volpe et al. 2010) is one such tool, specific to the aquaculture industry. GAPI is a framework that

quantitatively assesses each of the impacts (i.e. effluent, chemicals, energy, disease, feed, escapes and source of stock) and provides an overall ecological performance score (on a scale of 0-100) as well as disaggregated indicator scores. Overall performance describes the total impact of a country-species aquaculture product, while indicator scores allow for further understanding of areas for improvement and of superiority. These scores can be examined as either a normalized score (impact per unit of production) or a cumulative score (total impact of production).

Each step in the GAPI process is designed to be quantitative and objectively based. Indicator selection determines the key issues identified by leading aquaculture

sustainability initiatives. The repeated presence of ten concern areas (or indicators) by seafood sustainability initiatives suggests ten indicators are consistently identified as best quantifying the effects of aquaculture production. Subsequent scoring of indicators is data driven. It was important that these indicators explained as much of the performance as possible without adding factors that are not main drivers of ecological impact. For example, eutrophication potential of a water body will change the impact of nutrient effluent, but the major driver is the amount of excess nutrient being discharged. Indicator weighting is determined by statistical analysis, which distinguishes the indicators where greater differentiation is found. If an indicator has relatively similar scores across all aquaculture species, it is weighted lower as it has low discriminating power. This process allows the data to objectively determine the weighting of indicators and removes the opinions of investigators from the decision. Finally, targets are set at zero impact, which is decidedly unachievable, but will never shift. As the intention of the performance assessment tool is to drive increasingly improved performance, a target that can be surpassed will render the tool obsolete once production improves beyond the selected point. Although a target of no impact is recognized as being unachievable, it allows for continuous comparison against a stationary target.

(30)

The GAPI methodology produces several scores. Indicator scores provide a comparison tool for different impact areas. These can be used to detect where improvements can be made most easily and efficiently, or clarify areas where one species performs better than another. Once combined, the overall score can be examined as either a normalized (per tonne of fish produced) or cumulative score. Most certifications and policies are interested in normalized scores as they provide a method to compare species being produced at dissimilar scales. However, normalized scores are not always ecologically relevant. Cumulative scores are necessary to understand the total input, discharges and biological impact that occurs with a given amount of production. It is possible for high quantities of a well performing species (i.e. low impact per tonne) to result in a large ecological impact, while small-scale farming of a poor performing species (i.e. high impact per tonne) may not cause as much ecological damage.

Farm-level aquaculture performance index

Current ecological impact assessment methodologies are inappropriate for use at the aquaculture farm level (Aubin et al. 2009, Bunting 2002, Emerson et al. 2010, Roth et al. 2000, Singh et al. 2009, Tlusty 2011, Volpe et al. 2010, WWF 2007). Most general assessment tools do not include indicators specific enough to the industry to describe aquaculture impacts. GAPI was the first aquaculture specific sustainability index, but is comprised of indicators developed for data available at the country-species level. Although country-species performance is useful for describing the signal (or general trend) in ecological impact, the noise (or variation) from farms within a country is lost. GAPI is incapable of distinguishing between farms at either end of the performance spectrum within a country and thus is unable to inform management decisions at this scale.

To explore variation in performance and accurately describe ecological impact,

modifications must be made to allow for further differentiation with data available at the farm level. A farm level aquaculture performance index (FLAPI) would be able to differentiate ecological impact within a species, including distinguishing between

(31)

possible to understand the costs and benefits of pursuing each option. Rather than starting from scratch, it makes more sense to build upon the frameworks already in place. GAPI is an elegant tool, but is constrained somewhat by generating relative scores and the need to run principle component analysis with each addition or update to the data.

Monterey Bay Aquarium (MBA) Seafood Watch has recently updated their aquaculture assessment tool to incorporate most of the metrics and methods for assessing impact within their criteria from GAPI. MBA is the leading program in North America informing consumers and businesses about seafood sustainability for both farmed and capture fishery products. The primary distinction between the MBA tool and GAPI tool is the method for calculating indicator and final scores; all indicators from GAPI (with the exception of industrial energy) are included in the MBA assessment tool. A preliminary comparison between GAPI and MBA assessments shows a high degree of fidelity for species rankings and relative normalized performance. If the tool is to eventually be used by producers, the immense market buy-in of the MBA program along with the

transformation to a more quantitatively based framework makes it an ideal candidate for the starting point of a farm level assessment.

The first step in designing the necessary changes to the framework is to understand the current tool and examine areas for improvement. The eight categories identified by MBA are: data, effluent, habitat conversion and function, evidence or risk of chemical use, feed, escapes, pathogens/parasite interactions and source of stock. Within each of these are factors that affect performance within criteria as well as overall scoring.

Methods

To guide understanding of the MBA and FLAPI tools, the following is divided into Criteria (name), Rationale for inclusion (background and justification), Factors assessed (the proxy measures to assessing ecological impact), Input data requirements (what data points are needed to calculate a criterion) and Modifications for FLAPI (Factors assessed and Input data requirements will be added if required for modifications). The formulas required to calculate the intermediate and final scores for each criterion can be found in Appendix A – FLAPI Scoring.

(32)

1. Criterion: Data

Rationale for inclusion: Poor data quality or confidence in data sources can affect the ability to make an informed and robust assessment of a production system. The purpose of the data criterion is to assess the quality and relevance of data available for the rest of the criteria considered by the tool.

Factors assessed: Data relevance (Table A.1), Data quality (Table A.2) Input data requirements:

Data category

Industry or production statistics Effluent

Locations/habitats Predators and wildlife Chemical use

Feed

Escapes, animal movements Disease

Source of stock

Other – (e.g. energy use)

Modifications for FLAPI: As a stand-alone criterion, data quality is assigned the same ecological relevance as all other categories without directly identifying which areas are in need of improvement. To get around this, data quality can be used as a weighting factor. Using this scoring design, the data score for each criterion would be applied as a final step in the overall score calculation and as such would have an impact on each category. This change is necessary so that certainty can contribute to the criterion score. For example, if a species performs well in feed, but the data are not robust, the score is penalized accordingly.

2. Criterion: Effluent

Rationale for inclusion: The Effluent criterion is very similar to the GAPI Biochemical Oxygen Demand (BOD) indicator and is designed to measure the organic waste input from production. Nitrogen is used as a proxy to assess impact within the Allowable Zone of Effect (AZE), which is defined under the MBA tool as

(33)

30 meters. If the Data score for Effluent is 7.5 or higher the Rapid Assessment option can be used (Table A.3), which uses a ten-point scale from Critical to No Concern.

Factors assessed: Waste discharged per ton of fish (in Tonnes of nitrogen),

Production system discharge, Intent/content and Enforcement of effluent regulations and management measures

Input data requirements:

Factor Data input Unit or Reference Table

Waste discharged per ton of fish

Protein content of feed % Economic Feed Conversion Ratio

(eFCR) total feed used divided by total harvest of fish Fertilizer nitrogen input per ton

fish produced kg N t-‐1 Protein content of harvested

whole fish %

Production system discharge

Basic (unadjusted) production

System Discharge score Table A.4 Adjustment (if applicable) Table A.4

Regulation/ Management

Intent of effluent

regulations/management

measures Table A.6

Enforcement of effluent regulations/management

measures Table A.7

Modifications for FLAPI: The MBA Effluent criterion is designed fairly well to work at the farm-level. Instead of defaulting to the rapid assessment in the case of good quality data it should be used only in the instance of no external feed being applied to a system (extensive aquaculture production or shellfish production). If data quality is high, it should be even easier to complete the full assessment and therefore the detail should be included.

The management section included in this criterion is intended to incorporate the impacts of multiple farms. If the assessment is intended to specifically examine only farm level impacts, the set of questions needs to be refined. As all of the other criteria are influenced by management factors (even at the farm level) it makes sense

(34)

to divorce these from the two criteria that currently include them (Effluent and

Habitat) and combine them with the data quality weighting factor to be applied after

each criteria score is calculated. 3. Criterion: Habitat

Rationale for inclusion: Farms can be located in an array of aquatic and terrestrial habitats and their impact can vary as a result. The habitat criterion examines to what degree ecosystem services are affected by the presence of a farm depending on its location, scale and intensity.

Factors assessed: Habitat value, Maintaining/loss of habitat functionality,

Intent/content and Enforcement of habitat regulations and management measures, Exceptional factor – Wildlife & predator mortalities

Input data requirements: Factor

Data input Unit or Reference Table

Habitat value Habitat value Table A.8

Maintaining/loss of habitat functionality

Habitat functionality (if

applicable) Table A.9

Timeframe of habitat loss (if

applicable) Table A.10

Regulation/ Management

Intent of effluent

regulations/management

measures Table A.15

Enforcement of effluent regulations/management

measures Table A.16

Exceptional Factor Wildlife & predator mortalities Table A.17

Modifications for FLAPI: Despite the abundance of data that are available at the farm level, it would be difficult to design a habitat criterion that would be able to deal with every measurable impact, production system and location. As such, the current habitat criterion addresses all of these in a coarser fashion and has room for improvement upon the addition of farm-level impact data. The data gathered when farms are assessed can be compiled to create a more pointed tool for habitat impact and add other potential impact measures.

(35)

The habitat management and regulation section again attempts to address the

potential cumulative impacts of farms on an area. These questions remain important and should be modified to address this at the farm level and incorporated along with the effluent questions and pertinent management questions for each of the other criteria.

4. Criterion: Chemicals

Rationale for inclusion: The use and release of chemicals, the latter being of ecological significance, has been recognized as a major problem in aquaculture production and identified as the cause of impacts on non-target organisms and concern for development of resistance. This criterion uses a single scale to assess the evidence or risk involved with the spectrum of chemicals applied in aquaculture production. If the use or impacts of a chemical are unknown in a system, it is possible to defer to production system based scores within the guide (Table A.18). Factors assessed: Evidence or risk of chemicals use

Input data requirements: Evidence/risk of chemicals use (Table A.18)

Modifications for FLAPI: The Chemical criterion is intended to address all of the main classes of chemicals used in aquaculture production: antibiotics, paraciticides and biocides. The impact can vary greatly between and within each class of

chemical. It is necessary to account for these differences when measuring the performance of a farm. A farm level chemical criterion should track the number of uses of each chemical and its associated level of harm. This is one section of the assessment that would benefit in borrowing from the GAPI methodology. A

quantitative measure is necessary to prevent subjective decisions from clouding the results. During the development of GAPI, an expert workshop was convened to determine how to address the forms of waste produced by aquaculture. This expert group determined that the main chemicals of concern are antibiotics, parasiticides and biocides. The full development of scoring for different antibiotics, parasiticides and biocides can be found in Appendix A.

(36)

Factors assessed (modifications): System discharge, Antibiotic use, Parasiticide use, Biocide use

Input data requirements (modifications):

System discharge

Basic (unadjusted) production System

Discharge score Table A.19 Adjustment (if applicable) Table A.19 Antibiotic use Antibiotic (a, b, c…) # of treatments

Antibiotic risk score WHO/OIE ranking (Table A.20)

Parasiticide use

Parasticide (a, b, c…) # of treatments

Parasiticide risk score

0-‐10 relative to worst performing parasiticide using LC50 & persistence (Table A.22)

Biocide use Biocide (a, b, c…) # of treatments Biocide toxicity score

0-‐10 relative to copper toxicity using EC50/LC50 (Table A.23)

5. Criterion: Feed

Rationale for inclusion: Feed use in aquaculture has many factors that must be accounted for to understand the full impact on feed sources. This criterion includes the amount and type of raw ingredients, a measure of raw ingredient sustainability and nutritional gains or losses as well as the footprint required to produce the feed. Factors assessed: Wild fish use, Net protein gain or loss, Feed footprint

Input data requirements:

Wild fish use

Economic Feed Conversion Ratio (eFCR) total feed used divided by _{total
harvest
of
fish}

Fishmeal inclusion level %

Fishmeal from by-‐products % Fish oil inclusion level % Fish oil from by-‐products %

Sustainability of the source of wild fish unit less measure (Table A.26)

Net protein gain or loss

Protein content of feed %

Economic Feed Conversion Ratio (eFCR) total feed used divided by _{total
harvest
of
fish} Feed protein from non-‐edible sources %

The ecological limits of aquaculture: Comparative performance of salmon production systems

Supervisory Committee

Abstract

Table of Contents

List of Tables

List of Figures

Acknowledgments

Inputs

Outputs

Biological

Chapter 2 Ecological impact of aquaculture at the farm-level