Habitat preferences and fitness consequences for

fauna associated with novel marine environments

Luke T Barrett

orcid.org/0000-0002-2820-0421

Submitted in total fulfilment of the requirements of the degree of Doctor of

Philosophy

September 2018

School of BioSciences

University of Melbourne

i

ABSTRACT

The rapidly expanding reach of anthropogenic environmental change means that animals must

now navigate landscapes comprised largely of modified and degraded habitats. Individuals that

correctly perceive habitat quality will be best placed to survive and reproduce in novel

environments, but where environmental change outpaces the evolution of behavioural

responses, mismatches can arise between cues and the underlying value of habitats. These

mismatches can lead individuals to select habitats that offer relatively poor fitness outcomes,

creating ecological traps. In environments where ecological traps are likely to occur, data on

habitat preferences and fitness consequences can fundamentally change predictions of

metapopulation models and increase our understanding of the role that novel habitats play in

population persistence, but such data are rarely collected. In this thesis, I first conduct a global

meta-analysis to assess the state of knowledge on habitat preference and fitness metrics in

animal populations, using wildlife populations associated with aquaculture as a case study. My

findings reveal that responses to aquaculture vary widely across taxa and farming systems,

ranging from large increases in abundance to near complete displacement. However, the

influence of aquaculture on wildlife populations remains poorly understood, as researchers

rarely obtain appropriate measures of habitat preference, survival or reproductive success.

Accordingly, in subsequent chapters I apply the ecological trap framework to assess marine

habitats modified by aquaculture or invasive species. In the first application, I collect wild

Atlantic cod (a species known to be attracted to salmon farms) from areas of high and low

salmon farming intensity, and compare reproductive fitness via a captive spawning trial with

hatchery-rearing of offspring. I found limited negative effects of high farming intensity on quality

of offspring. In the second application, I show that the threat of predation by a native keystone

predator may limit the ability of an invasive seastar to exploit a food-rich habitat at shellfish

farms. In the third application, I show that an invasive canopy-forming marine macroalga

provides viable habitat for native fishes and may help to maintain fish biodiversity in areas where

urban impacts have driven a decline in native macroalgal canopy cover. Together, this thesis

demonstrates the utility of individual-level data on habitat preference and fitness outcomes—

via the application of the ecological trap conceptual framework—in assessing the impacts of

novel habitats on animals, and recommends greater use of this approach in future investigations

into the impacts of human-induced rapid environmental change in coastal marine ecosystems.

ii

DECLARATION

This is to certify that:

The thesis comprises only my original work towards the PhD except where indicated in the

Preface.

Due acknowledgement has been made in the text to all other material used.

The thesis is fewer than 100 000 words in length, exclusive of tables, maps, bibliographies and

appendices.

Luke Barrett

September 2018

Cover image: Mesocosm reef stocked with invasive wakame kelp (Undaria pinnatifida)

iii

PREFACE

I am the primary author and principle contributor on all chapters presented in this thesis. My

supervisors, Stephen E Swearer and Tim Dempster, are co-authors on all chapters.

Article publication status and author contributions

Chapter Two: Published by Reviews in Aquaculture on 14 Aug 2018. Co-authored by Tim

Dempster and Stephen E Swearer. LTB, TD and SES conceived and designed the experiment; LTB

conducted the experiment and collected data with assistance from technical staff and

volunteers; LTB analysed the data and wrote the manuscript; TD and SES provided editorial

comments.

Contributions: LTB 80 %, TD 10 %, SES 10 %

Chapter Three: Published by Aquaculture Environment Interactions on 16 Aug 2018. Co-

authored by Tim Dempster, Stephen E Swearer, Ørjan Karlsen, Torstein Harboe and Sonnich

Meier. LTB, TD, SES, ØK and TH conceived and designed the experiment; LTB conducted the

experiment and collected data with assistance from ØK, TH and SM, as well as technical staff at

the Norwegian Institute of Marine Research; LTB analysed the data and wrote the manuscript;

TD, SES and SM provided editorial comments.

Contributions: LTB 75 %, TD 5 %, SES 5 %, ØK 5 %, TH 5 %, SM 5 %

Chapter Four: Unpublished material not submitted for publication. Co-authored by Tim

Dempster and Stephen E Swearer. LTB, TD and SES conceived and designed the experiment; LTB

conducted the experiment and collected data with assistance from technical staff and

volunteers; LTB analysed the data and wrote the manuscript; TD and SES provided editorial

comments.

Contributions: LTB 80 %, TD 10 %, SES 10 %

Chapter Five: Unpublished material not submitted for publication. Co-authored by Stephen E

Swearer, Tim Dempster. LTB, SES and TD conceived and designed the experiment; LTB

conducted the experiment and collected data with assistance from technical staff and

volunteers; LTB analysed the data and wrote the manuscript; SES and TD provided editorial

comments.

Contributions: LTB 80 %, SES 10 %, TD 10 %

iv

This research was funded by grants from the Holsworth Wildlife Research Endowment (Chapters

Four and Five), the PADI Foundation (Chapter Five), the Victorian Environmental Assessment

Council (Chapter Five), the Sustainable Aquaculture Lab – Temperate and Tropical (all chapters),

the Research on the Ecology and Evolution (REEF) Lab (all chapters), and the Norwegian Seafood

Research Fund (Chapter Three). All animal research was conducted in accordance with the

animal ethics requirements of the University of Melbourne (Chapter Five: approval numbers

1413133 and 1413193) and Norwegian legislation on animal experimentation (Chapter Three:

approval number 8264). Permits were obtained from the Victorian state government for

collection and translocation of marine animals and algae for (Chapters Four and Five: RP919,

RP1185, NP280, NP282).

v

ACKNOWLEDGMENTS

I would firstly like to thank my supervisors, Steve Swearer and Tim Dempster, for their

unwavering support over the duration of my PhD. They have been everything I could have hoped

for in a pair of supervisors.

My friends and colleagues in the REEF and SALTT labs, past and present, provided helpful

discussions and comments on my thesis chapters, and were universally great company to have

a coffee or beer with, as were all my officemates in 131 and others around BioSciences 4. Special

mentions go to Simon, Emily, Ben, Fran, Qike, Valeriya, Tyler, Matt, James, Jack, Ollie and Fletch

for making me feel welcome in my first couple of years in Melbourne.

Many people combined forces with me to get fieldwork done on the cold and murky waters of

Port Phillip Bay. In alphabetical order: Dean Chamberlain, Seann Chia, Ben Cleveland, Emily

Fobert, Molly Fredle, Akiva Gebler, Kevin Jensen, Valeriya Komyakova, Nina Kriegisch, Kevin

Menzies, Rebecca Morris, Jack O’Connor, Simon Reeves, Juan Manuel Valero Rodriguez, Kyler

Tan, Chris Taylor, João Teixiera, Oliver Thomas and Rod Watson (Victorian Marine Science

Consortium). Apologies if I forgot anyone! Lance Wiffen provided access to his aquaculture

leases at Clifton Springs and Grassy Point. John Ahern and Tania Long averted a couple of

aquarium-related catastrophes in my absence, thanks and sorry!

Thanks to my Norwegian collaborators and surrogate supervisors during my time there: Torstein

Harboe, Ørjan Karlsen and Sonnich Meier. The work was made possible by numerous technical

staff, researchers, expert cod fishermen and all-round nice people. I’d especially like to thank

Margareth Møgster, Stig Ove Utskot, Theresa Aase, Michal Rejmer, Inger Semb Johansen, Nele

Gunkel-Sauer, Kristine Hovland Holm, Yvonne Rong, Terje van der Meeren, Tord Skår, Velimir

Nola and Glenn Sandtorv. I’d also like to thank the staff and students at Austevoll High School

for taking us to Brandasund and back with a boatload of live cod.

My family and friends back home in WA tolerated my long absence and my

incommunicativeness during the busy times, and largely stayed away from the question “when

will you be finished?”. Well played!

Finally, but most importantly, I’d like to thank my partner Marina, who provided constant love

and support, and made quite a few sacrifices to ensure that I had a clear run at this thesis. I hope

it’s been worth it!

vi

CONTENTS

List of Tables ....................................................................................................................... ix

List of Figures ....................................................................................................................... x

Chapter One | General introduction

General introduction ........................................................................................................... 1

References .......................................................................................................................... 9

Chapter Two | Impacts of marine and freshwater aquaculture on wildlife: a global meta-

analysis

Abstract ............................................................................................................................. 14

Introduction ...................................................................................................................... 14

Methods ............................................................................................................................ 17

Results ............................................................................................................................... 21

Discussion.......................................................................................................................... 28

References ........................................................................................................................ 36

Chapter Three | Limited evidence for differential reproductive fitness of wild Atlantic cod in

areas of high and low salmon farming density

Abstract ............................................................................................................................. 50

Introduction ...................................................................................................................... 50

Methods ............................................................................................................................ 53

Results ............................................................................................................................... 62

Discussion.......................................................................................................................... 66

References ........................................................................................................................ 71

vii

Chapter Four | Native predator prevents an invader from exploiting food-rich habitat

Abstract ............................................................................................................................. 79

Introduction ...................................................................................................................... 79

Methods ............................................................................................................................ 82

Results ............................................................................................................................... 86

Discussion.......................................................................................................................... 89

References ........................................................................................................................ 97

Chapter Five | An invasive habitat-former mitigates impacts of native habitat loss for endemic

reef fishes

Abstract ........................................................................................................................... 102

Introduction .................................................................................................................... 103

Methods .......................................................................................................................... 105

Results ............................................................................................................................. 113

Discussion........................................................................................................................ 119

References ...................................................................................................................... 123

Chapter Six | General discussion and conclusions

General discussion and conclusions ............................................................................... 128

References ...................................................................................................................... 137

Appendices

Appendix 2.1 ................................................................................................................... 142

Appendix 2.2 ................................................................................................................... 152

Appendix 2.3 ................................................................................................................... 153

Appendix 2.4 ................................................................................................................... 154

Appendix 2.5 ................................................................................................................... 155

viii

Appendix 3.1 ................................................................................................................... 156

Appendix 3.2 ................................................................................................................... 157

Appendix 3.3 ................................................................................................................... 158

Appendix 4.1 ................................................................................................................... 161

Appendix 4.2 ................................................................................................................... 164

Appendix 5.1 ................................................................................................................... 165

Appendix 5.2 ................................................................................................................... 166

Appendix 5.3 ................................................................................................................... 167

Appendix 5.4 ................................................................................................................... 168

Appendix 5.5 ................................................................................................................... 170

Appendix 5.6 ................................................................................................................... 171

Appendix 5.7 ................................................................................................................... 172

ix

LIST OF TABLES

Table 2.1. Mean effects of aquaculture sites on wildlife populations ........................................ 33

Table 3.1. Body size and condition metrics for low and high farm density groups .................... 62

Table 3.2. Egg and larval quality metrics for cod from low and high farm density areas ........... 67

Table 4.1. Population metrics for native (Coscinasterias muricata) and invasive (Asterias

amurensis) seastars inside and outside farms at Grassy Point and Clifton Springs Aquaculture

Fisheries Reserves ....................................................................................................................... 96

Table 5.1. Summary of fish recruitment to artificial reefs stocked with Undaria pinnatifida,

Ecklonia radiata, or left barren. ................................................................................................ 114

Table 5.2. Comparison of reef fish relative abundance estimated by diver catch per unit effort in

Undaria and Ecklonia habitats. ................................................................................................ 117

Table 5.3. Comparison of body condition and reproductive condition metrics in reef fishes

collected from Undaria and Ecklonia habitats. ......................................................................... 118

x

LIST OF FIGURES

Figure 1.1. Conceptual representation of habitat selection responses to habitats of varying

quality ............................................................................................................................................ 2

Figure 2.1. Distribution of research effort for studies that met the criteria for inclusion ......... 20

Figure 2.2. Distribution of research effort on interactions between aquaculture sites and wild

fauna among countries and territories ....................................................................................... 22

Figure 2.3. Summary statistics for log response ratios for each variable in our meta-analysis. 25

Figure 3.1. Map of collection sites relative to active salmon farms in south-western Norway . 54

Figure 3.2. Weight-at-length relationship for female and male cod .......................................... 61

Figure 3.3. Multidimensional scaling (MDS) plot showing dissimilarly (Euclidean distance) of

multivariate fatty acid profiles in Atlantic cod ovaries according to salmon farm density ........ 63

Figure 3.4. Daily egg production per tank during the captive spawning period ......................... 64

Figure 4.1. Population density of Asterias amurensis and Coscinasterias muricata inside and

outside the Clifton Springs and Grassy Point Fisheries Aquaculture Reserves ........................... 87

Figure 4.2. Armspan and gutted weight of Asterias amurensis inside and outside the Clifton

Springs Fisheries Aquaculture Reserve ....................................................................................... 88

Figure 4.3. Condition metrics for Asterias amurensis inside and outside the Clifton Springs

Fisheries Aquaculture Reserve .................................................................................................... 90

Figure 4.4. Density plot of Asterias amurensis size distribution inside and outside the farm

boundary at Clifton Springs Aquaculture Fisheries Reserve ....................................................... 91

Figure 4.5. Effect of prey, conspecifics and predators on habitat selection decisions by the

invasive seastar Asterias amurensis in laboratory trials ............................................................. 92

Figure 5.1. Archetypal examples of three rocky reef habitats in northern Port Phillip Bay ..... 103

Figure 5.2. Map of study locations in Port Phillip Bay, Australia .............................................. 106

Figure 5.3. Weight-at-length relationships for Heteroclinus perspicillatus, H. heptaeolus and

Neoodax balteatus .................................................................................................................... 113

Figure 5.4. Habitat choice trial results for Heteroclinus perspicillatus and Neoodax balteatus

................................................................................................................................................... 114

xi

Figure 5.5. Canonical analysis of principle coordinates (CAP) showing variation in fish

communities across underwater visual census (UVC) plots with and without Undaria pinnatifida

canopy ...................................................................................................................................... 120

Figure 5.6. Fish community metrics from underwater visual census plots .............................. 122

Figure 6.1. Conceptual diagram showing speculative positioning of responses of focal species to

novel habitats ............................................................................................................................ 130

Figure. 6.2. Conceptual representation of the ecological trap framework applied to an

assessment of the role of artificial reefs for fish production .................................................... 132

Figure. 6.3. Current distribution of Undaria pinnatifida within Port Phillip Bay in 2017 ......... 133

1

CHAPTER ONE: GENERAL INTRODUCTION

Anthropogenic impacts on the natural environment are occurring at an unprecedented rate

and spatial scale, with nearly every part of the globe affected by one or more human impacts

(Vitousek et al. 1997, Sanderson et al. 2002, Halpern et al. 2008, Vörösmarty et al. 2010). The

effects of these drivers are referred to collectively as human-induced rapid environmental

change (HIREC), and include those arising from habitat loss or habitat change, pollution,

species introductions, human harvesting, and climate change (Sih et al. 2011). As a result,

animals must now navigate a landscape comprised of modified or degraded environments that

have altered the ecological playing field from that of their evolutionary past (often to the

advantage of non-native over native species: Byers 2002, Crooks et al. 2011).

Impacts of HIREC on animal populations will depend on how individuals respond to novel cues

and environmental conditions. Animals will be best placed to survive and reproduce in

modified or degraded habitats if they exhibit adaptive behaviours, such as selecting the best

available habitats, choosing suitable food items and recognising and evading novel predators

(Sih et al. 2011, Sih 2013, Wong & Candolin 2014). However, HIREC frequently outpaces the

evolution of the indirect decision-making cues used by animals to assess current and future

conditions, which can result in individuals incorrectly evaluating risks and resources

(Schlaepfer et al. 2002).

Novel habitats, ecological traps and metapopulation ecology

These scenarios, in which ecological novelty leads individuals to choose behaviours or habitats

that lead to poor fitness outcomes, are termed evolutionary traps (Schlaepfer et al. 2002,

Robertson et al. 2013). Incorrect evaluations may arise through one of two sensory

mechanisms: a reliance on ‘outdated’ cues that no longer reliably predict fitness outcomes

because conditions have changed, or the introduction of novel cues that mimic or overpower

pre-existing cues but lead to poor fitness outcomes (Schlaepfer et al. 2002, Robertson et al.

2013, Wong & Candolin 2014).

A specific case of the evolutionary trap, the ecological trap, occurs when individuals either

prefer or fail to avoid low quality habitats when higher quality alternatives are available

(Robertson & Hutto 2006, Hale, Treml, et al. 2015). In doing so, they also fall into the corollary,

2

a perceptual trap, whereby individuals avoid or fail to prefer a relatively high quality habitat

(Fig. 1.1, Kokko and Sutherland 2001; Gilroy and Sutherland 2007; Patten and Kelly 2010).

Although the ecological trap concept concerns individual-level phenomena, traps directly drive

population-level processes, and can exacerbate population-level effects of environmental

change in impacted landscapes. Attractive population sinks caused by ecological traps can

impact metapopulation persistence disproportionately by drawing in animals that would

otherwise settle in source habitats (Battin 2004, Hale, Treml, et al. 2015). Similarly, perceptual

traps cause potential source habitats to be underutilised (Fig. 1.1), increasing effective habitat

loss in fragmented landscapes and potentially driving Allee effects for any individuals that do

choose to reside in the perceptual trap habitat (Kokko & Sutherland 2001, Gilroy & Sutherland

2007, Patten & Kelly 2010).

Habitat quality

High (potential source)

Low (potential sink)

Habitat selection

response

Preferred Correct decision Ecological trap

Avoided Perceptual trap Correct decision

Figure 1.1. Conceptual representation of the ecological trap framework in terms of possible

habitat selection responses to habitats of varying quality, sensu Patten and Kelly (2010).

In early source-sink metapopulation models, population growth was considered to be

approximately proportional to the spatial extent of source and sink habitats, with populations

in sink habitats maintained by passive dispersal from source habitats (Holt 1985). Later,

researchers began to incorporate active dispersal, allowing individuals to preferentially settle

in the best available habitats (population sources), and then spill over into more marginal

habitats in a density-dependent fashion (Pulliam 1988). Such models have significantly better

predictive ability than their predecessors and drove an important shift in thinking about

metapopulation ecology (Dias 1996). However, by assuming that individuals will tend to make

adaptive habitat selections, these models are vulnerable to making overly optimistic

predictions about population persistence whenever this assumption is violated, with

potentially dire consequences for the management of threatened animal populations (Battin

2004). The ecological trap concept provides a framework for assessing habitats through the

lens of individual habitat preferences and habitat quality (Fig. 1.1). This framework can, by

3

linking habitat-specific indices of habitat preference and fitness, distinguish between attractive

population sources (‘traditional’ population sources), attractive population sinks (ecological

traps), unattractive potential population sources (perceptual traps), and unattractive

population sinks (‘traditional’ population sinks) (Fig. 1.1, Patten and Kelly 2010; Hale et al.

2015; Hale and Swearer 2016). Incorporating such an assessment can improve the predictive

ability of source-sink models, and in some cases may fundamentally alter expectations of

population persistence in degraded or modified environments (Kokko & Sutherland 2001,

Battin 2004, Hale, Treml, et al. 2015). Despite this, uptake of the ecological trap assessment

framework has been relatively slow (Hale & Swearer 2016).

Current evidence for ecological traps

In their influential review, Robertson and Hutto (2006) proposed a set of criteria for

demonstrating an ecological trap: (1) there must be a suitable measure of habitat preference

indicating preference for, or non-avoidance of, the putative trap habitat relative to alternative

habitats, and (2) there must be a reasonable surrogate measure of fitness that is significantly

lower in the putative trap habitat relative to alternative habitats. Put simply, the preferred

habitat should not confer the best fitness outcomes. Habitat preference can be particularly

difficult to assess in the field, especially in systems where observed population densities are

determined by a combination of passive recruit supply, active habitat preference and early

stage survival (Underwood & Fairweather 1989, Hixon & Beets 1993, Shima 2001, Stevens

2003, Railsback et al. 2003). Robertson and Hutto (2006) proposed that estimates of habitat

preference should be based on one or more of the following lines of evidence: settlement

patterns, distribution of dominant individuals, temporal variance in population size, and

controlled choice experiments. Robertson and Hutto (2006) considered that only 5/45 putative

demonstrations at that time had provided sufficient evidence of habitat preference.

Terrestrial systems

While ecological traps may arise naturally (where a broadly adaptive preference is maladaptive

in certain situations), most proposed and demonstrated cases have been driven by HIREC (Hale

& Swearer 2016). The bulk of evidence for ecological traps comes from terrestrial systems,

4

with birds and ovipositing insects providing especially convenient model systems as fitness can

be easily compared between habitats by tracking clutch sizes and survivorship of adults or

offspring.

Habitat changes associated with human land use are perhaps the most well-documented

drivers of ecological traps. The creation of artificial materials such as plastic, asphalt, oils, glass,

and paint can provide hyperattractive cues that lead insects to choose unsuitable ovipositing

sites (e.g. Horváth et al. 1998; Kriska et al. 1998), while built structures may mimic natural/

nesting sites for birds, but offer poor fitness outcomes (Sumasgutner et al. 2014).

Similarly, human disturbance of native vegetation may mimic natural grasslands or disturbed

forest, but present novel risks for animals that choose to reside in these areas. Balme et al.

(2010) found that the efficacy of a wildlife reserve was reduced by a failure of leopards to

avoid adjacent areas of farmland, where they come into conflict with humans. Negative edge

effects were also observed in birds nesting in vegetation patches fragmented by human land

use: indigo buntings preferentially nested in patches with a high edge length to area ratio—an

adaptation that allows them to exploit natural disturbance—but experienced high chick

mortality in anthropogenic patches (Weldon & Haddad 2005).

Habitat structure is also influenced by exotic or invasive habitat-forming plants, and such

species can drive the creation of ecological traps by altering the quality of the habitat without

affecting associated attractive cues. Lloyd and Martin (2005) found that the presence of an

exotic grass did not alter the attractiveness of a native prairie habitat but was associated with

lower nest survival for longspurs, while Rodewald et al. (2011) reported lower reproductive

success in cardinals that selected exotic honeysuckle as nesting substrate in urban areas. Some

traps may arise due to phenology of exotic habitat-formers: returning migrant blackcaps

preferentially settled in plantations of an exotic tree, perhaps because the associated shrubs

develop spring foliage earlier, but experienced reduced nesting success (Remeš 2003).

Aquatic systems

Marine and freshwater systems have not escaped the reach of HIREC (Halpern et al. 2008,

Vörösmarty et al. 2010), yet there has been comparatively little research effort on behavioural

responses of aquatic fauna to habitat changes (Hale, Coleman, et al. 2015, Hale & Swearer

5

2016). This lack of research reflects difficulties in obtaining habitat preference and fitness

metrics in many aquatic animals, especially broadcast spawning fishes, but it is nonetheless

important to broaden the research effort beyond North American and European birds, as

terrestrial and aquatic fauna are likely to differ in their vulnerability to the various causes of

ecological traps (Hale & Swearer 2016).

Aquatic fauna are likely to be particularly vulnerable to ecological traps caused by pollution,

given the tendency for aquatic environments to transport and accumulate pollutants from

surrounding watersheds, intensifying exposure to toxins (Hale, Coleman, et al. 2015). Perhaps

as a result, pollution-driven ecological traps were among the earliest documented in the

aquatic environment. Juvenile flatfish detect and avoid heavily-oiled sediments, but not lightly-

oiled sediments (Moles et al. 1994). Likewise, juvenile crabs select preferred sediment grain

size regardless of the presence of oil, despite avoiding the oil when it was present on less

preferred sediments (Moles & Stone 2002). Negative fitness effects of oil contamination mean

that affected sediments are likely to function as ecological traps (Moles & Norcross 1998, Khan

2003). Other forms of pollution, such as noise or seismic pollution tend to be avoided,

although exposure prior to avoidance, or sub-avoidance levels of exposure, may have

deleterious effects that could lead to ecological trap formation to the extent that they co-

occur with particular habitats or locations (Codarin et al. 2009, Miller et al. 2014).

Much of the subsequent work on marine ecological traps has assessed habitat selection in

relation to fishing pressure. Natural floating debris signals oceanic convergence zones and high

food availability for pelagic predators, but there is now broad-scale use of artificial fish

aggregation devices (FADs) deployed by fishing fleets in non-convergence zones. These FADs

are associated with lower food availability and high fishing mortality for fish that are attracted

to them (Marsac et al. 2000, Hallier & Gaertner 2008). Similarly, casitas—concrete shelters

used to attract lobsters for harvesting—act as ecological traps for lobsters that find the casitas

more attractive than natural reefs (Gutzler et al. 2015). Fishing pressure may drive ecological

traps even without the need for attractive structures: Abrams et al. (2012) predicted that

some animals may find harvested areas to be more attractive than adjacent marine protected

areas because: (a) potential predators and competitors are rare due to harvesting pressure, (b)

prey species are abundant due to the rarity of predators, and (c) the risk of harvesting

mortality is undetectable. Indirect fishing-driven traps also arise for predators that compete

with fisheries for food. Verhulst et al. (2004) found that oystercatcher populations did not

redistribute following the establishment of a shellfish reserve, despite individuals outside

reserves having fewer shellfish in the diet, poorer condition indices and higher mortality rates.

6

Similarly, African penguins rely on sea surface temperature and primary productivity cues to

find optimum feeding areas, but a combination of fishing pressure and climate change has

driven large scale changes in the size and distribution of forage fish populations such that

these cues no longer reliably predict prey abundance (Sherley et al. 2017).

High densities of wildlife at some marine and freshwater aquaculture systems have prompted

research into the effects of association with farms on wildlife populations. Šigutová et al.

(2015) found that pond management regimes caused high mortality rates for endangered

dragonfly larvae, and recommended measures to make the ponds less attractive to ovipositing

dragonflies. Stocking of fish ponds or sea cages for aquaculture can also alter the availability of

food for piscivorous birds, leading to ecological traps where food availability is lower than

expected (Kloskowski 2012), where high densities of breeding waterbirds attracted to fish

ponds in turn attract nest predators (Broyer et al. 2017), or where culling is carried out (Quick

et al. 2004, Callier et al. 2017). Wild fish attracted to sea cage fish farms had higher condition

indices and a lower incidence of internal parasites, but also higher incidences of external

parasites (Dempster et al. 2011), while captive feeding trials indicate that consuming farm

waste may have deleterious effects on reproduction (Lavens et al. 1999, Mazorra et al. 2003,

Salze et al. 2005, Bogevik et al. 2012). Other studies indicate that marine infrastructure

(whether for aquaculture or other purposes) is probably both attractive and productive for

many fish and benthic invertebrate populations (Reubens et al. 2013, Borgert 2015).

Exotic or invasive species can have disastrous effects on native animal populations (Clavero &

García-Berthou 2005), and are drivers of several demonstrated terrestrial ecological traps

(Robertson & Hutto 2006, Hale & Swearer 2016). Despite this, there have been few

assessments of potential ecological traps driven by non-native aquatic species (candidates may

include novel predators, prey, competitors or habitats). One such case concerns the invasive

macroalga Caulerpa taxifolia, a habitat-forming species that functions as an ecosystem

engineer in soft sediment habitats. In Australia, native bivalves do not avoid C. taxifolia as a

substrate, and recruit in large numbers (Gribben & Wright 2006, Gribben et al. 2009). Recruits

fare well, but there is evidence that mortality risk over an individual’s lifetime is higher in the

invasive habitat (Gribben & Wright 2006, Wright & Gribben 2008, Byers et al. 2010).

7

Thesis aims and structure

Habitat association studies are a cornerstone of the animal ecology field, yet only a fraction

provide sufficient evidence to place habitats within a source–sink metapopulation framework,

let alone determine whether individuals utilising or avoiding habitats are making adaptive

habitat selection decisions (Hale & Swearer 2016). This is especially true for marine systems,

where most investigations provide only basic population- or community-level data. The

examples cited in the previous section constitute nearly all published investigations of

potential ecological traps in the marine environment. This is despite the ecological trap

concept being first articulated more than four decades ago (Dwernychuk & Boag 1972, Gates &

Gysel 1978), and refined over subsequent decades (Battin 2004, Robertson & Hutto 2006,

Gilroy & Sutherland 2007, Patten & Kelly 2010, Hale & Swearer 2016). Related calls have been

made regarding the analogous attraction–production controversy around fish populations on

artificial reefs (Bohnsack 1989, Osenberg et al. 2002, Brickhill et al. 2005), but studies that

separate these key population processes remain rare.

Accordingly, this thesis examines habitat selection and fitness consequences for animals in

several human-impacted coastal marine environments using the ecological trap assessment

framework, and aims to link these individual-level traits to the potential for population

persistence in degraded environments. The aim was not necessarily to demonstrate the

existence of ecological traps in these environments, but rather to demonstrate the application

of the ecological trap assessment framework to questions around the value of novel habitats

for fauna. I worked across three study systems modified by aquaculture and invasive species to

demonstrate the broad applicability of this individual-level approach.

Chapters 2-5 were written as standalone manuscripts, with minor changes to fit the thesis

format. As a result, some repetition of key concepts—particularly HIREC and ecological trap

theory—has been unavoidable. I have minimised unnecessary repetition by confining

descriptions of my study systems to the relevant chapters.

In Chapter Two, I conduct a global systematic review and meta-analysis of research on

interactions between all forms of aquaculture and vertebrate wildlife, with an emphasis on the

impacts of aquaculture on the distribution, fitness and population persistence of wildlife. I

demonstrate that there are considerable knowledge gaps around very basic questions,

including individual responses to aquaculture (i.e. the degree of attraction or repulsion relative

8

to natural habitats), and whether proximity to aquaculture is likely to have positive or negative

effects on fitness and population persistence of vertebrate wildlife.

Chapter Three concerns one of the key knowledge gaps highlighted by Chapter 2: the effect of

proximity to aquaculture on the fitness of wild fish. We have multiple lines of evidence that

wild fish are attracted to aquaculture (see Chapter 2), where they experience high food

availability (Dempster et al. 2011), but the effects of proximity to aquaculture on reproductive

fitness are poorly known. In this chapter, I collect adult Atlantic cod (Gadus morhua) in

spawning condition from sites within two areas of high and low salmon farming density, and

conduct a hatchery spawning experiment to compare egg production, egg quality, and larval

fitness metrics.

In Chapter Four, I study a complex interaction between shellfish aquaculture, a native

keystone predator (the eleven-arm seastar Coscinasterias muricata) and an invasive keystone

predator (the northern Pacific seastar Asterias amurensis). Shellfish aquaculture provides a

large trophic subsidy for benthic fauna, and has the potential to act as an important population

source for both species. However, as the native seastar has been reported to prey upon the

invasive seastar, there may be some predation pressure occurring that reduces the value of

the shellfish farm habitat for the invader, potentially leading to an ecological trap. Accordingly,

I investigate the behavioural responses of the invasive seastar to the shellfish farm habitat and

the native predator, and assess fitness metrics for seastars inside and outside the farms.

In Chapter Five, I consider the attractiveness and fitness value of habitat created by an

invasive ecosystem engineer (the wakame kelp Undaria pinnatifida) for native fish on urban-

impacted temperate reefs. This invasive habitat-forming kelp provides a seasonal canopy on

reefs where urchin-grazing and urban impacts have driven the decline of native kelp cover. I

combine a laboratory habitat choice experiment, recruitment data on mesocosm reefs, fish

community surveys on natural reefs, and fitness metrics to provide a holistic assessment of the

relative value of this novel habitat.

Finally, Chapter Six provides a general discussion of this body of work, and reflects on the

utility of the ecological trap framework in assessing the role of novel habitats in population

persistence of marine fauna.

9

REFERENCES

Abrams PA, Ruokolainen L, Shuter BJ, McCann KS (2012) Harvesting creates ecological traps:

consequences of invisible mortality risks in predator-prey metacommunities. Ecology

93:281–293

Balme GA, Slotow R, Hunter LTB (2010) Edge effects and the impact of non-protected areas in

carnivore conservation: leopards in the Phinda–Mkhuze Complex, South Africa. Anim

Conserv 13:315–323

Battin J (2004) When good animals love bad habitats: ecological traps and the conservation of

animal populations. Conserv Biol 18:1482–1491

Bogevik AS, Natário S, Karlsen Ø, Thorsen A, Hamre K, Rosenlund G, Norberg B (2012) The

effect of dietary lipid content and stress on egg quality in farmed Atlantic cod Gadus

morhua. J Fish Biol 81:1391–405

Bohnsack JA (1989) Are high-densities of fishes at artificial reefs the result of habitat limitation

or behavioral preference. Bull Mar Sci 44:631–645

Borgert RJ (2015) Value of off-bottom oyster aquaculture baskets as habitat for juvenile blue

crab. MSc thesis. Auburn University, Alabama

Brickhill MJ, Lee SY, Connolly RM (2005) Fishes associated with artificial reefs: attributing

changes to attraction or production using novel approaches. J Fish Biol 67:53–71

Broyer J, Chavas G, Chazal R (2017) The effects of cessation of fish farming on duck breeding in

French fishpond systems. Hydrobiologia 788:47–53

Byers JE (2002) Impact of non-indigenous species on natives enhanced by anthropogenic

alteration of selection regimes. Oikos 97:449–458

Byers JE, Wright JT, Gribben PE (2010) Variable direct and indirect effects of a habitat-

modifying invasive species on mortality of native fauna. Ecology 91:1787–1798

Callier MD, Byron CJ, Bengtson DA, Cranford PJ, Cross SF, Focken U, Jansen HM, Kamermans P,

Kiessling A, Landry T, O’Beirn F, Petersson E, Rheault RB, Strand Ø, Sundell K, Svåsand T,

Wikfors GH, McKindsey CW (2017) Attraction and repulsion of mobile wild organisms to

finfish and shellfish aquaculture: a review. Rev Aquac. DOI: 10.1111/raq.12208

Clavero M, García-Berthou E (2005) Invasive species are a leading cause of animal extinctions.

Trends Ecol Evol 20:110

Codarin A, Wysocki LE, Ladich F, Picciulin M (2009) Effects of ambient and boat noise on

hearing and communication in three fish species living in a marine protected area

(Miramare, Italy). Mar Pollut Bull 58:1880–1887

10

Crooks JA, Chang AL, Ruiz GM (2011) Aquatic pollution increases the relative success of

invasive species. Biol Invasions 13:165–176

Dempster T, Sanchez-Jerez P, Fernandez-Jover D, Bayle-Sempere JT, Nilsen R, Bjørn P-A, Uglem

I (2011) Proxy measures of fitness suggest coastal fish farms can act as population

sources and not ecological traps for wild gadoid fish. PLoS ONE 6:e15646

Dias PC (1996) Sources and sinks in population biology. Trends Ecol Evol 11:326–330

Dwernychuk LW, Boag DA (1972) Ducks nesting in association with gulls - an ecological trap?

Can J Zool 50:559–563

Gates JE, Gysel LW (1978) Avian nest dispersion and fledging success in field-forest ecotones.

Ecology 59:871–883

Gilroy JJ, Sutherland WJ (2007) Beyond ecological traps: perceptual errors and undervalued

resources. Trends Ecol Evol 22:351–356

Gribben PE, Wright JT (2006) Invasive seaweed enhances recruitment of a native bivalve: roles

of refuge from predation and the habitat choice of recruits. Mar Ecol Prog Ser 318:177–

185

Gribben PE, Wright JT, O’Connor WA, Steinberg P (2009) Larval settlement preference of a

native bivalve: the influence of an invasive alga versus native substrata. Aquat Biol 7:217–

227

Gutzler BC, Butler MJ, Behringer DC (2015) Casitas: a location-dependent ecological trap for

juvenile Caribbean spiny lobsters, Panulirus argus. ICES J Mar Sci 72:i177–i184

Hale R, Coleman R, Pettigrove V, Swearer SE (2015) Identifying, preventing and mitigating

ecological traps to improve the management of urban aquatic ecosystems. J Appl Ecol

52:928–939

Hale R, Swearer SE (2016) Ecological traps: current evidence and future directions. Proc R Soc B

Biol Sci 283:494–499

Hale R, Treml EA, Swearer SE (2015) Evaluating the metapopulation consequences of

ecological traps. Proc R Soc London B Biol Sci 282:20142930

Hallier J-P, Gaertner D (2008) Drifting fish aggregation devices could act as an ecological trap

for tropical tuna species. Mar Ecol Prog Ser 353:255–264

Halpern BS, Walbridge S, Selkoe KA, Kappel C V, Micheli F, D’Agrosa C, Bruno JF, Casey KS,

Ebert C, Fox HE, Fujita R, Heinemann D, Lenihan HS, Madin EMP, Perry MT, Selig ER,

Spalding M, Steneck R, Watson R (2008) A global map of human impact on marine

ecosystems. Science 319:948–952

Hixon MA, Beets JP (1993) Predation, prey refuges, and the structure of coral‐reef fish

11

assemblages. Ecol Monogr 63:77–101

Holt RD (1985) Population dynamics in two-patch environments: some anomalous

consequences of an optimal habitat distribution. Theor Popul Biol 28:181–208

Horváth G, Bernáth B, Molnár G (1998) Dragonflies find crude oil visually more attractive than

water: multiple-choice experiments on dragonfly polarotaxis. Naturwissenschaften

85:292–297

Khan RA (2003) Health of flatfish from localities in Placentia Bay, Newfoundland, contaminated

with petroleum and PCBs. Arch Environ Contam Toxicol 44:485–492

Kloskowski J (2012) Fish stocking creates an ecological trap for an avian predator via effects on

prey availability. Oikos 121:1567–1576

Kokko H, Sutherland WJ (2001) Ecological traps in changing environments: ecological and

evolutionary consequences of a behaviourally mediated Allee effect. Evol Ecol Res 3:537–

551

Kriska G, Horváth G, Andrikovics S (1998) Why do mayflies lay their eggs en masse on dry

asphalt roads? Water-imitating polarized light reflected from asphalt attracts

Ephemeroptera. J Exp Biol 201:2273–2286

Lavens P, Lebegue E, Jaunet H, Brunel A, Dhert P, Sorgeloos P (1999) Effect of dietary essential

fatty acids and vitamins on egg quality in turbot broodstocks. Aquac Int 7:225–240

Lloyd JD, Martin TE (2005) Reproductive success of chestnut-collared longspurs in native and

exotic grassland. Condor 107:363–374

Marsac F, Fonteneau A, Ménard F (2000) Drifting FADs used in tuna fisheries: an ecological

trap? In: Gall J-Y Le, Cayre P, Taquet M (eds) Pêche Thonière et Dispositifs de

Concentration de Poissons, 28th edn. Actes Colloq. - IFREMER, p 537–552

Mazorra C, Bruce M, Bell JG, Davie A, Alorend E, Jordan N, Rees J, Papanikos N, Porter M,

Bromage N (2003) Dietary lipid enhancement of broodstock reproductive performance

and egg and larval quality in Atlantic halibut (Hippoglossus hippoglossus). Aquaculture

227:21–33

Miller PJO, Antunes RN, Wensveen PJ, Samarra FIP, Catarina Alves A, Tyack PL, Kvadsheim PH,

Kleivane L, Lam F-PA, Ainslie MA, Thomas L (2014) Dose-response relationships for the

onset of avoidance of sonar by free-ranging killer whales. J Acoust Soc Am 135:975–993

Moles A, Norcross BL (1998) Effects of oil-laden sediments on growth and health of juvenile

flatfishes. Can J Fish Aquat Sci 55:605–610

Moles A, Rice S, Norcross BL (1994) Non-avoidance of hydrocarbon laden sediments by

juvenile flatfishes. Netherlands J Sea Res 32:361–367

12

Moles A, Stone RP (2002) Habitat preferences of juvenile tanner and red king crabs: substrate

and crude oil. In: Paul AJ, Dawe EG, Elner R, Jamieson GS, Kruse GH, Otto RS, Sainte-Marie

B, Shirley TC, Woodby D (eds) Crabs in Cold Water Regions: Biology, Management, and

Economics. University of Alaska Sea Grant, p 631–644

Osenberg C, Mary CM St., Wilson JA, Lindberg WJ (2002) A quantitative framework to evaluate

the attraction-production controversy. ICES J Mar Sci 59:S214–S221

Patten MA, Kelly JF (2010) Habitat selection and the perceptual trap. Ecol Appl 20:2148–2156

Pulliam HR (1988) Sources, sinks, and population regulation. Am Nat 132:652–661

Quick NJ, Middlemas SJ, Armstrong JD (2004) A survey of antipredator controls at marine

salmon farms in Scotland. Aquaculture 230:169–180

Railsback SF, Stauffer HB, Harvey BC (2003) What can habitat preference models tell us? Tests

using a virtual trout population. Ecol Appl 13:1580–1594

Remeš V (2003) Effects of exotic habitat on nesting success, territory density, and settlement

patterns in the blackcap (Sylvia atricapilla). Conserv Biol 17:1127–1133

Reubens JT, Vandendriessche S, Zenner AN, Degraer S, Vincx M (2013) Offshore wind farms as

productive sites or ecological traps for gadoid fishes? Impact on growth, condition index

and diet composition. Mar Environ Res 90:66–74

Robertson BA, Hutto RL (2006) A framework for understanding ecological traps and an

evaluation of existing evidence. Ecology 87:1075–1085

Robertson BA, Rehage JS, Sih A (2013) Ecological novelty and the emergence of evolutionary

traps. Trends Ecol Evol 28:552–560

Rodewald AD, Shustack DP, Jones TM (2011) Dynamic selective environments and evolutionary

traps in human‐dominated landscapes. Ecology 92:1781–1788

Salze G, Tocher DR, Roy WJ, Robertson DA (2005) Egg quality determinants in cod (Gadus

morhua L.): egg performance and lipids in eggs from farmed and wild broodstock. Aquac

Res 36:1488–1499

Sanderson EW, Jaiteh M, Levy MA, Redford KH, Wannebo A V, Woolmer G (2002) The human

footprint and the last of the wild. Bioscience 52:891–904

Schlaepfer MA, Runge MC, Sherman PW (2002) Ecological and evolutionary traps. Trends Ecol

Evol 17:474–480

Sherley RB, Ludynia K, Dyer BM, Lamont T, Makhado AB, Roux J-P, Scales KL, Underhill LG,

Votier SC (2017) Metapopulation tracking juvenile penguins reveals an ecosystem-wide

ecological trap.

Shima JS (2001) Recruitment of a coral reef fish: roles of settlement, habitat, and

13

postsettlement losses. Ecology 82:2190–2199

Šigutová H, Šigut M, Dolný A (2015) Intensive fish ponds as ecological traps for dragonflies: an

imminent threat to the endangered species Sympetrum depressiusculum (Odonata:

Libellulidae). J Insect Conserv 19:961–974

Sih A (2013) Understanding variation in behavioural responses to human-induced rapid

environmental change: a conceptual overview. Anim Behav 85:1077–1088

Sih A, Ferrari MCO, Harris DJ (2011) Evolution and behavioural responses to human-induced

rapid environmental change. Evol Appl 4:367–387

Stevens BG (2003) Settlement, substratum preference, and survival of red king crab

Paralithodes camtschaticus (Tilesius, 1815) glaucothoe on natural substrata in the

laboratory. J Exp Mar Biol Ecol 283:63–78

Sumasgutner P, Nemeth E, Tebb G, Krenn HW, Gamauf A (2014) Hard times in the city -

attractive nest sites but insufficient food supply lead to low reproduction rates in a bird

of prey. Front Zool 11:48

Underwood AJ, Fairweather PG (1989) Supply-side ecology and benthic marine assemblages.

Trends Ecol Evol 4:16–20

Verhulst S, Oosterbeek K, Rutten AL, Ens BJ (2004) Shellfish fishery severely reduces condition

and survival of oystercatchers despite creation of large marine protected areas. Ecol Soc

9:17

Vitousek PM, Mooney HA, Lubchenco J, Melillo JM (1997) Human domination of Earth’s

ecosystems. Science (80- ) 277:494–499

Vörösmarty CJ, McIntyre PB, Gessner MO, Dudgeon D, Prusevich A, Green P, Glidden S, Bunn

SE, Sullivan CA, Liermann CR, others (2010) Global threats to human water security and

river biodiversity. Nature 467:555–561

Weldon AJ, Haddad NM (2005) The effects of patch shape on indigo buntings: evidence for an

ecological trap. Ecology 86:1422–1431

Wong BBM, Candolin U (2014) Behavioral responses to changing environments. Behav Ecol

26:665–673

Wright JI, Gribben PE (2008) Predicting the impact of an invasive seaweed on the fitness of

native fauna. J Appl Ecol 45:1540–1549

14

CHAPTER TWO: IMPACTS OF MARINE AND FRESHWATER AQUACULTURE ON WILDLIFE: A

GLOBAL META-ANALYSIS

ABSTRACT

The global expansion of aquaculture has raised concerns about its environmental impacts,

including effects on wildlife. Aquaculture farms are thought to repel some species and function

as either attractive population sinks (‘ecological traps’) or population sources for others. We

conducted a systematic review and meta-analysis of empirical studies documenting

interactions between aquaculture operations and vertebrate wildlife. Farms were associated

with elevated local abundance and diversity of wildlife, although this overall effect was

strongly driven by aggregations of wild fish at sea cages and shellfish farms (abundance: 72x;

species richness: 2.0x). Birds were also more diverse at farms (1.1x), but other taxa showed

variable and comparatively small effects. Larger effects were reported when researchers

selected featureless or unstructured habitats as reference sites. Evidence for aggregation

‘hotspots’ is clear in some systems, but we cannot determine if farms act as ecological traps

for most taxa, as few studies assess either habitat preference or fitness in wildlife. Fish

collected near farms were larger and heavier with no change in body condition, but also faced

higher risk of disease and parasitism. Birds and mammals were frequently reported preying on

stock, but little data exists on the outcomes of such interactions for birds and mammals –

farms are likely to function as ecological traps for many species. We recommend researchers

measure survival and reproduction in farm-associated wildlife to make direct, causal links

between aquaculture and its effects on wildlife populations.

INTRODUCTION

Aquaculture infrastructure (farms hereafter) presents a novel environment for wild animal

populations. High stocking densities within farms aggregate biomass far beyond natural levels

(commonly 5-45 kg m-3 final biomass: FAO Fisheries and Aquaculture 2018), and in open

systems, provide considerable trophic subsidies for animals that take advantage of the

opportunity, potentially benefitting some wildlife. However, there are also deleterious effects

associated with proximity to farms, and the net impact of aquaculture on productivity and

15

persistence of wildlife populations will depend both on behavioural responses to farms and

the fitness consequences of those responses. Where individuals are attracted to a habitat that

confers poorer fitness outcomes than other available habitats, they have fallen into an

‘ecological trap’ (Robertson & Hutto 2006; Hale & Swearer 2016). While the concept is defined

at the individual level, trap habitats have population-level consequences by drawing

individuals from surrounding habitats into attractive population sinks (Hale et al. 2015). Even

in the absence of an ecological trap, changes in the abundance and spatial distribution of

influential species may indirectly affect other species and drive large-scale shifts in biodiversity

and ecosystem function (Gamfeldt et al. 2015).

A range of attractive and repulsive mechanisms for wildlife can occur simultaneously at farms

(Callier et al. 2017). The primary attractive mechanism in most systems is probably food

availability, either in the form of direct predation on stock, or an indirect trophic subsidy in the

form of farm waste (spilled feed, faeces and dead stock). Birds, pinnipeds and otters are well-

documented predators of stock at sea cage or pond fish farming systems (Carss 1993; Pitt &

Conover 1996; Adámek et al. 2003; Güçlüsoy & Savas 2003; Quick et al. 2004; Freitas et al.

2007; Dorr et al. 2012; Sepúlveda et al. 2015), while farm waste from sea cages also attracts

significant aggregations of opportunistic wild fish (Dempster et al. 2002, 2009; Tuya et al.

2006; Sanchez-Jerez et al. 2011). A high local abundance of fish is likely to lead to secondary

attraction of large predators, such as dolphins (Diaz López 2006; Piroddi et al. 2011). Shellfish

and algae farming do not require inputs of feed, but high densities of filter feeding shellfish in

farms do accumulate biomass, attracting wild fish and invertebrate species (Dealteris et al.

2004; Powers et al. 2007; McKindsey et al. 2011; Segvic-Bubic et al. 2011), while algae farming

attracts wild herbivores (Hehre & Meeuwig 2016). Farm structures themselves may also be

attractive, functioning in a similar manner to fish aggregation devices or artificial reefs

(Tallman & Forrester 2007; Sanchez-Jerez et al. 2011). Farm structures provide three-

dimensional habitat complexity, and associated light, noise and novel biofouling communities

may all be attractive to a range of wild taxa (Dumont et al. 2011; Callier et al. 2017).

Paradoxically, many of these environmental changes associated with farms, such as light,

noise, eutrophication and high densities of predators, may have repulsive effects on wary or

functionally specialised taxa (Markowitz et al. 2004; Becker et al. 2011).

Attraction to farms may increase or decrease the fitness of wildlife. One expectation is that

increased food availability will lead to faster growth, higher body condition and increased

reproductive output. Accordingly, there is some evidence that farm-associated wild fish have

higher body condition and reproductive investment indices than fish from reference sites

16

(Dempster et al. 2011), but little is known about potential benefits for other taxa. In broadcast

spawning taxa, high local population densities at farms are likely to confer greater mating

efficiency (Inglis & Gust 2003). Such benefits for farm-associated wildlife are likely to be at

least partially counteracted by potential deleterious fitness effects related to dietary shifts,

contamination, disease, parasitism, and elevated mortality rates. For example, a shift from fish

oils to terrestrially-derived ingredients in aquaculture feed may result in deficiencies of long-

chain polyunsaturated fatty acids in animals that feed regularly at farms (eg. Salze et al. 2005;

Fernandez-Jover et al. 2007a; Gonzalez-Silvera et al. 2017). Additionally, farm waste can create

an anoxic environment with significant effects on benthic and estuarine communities (Wu

1995; Yucel-Gier et al. 2007; Herbeck et al. 2013; Valdemarsen et al. 2015), while in some

areas, wildlife may also accumulate elevated tissue loadings of contaminants such as

antibiotics, pyrethroid parasiticides, metals and organohalogens (Samuelsen et al. 1992; Boyd

& Massaut 1999; Burridge et al. 2010; Bustnes et al. 2010) with potentially nontrivial effects

(e.g. Crump & Trudeau 2009; Berg et al. 2016). For fish, the primary concern may be the effect

of proximity to farms on disease and parasitism rates: high population densities within farms

create favourable conditions for outbreaks of diseases and parasites such as sea lice (Krkosek

et al. 2005, 2006; Costello 2009; Lafferty et al. 2015; Krkošek 2017). Wild fish populations may

also act as reservoirs for parasites and diseases, and as they move between cages to take

advantage of feeding opportunities, they act as potential transmission vectors that may

increase reinfection rates for farms, driving positive feedbacks (Uglem et al. 2009; Hayward et

al. 2011).

Despite this suite of environmental concerns, the aquaculture industry is the world’s fastest-

growing food production sector (FAO Fisheries and Aquaculture 2015). For this growth to be

sustainable in terms of environmental impacts and ‘social license’ to operate, the industry

must grapple with issues arising from interactions between aquaculture activities and the

natural environment and develop solutions to minimise negative effects on wildlife (and vice

versa). The first step should be to assess the state of knowledge on these issues and identify

the most severe effects. Recent reviews have outlined the range of interactions that occur

between aquaculture activities and wild fauna (e.g. Uglem et al. 2014; Taranger et al. 2015;

Glover et al. 2017; Callier et al. 2017), but there has been not yet been a quantitative global

synthesis of the impacts of aquaculture on wildlife. Here, we conduct a systematic review and

meta-analysis of studies documenting interactions between aquaculture activities and wildlife,

primarily to quantify the effects of these interactions on abundance, diversity, and fitness of

farm-associated wildlife, and secondarily to highlight potential drivers of conflict between

17

wildlife and aquaculture. Thereafter, we recommend directions for future research to address

key knowledge gaps in this area.

MATERIALS AND METHODS

Literature search and systematic review

Primary publications up to November 2017 were discovered by searching for the following

terms using the ISI Web of Science: (aquaculture OR mariculture OR "fish farm*" OR "shellfish

farm*" OR "mussel farm*" OR "oyster farm*" OR "sea cage*" OR "net pen*" OR "fish pond*"

OR "seaweed farm*" OR "macroalgal farm*" OR "algal farm*") AND (attract* OR avoid* OR

wild OR aggreg* OR impact* OR depredat* OR predat* OR disease) AND (wildlife OR animal*

OR fauna* OR fish* OR shark* OR mammal* OR dolphin* OR cetacean* OR otter* OR seal* OR

sea lion* OR bird* OR avian OR reptile* OR snake* OR amphibian* OR frog*). >9000 results

were manually screened on an individual basis, by title and abstract alone where the topic was

clearly irrelevant, or else after accessing the full text. Additional articles missed by our initial

search were discovered using informal exploratory searches using Google Scholar, and by

reading the reference lists of all relevant articles returned by our initial search. Our search

focused on interactions with vertebrate wildlife (defined here as fish, birds, mammals and

reptiles), as these animals are typically highly mobile and are therefore more able to make

decisions about whether to reside at and interact with farms.

For inclusion, publications were required to have provided empirical field data on at least one

of the following: (1) distribution, behaviour, condition, disease or mortality of wildlife in the

vicinity of aquaculture sites, or (2) direct interactions between wildlife and stock at

aquaculture sites (e.g. predation of stock). To minimise potential duplication of data, we only

included peer-reviewed English-language journal articles.

To document the distribution of research effort in the field, we recorded the year, country,

region, environment, culture system, culture taxa and the wild taxa for each study.

18

Meta-analysis

Studies were included in the subsequent meta-analysis if they provided quantitative data

sufficient to calculate effect sizes for variables at aquaculture sites relative to ‘natural’ or

‘reference’ sites. We extracted a range of quantitative variables that were representative of

the dominant types of interactions between aquaculture operations and wild vertebrates,

relating to spatial distribution (Abundance, Species Richness), size structure (Length, Weight),

food availability (Body Condition, Stomach Fullness), disease and parasite infection levels –

either infection loads on individuals or prevalence of infected individuals in the population

(Infection Level), as well as direct measures of Mortality and Fertility. Reproductive condition

metrics (e.g. relative gonad size) were considered a component of Body Condition.

Natural log response ratios were calculated for each variable: RR = ln(F/R), where F is the trait

mean at farm sites and R is the trait mean at reference sites. Taking the natural log of the

response ratio normalises the error distribution by reducing the influence of positive

responses (Hedges et al. 1999). Studies employed a variety of sampling designs, including

random or matched farm and reference sites, and stocked or unstocked farms. All were

treated as random for the purposes of this meta-analysis, with RR calculated from the mean

trait values across all farm and reference sites regardless of how sites were selected by the

authors. Where multiple complementary measures were available for a response variable, we

took the mean of those measures (for example, Fulton’s K, hepatosomatic index and

gonadosomatic index all contribute to the Body Condition variable). Where a study provided

data on a response variable from multiple species or sites, we combined data to provide a

single replicate, except where data spanned multiple culture systems (e.g. cages and ponds),

taxonomic classes (e.g. birds and mammals), environments (e.g. marine and freshwater), or

countries. No article contributed more than two studies to our database. This was done to

prevent studies that provided data on numerous species from having a disproportionate

influence on our findings, and to ensure spatial independence between replicates given the

high mobility of most species studied. Where data was provided for farms with and without

exclusion measures (e.g. fenced and unfenced sites), we used data from sites without

exclusion measures.

Some variables were not easily quantified for statistical analysis but were nonetheless

important in understanding interactions between farms and wild fauna. These included

changes in tissue fatty acid profiles, trace elements and stable isotopes, contamination from

19

antibiotics, heavy metals and other substances, and behavioural data such as residence time or

visitation rates. For these variables, we recorded the response ratio if possible, otherwise we

noted the direction or nature of the effect.

Statistical analyses

To test for a significant effect of farm-association on response variables, we checked normality

before conducting one sample t-tests on RR data (mean RR under null hypothesis of no farm

effect = 0) using R software (R Core Team 2017).

Exploratory model selection was used to determine which of the following factors best

predicted effects of farms on wildlife (abundance and species richness responses only, as

remaining responses had insufficient sample sizes for exploratory analysis): Year, Country,

Continent, Environment (Marine, Freshwater), Culture System (Cage, Pond, Longline, Rack,

Bed), Cultured Taxa (Fish, Shellfish, Crustacean, Alga), Wild Taxa (Fish, Bird, Mammal, Reptile,

Amphibian), and Reference Habitat (Structured, Unstructured). We fitted a global general

linear model using R, and employed the dredge() function in the MuMIn package (Barton 2016)

to compare the second-order Akaike’s Information Criterion (AICC) score of every possible

subset of the global model. AICC includes a correction for finite sample sizes and yields more

conservative models than AIC (Burnham and Anderson 2002). We selected the model with the

lowest AICC score, and then used the likelihood ratio to test whether the selected model

offered a significantly better fit than the null (intercept only) model, tested the significance of

model terms, and then conducted post-hoc tests with a Tukey correction to test pairwise

effects within significant model terms.

There was orders-of-magnitude variation in RRs for abundance and species richness among

studies and systems, and accordingly, the overall trends that we report may be strongly

influenced by a small number of studies with unusually large RRs. To test this possibility, we

conducted a sensitivity analysis by ranking studies (replicates) according to the absolute value

of the RR, removing the studies with the largest RR in a stepwise fashion, and rerunning the

model between each removal (Bancroft et al. 2007; Kroeker et al. 2010). We then report the

number of studies than can be removed from the dataset without altering the statistical

significance of the farm effect.

20

To test whether the geographic distribution of research effort on this topic corresponds to the

distribution of aquaculture production, we fitted a zero-inflated Poisson model (using the pscl

package for R: Zeileis et al. 2008) to compare the number of studies contributed by each

country with the reported aquaculture production (t) by that country (FAO Fisheries and

Aquaculture 2017). To account for the large disparity in peer-reviewed English language

research output between developed and developing nations, we also included the United

Nations Human Development Index as a model term (United Nations Development

Programme 2017).

Figure 2.1. Distribution of research effort for studies that met the criteria for inclusion in our

database, according to (A) Country, (B) Culture system, (C) Region, (D) Culture taxa, (E)

Environment, and (F) Wild taxa.

21

RESULTS

Our searches discovered 204 relevant studies across 191 articles published between 1978 and

2017 (Appendix 2.1). 91 studies provided comparative data on wildlife populations at farms

and reference sites suitable for inclusion in the meta-analysis of log response ratios (RR).

Distribution of research effort

There was a clear geographical bias in research effort within our database, with 114 peer-

reviewed English language studies conducted in Europe and 46 in North America (Fig. 2.1).

Among nations, Norway, the United States and Spain accounted for the most research (Fig.

2.1). Research effort across nations was significantly predicted by an interaction between the

size of the nation’s aquaculture industry and the developmental index of the nation (p = 0.03,

Appendix 2.2), wherein highly developed nations (especially those in Europe and North

America: Fig. 2.2) with large production contributed more studies than those with low

production (p <0.0001, Appendix 2.2). Several major aquaculture-producing nations were

either poorly represented or entirely absent from our database: most notably, mainland China

is by an order of magnitude the largest aquaculture producer in the world (FAO Fisheries and

Aquaculture 2017), yet was entirely absent from our database. Other leading producers,

namely Indonesia, India, Vietnam, Philippines and Bangladesh, were also either absent or

represented by only a single study.

Most studies in our database assessed interactions with wildlife in marine or estuarine

environments (Fig. 2.1), despite global animal aquaculture production being considerably

higher in freshwater environments (47 cf. 27 million t in 2014) (FAO Fisheries and Aquaculture

2015). 105/144 studies in the marine environment took place at sea cage farms, while 49/60

freshwater systems were pond-based (Fig. 2.1). Fish were the most common cultured taxa

studied (163 studies) – primarily salmonids (69 studies) in western Europe and the Americas,

and sea bream (Sparus aurata) and sea bass (Dicentrarchus labrax) in southern Europe (43

studies). The research effort on environmental effects of salmon farming is in line with the

predominance of salmonids in the marine fish farming sector, although freshwater cyprinid

culture is the most productive pisciculture sector overall (FAO Fisheries and Aquaculture

2015). Sea bream, sea bass and marine shellfish systems are overrepresented in our dataset

22

relative to the size of these sectors, perhaps due to their importance for nations with high

marine research activity (particularly Spain). Algal and crustacean culture (5 and 3 studies,

respectively) were dramatically underrepresented here relative to the size of the sectors (FAO

Fisheries and Aquaculture 2015).

Most studies reported interactions with wild fish (108 studies), followed by birds (53 studies),

mammals (38 studies), reptiles (3 studies) and amphibians (2 studies) (Fig. 2.1).

Figure 2.2. Distribution of research effort on interactions between aquaculture sites and wild

fauna among countries and territories. Production data taken from the Fishstatj database (FAO

Fisheries and Aquaculture Department 2017).

Effects on wildlife

Abundance

We discovered 65 studies that quantified the abundance of wildlife at aquaculture farming

sites compared to reference sites, using various forms of Control-Impact (CI), Before-After (BA)

and Control-Impact-Before-After (BACI) designs. These studies used a variety of sampling

methods, including visual census, catch-per-unit-effort and tagging/tracking. 17 studies

reported a lower abundance near farms, two no difference, and 46 a higher abundance. The

23

mean effect was a 49x increase in abundance near farms (RR = 1.05, t64 = 4.3, p <0.0001), but

this value was strongly influenced by a few outlier studies reporting very large aggregations of

wild fish around sea cages (for example, a mean 1327x increase over three sampling dates at

one Australian offshore farm compared to featureless mid-water reference sites: Dempster et

al., 2004). Fish demonstrated the largest abundance changes, while changes in bird and

mammal abundance were highly variable in both effect size and direction and not significantly

different to zero (Fig. 2.3, Table 1). We were not able to calculate RR for an additional six

studies reporting differential abundance at farms (fish: 2/2 higher; mammals: 1/2 higher; birds:

2/2 higher).

A sensitivity analysis revealed that it was possible to conduct stepwise removal of 25/65

replicates with the largest effect sizes without losing statistical significance, indicating that the

overall trend was robust. However, when studies that assessed changes in wild fish abundance

at sea cage systems were omitted from the analysis, the remaining studies did not provide

support for an overall effect of aquaculture on wildlife abundance (t38 = 0.81, p = 0.42),

indicating that wild fish aggregations around sea cages were largely responsible for this overall

effect.

Model selection indicated that differential abundance was best predicted by a model

containing Environment, Cultured Taxa and Reference Habitat (R2 = 0.33, F = 7.4, p <0.0001;

Appendix 2.3). The Cultured Taxa term was significant (p = 0.0001), as was Reference Habitat

(p = 0.003), while Environment was not (p = 0.12). Post-hoc testing revealed that increases in

abundance of wild fauna tended to be higher at fish farms than at shellfish farms (p = 0.001)

and that studies comparing abundance at farm sites to unstructured or featureless reference

sites (e,g, sandy seabed or open ocean) generally found a larger response than those that

chose natural reef, unstocked farms or other structured habitats as reference sites (p = 0.006;

Appendix 2.4).

Species richness

Most studies only assessed a limited number of target species, but 37 studies provided useful

data on species richness at farms and reference sites. Of these, all but six reported higher

species richness at farm sites, with a mean 1.7x increase (RR = 0.30, t36 = 3.1, p = 0.004). This

effect was strongest in fish (RR = 0.43) but was also significant in birds (RR = 0.13) (Table 1).

24

Only one study in our database quantified differential species richness in mammals, and one in

amphibians (Roycroft et al. 2004; Kloskowski 2010, Table 1).

There was large variation in effect size and direction across studies, but a sensitivity analysis

found the overall trend to be remarkably robust (25/37 studies removed without losing

statistical significance). As with abundance, the effect was not significant when sea cage

systems were omitted from the analysis (t21 = 1.6, p = 0.11).

Species richness effects were best predicted by a model containing Reference Habitat and Wild

Taxa (R2 = 0.27, F = 4.3, p = 0.007; Appendix 2.3). Post-hoc testing revealed that the only

significant pairwise effect was between fish and amphibians, with fish species richness

positively affected and amphibian species richness negatively affected by the presence of

aquaculture sites in their respective environments (p = 0.03; Appendix 2.5).

Size structure, body condition and stomach fullness

Wild fish collected near aquaculture sites were on average 1.2x larger and 1.7x heavier than

their counterparts from reference sites (Table 1), but no size comparisons were available for

non-fish taxa.

Most studies (11) reported trends toward higher condition metrics in farm-associated wildlife,

while two found no difference and two lower condition metrics at farms, although there was

no significant effect overall (Table 1). Similarly, 8/11 studies found higher rates of stomach

fullness in farm-associated wildlife, but these effects tended to be small and were not

significant overall (Table 1). All but two comparisons of body condition or stomach fullness

data concerned wild marine fish, while Gregory and Nelson (1991) estimated a 1.9x higher rate

of stomach fullness in snakes at fish hatcheries, and Kloskowski et al. (2017) reported higher

physiological stress indicators (=lower condition for our purposes) in grebes nesting on fish

ponds.

Physiological changes

All 16/17 studies that reported looking for physiological or dietary changes in farm-associated

wild fish relative to those from reference sites found evidence of dietary shifts, while the

25

remainder found only minor differences in stable isotopes (Johnston et al. 2010). Evidence for

dietary shift included farm feed pellets in the stomachs of farm-associated wild fish (Skog et al.

2003; Fernandez-Jover et al. 2011; Arechavala-Lopez et al. 2011), taxonomic changes in

stomach contents (Demétrio et al. 2012; Fernandez-Jover & Sanchez-Jerez 2015), higher tissue

fat content and altered tissue fatty acid profiles that reflected the terrestrial-origin of lipids in

farm feed (Skog et al. 2003b, Fernandez-Jover et al. 2007, 2011, Arechavala-Lopez et al. 2011,

2015, Abaad et al. 2016). Arechavala-Lopez et al. (2015) also reported differing trace element

profiles in saithe near and far from salmon farms, while two studies reported altered taste and

other metrics of quality (Skog et al. 2003; Bogdanović et al. 2012).

Figure 2.3. Summary statistics for log response ratios (RR) for each variable in our meta-

analysis. All taxa are included. Boxes denote median, lower (25 %) and upper (75 %) quartiles,

whiskers denote 1.5x interquartile range. Data points are ‘jittered’ for clarity. Asterisk indicates

variables for which higher RR corresponds to poorer outcomes.

Contamination

Comparisons of contaminant levels in the tissues of farm-associated and non-associated fish

revealed mixed results. All three studies that tested for antimicrobial contamination in farm-

associated wild fish at farms where antimicrobials were in use found evidence of antimicrobial

26

residue in the majority of fish sampled, including oxytetracycline (0.2-1.3 µg g-1 muscle tissue:

Björklund et al. 1990), oxolinic acid (0.4-4.4 µg g-1 muscle tissue at two farms: Samuelsen et al.

1992) and flumequine (1.0-4.9 µg g-1 muscle tissue: Ervik et al. 1994). In each case, mean

concentrations for positive samples exceed the current European Union limits for these

substances in skin and muscle of finfish for human consumption: oxytetracycline: 0.1 µg g-1;

oxolinic acid: 0.1 µg g-1; flumequine: 0.6 µg g-1 (European Union 2010). It should be noted that

the development of new vaccines has allowed fish farmers in some areas (e.g. salmonid farms

in Norway and Scotland) to largely cease antimicrobial use despite rapid expansion of the

industry, but use remains high in other regions (Watts et al. 2017). It remains unclear whether

antimicrobial residue impacts fitness in farm-associated wild fish, whether through toxicity,

loss of gut microbiota or antimicrobial resistance in pathogens.

There have also been assessments of organohalogens and metals in the tissues of farm-

associated wild fish. One study reported significantly higher levels of organochlorines and

polybrominated diphenyl ethers in farm-associated fish relative to those from reference sites

(1.5x higher in cod, 1.2x higher in saithe: Bustnes et al. 2010). Another reported higher levels

(2.1x) of mercury in tissues of farm-associated rockfish (Sebastes spp.), potentially related to

an increase in trophic level near farms (DeBruyn et al. 2006). In the most comprehensive study

to date, Bustnes et al. (2011) measured concentrations of 30 elements in cod and saithe livers

from three regions in Norway. In saithe, Hg (2.0x), U (1.4x), Cr (1.9x) and Mn (1.6x)

concentrations were significantly higher in farm-associated fish, while Se, Zn, Cd, Cs and As

were higher at reference sites. In cod, U (1.4x), Al (1.5x) and Ba (1.9x) were higher in farm-

associated fish, while Se, Zn, Cd, Cs and As were higher at reference sites. While there is

evidence that some metals accumulate in sediments under fish farms, there is little evidence

so far that farm-associated wild fish are accumulating high concentrations in their tissues.

Infection rates

We discovered 22 studies that empirically investigated viral, bacterial or parasite transmission

between farmed and wild populations. In all cases, the authors concluded that the risk of

infection was either unchanged or elevated by interactions between farms and wild fish

populations. Of the 11 studies that quantified changes in infection levels with the presence of

active fish farms, all found higher levels of infection in farm-associated wild fish, with a mean

16x increase overall (RR = 2.1, t10 = 5.5, p = 0.0003). This large effect was primarily driven by

27

eight studies of sea louse infection loads on wild salmonids near salmon farms (3.5-73x

increase, RR = 2.5). One study reported higher infection densities of external parasites but

lower densities of internal parasites in farm-associated gadids, probably as a result of

consuming fewer infected wild fish and invertebrates in favour of commercial feed (Dempster

et al. 2011). Three studies provided molecular evidence for likely viral or bacterial transmission

between cultured and wild fish in the Mediterranean Sea (Zlotkin et al. 1998; Diamant et al.

2000; Colorni et al. 2002), and a molecular analysis of stomach contents revealed that wild cod

consumed escaped salmon stock infected with piscine reovirus (Glover et al. 2013). However,

molecular evidence did not always support the transmission hypothesis: Mladineo et al. (2009)

reported that monogenean and isopod parasites were not transmitted between wild and

farmed fish at one Mediterranean Sea farm.

Survivorship and fertility rates

Only two studies in our database estimated differential mortality rates in farm-associated

fauna. Kilambi et al. (1978) used age structure to infer a 21 % increase in survivorship of

largemouth bass following the establishment of cage culture in a freshwater lake, while in

contrast, Broyer et al. (2017) recorded 39 % higher mortality of ducklings at fish ponds. In sea

cage systems, elevated external parasitism rates (especially sea louse infections on salmonids)

may increase mortality in farm-associated fish, but to our knowledge, differential mortality

between farm and reference sites has not yet been empirically demonstrated. A further six

studies quantified culling of numerous birds at farms but did not compare mortality rates at

farms to those at reference sites. Two others reported dolphins being accidentally drowned in

anti-predator nets (Kemper & Gibbs 2001; Diaz López & Bernal-Shirai 2007), but again, did not

benchmark these against natural mortality rates. Several studies noted higher fishing effort

adjacent to sea cages, although we are only aware of two studies that quantified fishing effort

and catch rates (Akyol & Ertosluk 2010; Bacher & Gordoa 2016), and none assessed fishing

mortality rates among farm-associated fish.

Estimates of fertility (i.e. reproductive success) for wildlife at farms are similarly rare, but two

recent examples were returned by our search, both documenting probable ecological traps:

Kloskowski (2012) reported that fledging rates of grebes nesting on fish ponds stocked with +1

carp were only 37 % of those nesting on unstocked ponds, while Broyer et al. (2017) found

28

that high food availability was outweighed by high predation rates for ducks nesting on

stocked ponds (Table 1).

Conflict with aquaculture operations

Birds were usually predators of stock. 45/53 studies that documented interactions with birds

considered predation on stock to be the major habitat use, whether in cages, ponds, shellfish

beds or longlines. The most common avian predators were cormorants and herons. 24/38

studies of interactions with wild mammals considered predation to be the major habitat use,

in most cases by otters in ponds or sea cages (12 studies) or pinnipeds in sea cages (10

studies). Five studies reported herbivorous fishes inhabiting algae farms, but only one

presented clear evidence of fish consuming algal crops (Anyango et al. 2017). One study

reported predation of farmed mussels by wild fish (Segvic-Bubic et al. 2011), while three

reported snakes taking stock from hatchery ponds (Plummer & Goy 1984; Gregory & Nelson

1991; Nelson & Gregory 2000).

Of the 77 studies that reported predation on stock or damage to infrastructure, only 11

quantified stock losses as a proportion of potential production, with a mean loss of 15 %

(range 0-50 %). The lower end of that range was due to mammals taking only dead or

moribund fish from hatcheries (Pitt & Conover 1996), while the upper was due to predation by

cormorants in fish ponds (Barlow & Bock 1984). Other studies quantified consumption of stock

by individual predators without placing it in the context of potential production (e.g. Glahn et

al. 1999). In addition to predating stock, pinnipeds were reported to damage nets and cause

fish escapes (e.g. Güçlüsoy & Savas 2003; Sepúlveda & Oliva 2005).

DISCUSSION

Responses to aquaculture by wildlife vary greatly across taxonomic groups and culturing

systems, but our systematic review and meta-analysis reveals several key and well-supported

trends within taxonomic groups and culturing systems and identifies clear knowledge gaps to

inform future research.

29

Are wildlife attracted to aquaculture?

Fish

Multiple lines of evidence suggest that many fish species prefer aquaculture sites over natural

habitats, and on average, farms are associated with a much higher density and diversity of wild

fish. The few available tracking studies indicate that farm-associated wild fish tend to be either

residents or regular visitors (Otterå & Skilbrei 2014; Arechavala-Lopez et al. 2015a; Loiseau et

al. 2016; Tsuyuki & Umino 2017), with spilled feed and waste likely to be the major attractive

cues driving wild fish aggregations (Tuya et al. 2006; Ballester-Moltó et al. 2015; Bacher et al.

2015). Effects on fish abundance and diversity are also likely to depend on the functional group

being assessed, with most surveys of fish populations at farms and reference sites targeting

mobile generalist carnivores (either by design or through choice of sampling method).

Birds

Studies of bird abundance revealed highly variable responses to farms, but our meta-analysis

indicates that aquaculture sites are associated with higher bird species richness overall.

Numerous studies documented large bird populations at farms without comparing them to

natural waterways, making it difficult to draw conclusions about the influence of farms on the

spatial distribution of wildlife. Furthermore, little work has been done to assess responses at

the individual level (i.e. migration or site fidelity) that can assist in inferring habitat preferences

(Robertson & Hutto 2006), but it is likely that many bird species (especially herons, cormorants

and gulls) find the availability of prey at fish and crustacean farms highly attractive (Barlow &

Bock 1984; Stickley et al. 1992, 1995; Carss 1993; Glahn et al. 1999; Harrison 2009). Shellfish

farms also increase local abundance of generalist or molluscivorous bird functional groups

(Roycroft et al. 2004; Kirk et al. 2007), but others, such as invertivorous wading birds, may be

displaced by shellfish farm infrastructure or associated ecological changes (Kelly et al. 1996;

Godet et al. 2009; Broyer & Calenge 2010).

Mammals

30

Marine mammals (pinnipeds and dolphins) also showed highly variable responses to the

presence of aquaculture, ranging from resident nuisance animals (Pemberton & Shaughnessy

1993; Hume et al. 2002; Güçlüsoy & Savas 2003; Sepúlveda and Oliva 2005), to periodic visitors

(Diaz López 2012, 2017; Díaz López and Methion 2017), to active avoidance of farms

(Markowitz et al. 2004; Watson-Capps & Mann 2005; Pearson 2009; Becker et al. 2011). Otters

were common at freshwater fish ponds (Kloskowski 2005; Kortan et al. 2007) and estuarine

sea cages in Europe (Freitas et al. 2007; Sales-Luis et al. 2009), but our search did not reveal

any data on abundance or attraction to farms relative to natural waterways.

How does aquaculture affect fitness of wildlife?

Fish

Our meta-analysis indicated that farm-associated fish tend to be larger and heavier, a finding

that is consistent with either aggregation of adult fish or higher growth rates due to a trophic

subsidy. This larger average size, together with greater abundance overall, results in a very

high local biomass of farm-associated wild fish. Despite this, farm-associated fish had similar or

higher body condition metrics and rates of stomach fullness than fish from reference sites

(Fernandez-Jover et al. 2007a; Dempster et al. 2011), indicating that farm environments may

have a higher carrying capacity for wild fish than reference sites. However, any potential

positive effects—such as higher reproductive potential—may be opposed by orders of

magnitude higher infection loads near farms (especially sea lice on salmonids: Krkošek 2017)

and possible impacts of a dietary shift from marine-derived to terrestrially-derived fatty acids

in commercial aquaculture feed (Lavens et al. 1999; Mazorra et al. 2003; Salze et al. 2005;

Bogevik et al. 2012; Arechavala-Lopez et al. 2015b). Little is known about how the plurality of

environmental changes at farms combine to influence survival and reproduction in wild fish.

Mortality rates are difficult to measure directly, but Kilambi et al. (1978) compared age

structure and recapture rates in a lake before and after the commencement of cage farming

and inferred that survivorship had increased with farming.

31

In this study, we only assessed direct interactions between aquaculture activities and wildlife,

but indirect interactions also occur, and are likely to have a considerable bearing on outcomes

for fish populations in farming areas. Dietary shifts may occur indirectly via benthic nutrient

loading and subsequent ecological changes across multiple trophic levels (Brown et al. 1987;

Wu 1995; Yucel-Gier et al. 2007; White et al. 2017), and potential deleterious effects of direct

or indirect dietary shifts or other changes may be most apparent in eggs or offspring of farm-

associated fish (Salze et al. 2005; Barrett et al. 2018). Aggregations of large predators around

sea cages may also reduce survivorship of fishes that inhabit the same area (Güçlüsoy & Savas

2003). Fish that escape from farms can reduce fitness in native populations through disease

transmission (Arechavala-López et al. 2013; Glover et al. 2013), genetic mixing (Glover et al.

2017), and interference with spawning or competition with offspring and adults (Jensen et al.

2010; Sundt-Hansen et al. 2015).

Birds

In birds, the effects of farm proximity on fitness are even less clear; only in a few cases were

we able to extract usable data on direct or indirect fitness metrics. Numerous studies reported

birds taking stock from ponds and cages, but none in our database compared feeding rates to

those on natural waterways. Nonetheless, we expect food availability to be high provided that

birds are able to access suitable food items (e.g. feed, stock or wild prey co-occurring at

farms). However, predatory birds also experience high mortality from culling and anti-predator

net entanglements where such methods are employed, potentially causing fish farms to act as

ecological traps for birds if mortality rates outweigh any benefits of higher food availability

(Carss 1994; Belant et al. 2000; Blackwell et al. 2000; Bechard & Márquez-Reyes 2003; Quick et

al. 2004). Negative effects will be exacerbated if food availability is lower than advertised, for

example if piscivores are attracted to fish ponds but cannot access fish due to anti-predator

nets, or if stocking regimes lead to cohorts of fish that are too large to be consumed. This latter

scenario was observed by Kloskowski (2012), who reported that European carp farms were

acting as ecological traps for red-necked grebes, as farmed fish were too large for fledglings to

consume leading to starvation. Predation risk for clutches may also be elevated at farms:

Broyer et al. (2017) observed high densities of breeding pairs and high food availability, but

also high offspring mortality – a probable ecological trap. Conversely, tuna ranching in

Australia was associated with a population boom for silver gulls and appears to be a clear case

32

of fish farms functioning as a strong population source for wildlife – reproductive success for

the gulls was dramatically increased by the trophic subsidy obtained by exploiting farm feed

(Harrison 2009). Similarly, long term trends in wading bird populations closely tracked the

scale of crayfish aquaculture in the southern USA (Fleury & Sherry 1995).

Mammals

Effects of aquaculture on mortality and reproduction of aquatic mammals are little known, but

as with piscivorous birds, are probably determined by the trade-off between high food

availability and high risk from culling and entanglements. Cetaceans may benefit from easy

food when they visit farms (Diaz López 2017) and culling and entanglements are relatively rare

(Diaz López & Bernal-Shirai 2007; Callier et al. 2017). As a result, attraction to farms may be an

adaptive trait that results in increased fitness on balance, although we lack direct evidence for

this. In contrast to cetaceans, pinnipeds experience heavy mortality from culling (Güçlüsoy &

Savas 2003; Quick et al. 2004; Callier et al. 2017) and are more vulnerable to accidental

entanglement (Callier et al. 2017). High mortality rates are likely to outweigh any increase in

food availability for a long-lived, slow breeding animal, such that seals that are attracted to

farms may be vulnerable to ecological traps driven by culling at farms.

Conflict and potential mutualism between aquaculture and wildlife

Our meta-analysis revealed that the nature of interactions between wild fauna and

aquaculture were highly dependent on the taxon. Wild fish generally do not interact directly

with stock unless small enough to enter sea cages through the mesh (although in rare cases

wild fish may damage nets: Moe et al. 2007, Sanchez-Jerez et al. 2008). Of more concern is the

role that wild populations play as reservoirs for pathogens and parasites, facilitating

reinfection of farms (Uglem et al. 2014b). This is an inevitable risk of farming in open systems,

but research is underway to lower infection rates by minimising spatiotemporal overlap

between stock and zones of high infection risk (Samsing et al. 2016; Wright et al. 2017).

Together with post-infection treatments, such measures also minimise the role that farms play

as amplifiers of pathogen and parasite populations.

33

Table 2.1. Mean effects of aquaculture sites on wildlife populations. F:W = mean at farms /

mean at reference sites. RR = ln(F:W). Positive RR indicates metric is higher at aquaculture

sites. t-stat and p refer to one sample t-test comparing RR data to null expectation of RR = 0.

Taxa are omitted where no comparative data are available. Asterisk indicates variables for

which higher RR corresponds to poorer outcomes.

N F:W RR ± SE t-stat p

Abundance

Fish 44 72 1.65 ± 0.29 5.7 <0.0001

Birds 13 1.8 0.13 ± 0.31 0.40 0.70

Mammals 7 1.1 -0.68 ± 0.67 -1.0 0.35

Amphibians 1 0.31 -1.17 – –

All taxa 65 49 1.05 ± 0.24 4.3 <0.0001

Species richness

Fish 28 2.0 0.43 ± 0.11 3.9 0.0005

Birds 7 1.1 0.13 ± 0.04 3.3 0.02

Mammals 1 0.50 -0.69 – –

Amphibians 1 0.32 -1.15 – –

All taxa 37 1.7 0.30 ± 0.10 3.1 0.004

Size (length)

Fish 18 1.2 0.15 ± 0.03 4.6 0.0002

Size (weight)

Fish 12 1.7 0.40 ± 0.13 3.0 0.01

Condition metrics

Fish 14 1.3 0.17 ± 0.09 1.9 0.08

Birds 1 0.7 -0.31 – –

All taxa 15 1.2 0.14 ± 0.09 1.5 0.15

Stomach fullness

Fish 10 1.4 0.04 ± 0.30 0.13 0.90

Reptiles 1 1.9 0.66 – –

All taxa 11 1.5 0.10 ± 0.28 0.35 0.73

Infection level+

Fish 11 16 2.09 ± 0.38 5.5 0.0003

Fertility

Birds 2 0.60 -0.60 ± 0.40 -1.5 0.37

Mortality+

Fish 1 0.82 -0.20 – –

Bird 1 1.4 0.33 – –

34

Most studies returned by our search concluded that predation or damage by birds and

mammals is an ongoing problem for managers, but stock losses were rarely quantified (but see

some recent examples: Sun et al. 2004; Sepúlveda & Oliva 2005; Morrison & Vogel 2009; Dorr

et al. 2012) Where suitably habituated, pinnipeds have a propensity to become ‘nuisance

animals’, damaging nets (leading to fish escapes) and consuming or stressing stock (Kemper et

al. 2003; Quick et al. 2004; Sepúlveda & Oliva 2005). Such problems are difficult to solve.

Culling is undesirable as it increases environmental impacts and negatively affects public

perceptions of aquaculture. Relocation is expensive and often ineffective (Hume et al. 2002)

and scaring devices have a limited effective lifespan before animals are desensitised. Exclusion

using steel mesh appears to be the only viable option in some cases (Pemberton &

Shaughnessy 1993).

While there tends to be a focus on negative interactions between farms and wild fauna, wild

fish can provide ecosystem services to aquaculture operations by increasing animal welfare

and reducing local environmental impacts of farming. Invertivorous fish that are small enough

to gain access to sea cages (such as wrasse and lumpfish in Norwegian salmon farms) can act

as cleaner fish and significantly reduce parasite loads on stock. Cleaner fish are now being

deployed in large numbers for this purpose (Imsland et al. 2014; Skiftesvik et al. 2014). Wild

fish and invertebrates ameliorate and disperse benthic nutrient loads by consuming spilled

feed, faeces and dead stock (Vita et al. 2004; Felsing et al. 2005; Fernandez-Jover et al. 2007b).

However, resident populations of large predators at fish farms may impede this waste

amelioration service by scaring or consuming wild fish (Diaz López 2006), resulting in more

severe benthic impacts, but such predators also prey on escaped fish, potentially reducing the

risk of genetic introgression from farmed to wild fish populations (Glover et al. 2017).

For fish farming to continue to grow, farmers need to demonstrate environmental

sustainability and good animal welfare standards. Protecting wild fish aggregations to take

advantage of the ecosystem services they provide may be an important part of achieving these

goals. Continued development of non-lethal bird and pinniped exclusion methods will be a

necessary step.

Recommendations for future research on impacts of aquaculture on wildlife

Simply documenting behaviour of wildlife at farms or changes in wildlife abundance provides

little information on the effects of aquaculture on persistence of wildlife populations.

35

Aquaculture can have qualitatively distinct effects on wildlife that are superficially

indistinguishable in the absence of data on habitat selection decisions, movement or fitness.

For example, an elevated density of wildlife at a farm relative to a reference site may support

various contradictory hypotheses, including but not limited to: (a) high survivorship or fertility

causing the farm to function as a productive population source, with or without strong

attraction, and typically with density-dependent spillover to surrounding areas (Pulliam 1988),

or (b) strong attraction to the farm habitat but high mortality rates or low reproductive success

for residents. The latter scenario describes an attractive population sink (ecological trap) that

draws animals in from surrounding areas and causes deleterious population effects

disproportionate to its area (Hale et al. 2015). Our meta-analysis reveals that in most cases we

do not have sufficient data on fitness outcomes, either direct or indirect, and as a result cannot

distinguish between attractive or productive population effects, or their resultant positive or

negative effects on wild populations.

Conceptual frameworks have been developed to distinguish between these two (non-

mutually-exclusive) processes on artificial reefs and fish aggregation devices (Osenberg et al.

2002; Brickhill et al. 2005; Reubens et al. 2013), and may be applied to aquaculture sites.

Evidence for attraction without significant production of wild fauna at aquaculture sites may

include: (i) rarity of younger cohorts relative to older cohorts, (ii) population declines at

adjacent reference sites corresponding to increases at farms, (iii) high mortality or

reproductive failure rates at farms, or (iv) tracking, microchemistry, tissue fatty acid or stable

isotope analysis indicating recent immigration to farms. Conversely, evidence for high

individual fitness leading to productive wild populations at farms may include, depending on

the taxa: (i) successful breeding pairs residing at farms, (ii) high densities of larvae or juveniles,

(iii) increases in abundance at farms followed by increases at adjacent reference sites

consistent with density-dependent spillover, (iv) tracking, microchemistry, tissue fatty acid and

stable isotope analysis indicating that most individuals are not recent immigrants.

Importantly, the above criteria for separating attraction and production are most relevant

when compared to reference habitats (i.e. is residing at a farm a good decision for an

individual, or likely to lead to an ecological trap?). Only 91/204 studies included in our

database allowed us to infer changes in at least one variable in farm-associated wildlife by

making comparison to reference sites or timepoints. In many cases, changes in distribution or

health of wildlife were not central to the study, but in other cases, there was a lost opportunity

to understand more about these interactions. Where relevant, we recommend that studies of

abundance or fitness of wild fauna at farms should benchmark their findings against reference

36

sites or timepoints (Underwood 1994; Osenberg et al. 2002; Brickhill et al. 2005). Reference

sites should be appropriate for the hypotheses being tested. For example, our meta-analysis

revealed that inferred increases in population densities at sea cage fish farms vary by orders of

magnitude depending on whether the reference habitat is a nearby natural reef or featureless

open water. Accordingly, researchers should be clear in their reasons for selecting a given

reference habitat.

Most importantly, we have highlighted the paucity of data on mortality rates and reproductive

success in farm-associated fauna. Such data are central to our understanding of the

environmental impacts of aquaculture but can be difficult to obtain. Population-level metrics

can be effective in closed or semi-closed systems (Kilambi et al. 1978), and researchers have

long been capable of tracking mortality and breeding success in birds, including at aquaculture

sites (Kloskowski 2012; Broyer et al. 2017). Open systems with highly mobile taxa (such as wild

fish associated with sea cage aquaculture) present a greater challenge, but in coastal

environments, acoustic tags in conjunction with external tags can provide excellent data on

spatiotemporal movement and mortality rates in areas with differing levels of farm activity

(Olsen & Moland 2011; Olsen et al. 2012; Fernández-Chacón et al. 2015).

It is now well established that wild fish are typically more abundant at sea cage fish farms than

reference sites, and that such fish are likely to consume farm waste, experience nutritional

shifts and depending on the system, be exposed to elevated parasite loads. The challenge now

is to develop an equivalent state of knowledge for other wild taxa and aquaculture systems,

and to obtain more direct measures of the effects of farm-association on wildlife populations.

REFERENCES

Abaad M, Tuset VM, Montero D, Lombarte A, Otero-Ferrer JL and Haroun R (2016) Phenotypic

plasticity in wild marine fishes associated with fish-cage aquaculture. Hydrobiologia 765:

343–358.

Adámek Z, Kortan D, Lepič P and Andreji J (2003) Impacts of otter (Lutra lutra L) predation on

fishponds: A study of fish remains at ponds in the Czech Republic. Aquaculture

International 11: 389–396.

Akyol O and Ertosluk O (2010) Fishing near sea-cage farms along the coast of the Turkish

Aegean Sea. Journal of Applied Ichthyology 26: 11–15.

Anyango JO, Mlewa CM and Mwaluma J (2017) Abundance diversity and trophic status of wild

37

fish around seaweed farms in Kibuyuni, South Coast Kenya. International Journal of

Fisheries and Aquatic Studies 5: 440–446.

Arechavala-Lopez P, Izquierdo-Gomez D, Uglem I and Sanchez-Jerez P (2015a) Aggregations of

bluefish Pomatomus saltatrix (L) at Mediterranean coastal fish farms: seasonal presence

daily patterns and influence of farming activity. Environmental Biology of Fishes 98: 499–

510.

Arechavala-Lopez P, Sæther B-S, Marhuenda-Egea F, Sanchez-Jerez P and Uglem I (2015b)

Assessing the influence of salmon farming through total lipids fatty acids and trace

elements in the liver and muscle of wild saithe Pollachius virens. Marine and Coastal

Fisheries 7: 59–67.

Arechavala-Lopez P, Sanchez-Jerez P, Bayle-Sempere JT, Fernandez-Jover D, Martinez-Rubio L,

Lopez-Jimenez JA and Martinez-Lopez FJ (2011) Direct interaction between wild fish

aggregations at fish farms and fisheries activity at fishing grounds: a case study with

Boops boops. Aquaculture Research 42: 996–1010.

Arechavala-López P, Sánchez-Jerez P, Bayle-Sempere JT, Uglem I and Mladineo I (2013) Reared

fish farmed escapees and wild fish stocks—a triangle of pathogen transmission of

concern to Mediterranean aquaculture management. Aquaculture Environment

Interactions 3: 153–161.

Bacher K and Gordoa A (2016) Does marine fish farming affect local small-scale fishery

catches? A case study in the NW Mediterranean Sea. Aquaculture Research 47: 2444–

2454.

Bacher K, Gordoa A and Sagué O (2015) Feeding activity strongly affects the variability of wild

fish aggregations within fish farms: a sea bream farm as a case study. Aquaculture

Research 46: 552–564.

Ballester-Moltó M, Sanchez-Jerez P, Garcia-Garcia B and Aguado-Giménez F (2015) Husbandry

and environmental conditions explain temporal variability of wild fish assemblages

aggregated around a Mediterranean fish farm. Aquaculture Environment Interactions 7:

193–203.

Bancroft BA, Baker NJ and Blaustein AR (2007) Effects of UVB radiation on marine and

freshwater organisms: a synthesis through meta-analysis. Ecology Letters 10: 332–345.

Barlow CG and Bock K (1984) Predation of fish in farm dams by cormorants Phalacrocorax spp.

Australian Wildlife Research 11: 559–566

Barrett LT, Swearer SE, Harboe T, Karlsen Ø, Meier S and Dempster T (2018) Limited evidence

for differential reproductive fitness of wild Atlantic cod in areas of high and low salmon

38

farming density. Aquaculture Environment Interactions doi: 10.3354/aei00275

Barton K (2016) MuMIn: multi-model inference R package version 1156. URL: https://cran.r-

project.org/package=MuMIn

Bechard MJ and Márquez-Reyes C (2003) Mortality of wintering ospreys and other birds at

aquaculture facilities in Colombia. Journal of Raptor Research 37: 292–298.

Becker BH, Press DT and Allen SG (2011) Evidence for long-term spatial displacement of

breeding and pupping harbour seals by shellfish aquaculture over three decades. Aquatic

Conservation: Marine and Freshwater Ecosystems 21: 247–260.

Belant JL, Tyson LA and Mastrangelo PA (2000) Effects of lethal control at aquaculture facilities

on populations of piscivorous birds. Wildlife Society Bulletin 28: 379–384.

Berg V, Kraugerud M, Nourizadeh-Lillabadi R et al. (2016) Endocrine effects of real-life

mixtures of persistent organic pollutants (POP) in experimental models and wild fish.

Journal of Toxicology and Environmental Health Part A 79: 538–548.

Björklund H, Bondestam J and Bylund G (1990) Residues of oxytetracycline in wild fish and

sediments from fish farms. Aquaculture 86: 359–367.

Blackwell BF, Dolbeer RA and Tyson LA (2000) Lethal control of piscivorous birds at

aquaculture facilities in the northeast United States: effects on populations. North

American Journal of Aquaculture 62: 300–307.

Bogdanović T, Šimat V, Frka-Roić A and Marković K (2012) Development and application of

quality index method scheme in a shelf-life study of wild and fish farm affected bogue

(Boops boops L). Journal of Food Science 77: S99–S106.

Bogevik AS, Natário S, Karlsen Ø, Thorsen A, Hamre K, Rosenlund G and Norberg B (2012) The

effect of dietary lipid content and stress on egg quality in farmed Atlantic cod Gadus

morhua. Journal of Fish Biology 81: 1391–405.

Boyd CE and Massaut L (1999) Risks associated with the use of chemicals in pond aquaculture.

Aquacultural Engineering 20: 113–132.

Brickhill MJ, Lee SY and Connolly RM (2005) Fishes associated with artificial reefs: attributing

changes to attraction or production using novel approaches. Journal of Fish Biology 67:

53–71.

Brown JR, Gowen RJ and McLusky DS (1987) The effect of salmon farming on the benthos of a

Scottish sea loch. Journal of Experimental Marine Biology and Ecology 109: 39–51.

Broyer J and Calenge C (2010) Influence of fish-farming management on duck breeding in

French fish pond systems. Hydrobiologia 637: 173–185.

Broyer J, Chavas G and Chazal R (2017) The effects of cessation of fish farming on duck

39

breeding in French fishpond systems. Hydrobiologia 788: 47–53.

Burnham KP and Anderson DR (2002) Model selection and multimodel inference: a practical

information-theoretic approach, 2nd edn. Springer, New York, NY.

Burridge L, Weis JS, Cabello F, Pizarro J and Bostick K (2010) Chemical use in salmon

aquaculture: a review of current practices and possible environmental effects.

Aquaculture 306: 7–23.

Bustnes JO, Lie E, Herzke D, Dempster T, Bjørn PA, Nygård T and Uglem I (2010) Salmon farms

as a source of organohalogenated contaminants in wild fish. Environmental Science &

Technology 44: 8736–43.

Bustnes JO, Nygard T, Dempster T, Ciesielski T, Jenssen BM, Bjorn PA and Uglem I (2011) Do

salmon farms increase the concentrations of mercury and other elements in wild fish?

Journal of Environmental Monitoring 13: 1687–1694.

Callier MD, Byron CJ, Bengtson DA et al. (2017) Attraction and repulsion of mobile wild

organisms to finfish and shellfish aquaculture: a review. Reviews in Aquaculture doi:

10.1111/raq.12208

Carss DN (1993) Grey heron Ardea cinerea L predation at cage fish farms in Argyll western

Scotland. Aquaculture and Fisheries Management 24: 29–45.

Carss DN (1994) Killing of piscivorous birds at Scottish fin fish farms, 1984-87. Biological

Conservation 68: 181–188.

Colorni A, Diamant A, Eldar A, Kvitt H and Zlotkin A (2002) Streptococcus iniae infections in Red

Sea cage-cultured and wild fishes. Diseases of Aquatic Organisms 49: 165–170.

Costello MJ (2009) How sea lice from salmon farms may cause wild salmonid declines in

Europe and North America and be a threat to fishes elsewhere. Proceedings of the Royal

Society B 276: 3385–3394.

Crump KL and Trudeau VL (2009) Mercury-induced reproductive impairment in fish.

Environmental Toxicology and Chemistry 28: 895.

Dealteris JT, Kilpatrick BD and Rheault RB (2004) A comparative evaluation of the habitat value

of shellfish aquaculture gear, submerged aquatic vegetation and a non-vegetated seabed.

Journal of Shellfish Research 23: 867–874.

DeBruyn AMH, Trudel M, Eyding N et al. (2006) Ecosystemic effects of salmon farming increase

mercury contamination in wild fish. Environmental Science & Technology 40: 3489–3493.

Demétrio JA, Gomes LC, Latini JD and Agostinho AA (2012) Influence of net cage farming on

the diet of associated wild fish in a Neotropical reservoir. Aquaculture 330: 172–178.

Dempster T, Sanchez-Jerez P, Bayle-Sempere JT and Giménez-Casalduero F (2002) Attraction of

40

wild fish to sea-cage fish farms in the south-western Mediterranean Sea: spatial and

short-term temporal variability. Marine Ecology Progress Series 242: 237–252.

Dempster T, Sanchez-Jerez P, Bayle-Sempere JT and Kingsford M (2004) Extensive aggregations

of wild fish at coastal sea-cage fish farms. Hydrobiologia 525: 245–248.

Dempster T, Sanchez-Jerez P, Fernandez-Jover D, Bayle-Sempere JT, Nilsen R, Bjørn PA and

Uglem I (2011) Proxy measures of fitness suggest coastal fish farms can act as population

sources and not ecological traps for wild gadoid fish. PLOS ONE 6: e15646.

Dempster T, Uglem I, Sanchez-Jerez P, Fernandez-Jover D, Bayle-Sempere J and Nilsen R (2009)

Coastal salmon farms attract large and persistent aggregations of wild fish: an ecosystem

effect. Marine Ecology Progress Series 385: 1–14.

Diamant A, Banet A, Ucko M, Colorni A, Knibb W and Kvitt H (2000) Mycobacteriosis in wild

rabbitfish Siganus rivulatus associated with cage farming in the Gulf of Eilat, Red Sea.

Diseases of Aquatic Organisms 39: 211–219.

Diaz López B (2006) Bottlenose dolphin (Tursiops truncatus) predation on a marine fin fish

farm: some underwater observations. Aquatic Mammals 32: 305–310.

Diaz López B (2012) Bottlenose dolphins and aquaculture: interaction and site fidelity on the

north-eastern coast of Sardinia (Italy). Marine Biology 159: 2161–2172.

Diaz López B (2017) Temporal variability in predator presence around a fin fish farm in the

Northwestern Mediterranean Sea. Marine Ecology 38: e12378.

Diaz López B and Bernal-Shirai JA (2007) Bottlenose dolphin (Tursiops truncatus) presence and

incidental capture in a marine fish farm on the north-eastern coast of Sardinia (Italy).

Journal of the Marine Biological Association of the UK 87: 113–117.

Díaz López B and Methion S (2017) The impact of shellfish farming on common bottlenose

dolphins’ use of habitat. Marine Biology 164: 83.

Dorr BS, Burger LW, Barras SC and Godwin KC (2012) Economic impact of double-crested

cormorant Phalacrocorax auritus depredation on channel catfish Ictalurus punctatus

aquaculture in Mississippi, USA. Journal of the World Aquaculture Society 43: 502–513.

Dumont CP, Gaymer CF and Thiel M (2011) Predation contributes to invasion resistance of

benthic communities against the non-indigenous tunicate Ciona intestinalis. Biological

Invasions 13: 2023–2034

Ervik A, Thorsen B, Eriksen V, Lunestad BT and Samuelsen OB (1994) Impact of administering

antibacterial agents on wild fish and blue mussels Mytilus edulis in the vicinity of fish

farms. Diseases of Aquatic Organisms 18: 45–51.

European Union (2010) Commission Regulation (EU) No 37/2010 of 22 December 2009 on

41

pharmacologically active substances and their classification regarding maximum residue

limits in foodstuffs of animal origin Brussels European Union. URL:

https://eceuropaeu/food/plant/pesticides/max_residue_levels_en

FAO Fisheries and Aquaculture (2015) FAO Yearbook - Fishery and Aquaculture Statistics

(2014).

FAO Fisheries and Aquaculture (2017) Fishery and Aquaculture Statistics - Global aquaculture

production 1950-2015 (FishstatJ). URL:

http://www.fao.org/fishery/statistics/software/fishstatj/en

FAO Fisheries and Aquaculture (2018) Cultured aquatic species fact sheets. URL:

http://www.fao.org/fishery/culturedspecies/search/en

Felsing M, Glencross B and Telfer T (2005) Preliminary study on the effects of exclusion of wild

fauna from aquaculture cages in a shallow marine environment. Aquaculture 243: 159–

174.

Fernández-Chacón A, Moland E, Espeland SH and Olsen EM (2015) Demographic effects of full

vs partial protection from harvesting: inference from an empirical before-after control-

impact study on Atlantic cod. Journal of Applied Ecology 52: 1206–1215.

Fernandez-Jover D, Jimenez JA, Sanchez-Jerez P, Bayle-Sempere JT, Giménez-Casalduero F,

Lopez FJM and Dempster T (2007a) Changes in body condition and fatty acid composition

of wild Mediterranean horse mackerel (Trachurus mediterraneus Steindachner 1868)

associated to sea cage fish farms. Marine Environmental Research 63: 1–18.

Fernandez-Jover D, Martinez-Rubio L, Sanchez-Jerez P et al. (2011) Waste feed from coastal

fish farms: a trophic subsidy with compositional side-effects for wild gadoids. Estuarine

Coastal and Shelf Science 91: 559–568.

Fernandez-Jover D and Sanchez-Jerez P (2015) Comparison of diet and otolith growth of

juvenile wild fish communities at fish farms and natural habitats. ICES Journal of Marine

Science 72: 916–929.

Fernandez-Jover D, Sanchez-Jerez P, Bayle-Sempere J, Carratala A and Leon VM (2007b)

Addition of dissolved nitrogen and dissolved organic carbon from wild fish faeces and

food around Mediterranean fish farms: Implications for waste-dispersal models. Journal

of Experimental Marine Biology and Ecology 340: 160–168.

Fleury BE and Sherry TW (1995) Long-term population trends of colonial wading birds in the

southern United States: the impact of crayfish aquaculture on Louisiana populations. Auk

112: 613–632.

Freitas D, Gomes J, Luis TS et al. (2007) Otters and fish farms in the Sado estuary: ecological

42

and socio-economic basis of a conflict. Hydrobiologia 587: 51–62.

Gamfeldt L, Lefcheck JS, Byrnes JEK, Cardinale BJ, Duffy JE and Griffin JN (2015) Marine

biodiversity and ecosystem functioning: what’s known and what’s next? Oikos 124: 252–

265.

Glahn JF, Tomsa T and Preusser KJ (1999) Impact of great blue heron predation at trout-rearing

facilities in the northeastern United States. North American Journal of Aquaculture 61:

349–354.

Glover KA, Solberg MF, McGinnity P et al. (2017) Half a century of genetic interaction between

farmed and wild Atlantic salmon: status of knowledge and unanswered questions. Fish

and Fisheries doi: 10.1111/faf.12214

Glover KA, Sørvik AGE, Karlsbakk E, Zhang Z and Skaala Ø (2013) Molecular genetic analysis of

stomach contents reveals wild Atlantic cod feeding on piscine reovirus (PRV) infected

Atlantic salmon originating from a commercial fish farm. PLOS ONE 8: e60924.

Godet L, Toupoint N, Fournier J, Le Mao P, Retiere C and Olivier F (2009) Clam farmers and

oystercatchers: effects of the degradation of Lanice conchilega beds by shellfish farming

on the spatial distribution of shorebirds. Marine Pollution Bulletin 58: 589–595.

Gonzalez-Silvera D, Guardiola FA, Cordero H, Cuesta A, Esteban MA, Martínez-López FJ and

López-Jiménez JA (2017) The short-term effects of farmed fish food consumed by wild

fish congregating outside the farms. Marine Pollution Bulletin 114: 689–698.

Gregory PT and Nelson KJ (1991) Predation on fish and intersite variation in the diet of

common garter snakes Thamnophis sirtalis on Vancouver Island. Canadian Journal of

Zoology 69: 988–994.

Güçlüsoy H and Savas Y (2003) Interaction between monk seals Monachus monachus

(Hermann 1779) and marine fish farms in the Turkish Aegean and management of the

problem. Aquaculture Research 34: 777–783.

Hale R and Swearer SE (2016) Ecological traps: current evidence and future directions.

Proceedings of the Royal Society B 283: 494–499.

Hale R, Treml EA and Swearer SE (2015) Evaluating the metapopulation consequences of

ecological traps. Proceedings of the Royal Society B 282: 20142930

Harrison SJ (2009) Interactions between Silver Gulls (Larus novaehollandiae) and Southern

Bluefin Tuna (Thunnus maccoyii) aquaculture in the Port Lincoln area. PhD thesis. Flinders

University, Australia.

Hayward CJ, Svane I, Lachimpadi SK, Itoh N, Bott NJ and Nowak BF (2011) Sea lice infections of

wild fishes near ranched southern bluefin tuna (Thunnus maccoyii) in South Australia.

43

Aquaculture 320: 178–182.

Hedges LV, Gurevitch J and Curtis PS (1999) The meta-analysis of response ratios in

experimental ecology. Ecology 80: 1150–1156.

Hehre EJ and Meeuwig JJ (2016) A global analysis of the relationship between farmed seaweed

production and herbivorous fish catch. PLOS ONE 11: e0148250.

Herbeck LS, Unger D, Wu Y and Jennerjahn TC (2013) Effluent nutrient and organic matter

export from shrimp and fish ponds causing eutrophication in coastal and back-reef

waters of NE Hainan, tropical China. Continental Shelf Research 57: 92–104.

Hume F, Pemberton D, Gales R, Brothers N and Greenwood M (2002) Trapping and relocating

seals from salmonid fish farms in Tasmania 1990-2000: was it a success? Papers and

Proceedings of the Royal Society of Tasmania 136: 1–6.

Imsland AK, Reynolds P, Eliassen G, Hangstad TA, Foss A, Vikingstad E and Elvegård TA (2014)

The use of lumpfish (Cyclopterus lumpus L) to control sea lice (Lepeophtheirus salmonis

Krøyer) infestations in intensively farmed Atlantic salmon (Salmo salar L). Aquaculture

424–425: 18–23.

Inglis GJ and Gust N (2003) Potential indirect effects of shellfish culture on the reproductive

success of benthic predators. Journal of Applied Ecology 40: 1077–1089.

Jensen Ø, Dempster T, Thorstad EB, Uglem I and Fredheim A (2010) Escapes of fishes from

Norwegian sea-cage aquaculture: causes consequences and prevention. Aquaculture

Environment Interactions 1: 71–83.

Johnston TA, Keir M and Power M (2010) Response of native and naturalized fish to salmonid

cage culture farms in northern Lake Huron, Canada. Transactions of the American

Fisheries Society 139: 660–670.

Kelly JP, Evens JG, Stallcup RW and Wimpfheimer D (1996) Effects of aquaculture on habitat

use by wintering shorebirds in Tomales Bay, California. California Fish and Game 82: 160–

174.

Kemper C, Pemberton D, Cawthorn M et al. (2003) Aquaculture and marine mammals: co-

existence or conflict? In: Gales N, Hindell M and Kirkwood R (ed.) Marine Mammals:

Fisheries Tourism and Management Issues, pp. 208–225. CSIRO Publishing, Collingwood.

Kemper CM and Gibbs SE (2001) Dolphin interactions with tuna feedlots at Port Lincoln, South

Australia and recommendations for minimising entanglements. Journal of Cetacean

Research and Management 3: 283–292.

Kilambi RV, Adams JC and Wickizer WA (1978) Effects of cage culture on growth abundance

and survival of resident largemouth bass (Micropterus salmoides). Journal of the Fisheries

44

Research Board of Canada 35: 157–160.

Kirk M, Esler D and Boyd WS (2007) Morphology and density of mussels on natural and

aquaculture structure habitats: implications for sea duck predators. Marine Ecology

Progress Series 346: 179–187.

Kloskowski J (2010) Fish farms as amphibian habitats: factors affecting amphibian species

richness and community structure at carp ponds in Poland. Environmental Conservation

37: 187–194.

Kloskowski J (2012) Fish stocking creates an ecological trap for an avian predator via effects on

prey availability. Oikos 121: 1567–1576.

Kloskowski J (2005) Otter Lutra lutra damage at farmed fisheries in southeastern Poland II:

exploitation of common carp Cyprinus carpio. Wildlife Biology 11: 257–261.

Kloskowski J, Kaczanowska E, Krogulec J and Grela P (2017) Hematological indicators of habitat

quality: erythrocyte parameters reflect greater parental effort of red-necked grebes

under ecological trap conditions. The Condor 119: 239–250.

Kortan D, Adámek Z and Poláková S (2007) Winter predation by otter Lutra lutra on carp pond

systems in South Bohemia (Czech Republic). Folia Zoologica 56: 416.

Krkošek M (2017) Population biology of infectious diseases shared by wild and farmed fish.

Canadian Journal of Fisheries and Aquatic Sciences 74: 620–628.

Krkosek M, Lewis MA, Morton A, Frazer LN and Volpe JP (2006) Epizootics of wild fish induced

by farm fish. Proceedings of the National Academy of Sciences of the USA 103: 15506–

15510.

Krkosek M, Lewis MA and Volpe JP (2005) Transmission dynamics of parasitic sea lice from

farm to wild salmon. Proceedings of the Royal Society B 272: 689–96.

Kroeker KJ, Kordas RL, Crim RN and Singh GG (2010) Meta-analysis reveals negative yet

variable effects of ocean acidification on marine organisms. Ecology Letters 13: 1419–

1434.

Lafferty KD, Harvell CD, Conrad JM et al. (2015) Infectious diseases affect marine fisheries and

aquaculture economics. Annual Review of Marine Science 7: 471–496.

Lavens P, Lebegue E, Jaunet H, Brunel A, Dhert P and Sorgeloos P (1999) Effect of dietary

essential fatty acids and vitamins on egg quality in turbot broodstocks. Aquaculture

International 7: 225–240.

Loiseau N, Kiszka JJ, Bouveroux T, Heithaus MR, Soria M and Chabanet P (2016) Using an

unbaited stationary video system to investigate the behaviour and interactions of bull

sharks Carcharhinus leucas under an aquaculture farm. African Journal of Marine Science

45

38: 73–79.

Markowitz TM, Harlin AD, Wursig B and Mcfadden CJ (2004) Dusky dolphin foraging habitat:

overlap with aquaculture in New Zealand. Aquatic Conservation: Marine and Freshwater

Ecosystems 14: 133–149.

Mazorra C, Bruce M, Bell JG et al. (2003) Dietary lipid enhancement of broodstock

reproductive performance and egg and larval quality in Atlantic halibut (Hippoglossus

hippoglossus). Aquaculture 227: 21–33.

McKindsey CW, Archambault P, Callier MD and Olivier F (2011) Influence of suspended and off-

bottom mussel culture on the sea bottom and benthic habitats: a review. Canadian

Journal of Zoology 89: 622–646.

Mladineo I, Segvic T and Grubisic L (2009) Molecular evidence for the lack of transmission of

the monogenean Sparicotyle chrysophrii (Monogenea, Polyopisthocotylea) and isopod

Ceratothoa oestroides (Crustacea, Cymothoidae) between wild bogue (Boops boops) and

cage-reared sea bream (Sparus aurata). Aquaculture 295: 160–167.

Moe H, Dempster T, Sunde LM, Winther U and Fredheim A (2007) Technological solutions and

operational measures to prevent escapes of Atlantic cod (Gadus morhua) from sea cages.

Aquaculture Research 38: 91–99.

Morrison SS and Vogel P (2009) Aquaculture ponds a Jamaican study: The impact of birds on

fish production. African Journal of Agricultural Research 4: 1447–1454.

Nelson KJ and Gregory PT (2000) Activity patterns of garter snakes Thamnophis sirtalis in

relation to weather conditions at a fish hatchery on Vancouver Island, British Columbia.

Journal of Herpetology 34: 32.

Olsen EM, Heupel MR, Simpfendorfer CA and Moland E (2012) Harvest selection on Atlantic

cod behavioral traits: implications for spatial management. Ecology and Evolution 2:

1549–1562.

Olsen EM and Moland E (2011) Fitness landscape of Atlantic cod shaped by harvest selection

and natural selection. Evolutionary Ecology 25: 695–710.

Osenberg C, St Mary CM, Wilson JA and Lindberg WJ (2002) A quantitative framework to

evaluate the attraction-production controversy. ICES Journal of Marine Science 59: S214–

S221.

Otterå H and Skilbrei OT (2014) Possible influence of salmon farming on long-term resident

behaviour of wild saithe (Pollachius virens L). ICES Journal of Marine Science 71: 2484–

2493.

Pearson HC (2009) Influences on dusky dolphin (Lagenorhynchus obscurus) fission-fusion

46

dynamics in Admiralty Bay, New Zealand. Behavioral Ecology and Sociobiology 63: 1437–

1446.

Pemberton D and Shaughnessy PD (1993) Interaction between seals and marine fish-farms in

Tasmania and management of the problem. Aquatic Conservation: Marine and

Freshwater Ecosystems 3: 149–158.

Piroddi C, Bearzi G and Christensen V (2011) Marine open cage aquaculture in the eastern

Mediterranean Sea: a new trophic resource for bottlenose dolphins. Marine Ecology

Progress Series 440: 255–266.

Pitt WC and Conover MR (1996) Predation at Intermountain West fish hatcheries. The Journal

of Wildlife Management 60: 616–624.

Plummer M V and Goy JM (1984) Ontogenetic dietary shift of water snakes (Nerodia

rhombifera) in a fish hatchery. Copeia 1984: 550.

Powers M, Peterson C, Summerson H and Powers S (2007) Macroalgal growth on bivalve

aquaculture netting enhances nursery habitat for mobile invertebrates and juvenile

fishes. Marine Ecology Progress Series 339: 109–122.

Pulliam HR (1988) Sources sinks and population regulation. The American Naturalist 132: 652–

661.

Quick NJ, Middlemas SJ and Armstrong JD (2004) A survey of antipredator controls at marine

salmon farms in Scotland. Aquaculture 230: 169–180.

R Core Team (2017) R: A language and environment for statistical computing. URL:

http://www.r-project.org/

Reubens JT, Vandendriessche S, Zenner AN, Degraer S and Vincx M (2013) Offshore wind farms

as productive sites or ecological traps for gadoid fishes? Impact on growth condition

index and diet composition. Marine Environmental Research 90: 66-74.

Robertson BA and Hutto RL (2006) A framework for understanding ecological traps and an

evaluation of existing evidence. Ecology 87: 1075–1085.

Roycroft D, Kelly TC and Lewis LJ (2004) Birds seals and the suspension culture of mussels in

Bantry Bay a non-seaduck area in Southwest Ireland. Estuarine Coastal and Shelf Science

61: 703–712.

Sales-Luis T, Freitas D and Santos-Reis M (2009) Key landscape factors for Eurasian otter Lutra

lutra visiting rates and fish loss in estuarine fish farms. European Journal of Wildlife

Research 55: 345–355.

Salze G, Tocher DR, Roy WJ and Robertson DA (2005) Egg quality determinants in cod (Gadus

morhua L): egg performance and lipids in eggs from farmed and wild broodstock.

47

Aquaculture Research 36: 1488–1499.

Samsing F, Johnsen I, Stien LH, Oppedal F, Albretsen J, Asplin L and Dempster T (2016)

Predicting the effectiveness of depth-based technologies to prevent salmon lice infection

using a dispersal model. Preventive Veterinary Medicine 129: 48–57.

Samuelsen OB, Lunestad BT, Husevag B, Holleland T and Ervik A (1992) Residues of oxolinic

acid in wild fauna following medication in fish farms. Diseases of Aquatic Organisms 12:

111–119.

Sanchez-Jerez P, Fernandez-Jover D, Bayle-Sempere J, Valle C, Dempster T, Tuya F and Juanes F

(2008) Interactions between bluefish Pomatomus saltatrix (L) and coastal sea-cage farms

in the Mediterranean Sea. Aquaculture 282: 61–67.

Sanchez-Jerez P, Fernandez-Jover D, Uglem I et al. (2011) Coastal fish farms as fish aggregation

devices (FADs). In: Bortone SA, Brandini FP, Fabi G and Otake S (ed.) Artificial Reefs in

Fishery Management, pp. 187–208. Taylor & Francis Group, Florida.

Segvic-Bubic T, Grubisic L, Karaman N, Ticina V, Jelavic KM and Katavic I (2011) Damages on

mussel farms potentially caused by fish predation – self service on the ropes?

Aquaculture 319: 497–504.

Sepúlveda M, Newsome SD, Pavez G, Oliva D, Costa DP and Hueckstaedt LA (2015) Using

satellite tracking and isotopic information to characterize the impact of South American

sea lions on salmonid aquaculture in southern Chile. PLOS ONE 10: e0134926–e0134926.

Sepúlveda M and Oliva D (2005) Interactions between South American sea lions Otaria

flavescens (Shaw) and salmon farms in southern Chile. Aquaculture Research 36: 1062–

1068.

Skiftesvik AB, Blom G, Agnalt A-L et al. (2014) Wrasse (Labridae) as cleaner fish in salmonid

aquaculture – the Hardangerfjord as a case study. Marine Biology Research 10: 289–300.

Skog T-E, Hylland K, Torstensen BE and Berntssen MHG (2003) Salmon farming affects the fatty

acid composition and taste of wild saithe Pollachius virens L. Aquaculture Research 34:

999–1007.

Stickley AR, Glahn JF, King JO and King DT (1995) Impact of great blue heron depredations on

channel catfish farms. Journal of the World Aquaculture Society 26: 194–199.

Stickley AR, Warrick GL and Glahn JF (1992) Impact of double‐crested cormorant depredations

on channel catfish farms. Journal of the World Aquaculture Society 23: 192–198.

Sun YH, Wu HJ and Wang Y (2004) Tawny fish-owl predation at fish farms in Taiwan. Journal of

Raptor Research 38: 326–333.

Sundt-Hansen L, Huisman J, Skoglund H and Hindar K (2015) Farmed Atlantic salmon Salmo

48

salar L parr may reduce early survival of wild fish. Journal of Fish Biology 86: 1699–1712.

Tallman JC and Forrester GE (2007) Oyster grow-out cages function as artificial reefs for

temperate fishes. Transactions of the American Fisheries Society 136: 790–799.

Taranger GL, Karlsen Ø, Bannister RJ et al. (2015) Risk assessment of the environmental impact

of Norwegian Atlantic salmon farming. ICES Journal of Marine Science 72: 997–1021.

Tsuyuki A and Umino T (2017) Spatial movement of black sea bream Acanthopagrus schlegelii

around the oyster farming area in Hiroshima Bay, Japan. Fisheries Science 83: 235–244.

Tuya F, Sanchez-Jerez P, Dempster T, Boyra A and Haroun RJ (2006) Changes in demersal wild

fish aggregations beneath a sea-cage fish farm after the cessation of farming. Journal of

Fish Biology 69: 682–697.

Uglem I, Dempster T, Bjørn PA, Sanchez-Jerez P and Økland F (2009) High connectivity of

salmon farms revealed by aggregation residence and repeated movements of wild fish

among farms. Marine Ecology Progress Series 384: 251–260.

Uglem I, Karlsen Ø, Sanchez-Jerez P and Sæther B (2014) Impacts of wild fishes attracted to

open-cage salmonid farms in Norway. Aquaculture Environment Interactions 6: 91–103

Underwood AJ (1994) On Beyond BACI: sampling designs that might reliably detect

environmental disturbances. Ecological Applications 4: 3–15.

United Nations Development Programme (2017) International Human Development Indicators.

Available at: http://hdrundporg/en/countries [Accessed October 30 2017].

Valdemarsen T, Hansen PK, Ervik A and Bannister RJ (2015) Impact of deep-water fish farms on

benthic macrofauna communities under different hydrodynamic conditions. Marine

Pollution Bulletin 101: 776–783.

Vita R, Marin A, Madrid JA, Jimenez-Brinquis B, Cesar A and Marin-Guirao L (2004) Effects of

wild fishes on waste exportation from a Mediterranean fish farm. Marine Ecology

Progress Series 277: 253–261.

Watson-Capps JJ and Mann J (2005) The effects of aquaculture on bottlenose dolphin (Tursiops

sp) ranging in Shark Bay, Western Australia. Biological Conservation 124: 519–526.

Watts JEM, Schreier HJ, Lanska L and Hale MS (2017) The rising tide of antimicrobial resistance

in aquaculture: sources sinks and solutions Marine Drugs 15: 158.

White CA, Bannister RJ, Dworjanyn SA, Husa V, Nichols PD, Kutti T and Dempster T (2017)

Consumption of aquaculture waste affects the fatty acid metabolism of a benthic

invertebrate. Science of The Total Environment 586: 1170–1181.

Wright DW, Stien LH, Dempster T, Vågseth T, Nola V, Fosseidengen J-E and Oppedal F (2017)

“Snorkel” lice barrier technology reduced two co-occurring parasites the salmon louse

49

(Lepeophtheirus salmonis) and the amoebic gill disease causing agent (Neoparamoeba

perurans) in commercial salmon sea-cages. Preventive Veterinary Medicine 140: 97–105.

Wu RSS (1995) The environmental impact of marine fish culture: towards a sustainable future.

Marine Pollution Bulletin 31: 159–166.

Yucel-Gier G, Kucuksezgin F and Kocak F (2007) Effects of fish farming on nutrients and benthic

community structure in the Eastern Aegean (Turkey). Aquaculture Research 38: 256–267.

Zeileis A, Kleiber C and Jackman S (2008) Regression models for count data in R. Journal of

Statistical Software 27. doi: 10.18637/jss.v027.i08

Zlotkin A, Hershko H and Eldar A (1998) Possible transmission of Streptococcus iniae from wild

fish to cultured marine fish. Applied and Environmental Microbiology 64: 4065–4067.

50

CHAPTER THREE: LIMITED EVIDENCE FOR DIFFERENTIAL REPRODUCTIVE FITNESS OF WILD

ATLANTIC COD IN AREAS OF HIGH AND LOW SALMON FARMING DENSITY

ABSTRACT

Sea cage fish aquaculture attracts large aggregations of wild fish that feed on farm waste. Fish

that associate closely with farms undergo physiological changes, and captive feeding trials

indicate possible negative effects on reproductive fitness. However, little is known about the

significance of this phenomenon for reproduction in wild fish over larger spatial scales. To

assess if coastal areas with intensive aquaculture impact the fitness of wild fish, we collected

Atlantic cod (Gadus morhua) from two areas of high and low salmon farming density (HFD and

LFD respectively) in south-western Norway, a region that hosts the world’s largest coastal fish

aquaculture industry. We conducted a captive spawning trial and compared a range of

reproductive fitness metrics. Two fatty acids that occur at high levels in commercial feed, oleic

and lineoleic acid, were strongly correlated in the ovaries of HFD fish, but a comparison of lipid

profiles between HFD and LFD fish showed no differences in total lipids or essential fatty acids.

Although HFD fish were slightly larger than LFD fish and had similar body condition, the volume

of eggs produced did not differ, indicating relatively lower reproductive investment by HFD

fish. HFD eggs were 5 % smaller than LFD eggs, which did not lead to differential hatching

success but may have contributed to HFD larvae being 8 % smaller than their LFD counterparts

at 40 days post spawning. The potential for cumulative effects of smaller eggs and larvae on

fitness at later life stages warrants further investigation.

INTRODUCTION

Intensive culture of fish within sea cages leads to considerable benthic nutrient loads via

spilled feed and waste, providing a trophic subsidy that attracts large and persistent

aggregations of ‘farm-associated’ wild fish (Dempster et al. 2002, 2009, 2010). Individuals may

reside in the vicinity of farms for months to years (Uglem et al. 2009, Otterå & Skilbrei 2014),

and during this time are exposed to a variety of environmental changes including elevated

infection risk from parasites and other pathogens (Dempster et al. 2011, Johansen et al. 2011,

Arechavala-López et al. 2013, Glover et al. 2013), artificial lighting regimes that may delay

51

maturation and alter behaviour (Porter et al. 1999, Hansen et al. 2001, McConnell et al. 2010,

Otterå and Skilbrei 2014, Skilbrei and Otterå 2016, reviewed in Trippel 2010), contamination

from chemicals and metals used in aquaculture (Burridge et al. 2010), elevated predation risk

due to the abundance of large predatory fish, and where permitted, fishing mortality (Akyol &

Ertosluk 2010, Bagdonas et al. 2012). However, perhaps the most obvious change is the high

availability of waste feed that typically results in higher somatic and gonadal condition indices

for farm-associated wild fish (e.g. Dempster et al. 2011). 67-90% of fish sampled near Spanish

sea cages had consumed pellets (Fernandez-Jover et al. 2008), while pellets made up 71 and

25%, respectively, of the diet for farm-associated saithe and cod in Norway (Dempster et al.

2011).

Superficially, this trophic subsidy appears to benefit wild fish, but it also results in a dietary

shift from marine-derived highly-unsaturated omega-3 fatty acids to terrestrially-derived

omega-6 fatty acids. This in turn translates to a compositional shift in tissues (Fernandez-Jover

et al. 2007, Fernandez-Jover et al. 2011, Olsen et al. 2014, Arechavala-López et al. 2015), and

given that dietary lipids are reflected in the egg stores (e.g. Lavens et al. 1999, Czesny et al.

2000, Salze et al. 2005), may result in deficiencies in several of the fatty acids required for

reproduction and development. For example, captive feeding trials have found that essential

fatty acids 20:5 n-3 (eicosapentaenoic acid, EPA) and 22:6 n-3 (docosahexaenoic acid, DHA), as

well as 20:4 n-6 (arachidonic acid, AA), are important for fecundity, egg and sperm quality,

hatching success and larval development in fish (Reitan et al. 1994, Rainuzzo et al. 1997,

Sargent et al. 1999, Sargent et al. 1999, Rahman et al. 2014). Dietary deficiencies in these

reproductive nutrients contribute to low fertilisation and hatching rates in Atlantic cod (Gadus

morhua) broodstock relative to their wild counterparts (Salze et al. 2005) and have been linked

to changes in egg quality metrics in turbot (Scophthalmus maximus) (Lavens et al. 1999),

Atlantic halibut (Hippoglossus hippoglossus) (Mazorra et al. 2003) and cod (Bogevik et al.

2012). Furthermore, while data are lacking for fish, evidence from other taxa indicate that

additional effects may become apparent only after fertilisation; sea urchins (Echinus acutus)

reared on commercial salmon feed had higher gonad indices but lower fertilization and larval

survival rates, leading to an overall reduction in reproductive fitness (White et al. 2016, White

et al. 2017a).

The net effect of farm proximity on fitness, whether positive or negative, has a considerable

bearing on wild fish populations in farming areas. If trophic subsidies and associated conditions

provide a net fitness benefit, farm-associated fish populations will experience higher

production than those in neighbouring areas, and farms will act as a population source.

52

However, because farms are highly attractive to many fish species, any decline in fitness in

farm-associated wild fish may cause farms to function as ‘ecological traps’ (Robertson & Hutto

2006, Hale & Swearer 2016). Ecological traps are attractive but low quality habitats that can

have significant metapopulation-level impacts by drawing in individuals from higher quality

adjacent habitats, thus acting as attractive population sinks (Hale et al. 2015, Hale & Swearer

2016). To demonstrate the existence of an ecological trap, we must show that (1) individuals

prefer or show equal preference for the putative trap habitat relative to other available

habitats, and (2) that fitness outcomes in the putative trap habitat are lower than they would

have been in the other available habitats (Robertson & Hutto 2006, Patten & Kelly 2010). We

have strong empirical evidence that ecological traps affect birds and mammals in modified

terrestrial and aquatic environments (e.g. Remeš 2003, Weldon & Haddad 2005, Balme et al.

2010, Kloskowski 2012), but to date there have been very few tests of the theory in the marine

environment (Hallier and Gaertner 2008, Dempster et al. 2011, Sherley et al. 2017), and none

that have directly assessed fitness.

Norway operates the largest coastal fish aquaculture industry in the world, with Atlantic

salmon production reaching ~1.4 million tons in 2016 (Norwegian Directorate of Fisheries

2016). Fjord cod stocks, which are distinct from the more mobile offshore cod stocks

(Robichaud & Rose 2004), coexist with salmon farming in southern Norway and are at

historically low levels, with recent recruitment rates also low (Knutsen et al. 2016). Given these

concerns, the population can ill afford a potential ecological trap scenario. The >300 active

salmon farms in south-western Norway have widespread effects on fish movement and

population distribution and represent hundreds of potential ecological traps for fjord cod.

Gadid populations are strongly attracted to salmon grow-out cages (Dempster et al. 2011) and

move from farm to farm (Uglem et al. 2009, Otterå & Skilbrei 2014). This satisfies the first

component of an ecological trap assessment by showing preference for the putative trap

habitat (Robertson & Hutto 2006). However, we have no data on direct fitness measures in any

farm-associated fish, including gadids.

Farms may impact fitness in wild fish by altering either the survival or reproductive success of

individuals. Both pathways are difficult to measure directly in marine fishes, particularly highly

mobile broadcast spawning species, and as a result most previous work has relied on proxies

such as body condition (Fernandez-Jover et al. 2007, Dempster et al. 2011). Here we employ a

more direct approach by conducting a captive spawning experiment with wild-caught Atlantic

cod, to investigate whether reproductive fitness differs between areas of high and low salmon

farming intensity in south-western Norway. We assess potential reproductive effects spanning

53

initial adult body condition, ovarian fatty acid profiles and volumetric egg production, through

to egg quality metrics, hatching rates, and larval quality metrics including growth rates,

deformity rates and behaviour.

MATERIALS AND METHODS

Experimental design and fish husbandry

The density of salmon farms throughout Norway means that there are no longer any true

reference sites for the impacts of salmon farming on highly mobile wild fish in most parts of

the country. Instead, we make a comparison between two areas with differential farm density,

reflecting the typical spectrum of farm exposure for wild fish populations. We collected two

groups of mature live wild Atlantic cod (Gadus morhua) from the outer fjords of Hordaland

county in south-western Norway during February 2016. The first group (High Farm Density;

HFD) was collected by commercial fishers and technical staff from six locations along an 8 km

stretch of coastline in the Austevoll archipelago and 4 km away at Reksteren (Fig. 3.1). HFD

collection locations were all in relatively close proximity (300–1200 m) to six active salmon

farms, in an area of generally dense farming activity. The collections took place over ten days

between 1-15 February. The second group (Low Farm Density; LFD) was collected by

commercial fishers from two locations within the Bømlo archipelago, a neighbouring area with

relatively little fish farming activity (Fig. 3.1), between 5–15 February. HFD fish were collected

from sites exposed to a mean of 4.0 ha (range 1.5–5.2 ha) of sea cage surface area within 4 km

of their collection location, while LFD collection sites were exposed to mean 0.8 ha (range 0–

3.2 ha) within 4 km. These estimates include all sea cages and holding pens, including

enclosures that were unstocked at times during 2015 (for locations of active salmon grow-out

cages relative to collection sites, see Fig. 3.1). 75 % of LFD fish were collected ~9 km from the

nearest salmon cages, although it should be noted that the remaining 25 % of LFD fish were

collected within 2 km of a small salmon farm (2340 t capacity). As all fish were pooled within

groups, this is likely to reduce overall effect sizes observed in this study. We judged the farm

density within 4 km of the collection site to be a suitable metric of farm exposure based on

telemetry data describing the movements of tagged wild coastal cod in Norway and Shetland

(Neat et al. 2006, Svåsand et al. 2008, Uglem et al. 2008), but even if some proportion of fish

move larger distances, a considerable difference in the average level of farm exposure

54

between HFD and LFD groups will be maintained (Fig. 3.1). These two areas are otherwise

comparable environments in terms of gross hydrology, geology and ecology. Several logistical

factors prevented us from including additional sampling areas. Specifically, our need to collect

both HFD and LFD fish from shallow water to minimise barotrauma reduced the number of

suitable locations, as farms are typically placed in much deeper water. We were also restricted

to areas with similar latitude and water temperature to minimise temporal effects on

spawning (Kjesbu et al. 1994). In consideration of this geographic restriction, we consider our

results to be representative of outer Hardangerfjord but with relevance for similar systems

worldwide.

Figure 3.1. Map of collection sites relative to active salmon farms in south-western Norway.

Active outgrowing locations are taken from Norwegian Directorate of Fisheries aquaculture

biomass geodata.

Fish were captured using gill nets and fyke nets over reef or mixed sand-reef substrates at 5 -

30 m depth, and held in marine net pens, unfed, until the commencement of the experiment.

The experiment was conducted at the Austevoll Research Station, Norwegian Institute of

Marine Research (IMR). On 24 February 2016, we sedated all fish using 20 g L-1 tricaine

methanesulfonate (MS-222: Finquel). Fish with unhealed injuries or other welfare concerns

55

were killed with a blow to the head while sedated. We recorded length, wet weight and sex of

the remaining fish, inserted a Passive Integrated Transponder (PIT) tag into the peritoneal

cavity of every fish (allowing us to track individual weight loss between the start and finish of

spawning), and took ovarian biopsies from all females for storage at -80 °C. Within LFD and

HFD groups, 54 females and 24 males were allocated randomly among six cylindrical 7 m3

tanks per group (nine females and four males per tank, total 12 tanks).

Tanks were supplied with 6-8 °C sand-filtered and aerated seawater from 168 m depth and

exposed to a natural photoperiod through light-reducing shades. Hatchery facilities at IMR

Austevoll are described in detail by Karlsen et al. (2015). Winter spawning typically occurs in

the dark, during the early hours of the morning every 2-3 days from February to April. Each

spawning tank was appended with a cylindrical 100 L egg collection tank that filtered the full

volume of the spawning tank outflow via outlets at the top and bottom of the water column. A

constant circular flow was maintained within the egg collector to prevent eggs from being

pressed against the filter and damaged. The egg collectors were emptied every morning for

the duration of spawning and the volume of floating and sinking eggs was recorded.

Fish in this experiment were not fed for the duration of captivity to better preserve any effects

of diet prior to capture and to prevent clogging of egg collectors with waste matter. Reduced

feeding and significant weight loss is typical for both wild and captive cod during spawning

(Lambert & Dutil 2000), but to improve animal welfare, we removed fish from the experiment

early if their body condition dropped below acceptable levels (indicated by loss of muscle

mass, cessation of normal swimming behaviour, or unhealed wounds). Eggs were present after

the first night of egg collection, indicating that spawning had already begun. Accordingly, some

caution must be applied to interpretations of egg production and quality. However, the two

groups were collected at similar latitudes and housed in almost identical temperatures (the

main determinant of spawning commencement time: Kjesbu 1994), while temporal trends in

egg production do not suggest differential start or end times between groups (Fig. 3.4). Egg

production declined gradually toward the end of the spawning season, with the season

considered complete when all tanks had failed to produce viable eggs for four consecutive

days. All fish were humanely killed with a blow to the head while sedated with 20 g L-1 MS-222

before recording final length and weight.

On two occasions during the spawning season (3-4 March and 24 March), we took eggs for

incubation and hatching. Up to 350 mL eggs per tank were disinfected in 400 mg L−1

56

glutaraldehyde for 8 min to limit harmful bacterial growth (Harboe et al. 1994), and

transferred to 70 L black polyethylene conical incubators (one per spawning tank) with 0.5 L

min-1 seawater flow at 6 °C. Each morning during incubation, dead eggs were drained from the

bottom of the incubator and measured volumetrically. On 22 March, when the majority of

eggs from the first collection had hatched, we took approximately 6000 larvae from each

incubator and divided them across duplicate 50 L larval feeding tanks (24 tanks in total) at 8 °C.

Larvae remaining in the incubators were killed with a lethal dose of MS-222. Larvae were fed

size-fractionated copepod nauplii collected from the IMR sea-pond facility at Svartatjern.

Techniques for harvesting and preparation of copepods for rearing of cod larvae are described

in detail by van der Meeren et al. (2014) and Karlsen et al. (2015). Twice daily, larval rearing

tanks received 150000 nauplii, with 1.5 mL of algal paste per tank added to improve feeding

performance (Naas et al. 1992). The experiment was concluded on 13 April, with early season

larvae at day 42 and late season larvae at day 21 post fertilisation.

Ovarian fatty acid composition

We randomly selected ten ovarian samples per group (LFD and HFD) from the biopsies stored at the

beginning of the experiment. The samples weighed 60-100 mg each. All samples were methylated

and the respective fatty acid methyl esters were analysed on a HP-7890A gas chromatograph

(Agilent, USA) with a flame ionization detector (GC-FID), according to the method described in

Meier et al. (2006). The fatty acid 19:0 was added as an internal standard and 2.5 M dry HCl in

methanol was used as a methylation reagent. The methyl esters were extracted using 22 mL of

hexane, and the solution diluted or concentrated to obtain a suitable chromatographic response. 1

μL was injected splitless (the split was opened after 2 min) with the injection temperature set to

270 °C. The column was a 25 m × 0.25 mm fused silica capillary, coated with polyethylene-glycol of

0.25 μm film thickness, CP-Wax 52 CB (Varian-Chrompack, Middelburg, The Netherlands). Helium

(99.9999 %) was used as mobile phase at 1 ml/min for 45 min, then 3 ml/min for 25 min. The

temperature of the flame ionization detector was set at 300 °C. The oven temperature was

programmed to hold at 90 °C for 2 min, then heated to 150 °C at 30 °C/min and then to 240 °C at

2.5 °C/min and held steady for 30 min. Total analysis time was 70 min. Seventy well-defined peaks

in the chromatogram were selected, and identified by comparing retention times with a fatty acid

methyl ester standard (GLC-463 from Nu-Chek Prep. Elysian, MN, USA) and retention index maps

and mass spectral libraries (GC-MS) (http://www.chrombox.org/index.html) performed under the

57

same chromatographic conditions as the GC-FID (Wasta & Mjøs 2013). Chromatographic peak areas

were corrected by empirical response factors calculated from the areas of the GLC-463 mixture.

The chromatograms were integrated using the EZChrom Elite software (Agilent Technologies). Only

the 39 fatty acids that contributed more than 0.1 % of the total fatty acid amount were included in

the calculation. The total amount of fatty acids and cholesterol was calculated using the internal

standard 19:0.

Reproductive fitness traits

Body condition and weight loss

The amount and quality of egg production is likely to depend on fish condition. To allow

quantification of any such relationship, we compared body condition between LFD and HFD

groups (within sexes) using the relative condition index Krel = 100*(W/ExpW), where W is the

measured wet weight of the individual and ExpW is the expected weight (LeCren 1951). The

expected weight was calculated using a power function of the form ExpW = aLb fitted to the

full dataset including both LFD and HFD fish (Fig. 3.2). In this case, a = 0.0272 and b = 2.76 (R2 =

0.89). We also calculated proportional weight loss between the start and end of the spawning

season. This provides a general index of relative reproductive investment that is typically

closely correlated with egg production (e.g. Kjesbu et al. 1996).

Egg production

We quantified volumetric egg production in terms of daily egg production per tank, both with

and without a correction for the size of the females within a given tank, as well as any loss of

females during the season. Corrected egg production (relative daily egg production; RDEP) was

calculated as follows: RDEP = eggV / CFL, where eggV is the volume (mL) of eggs collected

from the tank, and CFL is the combined length (cm) of all females in the tank at the time of egg

collection. Four females (all HFD) experienced ≤10% weight loss, indicating little or no egg

release.

Egg quality

58

Viability, fertilization and early development

Egg viability rates were estimated from the proportion of eggs that were floating when the egg

collectors were emptied each morning. Fertilisation rates were assessed in subsamples of 100

eggs per tank on four occasions (three occasions for one HFD tank that finished spawning

early) during the spawning season (early season: 3-4 and 8-9 March, late season: 17 and 24-25

March), as egg quality typically declines during the spawning season (e.g. Kjesbu et al. 1996).

Where egg production was low on a given occasion, we combined fertilisation estimates from

two successive days. Eggs were scored as fertilised and normal (symmetrical cell divisions),

fertilised and abnormal (asymmetry or other abnormalities), or unfertilised.

Egg size and variability

Mean egg size and variability was measured on two occasions (3-4 March and 11 April) during

the spawning season. Eggs were taken from the egg collectors, stored in 6 °C seawater for 2-3

hours, and placed on a petri dish and photographed using a digital camera mounted on a light

microscope. The images had a 19 mm field of view at a resolution of 1024x768 pixels. We

measured the diameter of up to 50 eggs per tank (range 18-50) using the image analysis

software package ImageJ (Schneider et al. 2012), calibrated against a micrometer slide.

Hatching success

To estimate the proportion of viable eggs that successfully hatch, duplicate egg subsamples

(>100 eggs) were taken from incubators on two occasions, early and late in the season (early

season: 10 March, Day 8 of incubation; late season: 1 April, Day 9 of incubation) for a hatching

trial. Eggs were rinsed with filtered seawater, placed in covered containers filled with 200 mL

of filtered seawater, and maintained at 6-7 °C until hatching was complete. Dead eggs and live

and dead hatchlings were scored and removed daily until no viable embryos remained. As

there is large variation in growth rates between individual larvae, cannibalism makes it

59

impossible to reliably estimate impacts of farm density on larval survival in this system beyond

first feeding.

Larval quality

Larval development and growth rates

Subsamples of at least 30 larvae per tank were collected 40 days after spawning (23-28 days

post hatching), killed by a lethal dose of MS-222 and stored at 6 °C until required for

photography (<4 hrs). Larvae were transferred to a petri dish with a thin layer of seawater and

photographed laterally under dark field illumination on a light microscope with mounted

digital camera. The images provided a 12 mm field of view at a resolution of 1024x768 pixels.

Larvae were measured for length using the ImageJ measuring tool calibrated against a

micrometer slide. The measurement was taken along a polyline running from the tip of the

snout to the cranial vertebra and along the spine to the end of the caudal peduncle.

Phototaxis

Development of visual and cognitive systems in larvae (at day 42, ca. 23 days since first

feeding) was compared by means of a phototaxis trial in which we tested the proportion of

larvae exhibiting behavioural responses to a light gradient. Phototactic responses in this

context correlate with larval fitness (e.g. Karlsen and Mangor-Jensen 2001, Forsgren et al.

2013). Approximately 70 larvae (± 16.9 SD) were placed in a 60 x 10 x 5 cm tank and allowed to

disperse. We then completely covered the tank with black plastic to block any light. After 8

mins, we removed the covering on a 10 x 10 x 5 cm end section of the tank (thus exposing it to

ambient light from a fluorescent bulb) and scored the number of larvae already present in the

end section. This was treated as the control score: the number of larvae present in the end

section due to random dispersal without stimulus. The tank was then left for a further 5 mins,

and the number of larvae in the end section scored again. We then removed the rest of the

covering and counted the larvae that had not showed preference for light. The trial was

conducted for each of the 12 spawning tanks. The change in spatial distribution (the difference

60

in the proportion of larvae in the end section before and after the light treatment) provided a

measure of responsiveness to this environmental stimulus.

Statistical analysis

We compared initial fish size, weight-at-length and condition (Krel) metrics across LFD and HFD

groups (Group), using linear analysis of variance models constructed using the lm function in R

(R Core Team 2016). Data were log transformed as necessary to improve normality (and to

linearise weight-at-length curves). Proportional weight loss was compared between LFD and

HFD groups using a beta regression generalised linear model fitted using the betareg package

for R (Cribari-Neto & Zeileis 2010).

We conducted several ovarian fatty acid analyses: initally, we made univariate comparisons of

total lipids, cholesterol, and aquafeed markers oleic and linoleic acid across LFD and HFD

groups using linear analysis of variance models. We then compared the entire suite of fatty

acids across groups using a multivariate permutational analysis of variance (Permanova) fitted

to a Euclidean dissimilarity matrix using Primer 6 software (Anderson et al. 2008, White et al.

2017c).

Egg production metrics were compared between treatments and over time using negative

binomial generalised linear mixed models fitted using the glmmTMB package (Brooks et al.

2017). Group (farm density group) and Day (sampling day) were fitted as fixed effects. As tanks

were sampled repeatedly, nonindependence between samples was addressed by including a

random intercept term with TankID (tank identity) nested within Group. We report significance

of individual model terms within glmmTMB models by comparing the fit of the full model and

a null model with the term removed (X2 test on 1 df).

Effects of farm density on egg quality metrics were tested by fitting linear mixed models, also

using the glmmTMB package. Egg size data were best fitted with a gaussian model family,

while proportion data were fitted using a beta regression family. We included three fixed

terms: Group, Time (early or late season collections) and MeanFL (mean female length at the

tank level, to account for possible effects of female age on egg quality). As with egg production

models, we included the nested random intercept term Group/TankID. Proportion or rate

responses were analysed at the tank level. Egg size data were analysed at the egg level, with

Group, Time and MeanFL as fixed terms and Group/TankID as a random intercept term.

61

Following the full analysis described above, egg quality datasets were split into early and late

season collections and re-analysed using beta regression models fitted using the betareg

package. These results are reported alongside those from the full analyses (Table 3.2), using a

z-test of significance for the Group term.

Data on deformity rates, maximum sizes and phototaxis were analysed at the tank level using

beta regression models fitted with Group and MeanFL as fixed terms (betareg package). The

Time and TankID terms were not necessary as all larvae were reared from eggs collected over a

two day period early in the season. Larval size data were analysed at the level of individual

larvae using a linear mixed model fitted by glmmTMB, with Group and MeanFL as fixed terms

and Group/TankID as a random intercept term.

As tank-level analyses come with a cost to statistical power, we calculated Cohen’s d effect

sizes (calculated using individuals as replicates for fish size and condition, and tanks as

replicates for all egg and larval quality data) to provide an estimate of effects independent of

sample size and statistical significance (Cohen 1988).

Figure 3.2. Weight-at-length relationship for female (left panel) and male (right panel) cod

used in this study. The best fitting power function (y = axb) is shown for each combination of

sex and group (HFD: red triangles, solid red fitted line; LFD: blue circles, dashed blue fitted

line).

62

RESULTS

Body condition and weight loss

Female Atlantic cod collected from the high farm density environment (HFD) were significantly

longer (7%) and heavier (16%) than fish from the area of low farm density (LFD), with no

difference in body condition between females from HFD and LFD sites (Table 3.1). HFD males

were not significantly different in length or weight from LFD males, but did have a higher body

condition than LFD males (Table 3.1). The log-transformed weight-at-length slope was

significantly higher for HFD females than LFD females (although the effect was small: partial R2

= 0.06), while weight-at-length did not differ for HFD and LFD males (Appendix 3.2, Fig. 3.2).

During the captive spawning season (mean 37 days LFD, 39 days HFD), females lost 29 ± 9%

body weight, while males lost 18 ± 6% (mean ± SD). LFD females lost more weight than HFD

females (mean ± SE: 30 ± 1.1% and 27 ± 1.2% respectively, z1,9 = 2.4, p = 0.02), consistent with

greater reproductive effort by LFD females.

Table 3.1. Body size and condition metrics for low (LFD) and high (HFD) farm density groups at

the commencement of the experiment (group means ± standard error). Significant p values are

highlighted in bold. Positive direction indicates that metric was higher in HFD fish.

LFD HFD Statdf p Cohen’s d

Females n = 54 n = 54 Length (cm) 60.1 ± 1.1 64.6 ± 1.2 F1,106 = 7.2 0.008 +0.49 Weight (g) 2442 ± 145 2839 ± 164 F1,106 = 4.2 0.043 +0.35 Condition (Krel) 1.03 ± 0.03 0.98 ± 0.01 F1,106 = 3.1 0.08 -0.34 Males n = 24 n = 24 Length (cm) 60.3 ± 1.2 62.8 ± 2.0 F1,46 = 0.9 0.34 +0.29 Weight (g) 2204 ± 148 2624 ± 243 F1,46 = 1.7 0.20 +0.37 Condition (Krel) 0.98 ± 0.02 1.04 ± 0.01 F1,46 = 4.7 0.04 +0.63

Ovarian fatty acid composition

We found no evidence that ovarian lipid modification reflected the density of salmon farms,

with no significant difference in total fatty acid content or cholesterol content, nor in total

saturated, monounsaturated or polyunsaturated fatty acids (Appendix 3.1). Likewise,

multivariate fatty acid composition was not significantly different between LFD and HFD

63

groups (pseudo-F18 = 0.58, p = 0.71; Fig. 3.3). Multidimensional scaling (MDS) revealed that

three individuals—one LFD and two HFD—were separated from the main cluster. These three

fish were not remarkable in size or body condition (57-72 cm, 2022-3620 g, Krel 0.99-1.11), but

both HFD fish were high in oleic acid 18:1 (n-9) (15.2 and 17.5%) and one was also high in

linoleic acid 18:2 (n-6) (3.7%). The outlying LFD fish was also high in linoleic acid (2.5%). Both

oleic and linoleic acid can indicate consumption of commercial feed (White et al. 2017c), and

were marginally higher in the HFD group on average (6% and 18% increases, respectively), but

this effect was not significant for either oleic acid (F1,18 = 1.9, p = 0.2) or linoleic acid (F1,18 = 0.4,

p = 0.6). Levels of these fatty acids were strongly positively correlated with each other in the

HFD group (Pearson’s r = 0.90, t8 = 6.0, p <0.001), while this effect was nonsignificant in the

LFD group (r = 0.44, t8 = 1.4, p = 0.2). The highest levels of both oleic and linoleic acid (17.5%

and 3.7% respectively) occurred in one individual from the HFD group.

Figure 3.3. Multidimensional scaling (MDS) plot showing dissimilarly (Euclidean distance) of

multivariate fatty acid profiles in Atlantic cod ovaries according to salmon farm density.

Individuals are grouped by high (HFD; red triangles) and low (LFD; blue circles) salmon farm

density. Model stress is 0.08.

Egg production

Twelve spawning tanks produced a total of 137 L of eggs over the spawning season, with no

difference in the total egg volume from LFD and HFD tanks (11.3 ± 0.9 L and 11.3 ± 0.7 L,

64

respectively; Fig. 3.4). A model comparing raw daily egg production between HFD and LFD

tanks over time found no effect of the farm density factor (p = 0.08; Table S3). There was a

significant temporal decline in daily egg production (p <0.0001, Fig. 3.4), and a positive effect

of total female length in the tank on daily egg production (p <0.0001).

The relative daily egg production metric (daily egg production corrected for total female

length) also did not significantly differ across groups (LFD: 64.6 ± 6.2 cf. HFD: 57.2 ± 3.4 mL

female m-1 day-1 tank-1; p = 0.52), but again, there was a significant decline over time (p

<0.0001, Fig. 3.4).

Initial female body condition (tank mean values) did not significantly predict either raw daily

egg production or relative daily egg production metrics overall (p >0.10 in both cases), or

within farm density groups (p >0.05 in each case). However, mean female weight loss during

the season was strongly correlated with both egg production metrics (p <0.0004 in each case).

Figure 3.4. Daily egg production per tank during the captive spawning period (25 Feb 2016 – 11

April 2016). Tanks are grouped by high (HFD; red triangles, solid red fitted line) and low (LFD;

blue circles, dashed blue fitted line) farm density. Production is quantified by (left panel) raw

daily egg volume per tank and (right panel) with a correction for the total length of females in

the tank. The temporal trend within groups is fitted by third order polynomial functions.

Egg quality and survivorship

The proportion of viable (floating) eggs was almost identical across farm density groups (Table

3.2; Table S3), with the viable proportion significantly declining over the season (p <0.0001).

65

Neither mean female length (p = 0.18) nor body condition (p = 0.60) were correlated with egg

viability rates at the tank level, regardless of farm density group.

The LFD group produced larger eggs over the duration of the season (Table 3.2; Appendix 3.3).

LFD eggs were 4.8 % larger by diameter at the time of the early season collections, with no

significant difference later in the season (Table 3.2). We did not test for a temporal decline in

egg size, as late but not early season samples were fixed in formalin. Again, neither mean

female length (p = 0.065) nor body condition (p = 0.32) were significantly correlated with egg

size, regardless of farm density group.

Rates of asymmetrical cell division in pre-blastula eggs did not differ between LFD and HFD

groups in either early or late season collections (Table 3.2; Appendix 3.3). Neither mean female

size nor body condition were correlated with egg symmetry (p >0.06 in each case). Mean

asymmetry rates did not exceed 10% in either group on any collection date, although samples

ranged from 0-53% asymmetry.

Early season eggs hatched 12-17 days after collection, while late season eggs hatched 14-18

days after collection. Egg survival during incubation prior to the hatching trial did not differ

significantly between groups (Table 3.2; Appendix 3.3). Declines in survivorship between early

and late season collections did not differ (p >0.10), nor did the effect of mean female length or

body condition (p >0.10 in each case).

Hatching success rates were similar between groups (Table 3.2; Appendix 3.3), with no decline

between early and late season collections. One HFD tank ceased spawning before the late

season egg sampling, so late season HFD hatching data comes from only five tanks.

Larval quality

Larvae from LFD tanks were 8 % larger than those from HFD tanks (Table 3.2; Appendix 3.3).

Neither mean female size (p = 0.97) nor body condition (p = 0.34) were correlated with larval

size either overall or within farm density groups. The size of the largest larvae from each tank

did not differ across farm density groups, nor did the rate of deformities (Table 3.2; Appendix

3.3). Neither were significantly correlated with mean female length or body condition (p >0.10

in each case). Overall rates of larval deformity were 48% at day 40 from both LFD and HFD

groups (Table 3.2). Most deformities were of the spine (lordosis, kyphosis or vertebral

misalignment) (152/173 deformities), followed by deformities of the jaw (15/173) and skull

66

(6/173). Larvae from both groups exhibited phototaxis in response to a horizontal light

gradient (proportion in end section before exposure: 0.25 ± 0.03 cf. after exposure: 0.69 ±

0.04; t-test: t22 = -9.5, p <0.0001), but we found no evidence that the farm density groups

differed in the extent of the phototactic response (Table 3.2; Appendix 3.3), nor any strong

evidence for an effect of mean female length or body condition (p >0.13 in each case).

DISCUSSION

This study presents limited evidence for negative impacts of high salmon farm density on

reproductive fitness in the studied Atlantic cod (Gadus morhua) population. Female cod

collected from locations with a high density of salmon farms (HFD) were larger and heavier

than fish from the low farm density location (LFD), but with no consistent changes in female

body condition. Effects on ovarian lipid composition were also small and largely limited to two

HFD individuals. There was no significant difference in egg production, viability, fertilisation

rates, symmetry or hatching success, but eggs from HFD tanks were 5 % smaller and this likely

contributed to the observed -8 % disparity in mean larval length at 40 days post spawning for

the HFD group relative to the LFD group.

Among the viable proportion, fertilisation rates were similar for LFD and HFD tanks

respectively (Table 3.2; Appendix 3.3), but declined during the season (p = 0.009). Neither

mean female length nor body condition significantly predicted fertilisation rates overall (p

>0.06 in each case), although the effect of female length was significant within the LFD group

(p = 0.04). Given the very small difference in egg viability and fertilisation rates between

groups, and nonsignificant effects of male length and body condition on fertilisation rates

either overall or within groups (p >0.13 in each case), we did not test for differences in sperm

quality.

Gadid fishes accumulate significant energy reserves for reproduction, with lipids stored

primarily in the liver and proteins in muscle tissue (Kjesbu et al. 1991, Lambert & Dutil 1997).

Accordingly, body condition indices during vitellogenesis are typically good predictors of

fecundity in coastal cod from this region (Skjæraasen et al. 2006). In our case, mean female

size and percentage weight loss during the season positively tracked egg production at the

tank level; both were better predictors of egg production than initial female body condition.

Table 3.2. Egg and larval quality metrics for cod from low (LFD) and high (HFD) farm density

areas (group mean ± standard error). Positive effect sizes indicate that quality was higher in

67

the HFD group. Significance has not been corrected for false discovery rate of multiple

comparisons.

LFD HFD N (HFD, LFD) Stat p Cohen’s d

Viability rate

Early season 0.96 ± 0.01 0.94 ± 0.03 102, 102 z = 0.21 0.83 -0.25

Late season 0.78 ± 0.15 0.76 ± 0.11 70, 74 z = 0.56 0.56 -0.19

Overall 0.87 ± 0.04 0.85 ± 0.04 172, 176 X2 = 0.23 0.61 -0.14

Egg diameter (μm) (of viable eggs)

Early season 1183 ± 16 1129 ± 16 257, 289 X2 = 5.8 0.016 -1.36

Late season* 1261 ± 30 1210 ± 13 221, 271 X2 = 5.3 0.07 -0.92

Overall 1222 ± 20 1166 ± 16 478, 560 X2 = 6.5 0.011 -0.90

Fertilisation rate (of viable eggs)

Early season 0.75 ± 0.04 0.77 ± 0.03 19, 20 z = 0.68 0.49 +0.27

Late season 0.59 ± 0.09 0.68 ± 0.06 15, 18 z = 0.47 0.63 +0.46

Overall 0.67 ± 0.08 0.73 ± 0.04 34, 38 X2 = 1.0 0.60 +0.36

Egg symmetry rate (of fertilised viable eggs)

Early season 0.92 ± 0.01 0.90 ± 0.02 19, 20 z = 0.69 0.49 -0.29

Late season 0.88 ± 0.18 0.88 ± 0.08 15, 18 z = 0.48 0.63 -0.10

Overall 0.90 ± 0.03 0.89 ± 0.04 34, 38 X2 = 0.35 0.84 -0.08

Egg survival rate during incubation

Early season 0.85 ± 0.03 0.83 ± 0.05 6, 6 z = 0.66 0.49 -0.23

Late season 0.56 ± 0.11 0.68 ± 0.09 5, 6 z = 0.48 0.63 +0.41

Overall 0.70 ± 0.08 0.76 ± 0.06 11, 12 X2 = 0.01 0.95 +0.23

Hatching success rate

Early season 0.90 ± 0.03 0.85 ± 0.03 12, 12 z = 1.56 0.12 -0.64

Late season 0.92 ± 0.03 0.88 ± 0.08 11, 12 z = 0.15 0.88 -0.07

Overall 0.91 ± 0.02 0.86 ± 0.04 23, 24 X2 = 0.18 0.67 -0.48

Larval length (mm)

Early season 6.46 ± 0.05 5.98 ± 0.18 180, 176 X2 = 3.9 0.048 -1.46

Max larval length (mm)

Early season 8.05 ± 0.17 7.68 ± 0.21 6, 6 t1, 10 = 1.32 0.21 -0.77

Larval deformity rate

Early season 0.48 ± 0.06 0.48 ± 0.07 6, 6 z1,9 = 0.01 0.99 0

Larval phototaxis rate

Early season 0.71 ± 0.07 0.67 ± 0.04 6, 6 z1,9 = 0.79 0.43 -0.28

*Late season eggs were fixed in formalin prior to examination, which may have affected egg diameter.

68

However, the direction of the body condition trend was positive, and together, our findings are

consistent with greater reproductive investment relative to body length by LFD females.

Egg quality metrics revealed similar quality overall in LFD and HFD groups, but with

significantly smaller eggs from HFD fish. This was somewhat unexpected, as HFD females were

slightly larger; female size and age in cod is usually positively correlated with egg size (e.g.

Marteinsdottir & Steinarsson 1998, Vallin & Nissling 2000), although body condition can be

equally important (Chambers & Waiwood 1996, Marteinsdottir & Steinarsson 1998). The egg

size effect that we observed did not correspond to any significant decrease in other metrics of

egg quality, and the cumulative effect of the more direct metrics (viability, fertilisation,

developmental symmetry, survival during incubation and hatching success) was such that eggs

collected from HFD and LFD tanks did not differ substantially in their likelihood of successfully

hatching (HFD: 40 % cf. LFD: 37 %). However, previous studies have indicated that egg size can

predict larval quality, with larger size at hatching, faster growth rates and successful

development of the swim bladder leading to a survival advantage for larvae from larger eggs

(Knutsen & Tilseth 1985, Marteinsdottir & Steinarsson 1998). This prediction was consistent

with our larval size data; HFD larvae were 8 % smaller on average than their LFD counterparts

40 days after spawning.

Larger larvae often exhibit differing responses to stimuli such as light (e.g. Colton & Hurst

2010). In some fish, phototaxis along a horizontal light gradient correlates positively with other

metrics of development (e.g. first feeding: Karlsen & Mangor-Jensen 2001) and is affected by

the environment (Forsgren et al. 2013). In our case, LFD and HFD fish did not differ in their

phototactic response to a horizontal light gradient, although the effect was in the direction of

the greater phototaxis by LFD larvae (Cohen’s d = 0.28). Regardless, even small differences in

hatching size and larval growth rates can affect survival during the planktonic stage and even

influence post settlement fitness (e.g. Sclafani et al. 1993, Shima & Swearer 2010), so it is likely

that the HFD larvae in this study would have experienced nontrivial negative fitness effects

later in development.

Condition indices for fish in the present study were similar to those from wild caught fish in the

same region prior to the expansion of salmon farming (Botros 1962, cited in Kjesbu 1989). Egg

sizes in our study (mean 1.2 mm) were at the lower end of those reported by previous captive

spawning studies with Norwegian coastal cod (1.2-1.4 mm: Kjesbu et al. 1996, Otterå et al.

2006, van der Meeren & Ivannikov 2006, Bogevik et al. 2012, Karlsen et al. 2015), but this may

be related to the lack of feed during spawning rather than condition at the time of collection

(Kjesbu et al. 1990). Egg fertilisation, symmetry and hatching rates were all within the range of

69

those reported by previous studies (Morgan et al. 1999, van der Meeren & Ivannikov 2006,

Bogevik et al. 2012, Karlsen et al. 2015), while larval growth rates (6.2 mm cf. 7.4 mm at 25

days post hatching) were slightly below those reported by Karlsen et al. (2015).

While previous studies have found significant shifts in ovarian fatty acid composition in

captive-fed sea bream (Cejas et al. 2003) and farm-associated bogue (L. Martinez-Rubio

unpubl. data, cited in Fernandez-Jover et al. 2011b), the absence of such clear effects in this

study should not be taken as strong evidence that these fish have not fed at salmon farms.

Lipid profiles of the gonads are less affected by diet than those of the liver or muscle tissue, as

the gonads are composed almost entirely of phospholipids rather than dietary fatty acids.

Accordingly, our analysis of ovarian lipid composition may only have detected effects of a large

dietary shift, while the presence of a single active salmon farm near one LFD location (affecting

25 % of LFD fish) may have reduced overall differences in fatty acid profiles between HFD and

LFD groups. Indirect intake of farm waste via predation of farm-associated invertebrates and

fish, combined with dietary sparing and biosynthesis, may also weaken or mask fatty acid

signals in fish that are not strongly farm-associated (White et al. 2017c). The strong correlation

between levels of oleic and linoleic acids in the HFD group (less so in the LFD group) is

consistent with (but not necessarily strong evidence for) a spectrum of farm association, with

the most strongly farm-associated individuals in the HFD group having higher levels of oleic

and linoleic acids than their counterparts in the LFD group. Both oleic and linoleic acid are

present in the natural diet but are especially abundant in commercial salmon feed, and captive

feeding trials revealed a strong correlation between these two fatty acids in muscle, liver and

gonad tissue of saithe fed on salmon feed (Karlsen et al., unpublished data).

Our collection sites were selected to represent levels of farm exposure experienced by cod

populations in the south-western fjords (Fig. 3.1). Importantly, cod do not associate with farms

as closely or persistently as other species (e.g. saithe), and likely do not feed exclusively at

farms for extended periods (Dempster et al. 2011), so our experiment was designed to test for

effects of farm density on cod whose home ranges at the time of collection overlap with areas

of farming influence. Salmon farms have relatively localised nutrient footprints, with acute

deposition within 50-250 m and very diffuse deposition beyond 500 m (Kutti et al. 2007,

Bannister et al. 2016, White et al. 2017b). Accordingly, our collection sites are likely to reflect

the two dominant types of farm exposure for wild Norwegian fjord cod: temporally dynamic or

partial association versus little or no association. Tissue fatty acid profiles in cod are altered

within 3 weeks of a dietary shift (Kirsch et al. 1998), with vitellogenesis commencing 3-4

months prior to spawning (Skjæraasen et al. 2006). Individuals that reside within farm

70

footprints throughout vitellogenesis are likely to show the largest shifts in ovarian fatty acid

profiles and concomitant effects on reproductive physiology and output. We do not know how

long HFD and LFD cod have resided at their capture locations, but available telemetry and

mark-recapture data indicate that wild coastal cod have relatively restricted home ranges over

a scale of weeks and months. 87 % of tag returns for wild cod released in Heimarkspollen, a 2.9

km2 semi-enclosed fjord in the Austevoll archipelago, were recaptured within the fjord

(Svåsand 1990), while in Balsfjord in northern Norway, the majority of wild cod tagged and

released at a farm were still present at the same farm 9-12 weeks later (Uglem et al. 2008). In

the Shetland Isles, 133 wild cod were tagged and released, with 37/39 recaptures over a two

year period occurring within 15 km of the release site (Neat et al. 2006). While some

individuals move larger distances and may have spent time in areas that differ from their

capture location in terms of farm density, our study only assumes that fish collected from LFD

locations will, on average, be less affected by farms than fish from HFD locations.

Previous captive feeding experiments have shown that gadids and other fishes fed commercial

diets experience changes in reproductive fitness (e.g. Salze et al. 2005, Bogevik et al. 2012),

but very little work has been done to assess potential impacts in a real world ecological

context. Taken together, our findings indicate that salmon farming in this region has some

negative effects on the reproductive physiology of Atlantic cod on a fjord-wide scale, with

potential cumulative effects of egg and larval size on later developmental stages. More work is

needed to track fitness effects later in development, including potential effects of decreased

egg and larval size on later life stages. In addition, as our study was restricted in its geographic

extent and spatial replication, we are cautious of generalising these findings beyond our study

environment. We encourage others to replicate and extend this important line of research.

Future work may also consider the other potential pathway for fitness impacts: individual

mortality. Mortality may decline due to the provision of a trophic subsidy (Kilambi et al. 1978),

or increase due to elevated levels of contamination, infection, or predation at farms. Wild fish

aggregations are also an easy target for fishers, and fishing mortality will take on greater

importance if the current 100 m fishing restriction around Norwegian farms is lifted (Bagdonas

et al. 2012). Previous studies have employed acoustic tracking to good effect in comparing

spatiotemporal movement and mortality rates of Atlantic cod individuals across habitats

(Olsen et al. 2011, Olsen et al. 2012, Fernández-Chacón et al. 2015), and it would be entirely

feasible to apply the same approach to quantify differential mortality rates in wild fish across

multiple farm-affected and non-affected areas.

71

ACKNOWLEDGMENTS

We thank the staff at IMR Austevoll for their expert advice and assistance in the collection and

maintenance of the spawning fish, egg incubation and feeding of larvae: Margareth Møgster,

Stig Ove Utskot, Michal Rejmer, Inger Semb Johansen, Nele Gunkel-Sauer, Kristine Hovland

Holm, Yvonne Rong, Terje van der Meeren, Tord Skår, Velimir Nola and Glenn Sandtorv. We

also thank the staff and students at Austevoll High School for their assistance in transporting

live fish to the research station. Theresa Aase prepared the fatty acid analysis. Camille White

provided useful comments on lipid data. Francisca Samsing assisted with the creation of

figures. The manuscript was improved by comments from three anonymous reviewers. All

procedures were conducted in accordance with Norwegian animal welfare regulations by

experienced personnel.

This project was funded by the Sustainable Aquaculture Laboratory, University of Melbourne,

and the Norwegian Seafood Research Fund (Fiskeri og Havbruksnæringens Forskningsfond).

LITERATURE CITED

Akyol O, Ertosluk O (2010) Fishing near sea-cage farms along the coast of the Turkish Aegean

Sea. J Appl Ichthyol 26:11–15

Anderson M, Gorley R, Clarke R (2008) Permanova+ for Primer: guide to software and

statistical methods. Primer-E Limited, Plymouth.

Arechavala- López P, Sæther B-S, Marhuenda-Egea F, Sánchez-Jerez P, Uglem I (2015)

Assessing the influence of salmon farming through total lipids, fatty acids, and trace

elements in the liver and muscle of wild saithe Pollachius virens. Mar Coast Fish 7:59–

67

Arechavala-López P, Sánchez-Jerez P, Bayle-Sempere JT, Uglem I, Mladineo I (2013) Reared

fish, farmed escapees and wild fish stocks—a triangle of pathogen transmission of

concern to Mediterranean aquaculture management. Aquac Environ Interact 3:153–

161

Bagdonas K, Humborstad O-B, Løkkeborg S (2012) Capture of wild saithe (Pollachius virens)

and cod (Gadus morhua) in the vicinity of salmon farms: Three pot types compared.

Fish Res 134–136:1–5

72

Balme GA, Slotow R, Hunter LTB (2010) Edge effects and the impact of non-protected areas

in carnivore conservation: leopards in the Phinda–Mkhuze Complex, South Africa.

Anim Conserv 13:315–323

Bannister RJ, Johnsen IA, Hansen PK, Kutti T, Asplin L (2016) Near- and far-field dispersal

modelling of organic waste from Atlantic salmon aquaculture in fjord systems. ICES J

Mar Sci 73:2408–2419

Bates D, Maechler M, Bolker B, Walker S (2015) Fitting linear mixed effects models using

lme4. J Stat Softwar 67:1–48

Bogevik AS, Natário S, Karlsen Ø, Thorsen A, Hamre K, Rosenlund G, Norberg B (2012) The

effect of dietary lipid content and stress on egg quality in farmed Atlantic cod Gadus

morhua. J Fish Biol 81:1391–405

Botros G (1962) Die Fruchtbarkeit des Dorsches (Gadus morhua L.) in der westlichen Ostsee

und den westnorwegischen Gewassern. Kieler Meeresforsch 18:67–80

Brooks ME, Kristensen K, van Benthem KJ, Magnusson A, Berg CW, Nielsen A, Skaug HJ,

Maechler M, Bolker BM (2017) Modeling zero-inflated count data with glmmTMB.

bioRxiv preprint bioRxiv:132753.

URL: https://www.biorxiv.org/content/early/2017/05/01/132753

Burridge L, Weis JS, Cabello F, Pizarro J, Bostick K (2010) Chemical use in salmon aquaculture:

A review of current practices and possible environmental effects. Aquaculture 306:7–

23

Cejas JR, Almansa E, Villamandos JE, Badıa P, Bolaños A, Lorenzo A (2003) Lipid and fatty acid

composition of ovaries from wild fish and ovaries and eggs from captive fish of white

sea bream (Diplodus sargus). Aquaculture 216:299–313

Chambers RC, Waiwood KG (1996) Maternal and seasonal differences in egg sizes and

spawning characteristics of captive Atlantic cod, Gadus morhua. Can J Fish Aquat Sci

53:1986–2003

Cohen J (1988) Statistical power analysis for the behavioral sciences (2nd edn). Erlbaum,

Hillsdale

Colton AR, Hurst TP (2010) Behavioral responses to light gradients, olfactory cues, and prey

in larvae of two North Pacific gadids (Gadus macrocephalus and Theragra

chalcogramma). Environ Biol Fishes 88:39–49

Cribari-Neto F, Zeileis A (2010) Beta egression in R. J Stat Softw 34:1–24

73

Czesny S, Dabrowski K, Christensen JE, Eenennaam J Van, Doroshov S (2000) Discrimination

of wild and domestic origin of sturgeon ova based on lipids and fatty acid analysis.

Aquaculture 189:145–153

Dempster T, Sánchez-Jerez P, Bayle-Sempere JT, Giménez-Casalduero F (2002) Attraction of

wild fish to sea-cage fish farms in the south-western Mediterranean Sea: spatial and

short-term temporal variability. Mar Ecol Prog Ser 242:237–252

Dempster T, Sánchez-Jerez P, Fernandez-Jover D, Bayle-Sempere JT, Nilsen R, Bjørn P-A,

Uglem I (2011) Proxy measures of fitness suggest coastal fish farms can act as

population sources and not ecological traps for wild gadoid fish. PLOS ONE 6:e15646

Dempster T, Sánchez-Jerez P, Uglem I, Bjørn P-A (2010) Species-specific patterns of

aggregation of wild fish around fish farms. Estuar Coast Shelf Sci 86:271–275

Dempster T, Uglem I, Sánchez-Jerez P, Fernandez-Jover D, Bayle-Sempere J, Nilsen R (2009)

Coastal salmon farms attract large and persistent aggregations of wild fish: an

ecosystem effect. Mar Ecol Prog Ser 385:1–14

Fernández-Chacón A, Moland E, Espeland SH, Olsen EM (2015) Demographic effects of full vs.

partial protection from harvesting: inference from an empirical before-after control-

impact study on Atlantic cod. J Appl Ecol 52:1206–1215

Fernandez-Jover D, Arechavala-López P, Martinez-Rubio L, Tocher D, Bayle-Sempere J, López-

Jiménez J, Martinez-López F, Sánchez-Jerez P (2011) Monitoring the influence of

marine aquaculture on wild fish communities: benefits and limitations of fatty acid

profiles. Aquac Environ Interact 2:39–47

Fernandez-Jover D, Jimenez JAL, Sánchez-Jerez P, Bayle-Sempere JT, Giménez-Casalduero F,

López FJM, Dempster T (2007) Changes in body condition and fatty acid composition of

wild Mediterranean horse mackerel (Trachurus mediterraneus, Steindachner, 1868)

associated to sea cage fish farms. Mar Environ Res 63:1–18

Fernandez-Jover D, Martinez-Rubio L, Sánchez-Jerez P, Bayle-Sempere JT, Lopez Jimenez JA,

Martínez López FJ, Bjørn P-A, Uglem I, Dempster T (2011) Waste feed from coastal fish

farms: a trophic subsidy with compositional side-effects for wild gadoids. Estuar Coast

Shelf Sci 91:559–568

Fernandez-Jover D, Sánchez-Jerez P, Bayle-Sempere JT, Valle C, Dempster T (2008) Seasonal

patterns and diets of wild fish assemblages associated with Mediterranean coastal fish

farms. ICES J Mar Sci 65:1153–1160

Forsgren E, Dupont S, Jutfelt F, Amundsen, T (2013) Elevated CO2 affects embryonic

development and larval phototaxis in a temperate marine fish. Ecol Evol 3:3637–3646

74

Glover KA, Sørvik AGE, Karlsbakk E, Zhang Z, Skaala Ø (2013) Molecular genetic analysis of

stomach contents reveals wild Atlantic cod feeding on piscine reovirus (PRV) infected

Atlantic salmon originating from a commercial fish farm. PLOS ONE 8:e60924

Hale R, Swearer SE (2016) Ecological traps: current evidence and future directions. Proc R Soc

B Biol Sci 283:494–499

Hale R, Treml EA, Swearer SE (2015) Evaluating the metapopulation consequences of

ecological traps. Proc R Soc London B Biol Sci 282

Hansen T, Karlsen Ø, Taranger GL, Hemre G-I, Holm JC, Kjesbu OS (2001) Growth, gonadal

development and spawning time of Atlantic cod (Gadus morhua) reared under

different photoperiods. Aquaculture 203:51–67

Hallier J-P, Gaertner D (2008) Drifting fish aggregation devices could act as an ecological trap

for tropical tuna species. Mar Ecol Prog Ser 353:255–264

Harboe T, Huse I, Øie G (1994) Effects of egg disinfection on yolk sac and first feeding stages

of halibut (Hippoglossus hippoglossus L.) larvae. Aquaculture 119:157–165

Johansen L-H, Jensen I, Mikkelsen H, Bjørn P-A, Jansen PA, Bergh Ø (2011) Disease

interaction and pathogens exchange between wild and farmed fish populations with

special reference to Norway. Aquaculture 315:167–186

Karlsen Ø, Mangor-Jensen A (2001) A correlation between phototactic response and first-

feeding of Atlantic halibut (Hippoglossus hippoglossus L.) larvae. Aquacult Res 32: 907–

912

Karlsen Ø, Meeren T van der, Rønnestad I, Mangor-Jensen A, Galloway TF, Kjørsvik E, Hamre

K (2015) Copepods enhance nutritional status, growth and development in Atlantic

cod (Gadus morhua L.) larvae - can we identify the underlying factors? PeerJ 3:e902

Kilambi RV, Adams JC, Wickizer WA (1978) Effects of cage culture on growth, abundance, and

survival of resident largemouth bass (Micropterus salmoides). J Fish Res Board Canada

35:157–160

Kirsch PE, Iverson SJ, Bowen WD, Kerr SR, Ackman RG (1998) Dietary effects on the fatty acid

signature of whole Atlantic cod (Gadus morhua). Can J Fish Aquat Sci 55:1378–1386

Kjesbu OS (1989) The spawning activity of cod, Gadus morhua L. J Fish Biol 34:195–206

Kjesbu OS (1994) Time of start of spawning in Atlantic cod (Gadus morhua) females in

relation to vitellogenic oocyte diameter, temperature, fish length and condition. J Fish

Biol 45:719–735

75

Kjesbu OS, Klungsøyr J, Kryvi H, Witthames PR, Walker MG (1991) Fecundity, atresia, and egg

size of captive Atlantic cod (Gadus morhua) in relation to proximate body composition.

Can J Fish Aquat Sci 48:2333–2343

Kjesbu OS, Solemdal P, Bratland P, Fonn M (1996) Variation in annual egg production in

individual captive Atlantic cod (Gadus morhua). Can J Fish Aquat Sci 53:610–620

Kjesbu OS, Witthames PR, Solemdal P, Walker MG (1990) Ovulatory rhythm and a method to

determine the stage of spawning in Atlantic cod (Gadus morhua). Can J Fish Aquat Sci

47:1185–1193

Kloskowski J (2012) Fish stocking creates an ecological trap for an avian predator via effects

on prey availability. Oikos 121:1567–1576

Knutsen H, Olsen EM, Espeland SH (2016) Norwegian coastal cod south of 62° N (in

Norwegian). In: Institute for Marine Research Report (Havforskningsrapporten).

Norwegian Institute for Marine Research (Havforskningsinstituttet)

Knutsen GM, Tilseth S (1985) Growth, development, and feeding success of Atlantic cod

larvae Gadus morhua related to egg size. Trans Am Fish Soc 114:507–511

Kutti T, Ervik A, Hansen PK (2007) Effects of organic effluents from a salmon farm on a fjord

system. I. Vertical export and dispersal processes. Aquaculture 262:367–381

Lambert Y, Dutil J-D (1997) Can simple condition indices be used to monitor and quantify

seasonal changes in the energy reserves of cod (Gadus morhua)? Can J Fish Aquat Sci

54:104–112

Lambert Y, Dutil J-D (2000) Energetic consequences of reproduction in Atlantic cod (Gadus

morhua) in relation to spawning level of somatic energy reserves. Can J Fish Aquat Sci

57:815–825

Lavens P, Lebegue E, Jaunet H, Brunel A, Dhert P, Sorgeloos P (1999) Effect of dietary

essential fatty acids and vitamins on egg quality in turbot broodstocks. Aquac Int

7:225–240

LeCren ED (1951) The length-weight relationship and seasonal cycle in gonad weight and

condition in the perch (Perca fluviatilis). J Anim Ecol 20: 201–219

Marteinsdottir G, Steinarsson A (1998) Maternal influence on the size and viability of Iceland

cod Gadus morhua eggs and larvae. J Fish Biol 52:1241–1258

Mazorra C, Bruce M, Bell JG, Davie A, Alorend E, Jordan N, Rees J, Papanikos N, Porter M,

Bromage N (2003) Dietary lipid enhancement of broodstock reproductive performance

and egg and larval quality in Atlantic halibut (Hippoglossus hippoglossus). Aquaculture

227:21–33

76

McConnell A, Routledge R, Connors B (2010) Effect of artificial light on marine invertebrate

and fish abundance in an area of salmon farming. Mar Ecol Prog Ser 419: 147–156

Meeren T van der, Ivannikov VP (2006) Seasonal shift in spawning of Atlantic cod (Gadus

morhua L.) by photoperiod manipulation: egg quality in relation to temperature and

intensive larval rearing. Aquac Res 37:898–913

Meeren T van der, Karlsen Ø, Liebig PL, Mangor-Jensen A (2014) Copepod production in a

saltwater pond system: A reliable method for achievement of natural prey in start-

feeding of marine fish larvae. Aquac Eng 62:17–27

Meier S, Mjøs SA, Joensen H, Grahl-Nielsen O (2006) Validation of a one-step

extraction/methylation method for determination of fatty acids and cholesterol in

marine tissues. J Chromatogr A 1104:291–8

Morgan MJ, Wilson CE, Crim LW (1999) The effect of stress on reproduction in Atlantic cod. J

Fish Biol 54:477–488

Naas KE, Næss T, Harboe T (1992) Enhanced first feeding of halibut larvae (Hippoglossus

hippoglossus L.) in green water. Aquaculture 105:143–156

Neat FC, Wright PJ, Zuur AF, Gibb IM, Gibb FM, Tulett D, Righton DA, Turner RJ (2006)

Residency and depth movements of a coastal group of Atlantic cod (Gadus morhua L.).

Mar Biol 148:643–654

Olsen SA, Hansen PK, Givskud H, Ervik A, Samuelsen OB (2014) Changes in fatty acid

composition and stable isotope signature of Atlantic cod (Gadus morhua) in response

to laboratory dietary shifts. Aquaculture 435:277–285

Olsen EM, Heupel MR, Simpfendorfer CA, Moland E (2012) Harvest selection on Atlantic cod

behavioral traits: implications for spatial management. Ecol Evol 2:1549–1562

Olsen EM, Moland E (2011) Fitness landscape of Atlantic cod shaped by harvest selection and

natural selection. Evol Ecol 25:695–710

Otterå H, Agnalt A, Jorstad K (2006) Differences in spawning time of captive Atlantic cod from

four regions of Norway, kept under identical conditions. ICES J Mar Sci 63:216–223

Otterå H, Skilbrei OT (2014) Possible influence of salmon farming on long-term resident

behaviour of wild saithe (Pollachius virens L.). ICES J Mar Sci 71:2484–2493

Patten MA, Kelly JF (2010) Habitat selection and the perceptual trap. Ecol Appl 20:2148–2156

Porter MJ., Duncan N., Mitchell D, Bromagea N. (1999) The use of cage lighting to reduce

plasma melatonin in Atlantic salmon (Salmo salar) and its effects on the inhibition of

grilsing. Aquaculture 176:237–244

R Core Team (2016) R: A language and environment for statistical computing.

77

Rahman MM, Gasparini C, Turchini GM, Evans JP (2014) Experimental reduction in dietary

omega-3 polyunsaturated fatty acids depresses sperm competitiveness. Biol Lett 10

Rainuzzo JR, Reitan KI, Olsen Y (1997) The significance of lipids at early stages of marine fish:

a review. Aquaculture 155:103–115

Reitan KI, Rainuzzo JR, Olsen Y (1994) Influence of lipid composition of live feed on growth,

survival and pigmentation of turbot larvae. Aquac Int 2:33–48

Remeš V (2003) Effects of exotic habitat on nesting success, territory density, and settlement

patterns in the blackcap (Sylvia atricapilla). Conserv Biol 17:1127–1133

Robertson BA, Hutto RL (2006) A framework for understanding ecological traps and an

evaluation of existing evidence. Ecology 87:1075–1085

Robichaud D, Rose GA (2004) Migratory behaviour and range in Atlantic cod: inference from

a century of tagging. Fish Fish 5:185–214Salze G, Tocher DR, Roy WJ, Robertson DA

(2005) Egg quality determinants in cod (Gadus morhua L.): egg performance and lipids

in eggs from farmed and wild broodstock. Aquac Res 36:1488–1499

Sargent J, Bell G, McEvoy L, Tocher D, Estevez A (1999) Recent developments in the essential

fatty acid nutrition of fish. Aquaculture 177:191–199

Sargent J, McEvoy L, Estevez A, Bell G, Bell M, Henderson J, Tocher D (1999) Lipid nutrition of

marine fish during early development: current status and future directions.

Aquaculture 179:217–229

Schneider CA, Rasband WS, Eliceiri KW (2012) NIH Image to ImageJ: 25 years of image

analysis. Nat Methods 9:671–675

Sclafani M, Taggart CT, Thompson KR (1993) Condition, buoyancy and the distribution of

larval fish: implications for vertical migration and retention. J Plankton Res 15:413–435

Sherley RB, Ludynia K, Dyer BM, Lamont T, Makhado AB, Roux J-P, Scales KL, Underhill LG,

Votier SC (2017) Metapopulation tracking juvenile penguins reveals an ecosystem-

wide ecological trap. Curr Biol 27:563–568

Shima JS, Swearer SE (2010) The legacy of dispersal: larval experience shapes persistence

later in the life of a reef fish. J Anim Ecol 79:1308–1314

Skilbrei OT, Otterå H (2016) Vertical distribution of saithe (Pollachius virens) aggregating

around fish farms. ICES J Mar Sci 73:1186–1195

Skjæraasen JE, Nilsen T, Kjesbu OS (2006) Timing and determination of potential fecundity in

Atlantic cod (Gadus morhua). Can J Fish Aquat Sci 63:310–320

Vallin L, Nissling A (2000) Maternal effects on egg size and egg buoyancy of Baltic cod, Gadus

morhua: implications for stock structure effects on recruitment. Fish Res 49:21–37

78

Svåsand T (1990) Comparisons of migration patterns of wild and recaptured reared coastal

cod, Gadus morhua L., released in a small fjord in western Norway. Aquac Res 21:491–

495

Uglem I, Bjørn PA, Dale T, Kerwath S, Økland F, Nilsen R, Aas K, Fleming I, McKinley RS (2008)

Movements and spatiotemporal distribution of escaped farmed and local wild Atlantic

cod (Gadus morhua L.). Aquac Res 39:158–170

Uglem I, Dempster T, Bjørn P-A, Sánchez-Jerez P, Økland F (2009) High connectivity of salmon

farms revealed by aggregation, residence and repeated movements of wild fish among

farms. Mar Ecol Prog Ser 384:251–260

Wasta Z, Mjøs SA (2013) A database of chromatographic properties and mass spectra of fatty

acid methyl esters from omega-3 products. J Chromatogr A 1299:94–102

Weldon AJ, Haddad NM (2005) The effects of patch shape on indigo buntings: evidence for

an ecological trap. Ecology 86:1422–1431

White CA, Bannister RJ, Dworjanyn SA, Husa V, Nichols PD, Kutti T, Dempster T (2017a)

Consumption of aquaculture waste affects the fatty acid metabolism of a benthic

invertebrate. Sci Total Environ 586:1170–1181

White CA, Dworjanyn SA, Nichols PD, Mos B, Dempster T (2016) Future aquafeeds may

compromise reproductive fitness in a marine invertebrate. Mar Environ Res 122:67–75

White CA, Nichols PD, Ross DJ, Dempster T (2017b) Dispersal and assimilation of an

aquaculture waste subsidy in a low productivity coastal environment. Mar Pollut Bull

120:309-321

White CA, Woodcock SH, Bannister RJ, Nichols PD (2017) Terrestrial fatty acids as tracers of

finfish aquaculture waste in the marine environment. Rev Aquac. Early view:

https://doi.org/10.1111/raq.12230

79

CHAPTER FOUR: NATIVE PREDATOR PREVENTS AN INVADER FROM EXPLOITING

FOOD-RICH HABITAT

ABSTRACT

Biodiverse ecosystems may be inherently resistant to invasion, but environmental change can

facilitate invasion by disturbing natural communities and providing resources that are

underutilised by native species. In such cases, abundant native predators may be crucial in

limiting invasive population growth. We studied invasive seastars feeding at shellfish farms

together with a larger native predatory seastar, to examine whether farms are reproductive

hotspots for the invader or whether the native predator causes farms to function as ecological

traps. Invaders were not more abundant at farms, despite individuals residing at farms having

higher body condition and reproductive investment than those on reference habitats. The

native predator was 25x more abundant at farms than surrounding habitats and was

negatively correlated with invader abundance. We also observed several native-on-invasive

predation events and a conspicuous absence of small invaders at farms despite high larval

recruitment to farms, consistent with biotic control by the native predator. In a choice

experiment, invaders were strongly attracted to mussels, but the presence of a single native

predator nullified the attractive effect, indicating that the invader recognises and avoids the

native predator. Attractive habitats add spatial structure and may leave invasive populations

more susceptible to top-down control. We provide evidence that the presence of a native

predator prevents an invader from exploiting an attractive rich food resource.

INTRODUCTION

Invasive species are economically costly and an important contributor to biodiversity loss

worldwide (Clavero and García-Berthou, 2005; Pimentel et al., 2001). Species introductions are

increasing in frequency as a result of human transport, with the likelihood of a newly

introduced species establishing and becoming invasive depending on multiple interacting

factors including environmental conditions, ecosystem characteristics and species traits

(Carlton, 1996; Keane and Crawley, 2002; Kennedy et al., 2002; Kimbro et al., 2013; Papacostas

et al., 2017).

80

Intact, biodiverse ecosystems sometimes have inherent biotic invasion resistance (Elton,

1958), conferred by low niche vacancy rates and strong competition for resources (Kennedy et

al., 2002), as well as a higher likelihood of the new arrival encountering predators or other

enemy species (Keane and Crawley, 2002; Kimbro et al., 2013). However, ecosystem-level

invasion resistance is being eroded by large scale habitat disturbance and loss of biodiversity

associated with human-induced rapid environmental change (HIREC: Vitousek et al. 1997;

Sanderson et al. 2002; Halpern et al. 2008; Vörösmarty et al. 2010; Sih et al. 2011). Non-native

species can benefit from the impacts of HIREC by taking advantage of man-made or degraded

habitats that are low in biodiversity and have resources that are underutilised by native

competitors or predators (Byers, 2002; Didham et al., 2005; MacDougall and Turkington, 2005;

Mack et al., 2000). In the early stages of invasion, or at range limits, these refuge habitats can

become invasion hubs: satellite populations that facilitate persistence and expansion (Carlton,

1996; Letnic et al., 2014; Russell et al., 2011; Suarez et al., 2001; With, 2002). Later, they may

function as reproductive hotspots for both native and invasive species, for example by

permitting high local population densities that overcome Allee effects (Inglis & Gust 2003; Ling

et al. 2012).

A single abundant predator species can exert considerable top-down control and play an

important role in limiting the eventual population size and distribution of an invader,

essentially providing invasion resilience when invasion resistance has failed (DeRivera et al.,

2005; Kimbro et al., 2013; Papacostas et al., 2017). Predator-mediated biotic control may be

particularly effective in cases where the invader does not recognise or respond appropriately

to the native predator; a scenario that is most likely to occur when the invader has no

evolutionary history with the native predator or its relatives (Cox and Lima, 2006; Pintor and

Byers, 2015; Sih et al., 2010). Such conditions favour the formation of an evolutionary trap:

where an individual chooses a habitat (ecological trap: Robertson & Hutto 2006; Hale &

Swearer 2016) or behaviour that leads to poor fitness outcomes (Robertson et al., 2013). Traps

usually arise as a result of environmental change outpacing the evolution of behaviour, and

account for some of the disastrous effects of invasive predators on naïve native prey (e.g.

juvenile fish avoid reefs with a resident native predator but not those with a resident invasive

predator: Benkwitt 2017). Most work has focused on mitigating traps that affect native

species, but evolutionary traps also have potential as tools in the management of non-native

species (Letnic et al., 2015; Robertson et al., 2017). One potential approach is to increase

mortality rates at attractive sites to create attractive sinks – Letnic et al. (2015) recently

demonstrated that excluding invasive toads from water sources converts invasion hubs into

81

ecological traps. Native predators may also contribute to the formation of ecological traps for

non-native species, but such traps have not yet been empirically demonstrated.

In this study, we employ the ecological trap framework to study interactions between native

and invasive seastars at shellfish farms in southern Australia. To demonstrate an ecological

trap, it is necessary that individuals either prefer or fail to avoid the putative trap habitat when

better options are available, and that fitness is reduced in the putative trap habitat (Robertson

and Hutto, 2006). Shellfish farms are an excellent model for a range of anthropogenic marine

habitats worldwide that are simultaneously highly disturbed and rich in food for benthic

predators (D’Amours et al., 2008; Inglis and Gust, 2003; McKindsey et al., 2011). This

combination of traits may cause such habitats to be highly vulnerable to exploitation by

invasive species (e.g. Ling et al. 2012). The invasive seastar Asterias amurensis (Asterias

hereafter) was introduced to Tasmania during the 1980s (Grannum et al., 1996), and mainland

Australia in 1995 (Parry and Cohen, 2001), with the mainland population reaching 75 million by

2000 (Parry and Cohen, 2001). The invader has decimated naïve bivalve populations (Hutson et

al., 2005; Ross et al., 2003), and has few predators in its new environment, but there have

been reports of predation by a native predator, the eleven-arm seastar Coscinasterias

muricata (Coscinasterias hereafter) (Byrne et al., 2013; Parry, 2017). The trophic subsidy

provided by shellfish farms attracts aggregations of both species, but the ability of the invader

to benefit from the high food availability will depend on competition and predation risk from

the native predator (Appendix 4.1). If the native predator does not impede the ability of the

invader to access the food resource, then such habitats are likely to act as population sources

(Letnic et al., 2014; Ling et al., 2012). Alternatively, if the invader is attracted to these sites but

then suffers high predation rates or competitive exclusion, there is potential for an ecological

trap to arise (Hale and Swearer, 2016). We combine data on population distribution, individual

body condition, reproductive investment and habitat preference to assess the likely role of

two coastal mussel farms for populations of the native and invasive seastars. In doing so, we

provide evidence indicating that an abundant native predatory seastar (a) prevents an

established invader from accessing a disturbed but food-rich habitat.

82

METHODS

Study sites and focal species

Port Phillip Bay is a 1930 km2 semi-enclosed marine water body in south-eastern Australia

subject to multiple impacts from invasive species and urban inputs (Hewitt et al., 2004;

Sampson et al., 2014). We studied native (Coscinasterias) and invasive (Asterias) seastar

populations at two shellfish aquaculture reserves 12 km apart in the south-western part of the

bay: Clifton Springs and Grassy Point. These leases primarily produce blue mussels (Mytilus

galloprovincialis) grown on suspended longlines. Clifton Springs and Grassy Point contain

approximately 81 and 42 ha of active bivalve aquaculture areas, respectively. Active areas

consist of parallel longlines approximately 20 m apart, which are stocked year-round with

short turnaround times between harvesting and restocking. Live mussels (typically 40-60 mm

shell length) were abundant on the seabed throughout the study (pers. obs.), with the natural

substrate entirely replaced by mussel shell debris within active areas. Both reserves are

located on soft sediment between 8-14 m depth, with constant mixing and exchange of water

by tidal cycling and wind-generated currents. Clifton Springs is the more sheltered of the two

locations and is dominated by silty substrate, while Grassy Point is characterised by sandy

substrate and lower turbidity. The natural benthos within farms and surrounding areas at both

locations is dominated by Halophila australis seagrass beds, with paper mussels (Electroma

georgiana) attached to available seagrass or macroalgae. Caulerpa spp. beds are common at

Grassy Point. Seastars are the predominant benthic predators within and around both

reserves.

Coscinasterias muricata (Asteroidea: Asteriidae) are a large (armspan commonly 30-45 cm)

subtidal reef dwelling invertivore. Bivalves are preferentially consumed, especially the mussels

that are abundant on reefs and marine infrastructure in southern Australia, but will

opportunistically consume a variety of invertebrate prey (Day et al., 1995; Parry, 2017).

Asterias amurensis (Asteroidea: Asteriidae) are smaller (armspan commonly 20-30 cm) but

highly mobile invertivores, and are typically more abundant on soft sediment habitats where

they consume primarily bivalve prey (e.g. clams, scallops and mussels) (Ross et al., 2003; Ross

and Johnson, 2002).

83

Experiment 1: Effect of shellfish farming on population density of native and invasive

seastars

We surveyed Asterias and Coscinasterias population densities along diver transects inside and

outside the two aquaculture reserves: Clifton Springs on six occasions between July 2014 and

June 2016, and Grassy Point on seven occasions between July 2014 and June 2015. On each

sampling occasion, we surveyed three replicate 2 x 25 m transects (extended to 2 x 50 where

<10 seastars were recorded) inside and outside the reserve boundaries (a total of six transects

per reserve per sampling date). ‘Inside’ transects were haphazardly placed within active

mussel longline areas where live mussels were present on the seabed. ‘Outside’ transects were

haphazardly placed on natural soft sediment habitat within a ~6 km2 area 1-2 km outside the

reserve boundary. Inside and Outside habitats were comparable in terms of depth (10-13 m),

water flow and sediment characteristics. Each transect consisted of the diver laying out a

transect centreline on a predetermined random heading and counting all seastars with their

central point within 1 m of the transect centreline. On rare occasions where a transect strayed

beyond the area of fallen mussel debris (typically >20 m from the outermost longline) we

altered the heading of the remainder of transect. We placed transects at least 50 m apart to

ensure that no seastars were sampled twice.

To test for effects of shellfish farms on seastar population density, we constructed generalised

linear mixed models (GLMMs) for each species using the glmmTMB package for R (Brooks et

al., 2017; R Core Team, 2017). A negative binomial model family was specified as there was a

higher frequency of low or zero counts than would be contained within a Poisson distribution.

We included Habitat (Inside or Outside) and Location (Clifton Springs or Grassy Point) as fixed

effects and Date as a random intercept term to account for temporal variation. As there was a

very strong effect of Location (largely driven by the rarity of Asterias at Grassy Point), we then

split the dataset by Location and re-ran the models without the Location term, reporting both

sets of results. In each case, we tested the significance of the Habitat effect by comparing the

fit of models with and without the Habitat term using a likelihood ratio test. To test for a

potential effect of Coscinasterias abundance on Asterias abundance, we fitted a negative

binomial GLMM in which Asterias density is regressed against Coscinasterias density, with

Habitat and Location also as fixed effects and Date as a random intercept term.

84

Experiment 2: Effect of shellfish farming on size and condition of native and invasive seastars

To track any differences in size between habitats, we measured the arm span of all seastars

encountered on each transect (or the first 30 of each species) and took the mean arm span for

each species on each transect.

We compared condition metrics between habitats by collecting up to 20 individuals per

species per farm per sampling occasion (collected on transects and supplemented with off-

transect collections when necessary). When fewer than 20 individuals were available in one

habitat on a given day, we limited our sample size in the other habitat to avoid temporal bias

caused by seasonal dynamics in condition metrics. Drained and gutted carcass weight provided

a measure of size that was largely independent of body condition, as seastars store nutrients

primarily in the coelomic fluid and digestive organs. The relative size of the pyloric caeca

(digestive organs that increase in size with feeding activity) provided a measure of body

condition and an indirect measure of food availability inside and outside the aquaculture

reserves (Oudejans et al., 1979; Xu and Barker, 1990). We calculated the pyloric caeca index

relative to somatic weight: PCI = PCW/CW, where PC = pyloric caeca wet weight and CW =

gutted carcass weight drained of coelomic fluid. We considered reproductive condition in

terms of two correlated metrics: relative reproductive investment (gonadosomatic index: GSI =

GW/CW, where GW = gonad weight) and potential fecundity (gonad size). A pilot study

demonstrated that PCI and GSI can be accurately estimated from a single limb (whole-body vs.

limb-only linear R2 with n = 20 for Asterias PCI = 0.79, GSI = 0.90; with n = 23 for Coscinasterias

PCI = 0.88, GSI = 0.78). This permitted the release of the native Coscinasterias after the

removal of a single limb. As Asterias is a declared noxious aquatic species, entire Asterias

individuals were taken, but only a single limb dissected. Asterias are sufficiently abundant and

mobile that the removal of up to 20 individuals per farm per sampling occasion likely had a

negligible effect on numbers in later surveys (Andrews et al., 1996). Limbs were cut open

longitudinally, drained of coelomic fluid and weighed. The pyloric caeca and gonads were then

removed and weighed separately. Individuals were sexed based on the colour of the gonads, a

method that was validated by microscopy during a pilot study (n = 20 per species, 100 %

accuracy).

To test for effects of shellfish farms on size and condition, we fitted linear mixed models with

Habitat as a fixed term and Date and Sex as a random intercept term. We tested the Habitat

85

term as per Experiment 1. Responses were transformed as necessary to improve linearity and

normality: log10 for Asterias armspan and 4th root for PCI and GSI.

Experiment 3: Interactions between native and invasive seastars

Any fitness benefits of high food availability may be offset if there is an elevated predation risk.

We recorded evidence of predation on transects as for population surveys in Experiment 1,

including predation events in progress and sublethal damage that is unlikely to be explained by

asexual reproduction (e.g. damage to arms consistent with crab or fish predation: Ling &

Johnson 2013). Predation may influence the size distribution of a prey species whenever

predation events and/or success rates depend on the relative size of the predator and prey.

Coscinasterias predation on Asterias is strongly size dependent, with a higher likelihood of prey

escaping when relative sizes are close (Parry 2017). Cannibalism of small Asterias by larger

conspecifics also occurs (pers. obs.). In light of this, we compared the size distributions of

Asterias inside and outside farms for evidence of size-dependent predation within and/or

avoidance of the farm environment.

Given evidence that the native seastar does prey on the invasive seastar, we devised a

laboratory-based choice experiment to test the strength of the invader’s attraction to fallen

mussels, and whether this attraction is moderated by the presence of either conspecifics or

native predators. We conducted the experiment in a 72 L flume tank (120 x 30 x 20 cm),

housed within a recirculating seawater facility containing ca. 12000 L of continuously filtered

and UV-treated seawater. A flow rate of 4 L min-1 was divided between the two chambers at

the head of the tank, with both chambers opening into an undivided downstream portion of

the tank containing the Asterias subject. We compared rates of attraction to four treatments:

a seawater control with no predatory nor prey cues (‘SW’), five thawed mussels (40-60 mm

shell length) placed in a perforated compartment at the upstream end of the chamber (‘M’),

five thawed mussels with three conspecifics in the chamber with the mussels (‘AM’), and five

thawed mussels with a large Coscinasterias individual also present in the chamber (‘CM’). All

seastars placed in the treatment chamber for AM and CM treatments remained by the end of

the trial. We applied treatments singly, with the cue randomly assigned to one of the two

chambers (all treatments were tested relative to a seawater control). We left the subjects

overnight (16 hrs), and considered them to be attracted to the habitat cue if they were present

inside the corresponding chamber at the completion of the trial, providing a binary outcome.

86

The flume tank was cleaned and flushed between trials. Subjects were randomly selected from

a large captive population (>200) collected from the wild within 2 months of the trial, fasted

for 2-4 days prior to the trial, and randomly assigned to a treatment. The size range of subjects

was 4-17 cm in every treatment group (mean 11 ± 0.2 cm).

We tested for differential attraction to treatments by fitting a binomial generalised linear

model using the glm function in R, before conducting post-hoc pair-wise X2 tests of proportions

using the prop.test function.

RESULTS

Experiment 1: Effect of shellfish farming on population density of native and invasive

seastars

The invasive seastar Asterias was common on silty substrate and mussel shell debris both

inside and outside the farm boundary at Clifton Springs, with very high abundance inside

throughout winter 2014 (Fig. 4.1). However, Asterias population density was highly variable,

with no significant effect of the farm habitat over the full study duration (2014-2016) (Table

4.1). Asterias was nearly absent from Grassy Point, being recorded on a single transect only

inside the Grassy Point boundary in winter 2014. (Fig. 4.1).

The native seastar Coscinasterias showed contrasting habitat preference (Fig. 4.1). At Grassy

Point, Coscinasterias was common inside and outside the Grassy Point farm boundary, with a

14x higher population density inside (Table 4.1). They were also common inside the boundary

at Clifton Springs, but absent on our transects outside the Clifton Springs boundary (Fig. 4.1).

Across the two locations, Coscinasterias occurred at a higher density (mean 25x) inside farms

on every sampling occasion between 2014-2016.

If both species were distributed according to availability of a common food source, we would

expect a positive correlation between the abundance of the two species. Instead, our

population density data reveal that Asterias density is negatively dependent on Coscinasterias

density (negative binomial mixed effects model with Coscinasterias density, Habitat and

Location as fixed effects and Date as random effect: X2 = 4.1, p = 0.04).

87

Experiment 2: Effect of shellfish farming on size and condition of native and invasive seastars

Both seastar species were larger on average inside farm boundaries. Due to the rarity of

Asterias outside Grassy Point and the rarity of Coscinasterias outside Clifton Springs, we

compared the size distribution of both species inside and outside a single farm (Asterias:

Clifton Springs, Coscinasterias: Grassy Point).

Figure 4.1. Population density of Asterias amurensis (plots A and B) and Coscinasterias

muricata (plots C and D) inside and outside the Clifton Springs and Grassy Point Fisheries

Aquaculture Reserves. Boxes denote median, lower (25 %) and upper (75 %) quartiles,

whiskers denote 1.5x interquartile range.

Asterias inside the farm had a 25 % wider arm span than their counterparts outside,

corresponding to a 92 % increase in gutted weight (Table 4.1). This effect was consistent across

88

all but one sampling date (Fig. 4.2). The difference in mean arm span of Asterias was driven by

a contraction at the lower end of the size distribution inside the Clifton Springs farm, rather

than the presence of larger individuals (Fig. 4.4). All body and reproductive condition metrics

that we assessed in Asterias were higher inside the farm than outside: drained weight was 77

% higher, gonad weight 57 % higher, PCI 26 % higher and GSI 57 % higher (Table 4.1). The PCI

effect was highly consistent over the study duration, while differences in GSI varied over time

(Fig. 4.3).

Figure 4.2. Armspan (A) and gutted weight (B) of Asterias amurensis inside and outside the

Clifton Springs Fisheries Aquaculture Reserve.

Likewise, Coscinasterias inside the farm had a 7 % wider arm span and 42 % heavier gutted

weight than those outside (Table 4.1). Individuals inside the farm also had higher body

condition (11 %) but there was no difference in either gonad size or GSI (Table 4.1).

Experiment 3: Interactions between native and invasive seastars

We observed three native-on-invasive seastar predation events on our transects inside the

Clifton Springs farm (total 3325 m2 surveyed, 160 Asterias scored), with none observed on the

same number of transects outside (3900 m2 surveyed, 126 Asterias scored). We also noted a

further ten predation events off-transect, as well as five skeletons of recently-consumed

89

Asterias, all inside the farm boundary. We observed one cannibalism event (a large Asterias

consuming a smaller conspecific), also on a transect within the Clifton Springs farm.

We found strong effects of prey and predator cues on habitat selection decisions by the

invasive Asterias in captive behavioural trials (n = 40 per treatment, p <0.0001) (Fig. 4.5). The

mussel cue (M) attracted Asterias subjects at a significantly higher rate than the seawater

control (SW) (93 vs. 28 % respectively, X2 = 32, p <0.0001). Mussels remained more attractive

than seawater when three Asterias conspecifics were present feeding on the mussels (AM) (88

%, X2 = 27, p <0.0001), and there was no difference between rates of attraction to M or AM

cues (93 % vs. 88 % respectively, X2 = 0.1, p = 0.7). In contrast, the presence of a single

Coscinasterias individual feeding on the mussels (CM) resulted in a dramatic decline in

attraction rates relative to both the M (28 vs. 93 % respectively, X2 = 32, p <0.0001) and AM

cues (28 vs. 88 % respectively, X2 = 27, p <0.0001). 2/11 individuals that selected the chamber

with the CM cue were predated by the Coscinasterias (one fully, one partially consumed

overnight). Asterias individuals that avoided the chamber with the CM cue exhibited a variety

of behaviours, such as moving immediately to the cue-free (clean seawater) chamber and

remaining there for the duration of the study, circling the downstream portion of the tank, or

attempting to reach the mussels before evading a predation attempt by Coscinasterias. We

found no evidence that the size of the subject influenced their response to the presence of

either conspecifics (z38 = 1.1, p = 0.3) or Coscinasterias (z38 = 1.0, p = 0.3). The size differential

between the subject and the three conspecifics in the AM treatment (natural log of subject

arm span / mean conspecific arm span) also did not influence the likelihood of attraction to

the mussels (z38 = 0.6, p = 0.5).

DISCUSSION

In disturbed ecosystems, abundant native predators may be an important line of defence

against invaders, limiting the probability of establishment and subsequent population growth.

In Experiment 1 we showed that a native predator occurs at 25x higher abundance inside two

shellfish farms relative to reference habitats, with the shellfish farm accounting for almost the

entire population in an area with non-preferred silty substrate. In contrast, the invasive species

was not significantly more abundant inside the farms and was negatively associated with the

native predator. Experiment 2 revealed that invasive seastars residing inside one farm

possessed considerably higher body condition and reproductive condition metrics than

90

conspecifics at reference sites, supporting the hypothesis that this shellfish farm may be a

reproductive hotspot for the invader. When contrasted with the lack of a clear effect on

abundance in Experiment 1, this suggests a failure or inability to take advantage of a high

quality food resource. Accordingly, in Experiment 3 we provided observational field evidence

of elevated predation risk for the invasive species inside shellfish farms, and found that in

captivity, the invader recognises and avoids the native predator even if it means forgoing a

favoured prey item. Taken together, these findings raise the possibility that the presence of

the native predator is preventing the invader from accessing a highly disturbed food-rich

habitat.

Figure 4.3. Condition metrics for Asterias amurensis inside and outside the Clifton Springs

Fisheries Aquaculture Reserve: drained carcass weight (A), gonad weight (B), gonadosomatic

index (C), and pyloric caeca index (D).

In the absence of abundant native predators, man-made habitats such as shellfish farms, fish

farms and mussel-fouled infrastructure can function as reproductive hotspots for invasive

91

Asterias amurensis populations, simultaneously supporting orders-of-magnitude higher

population density and elevated individual reproductive output (Ling et al. 2012). Within their

native range, Asterias species also have highly elevated population densities at shellfish farms,

perhaps indicating release from native predators such as crabs and sunstars (e.g. Olaso Toca

1979, 1982; Saranchova & Kulakovskii 1982, cited in McKindsey et al. 2011; D’Amours et al.

2008). The present study indicates that top-down control may alter the role played by these

food-rich habitats.

Figure 4.4. Density plot of Asterias amurensis size distribution inside and outside the farm

boundary at Clifton Springs Aquaculture Fisheries Reserve. N = 165 (inside) + 107 (outside).

Effective top-down control of non-native animals by native predators has been reported in

numerous marine, freshwater and terrestrial systems (e.g. DeRivera et al. 2005; Cheng & Hovel

2010; Tetzlaff et al. 2011; Freed & Leisnham 2014), but to our knowledge, this is the first clear

evidence of a native predator preventing invaders from accessing a high value food resource.

Asterias is thought to be primarily food limited in its new range: growth rates declined during

the first three years following its introduction to mainland Australia, coinciding with increasing

population densities and large declines in prey abundance (Parry and Cohen, 2001). Our body

condition and reproductive investment data show elevated condition inside farms, indicating

that food limitation still exists outside farm habitats. This means that regardless of the

frequency of predation events, if Coscinasterias discourage Asterias from accessing farms, the

92

native predator may exert indirect biotic control on the invasive population by maintaining

food limitation and thus a lower equilibrium population size.

Figure 4.5. Effect of prey, conspecifics and predators on habitat selection decisions by the

invasive seastar Asterias amurensis in laboratory trials. Treatments: SW = seawater control, M

= mussels, AM = mussels + Asterias conspecifics, CM = mussels + Coscinasterias. Bars show the

proportion of subjects that are attracted to the cue chamber (‘Y’) or not (‘N’) according to the

treatment. Individuals that were not attracted to the cue chamber either remained in the

downstream portion or moved into the control chamber. N = 160 (40 per treatment). Matching

letters indicate proportions that do not differ statistically.

Our data also raise the possibility that juveniles of the non-native species may be falling into a

predation-mediated ecological trap at shellfish farms. Asterias has a long pelagic larval

duration (79-112 days: Bruce et al., 1995), and in Australia, planktonic Asterias recruit in large

numbers directly to mussel lines and other suspended structures (e.g. Dommisse & Hough

2004). This creates the conditions for an ecological trap affecting juvenile Asterias, as juveniles

living on mussel lines are (a) unlikely to reach sexual maturity before the mussel lines are

harvested, and (b) likely to experience high predation risk if they fall to a benthic environment

populated by >2000 Coscinasterias ha-1 (this study). The truncated size distribution for Asterias

on the seabed under mussel farms is consistent the hypothesis that juvenile Asterias falling to

the seabed from mussel lines either leave the area or are consumed, although we cannot rule

out individuals leaving or avoiding the farm environment due to innate preferences (such as

93

for soft sediment rather than biofouled mussel shell substrate). The risk of predation and

potential for an ecological trap may be increased if juveniles do not have sufficiently strong

predator avoidance behaviour (Cox and Lima, 2006; Sih et al., 2010); in our habitat choice

experiment, small individuals were no more likely to avoid the native predator than large

individuals, despite being at higher risk of predation than larger conspecifics (size-dependent

predation success: Parry 2017).

As adults, the potential for an ecological trap at shellfish farms depends on whether individuals

are attracted to the farm (i.e. the strength of their preference for bivalve prey versus their

avoidance response to the native predator) and the net fitness effects of any preference for

the farm (i.e. high food availability versus high predation risk). Our choice experiment found

that Asterias were strongly attracted to bivalves, but only in the absence of the native

predators, and while actual predation rates are almost certainly higher inside the farms (this

study; Parry 2017), so is reproductive output (this study; Ling et al. 2012). Further work is

needed to calculate the net effect of these opposing factors, and to test whether Asterias

individuals are able to make informed habitat selection decisions at ecologically-relevant

distances. The high densities of mussels and predators at farms provide a relatively strong

scent trail, and all else being equal, even random movement of seastars will tend to result in

higher densities in preferred habitats if individuals move less after arrival (Benhamou 2011;

see also elevated density of Coscinasterias inside farms, this study).

Invading Asterias benefitted from the naivety of Australian bivalves (Hutson et al., 2005; Ross

et al., 2003), but to date there has been little evidence that Asterias responds naïvely to novel

predators, despite theoretical expectations arising from a lack of shared evolutionary history

(Cox and Lima, 2006; Sih et al., 2010). Behavioural data on predator-prey interactions in the

early days of invasion are not available, but the establishment and expansion of Asterias in

Australia may have been assisted by pre-existing predator cues evolved in response to

functionally similar predators in its native range (Sih et al., 2010). In particular, the sun star

Solaster paxillatus and king crab Paralithodes camtschaticus of the northern Pacific region are

notably similar in form and functional role to the major invertebrate predators in its new

range: the seastar Coscinasterias and the spider crab Leptomithrax gaimardii (Byrne et al.,

2013; Ling and Johnson, 2013), and may emit similar sensory cues.

Little is known about how the Asterias invasion affected native predators. A meta-analysis by

Pintor and Byers (2015) found that simultaneous access to native and non-native prey was

associated with increased native predator abundance, but exceptions are common where

native and non-native species rely on a shared food resource or where abundant non-native

94

species drive severe ecological change (e.g. Pothoven et al. 2001; Suarez & Case 2002). Here,

the invader competes with the native species for mussel prey, but localised high densities of

mussels, such as those on rocky reefs, shellfish farms and marine infrastructure may suit the

native species by improving their ability to exclude Asterias from the food resource, as well as

supplement the mussel diet with opportunistic intraguild predation of the invader.

Eradication of this invader is not feasible, but any top-down pressure may have ecological

benefits by partially releasing vulnerable prey populations from predation and reducing the

likelihood of new populations establishing (DeRivera et al., 2005; Letnic et al., 2009; Mack et

al., 2000). Conversely however, in cases where a non-native prey species provides a large

trophic subsidy for native predators, native predators can become overabundant and severely

impact their native prey species (Noonburg and Byers, 2005). It is unclear whether the

presence of Asterias at farms is sufficient to maintain locally elevated Coscinasterias

populations, but such predation events could conceivably be important during rare periods

when preferred prey (mussels) are scarce.

A range of harvesting methods have been employed to control seastars globally, but none are

cost-effective (Barkhouse et al., 2007). In particular, a lack of spatial population structure on

homogeneous landscapes limits the efficiency of control methods, but attractive habitats such

as shellfish farms and other marine infrastructure add spatial structure and may leave invasive

populations more vulnerable if mortality can be induced at such sites (Bascompte et al., 2002;

Letnic et al., 2015; Russell et al., 2011). In our study, biotic control by the native predator

apparently limits the ability of the invader to exploit a food-rich resource, and we provide

evidence that these habitats may function as ecological traps for the invader by combining

attractive cues from bivalve prey with a high predation risk from the native predator. In such a

case, the native predator provides a potentially valuable ecosystem service by suppressing

invasive populations. It may be feasible to improve the impact of this service by protecting and

augmenting predator populations in key areas such as localised sites of new introductions

(Parry, 2017); an approach has been used with some success in other invaded marine

environments (Atalah et al., 2016, 2013).

ACKNOWLEDGMENTS

This work was funded by a Holsworth Wildlife Research Endowment grant awarded to LB.

Lance Wiffen provided access to aquaculture leases, while Chris Taylor, Simon Reeves, Emily

95

Fobert, Dean Chamberlain, Jack O'Connor, Ben Cleveland, Kevin Jensen, Kevin Menzies, Oliver

Thomas and Rod Watson (Victorian Marine Science Consortium) assisted with fieldwork.

96

Table 4.1. Population metrics for native (Coscinasterias muricata) and invasive (Asterias amurensis) seastars inside and outside farms at Grassy Point

and Clifton Springs Aquaculture Fisheries Reserves. The Habitat (Inside vs. Outside) effect is tested by comparison of linear mixed effects model fit

with and without the Habitat term. Data are missing for Asterias inside and outside Grassy Point and Coscinasterias outside Clifton Springs because

too few individuals were present in these habitats. Sample sizes are reported as follows: n = Inside n, Outside n.

Clifton Springs Grassy Point

Inside Outside % n X2 p Inside Outside % n X2 p

Asterias amurensis Population density (ha-

1)a 1300 ± 311 755 ± 92 72 20, 20 1.4 0.23 22 ± 22 0 ± 0 ∞ 18, 18 1.4 0.23

Armspan (cm) 19.5 ± 0.6 15.6 ± 0.8 25 20, 20 21 <0.0001 12.6 ± 12.6 – – 1, 0 – –

Gutted wt (g) 9.0 ± 0.5 4.7 ± 0.3 92 96, 96 71 <0.0001 – – – 0, 0 – –

Drained wt (g) 15.2 ± 0.7 8.6 ± 0.6 77 96, 96 59 <0.0001 – – – 0, 0 – –

PCI 0.24 ± 0.01 0.19 ± 0.01 26 96, 96 16 <0.0001 – – – 0, 0 – –

Gonad wt (g) 3.6 ± 0.5 2.3 ± 0.5 57 96, 96 12 0.0005 – – – 0, 0 – –

GSI 0.22 ± 0.02 0.14 ± 0.03 57 96, 96 22 <0.0001 – – – 0, 0 – –

Coscinasterias muricata

Population density (ha-

1)b 2240 ± 460 0 ± 0 ∞ 20, 20 72 <0.0001 3194 ± 829 233 ± 55 14 18, 18 21 <0.0001

Armspan (cm) 30.1 ± 0.7 – – 20, 0 – – 30.7 ± 0.4 28.6 ± 0.7 7 18, 18 8.8 0.003

Gutted wt (g) 16.5 ± 0.4 – – 0, 0 – – 16.5 ± 0.3 15.5 ± 0.3 6 79, 79 13 0.0003

Drained wt (g) 23.5 ± 0.7 – – 104, 0 – – 21.8 ± 0.6 19.1 ± 0.8 14 79, 79 12 0.0005

PCI 0.21 ± 0.01 – – 104, 0 – – 20 ± 0.01 0.18 ± 0.01 11 79, 79 6.5 0.01

Gonad wt (g) 3.5 ± 0.3 – – 104, 0 – – 1.9 ± 0.2 1.9 ± 0.2 0 79, 79 0.0 1.0

GSI 0.20 ± 0.02 – – 104, 0 – – 0.11 ± 0.01 0.12 ± 0.01 -9 79, 79 0.1 0.8 aFull dataset (Clifton Springs + Grassy Point) with location as fixed term: n = 38, 38, X2 = 1.8, p = 0.18 bFull dataset (Clifton Springs + Grassy Point) with location as fixed term: n = 38, 38, X2 = 43, p <0.0001

97

REFERENCES

Andrews, D., Whayman, G., Edgar, G., 1996. Assessment of optimal trapping techniques to

control densities north Pacific seastars on marine farm leases. Tasmanian Department of

Primary Industries and Fisheries, Hobart.

Atalah, J., Hopkins, G.A., Forrest, B.M., 2013. Augmentative biocontrol in natural marine

habitats: persistence, spread and non-target effects of the sea urchin Evechinus

chloroticus. PLoS One 8, 1–10

Atalah, J., Newcombe, E.M., Zaiko, A., 2016. Biocontrol of fouling pests: effect of diversity,

identity and density of control agents. Mar. Environ. Res. 115, 20–27

Barkhouse, C., Niles, M., Davidson, L.-A., 2007. A literature review of sea star control methods

for bottom and off bottom shellfish cultures. Can. Ind. Rep. Fish. Aquat. Sci. 279, vii + 35p.

Bascompte, J., Possingham, H., Roughgarden, J., 2002. Patchy populations in stochastic

environments: critical number of patches for persistence. Am. Nat. 159, 128–137

Benhamou, S., 2011. Dynamic approach to space and habitat use based on biased random

bridges. PLoS One 6, e14592

Benkwitt, C.E., 2017. Predator effects on reef fish settlement depend on predator origin and

recruit density. Ecology 98, 896–902

Brooks, M.E., Kristensen, K., van Benthem, K.J., Magnusson, A., Berg, C.W., Nielsen, A., Skaug,

H.J., Maechler, M., Bolker, B.M., 2017. Modeling zero-inflated count data with glmmTMB.

bioRxiv 132753. doi:10.1101/132753

Bruce, B.D., Sutton, C.A., Lyne, V.D., 1995. Laboratory and field studies of the larval distribution

and duration of the introduced seastar Asterias amurensis with updated and improved

prediction of the species spread based on a larval dispersal model: final report to

Fisheries Research and Development Corporation, Canberra.

Byers, J.E., 2002. Impact of non-indigenous species on natives enhanced by anthropogenic

alteration of selection regimes. Oikos 97, 449–458

Byrne, M., O’Hara, T.D., Lawrence, J.M., 2013. Asterias amurensis, in: Lawrence, J.M. (Ed.),

Starfish: Biology and Ecology of the Asteroidea. The John Hopkins University Press,

Baltimore, MA, pp. 174–180.

Carlton, J.T., 1996. Pattern, process, and prediction in marine invasion ecology. Biol. Conserv.

78, 97–106.

Cheng, B.S., Hovel, K.A., 2010. Biotic resistance to invasion along an estuarine gradient.

Oecologia 164, 1049–1059

98

Clavero, M., García-Berthou, E., 2005. Invasive species are a leading cause of animal

extinctions. Trends Ecol. Evol. 20, 110

Cox, J.G., Lima, S.L., 2006. Naiveté and an aquatic-terrestrial dichotomy in the effects of

introduced predators. Trends Ecol. Evol. 21, 674–680

D’Amours, O., Archambault, P., McKindsey, C.W., Johnson, L.E., 2008. Local enhancement of

epibenthic macrofauna by aquaculture activities. Mar. Ecol. Prog. Ser. 371, 73–84.

Day, R., Dowell, A., Sant, G., Klemke, J., Shaw, C., 1995. Patchy predation: foraging behaviour

of Coscinasterias calamaria and escape responses of Haliotis rubra. Mar. Freshw. Behav.

Physiol. 26, 11–33

DeRivera, C.E., Ruiz, G.M., Hines, A.H., Jivoff, P., 2005. Biotic resistance to invasion: native

predator limits abundance and distribution of an introduced crab. Ecology 86, 3364–3376

Didham, R.K., Tylianakis, J.M., Hutchison, M.A., Ewers, R.M., Gemmell, N.J., 2005. Are invasive

species the drivers of ecological change? Trends Ecol. Evol. 20, 470–474

Dommisse, M., Hough, D., 2004. Controlling the northern Pacific seastar (Asterias amurensis)

in Australia. Final report for the Australian Government Department of the Environment

and Heritage, Canberra.

Elton, C.S., 1958. The ecology of invasions by animals and plants., The ecology of invasions by

animals and plants. Methuen & Co. Ltd., London.

Freed, T.Z., Leisnham, P.T., 2014. Roles of spatial partitioning, competition, and predation in

the North American invasion of an exotic mosquito. Oecologia 175, 601–611

Grannum, R.K., Murfet, N.B., Ritz, D.A., Turner, E., 1996. The distribution and impact of the

exotic seastar, Asterias amurensis (Lutken) in Tasmania.

Hale, R., Swearer, S.E., 2016. Ecological traps: current evidence and future directions. Proc. R.

Soc. B Biol. Sci. 283, 494–499

Halpern, B.S., Walbridge, S., Selkoe, K.A., Kappel, C. V, Micheli, F., D’Agrosa, C., Bruno, J.F.,

Casey, K.S., Ebert, C., Fox, H.E., Fujita, R., Heinemann, D., Lenihan, H.S., Madin, E.M.P.,

Perry, M.T., Selig, E.R., Spalding, M., Steneck, R., Watson, R., 2008. A global map of

human impact on marine ecosystems. Science 319, 948–952

Hewitt, C.L., Campbell, M.L., Thresher, R.E., Martin, R.B., Boyd, S., Cohen, B.F., Currie, D.R.,

Gomon, M.F., Keough, M.J., Lewis, J.A., Lockett, M.M., Mays, N., McArthur, M.A., O’Hara,

T.D., Poore, G.C.B., Ross, J.D., Storey, M.J., Watson, J.E., Wilson, R.S., 2004. Introduced

and cryptogenic species in Port Phillip Bay, Victoria, Australia. Mar. Biol. 144, 183–202

Hutson, K.S., Ross, D.J., Day, R.W., Ahern, J.J., 2005. Australian scallops do not recognise the

introduced predatory seastar Asterias amurensis. Mar. Ecol. Prog. Ser. 298, 305–309

99

Inglis, G.J., Gust, N., 2003. Potential indirect effects of shellfish culture on the reproductive

success of benthic predators. J. Appl. Ecol. 40, 1077–1089

Keane, R.M., Crawley, M.J., 2002. Exotic plant invasions and the enemy release hypothesis.

Trends Ecol. Evol. 17, 164–170

Kennedy, T.A., Naeem, S., Howe, K.M., Knops, J.M.H., Tilman, D., Reich, P., 2002. Biodiversity

as a barrier to ecological invasion. Nature 417, 636

Kimbro, D.L., Cheng, B.S., Grosholz, E.D., 2013. Biotic resistance in marine environments. Ecol.

Lett. 16, 821–833

Letnic, M., Crowther, M.S., Koch, F., 2009. Does a top-predator provide an endangered rodent

with refuge from an invasive mesopredator? Anim. Conserv. 12, 302–312

Letnic, M., Webb, J.K., Jessop, T.S., Dempster, T., 2015. Restricting access to invasion hubs

enables sustained control of an invasive vertebrate. J. Appl. Ecol. 52, 341–347

Letnic, M., Webb, J.K., Jessop, T.S., Florance, D., Dempster, T., 2014. Artificial water points

facilitate the spread of an invasive vertebrate in arid Australia. J. Appl. Ecol. 51, 795–803

Ling, S.D., Johnson, C.R., 2013. Native spider crab causes high incidence of sub-lethal injury to

the introduced seastar Asterias amurensis, in: Echinoderms in a Changing World:

Proceedings of the 13th International Echinoderm Conference, January 5-9 2009,

University of Tasmania, Hobart. p. 195.

Ling, S.D., Johnson, C.R., Mundy, C.N., Morris, A., Ross, D.J., 2012. Hotspots of exotic free-

spawning sex: man-made environment facilitates success of an invasive seastar. J. Appl.

Ecol. 49, 733–741

MacDougall, A.S., Turkington, R., 2005. Are invasive species the drivers of change or

passengers of change in degraded ecosystems? Ecology 86, 42–55

Mack, R.N., Simberloff, D., Mark Lonsdale, W., Evans, H., Clout, M., Bazzaz, F.A., 2000. Biotic

invasions: causes, epidemiology, global consequences, and control. Ecol. Appl. 10, 689–

710

McKindsey, C.W., Archambault, P., Callier, M.D., Olivier, F., 2011. Influence of suspended and

off-bottom mussel culture on the sea bottom and benthic habitats: a review. Can. J. Zool.

89, 622–646

Noonburg, E.G., Byers, J.E., 2005. More harm than good: when invader vulnerability to

predators enhances impact on native species. Ecology 86, 2555–2560

Olaso Toca, I., 1982. Ecologia de los equinodermos de la Ría de Arosa. Bol. Inst. Esp. Ocean. 7,

4–29

Olaso Toca, I., 1979. Biologia de los equinodermos de la Ría de Arosa. Bol. Inst. Esp. Ocean. 5,

100

81–127

Oudejans, R.C.H.M., van der Sluis, I., van der Plas, A.J., 1979. Changes in the biochemical

composition of the pyloric caeca of female seastars, Asterias rubens, during their annual

reproductive cycle. Mar. Biol. 53, 231–238

Papacostas, K.J., Rielly-Carroll, E.W., Georgian, S.E., Lomg, D.J., Princiotta, S.D., Quattrini, A.M.,

Reuter, K.E., Freestone, A.L., 2017. Biological mechanisms of marine invasions. Mar. Ecol.

Prog. Ser. 565, 251–268

Parry, G.D., 2017. Potential for biocontrol of the exotic starfish, Asterias amurensis, using a

native starfish. Biol. Invasions 19, 2185–2196

Parry, G.D., Cohen, B.F., 2001. The distribution, abundance and population dynamics of the

exotic seastar Asterias amurensis during the first three years of its invasion of Port Phillip

Bay, (incorporating a Report on the Bay Pest Day, 2 April 2000). Marine and Freshwater

Resources Institute, Melbourne.

Pimentel, D., McNair, S., Janecka, J., Wightman, J., Simmonds, C., O’Connell, C., Wong, E.,

Russel, L., Zern, J., Aquino, T., Tsomondo, T., 2001. Economic and environmental threats

of alien plant, animal, and microbe invasions. Agric. Ecosyst. Environ. 84, 1–20

Pintor, L.M., Byers, J.E., 2015. Do native predators benefit from non‐native prey? Ecol. Lett. 18,

1174–1180

Pothoven, S.A., Nalepa, T.F., Schneeberger, P.J., Brandt, S.B., 2001. Changes in diet and body

condition of lake whitefish in southern Lake Michigan associated with changes in

benthos. North Am. J. Fish. Manag. 21, 876–883

R Core Team, 2017. R: A language and environment for statistical computing. Vienna.

Robertson, B.A., Hutto, R.L., 2006. A framework for understanding ecological traps and an

evaluation of existing evidence. Ecology 87, 1075–1085

Robertson, B.A., Ostfeld, R.S., Keesing, F., 2017. Trojan females and Judas goats: evolutionary

traps as tools in wildlife management. Bioscience 67, 983–994

Robertson, B.A., Rehage, J.S., Sih, A., 2013. Ecological novelty and the emergence of

evolutionary traps. Trends Ecol. Evol. 28, 552–560

Ross, D.J., Johnson, C.R., 2002. Impact of introduced seastars Asterias amurensis on

survivorship of juvenile commercial bivalves Fulvia tenuicostata. Mar. Ecol. Prog. Ser.

241, 99–112

Ross, D.J., Johnson, C.R., Hewitt, C.L., 2003. Variability in the impact of an introduced predator

(Asterias amurensis: Asteroidea) on soft-sediment assemblages. J. Exp. Mar. Biol. Ecol.

288, 257–278

101

Russell, B.G., Letnic, M., Fleming, P.J.S., 2011. Managing feral goat impacts by manipulating

their access to water in the rangelands. Rangel. J. 33, 143–152

Sampson, J., Easton, A., Singh, M., 2014. Port Phillip Bay, in: Wolanski, E. (Ed.), Estuaries of

Australia in 2050 and Beyond. Springer, pp. 49–68

Sanderson, E.W., Jaiteh, M., Levy, M.A., Redford, K.H., Wannebo, A. V, Woolmer, G., 2002. The

human footprint and the last of the wild. Bioscience 52, 891–904

Saranchova, O.L., Kulakovskii, É.E., 1982. Effect of environmental salinity on different

developmental stages of sea stars Asterias rubens and mussels Mytilus edulis. Sov. J. Mar.

Biol. 8, 32–36

Sih, A., Bolnick, D.I., Luttbeg, B., Orrock, J.L., Peacor, S.D., Pintor, L.M., Preisser, E., Rehage, J.S.,

Vonesh, J.R., 2010. Predator-prey naïveté, antipredator behavior, and the ecology of

predator invasions. Oikos 119, 610–621

Sih, A., Ferrari, M.C.O., Harris, D.J., 2011. Evolution and behavioural responses to human-

induced rapid environmental change. Evol. Appl. 4, 367–387

Suarez, A. V, Case, T.J., 2002. Bottom-up effects on persistence of a specialist predator: ant

invasions and horned lizards. Ecol. Appl. 12, 291–298

Suarez, A. V, Holway, D.A., Case, T.J., 2001. Patterns of spread in biological invasions

dominated by long-distance jump dispersal: insights from Argentine ants. Proc. Natl.

Acad. Sci. USA 98, 1095–1100

Tetzlaff, J.C., Roth, B.M., Weidel, B.C., Kitchell, J.F., 2011. Predation by native sunfishes

(Centrarchidae) on the invasive crayfish Orconectes rusticus in four northern Wisconsin

lakes. Ecol. Freshw. Fish 20, 133–143

Vitousek, P.M., Mooney, H.A., Lubchenco, J., Melillo, J.M., 1997. Human domination of Earth’s

ecosystems. Science 277, 494–499

Vörösmarty, C.J., McIntyre, P.B., Gessner, M.O., Dudgeon, D., Prusevich, A., Green, P., Glidden,

S., Bunn, S.E., Sullivan, C.A., Liermann, C.R., others, 2010. Global threats to human water

security and river biodiversity. Nature 467, 555–561

With, K.A., 2002. The landscape ecology of invasive spread. Conserv. Biol. 16, 1192–1203

Xu, R.A., Barker, M.F., 1990. Laboratory experiments on the effects of diet on the gonad and

pyloric caeca indices and biochemical composition of tissues of the New Zealand starfish

Sclerasterias mollis (Hutton) (Echinodermata: Asteroidea). J. Exp. Mar. Biol. Ecol. 136, 23–

45

102

CHAPTER FIVE: AN INVASIVE HABITAT-FORMER MITIGATES IMPACTS OF NATIVE

HABITAT LOSS FOR ENDEMIC REEF FISHES

ABSTRACT

Impacts of environmental change on animal populations depend on how individuals respond

to novel or degraded habitats. Animals that make adaptive habitat selection choices can

maximise their fitness in altered landscapes, while maladaptive selection of low quality

habitats can lead to the formation of ‘ecological traps’ with increased mortality or lowered

reproductive success. Habitat-forming invasive species provide evolutionarily unfamiliar

habitat with the potential to create ecological traps. In southern Australia, urchin grazing has

driven a decline in native kelps enabling an invasive kelp (Undaria pinnatifida) to fill the vacant

niche and provide replacement canopy with unknown effects. We assessed the value of this

novel habitat for macroalgae-associated fishes by combining laboratory habitat choice

experiments, an artificial reef experiment, before-after-control-impact field surveys and

estimates of fitness. In captive choice experiments, macroalgae-associated fishes preferred

rock with macroalgal cover to rock without, but did not distinguish between invasive and

native macroalgae. Cryptobenthic reef fishes occurred in higher abundance and diversity on

urchin-grazed reefs with invasive kelp than those without, while the invader did not affect

communities of large or highly mobile fishes. More fish recruited to artificial reefs that were

stocked with kelp compared to barren reefs, with no preference between native and invasive

kelp canopy. Fitness metrics, including stomach contents, body condition, liver and gonad

indices and fecundity in fish collected from invasive kelp habitats were similar to those in fish

from adjacent native kelp patches, indicating that this invader can function as a partial

replacement for lost native habitat-forming species. These findings illustrate how invasive

habitat-formers can play a role in maintaining biodiversity in heavily impacted ecosystems, and

reinforces the need for managers to consider effects of habitat-forming invasive species within

the broader context of environmental change.

103

INTRODUCTION

Ecosystems will continue to undergo dramatic anthropogenic shifts in the coming century,

with impacts on fauna likely to depend on how individuals respond to modified or degraded

habitats. Animals that make adaptive decisions by responding appropriately to novel risks and

resources and that prefer habitats conferring the best fitness outcomes will be most able to

persist in impacted landscapes (Sih et al. 2011, Wong & Candolin 2014). However, a lack of

evolutionary history in altered environments can create a mismatch between cues and

outcomes, causing individuals to make maladaptive decisions (evolutionary trap: Robertson,

Rehage & Sih 2013). In the case of habitat selection decisions, this may result in individuals

avoiding high quality but unattractive habitats (perceptual traps) or selecting low quality but

attractive habitats (ecological traps) (Robertson & Hutto 2006; Patten & Kelly 2010; Hale &

Swearer 2016). In impacted landscapes, such individual-level responses to altered habitats

could exacerbate population-level effects of environmental change, as animals are either

drawn into attractive population sinks from surrounding higher quality habitats, or fail to take

advantage of remaining viable habitats in fragmented landscapes because habitat cues have

changed (Hale et al. 2015).

Traditional density-based habitat association studies may come to erroneous conclusions

about the ecological value of a habitat if individual-level processes are not considered (van

Horne 1983). Density-dependent source-sink population models account for differential fitness

between habitats (Pulliam 1988), but such models typically assume (often incorrectly) that

individuals will choose the best available habitat, and thus ignore the potential for attractive

sinks or underutilised sources (Kokko & Sutherland 2001, Battin 2004). By linking individual

habitat selection and fitness outcomes, the ecological trap framework provides a more

informative approach to assessing the role of habitats in population persistence (Hale et al.

2015, Hale & Swearer 2016).

Invasive habitat-forming ecosystem engineers, such as plants, algae and sessile marine fauna,

can alter ecosystems by competing for space with native habitat-formers and by changing the

availability of food and shelter for animals (Crooks 2002, MacDougall & Turkington 2005,

Gribben & Wright 2006, Pyšek et al. 2012). It is often assumed that invasive species reduce

biodiversity in invaded areas, but the body of empirical evidence is equivocal, particularly for

animals (Gribben & Wright 2006, Pyšek et al. 2012, Dijkstra et al. 2017). Effects on biodiversity

will depend on a range of conditions, including the state of the ecosystem prior to invasion,

104

the quality of the novel habitat for native fauna, and on behavioural responses by native fauna

to the novel habitat.

Figure 5.1. Archetypal examples of three rocky reef habitats in northern Port Phillip Bay,

Australia: (A) dense canopy of the native kelp Ecklonia radiata, (B) urchin-grazed barren reef,

and (C) spring growth of the invasive kelp Undaria pinnatifida on urchin-grazed reef.

Where anthropogenic stressors lead to declines in native habitat-forming species, invasive

habitat-formers may take advantage of vacant niches (MacDougall & Turkington 2005). A

marine macroalga, the Japanese kelp (Undaria pinnatifida), is one such invader. Native to the

north-west Pacific Ocean, and has now established in Europe, the Americas, New Zealand and

Australia. However, despite its high profile as a serial invader and a considerable body of

literature on its physiology and reproductive ecology (Schaffelke & Hewitt 2007, Davidson et

al. 2015), very little is known about its impacts on animal populations in invaded ecosystems

(Raffo et al. 2009, 2014, Thomsen et al. 2009, Irigoyen et al. 2011, Howland 2012). In south-

eastern Australia, Undaria fills a macroalgal niche on artificial substrates and degraded reefs

where poor water quality and urchin grazing have driven the decline of native kelps such as

Ecklonia radiata (Fig. 5.1). Undaria is a weak competitor in undisturbed macroalgal

communities, especially those with laminarian kelps (Valentine & Johnson 2003, Edgar et al.

2004, Farrell & Fletcher 2006, de Leij et al. 2017), but quickly takes advantage of bare

substrate and appears better able to persist on degraded reefs (Campbell & Burridge 1998,

Valentine & Johnson 2003, Edgar et al. 2004, South & Thomsen 2016).

Seasonal Undaria growth may mitigate loss of fish biodiversity on urchin-grazed rocky reefs if

native fish utilize the potential refugia provided by the invasive habitat, and the invasive

habitat is of higher quality than the urchin barren habitats that it replaces. Potential benefits

105

will be maximised if the invasive habitat does not act as an ecological trap by attracting fish

that might otherwise settle in higher quality native habitats (Robertson & Hutto 2006, Patten

& Kelly 2010, Hale & Swearer 2016). In this study, we employ the ecological trap framework to

test three predictions about the effects of the invasive habitat-former on native fauna, by

surveying fish populations on comparable natural rocky reefs and artificial boulder reefs with

differing canopy cover. The directions of our predictions are based on a common expectation

that invaded habitats will be both less attractive and of poorer quality than uninvaded

habitats, but that the invasive habitat-former may offer better habitat than urchin barrens:

native fish will prefer native over invasive kelps over urchin barrens (Prediction 1); native fish

communities will differ and be more abundant and diverse in invasive kelp than urchin barrens

(Prediction 2); and measures of fitness will be higher for fish in native than invasive kelp

habitat (Prediction 3). In testing these predictions, we inform management responses to this

established invader, and demonstrate how the ecological trap framework can be employed in

future assessments of the effects of invasive habitat-formers on native fauna.

METHODS

Study system

We conducted this study at five locations in Port Phillip Bay (Fig. 5.2; see Appendix 5.1 for

location descriptions), a 1930 km2 semi-enclosed marine embayment in south-eastern

Australia. The bay is affected by numerous anthropogenic impacts arising from population

centres and industrial inputs (Sampson et al. 2014). Soft sediment habitats are predominant,

with scattered rocky shorelines and reef patches in 0-8 m depth. Rocky reef habitats in the

northern half of the bay have been degraded by a combination of abiotic conditions and urchin

grazing, leading to large areas of urchin barrens and epilithic sediment matrix with turfing

algae (Ling et al. 2010, Filbee-Dexter & Wernberg 2018). Canopy-forming macroalgae such as

the native common kelp (Ecklonia radiata) have been in decline for decades (Jung et al. 2011).

Concurrently, the invasive kelp Undaria pinnatifida has been spreading since its introduction in

the late 1980s, and is now widespread on rocky reefs and marine infrastructure throughout

the bay, with high densities in the northern half (pers. obs.). Throughout its native range and

most of its introduced range (including Australia), Undaria is a winter annual, forming dense

stands during winter-spring and senescing in summer.

106

Figure 5.2. Map of study locations in Port Phillip Bay, Australia. Round markers denote

locations with established populations of Undaria pinnatifida. Governor Reef (square marker)

is a comparatively unimpacted reef with dense Ecklonia radiata canopy.

Prediction 1: Native fish will prefer native over invasive kelps and urchin barrens habitat

The first criterion for demonstrating an ecological trap is to show that animals either prefer or

do not avoid the novel habitat (Robertson & Hutto 2006).

Habitat choice experiment

To investigate whether native reef fish prefer native over invasive kelp, we collected common

weedfish (Clinidae: Heteroclinus perspicillatus) and little weed whiting (Odacidae: Neoodax

balteatus) individuals from rocky reef habitats and subjected them to a habitat choice trial.

These species are among the most common macroalgae-associated fishes in the Bay and are

easily housed in aquaria for the duration of the experiment. Both species use macroalgae and

seaweed structure as shelter. Clinids are predators of small fish and motile invertebrates and

use available cover to stalk and ambush prey. N. balteatus are more active in browsing

macroalgal surfaces for epifaunal invertebrate prey such as molluscs and crustaceans.

107

We placed individuals in the centre of a four-chambered crucifix-shaped tank (McDermott and

Shima 2006; Appendix 5.2) and offered them a simultaneous choice between one reference

(bare rock, simulating an urchin barren) and three macroalgal habitat cues (similarly-sized

rocks with a Undaria, Ecklonia or Sargassum linearifolium thallus attached via a black cable tie

on the holdfast). We provided equal volumes of each species of macroalgae (measured by

water displacement) and randomised their positions within the tank for each replicate fish.

Undaria and Ecklonia are very similar in height and gross morphology at a given volume, with

large fronds making up the bulk of the volume and spreading to cover a large area of

substrate. Sargassum tends to be more upright, with gas-filled bladders holding foliage above

the substrate. At least 10 thalli of each macroalgal species were used throughout the

experiment; for each replicate trial, a thallus was randomly selected and attached to a

randomly-selected rock. We held fish in a transparent plastic cylinder in the centre of the tank

for 5 min to visually survey the habitat options prior to the commencement of the trial. The

cylinder was then raised by a string-and-pulley system with the researcher out of sight (but

able to view the fish through a small hole in a screen), and the fish allowed to swim freely

around the choice tank. As measures of preference, we recorded both the initial choice of the

fish, which was likely to be a primarily vision-based decision (first chamber entered after

release from the cylinder) and the location of the fish after 20 mins, which is expected to be a

decision based on multiple senses. No fish were recorded in the centre of the tank at the 20

min mark.

Recruitment to artificial reefs

To complement the laboratory habitat choice experiment, we conducted an artificial reef

experiment at Half Moon Bay to test the effect of native and invasive kelp canopies on fish

recruitment under controlled field conditions. Half Moon Bay was selected as it offered a large

expanse of sandy substrate at a suitable depth and in an area of the Bay where Undaria has

been established since at least 2009 (Primo et al. 2010). During September 2014, we

constructed 20 replicate boulder reefs (1 m2) on sandy substrate, arrayed 20 m apart in a 4 x 5

grid pattern, and randomly assigned each reef to one of three treatments (n = 7 with Undaria

canopy, n = 7 with E. radiata canopy and n = 6 as unstocked controls) using kelp thalli collected

from an adjacent subtidal reef and attached to the artificial reef using rope and cable ties. We

stocked kelp treatments in late July and early August 2015 and recorded natural recruitment of

108

reef fishes on three occasions throughout spring (16 September, 26 October and 31 November

2015). Kelp canopies were replenished as necessary to maintain similar percent coverage

(Undaria treatments required regular replenishment as thalli regressed during October and

November). As artificial reefs were initially devoid of fish, relative recruitment was quantified

in terms of the abundance and diversity of fish present on the reef at each survey date.

Prediction 2: Native fish will be more abundant and diverse in invasive kelp than urchin

barrens

To test whether habitat preferences in experimental conditions correspond to patterns of fish

abundance in the field, we compared fish communities on heavily urchin-grazed rocky reefs

with and without seasonal Undaria canopy using the underwater visual census (UVC) survey

method for benthic fish species and baited remote underwater video (BRUV) for larger, more

mobile fish species. Survey locations were selected based on the presence of Undaria patches

on urchin-grazed subtidal reefs. We primarily focused on a comparison of fish populations in

invasive Undaria habitat relative to urchin barrens—rather than native Ecklonia habitat—

because (a) the small size and rarity of remaining Ecklonia habitats in northern Port Phillip Bay

limited the potential for spatial replication of comparable sites, with Williamstown currently

the only known location where urchin barrens, Undaria patches and remnant Ecklonia patches

co-occur on rocky reef, and (b) this invader fills vacant habitat on degraded reefs or artificial

substrates rather than directly outcompeting native kelps, making the comparison to barren

reefs more ecologically-relevant. Nonetheless, we did compare fish populations in Undaria and

Ecklonia beds using a diver catch-per-unit-effort metric at one location (Williamstown), as well

as BRUV deployments at two locations (Williamstown and St Leonards) (Fig. 5.2). All surveys

and collections were conducted by the same researcher (LB) under calm weather conditions

(<15 knots) between 1000-1600 hours. Survey and collection efforts alternated randomly

between the two habitat types to avoid temporal bias.

Underwater visual census survey

The UVC was conducted on paired circular plots of 12.6 m2 (2 m radius) with comparable

rugosity and depth but differing in their coverage of Undaria. The paired design only extended

109

to placement of plots to maximise comparability of plots with and without Undaria, rather

than a paired statistical analysis. We employed a before-after-control-impact (BACI) design by

surveying plots with and without seasonal Undaria canopy, during and after the Undaria

growing season (when the canopy was at its most dense and after it had completely

regressed). In this environment, adult Undaria undergo a complete die-off over summer, with

no holdfasts remaining by the time of the “after” surveys. We haphazardly placed paired plots

10-15 m apart, and >30 m from adjacent pairs at that location. Populations of small benthic

and cryptic fishes are strongly influenced by habitat characteristics at this scale (Willis &

Anderson 2003). We dropped a weight to mark the middle of the plot and surveyed larger

benthic and benthopelagic fish initially by swimming a circular path outside the plot. Small or

cryptic species were detected by a diver moving in concentric circular paths from the outer

edge to the centre of the plot, searching within the reef structure and among algal cover. If fish

of the same species were observed multiple times within the plot, they were counted as a

single fish unless they were clearly distinct individuals (size, sex or markings). Several members

of the Heteroclinus genus (H. perspicillatus, H. heptaeolus, H. adelaidae, H. wilsoni, H.

eckloniae, H. macrophthalmus) are difficult to reliably distinguish in situ, so were treated as a

single species. Accordingly, the species richness metric is likely underestimating the true

diversity of the Heteroclinus genus. Upon completing the fish census, we counted the number

of urchins present within the plot and estimated percent coverage of kelp, all macroalgae, and

algal turf.

Baited video survey

The BRUV system consisted of a GoPro Hero3+ camera mounted on a weighted milk crate with

a 1 m long bait arm made from 20mm PVC conduit. Two pilchards (Sardinops sagax) were cut

in half and placed in a 20 x 20 cm plastic mesh bag attached to the end of the bait arm. Larger

baits are typically used in BRUV systems, but we wished to minimise the spread of the bait

plume as elevated rocky reefs in the bay are often very small (10-100 m). BRUV deployments

were made at least 50 m apart on comparable reef substrate at depths of 2-4.5 m. Fish

observations were made for 20 min from the time when the BRUV unit settled on the sea

floor. To assess relative abundance of each species at each site while preventing potential

double counting of individual fish, we used a conservative metric termed maxN, where maxN is

the maximum number of individuals of a given species occurring simultaneously in the video

110

field of view (Willis & Babcock 2000). In the case of sexually dimorphic species or juvenile

fishes, we summed the maxN counts for males, females and juveniles of that species. Where

an individual could not be identified to species level, we identified it to the lowest taxonomic

rank possible (usually Family or Genus). We included the southern calamari squid (Sepioteuthis

australis) in the analysis as it is similar in habit to many inshore fishes.

Catch-per-unit-effort

Collection efforts for N. balteatus, H. perspicillatus and H. heptaeolus in Ecklonia and Undaria

habitat patches at Williamstown were timed to allow quantification of catch per unit effort

(CPUE). The Williamstown location contained a mosaic of seasonal Undaria growing on urchin-

grazed reefs between well-defined remnant patches of Ecklonia and Sargassum. Collections

were made by hand net on SCUBA. We calculated CPUE on a per species basis, with each

collection dive treated as a single statistical replicate where CPUE = N fish collected / duration

of dive. Collection dives alternated between habitats over the course of nine days between 24

October to 28 November 2016.

Prediction 3: Measures of fitness will be higher for fish in native than invasive kelp habitat

The second criterion for demonstrating an ecological trap is to show that a suitable measure of

fitness is lower in the putative trap habitat than other habitat options (Robertson & Hutto

2006). We expect that the native kelp habitat will be of higher quality and may lead to higher

fitness metrics, although individual fitness (Prediction 2) may not be independent of

population density (Prediction 1); population density tends to be higher in high quality

habitats, leading to equalisation of individual fitness between habitats of differing quality

(ideal free distribution: Fretwell & Lucas 1969).

Fish collected from natural reefs

We compared body condition metrics in N. balteatus, H. perspicillatus and Ogilby's weedfish

(Heteroclinus heptaeolus) collected at Williamstown from Undaria and Ecklonia habitats

111

(collection efforts described above). The habitat occupied by each individual was defined by

the dominant kelp species within 2 m of the collection site, with percentage coverage of

macroalgal species estimated visually. Fish were collected using handnets and killed using

clove oil, and placed in sealed plastic bags in an ice slurry. Within 24 hours of collection,

specimens were weighed (wet weight), measured (total length) and dissected to determine sex

and weight of the stomach contents, liver and gonads. Overall body condition was quantified

using the relative condition metric recommended by (Le Cren 1951): Krel = W/Wexp, where W is

the measured gutted weight and Wexp is the gutted weight predicted by the weight-at-length

power curve fitted to all available samples (Fig. 5.3). Individuals with Krel values >1 are heavier

than the population mean for their length. One particularly large H. perspicillatus individual

from Undaria habitat was removed from the Krel analysis (145 mm, 4.7 standard deviations

above the mean). Liver and gonad condition was quantified using hepatosomatic and

gonadosomatic indices respectively (HSI or GSI = OW/GW, where OW = wet organ weight and

GW = gutted weight). For gravid females, we also photographed and counted subsamples of

ova (N. balteatus) or embryos (Heteroclinus spp.) to assess fecundity (Neoodax are broadcast

spawners, while Heteroclinus are livebearers).

Fish collected from artificial reefs

We collected H. perspicillatus individuals from artificial reefs (see Prediction 1) from Jan-Nov

2015 and compared body condition and reproductive investment metrics in fish from reefs

stocked with Undaria or Ecklonia. Collection methods were identical to those on natural reefs,

with collection efforts alternating between Undaria and Ecklonia stocking treatments. There

were not enough H. perspicillatus recruits on control (unstocked) reefs for meaningful

comparisons of condition on reefs with and without kelp.

Statistical analysis

Choice data (Prediction 1) were tested for habitat preference using a X2 test of proportions

implemented in R (R Core Team 2017), with expected (null) proportions equally distributed

across the four habitat cues. Artificial reef recruit abundance and species richness data

(Prediction 1) were compared across kelp treatments using Poisson generalised linear mixed

112

effects models implemented in the lme4 package for R (Bates et al. 2015). We included

Treatment (stocked kelp species: Undaria, Ecklonia or barren) as a fixed term, KelpCover

(percentage coverage of canopy) as a covariate, and as reefs were surveyed repeatedly, a reef

identity random intercept term (ReefID) nested within Treatment. There was some data

overdispersion, so we included an observation-level random term to account for excess

variation and avoid overestimating the predictive ability of model terms (Harrison 2014). We

tested for a significant Treatment effect by comparing the fit of models with and without the

Treatment term, while the lsmeans package for R provided Tukey’s pairwise post-hoc

comparisons of the three treatments (Lenth 2016).

Fish community data from UVC and BRUV surveys (Prediction 2) were fitted to a permutational

multivariate ANOVA model implemented in PRIMER 6 with the PERMANOVA+ add-on

(Anderson et al. 2008). Data were log(x+1) transformed to reduce the influence of a few highly

abundant species and fitted to a Bray-Curtis similarity resemblance matrix with a dummy

variable of 1. The models contained two fixed factors, Habitat and Location, as well as an

interaction term (Habitat×Location). The Location factor contained two levels (North and

West), reflecting environmental differences between sites in the north (Point Cook, Altona

Bay, Williamstown, Half Moon Bay) and west (Kirk Point) of the Bay. The models were fitted

using Type III sums of squares, with unrestricted permutation of raw data and 9999

permutations. We also extracted diversity metrics using the Diversity function and identified

species that were associated most strongly with the observed differences between habitats

and locations using the similarity percentage (SIMPER) function. Finally, data were visualised

using multidimensional scaling (MDS) and canonical analysis of principal coordinates (CAP)

plots.

Catch per unit effort data were log(x+1) transformed to improve normality, and compared

across Ecklonia and Undaria habitats using a linear analysis of covariance (lm function in R)

with Habitat as a fixed factor and sampling Date as a temporal covariate.

We compared fitness metrics (Prediction 3) in native and invasive kelp habitats using a series

of univariate linear models implemented in R. Response variables were checked for normality

and equality of variance and transformed as necessary. We analysed both weedfish species (H.

perspicillatus and H. heptaeolus) with a single model and included terms for kelp habitat type

(‘Habitat’), species (‘Species’), day of the season (‘Day’), and sex (‘Sex’). Models for the

sequentially hermaphroditic N. balteatus included only Habitat and Day terms. Fitness metrics

for H. perspicillatus recruits collected from artificial reefs (Prediction 3) were compared across

kelp treatments using univariate linear mixed models implemented in lme4, and included

113

Treatment, KelpCover and Sex as fixed terms and Treatment/ReefID as a random intercept

term. In all cases, we tested the significance of the effect of interest by comparing the fit of

models with and without the relevant term.

Plots were produced using the ggplot2 package for R (Wickham 2009).

Figure 5.3. Weight-at-length relationships for (A) Heteroclinus perspicillatus and H. heptaeolus

and (B) Neoodax balteatus individuals collected from Undaria pinnatifida and Ecklonia radiata

habitats in Port Phillip Bay.

RESULTS

Prediction 1: Native fish will prefer native over invasive kelps and urchin barrens habitat

Habitat choice experiment

Common weedfish (Heteroclinus perspicillatus) were more likely to select macroalgal cover

than barren rock, both initially (1.25x more than expected; p <0.0001) and after 20 mins (1.22x

more than expected; p <0.0001) (Fig. 5.4A; Appendix 5.3). Of those that chose macroalgae

during their initial decision, Undaria was preferable to Ecklonia or Sargassum (2.0x and 2.7x

more likely, respectively; p = 0.015), but after 20 mins, individuals were evenly distributed

across the three macroalgal options (p = 0.9) (Fig. 5.4A; Appendix 5.3).

114

Figure 5.4. Habitat choice trial results for (A) common weedfish (Heteroclinus perspicillatus): N

= 48; (B) little weed whiting (Neoodax balteatus): N = 23.

Little weed whiting (N. balteatus) also preferred macroalgae to bare rock, both initially (1.16x

more than expected; p = 0.0004) and after 20 mins (1.22x more than expected; p <0.0001) (Fig.

5.4B; Appendix 5.3). However, we observed no clear preference among macroalgal options,

either initially (p = 0.12) or after 20 mins (p = 0.4) (Fig. 5.4B; Appendix 5.3).

Table 5.1. Summary of fish recruitment to artificial reefs stocked with Undaria pinnatifida,

Ecklonia radiata, or left barren.

Treatment Kelp canopy cover (%) Abundance reef-1 Species richness reef-1 N surveys

Undaria 42 ± 3 0.9 ± 0.1 0.8 ± 0.1 89

Ecklonia 73 ± 2 1.1 ± 0.1 1.1 ± 0.1 89

Barren 5 ± 5 0.4 ± 0.2 0.4 ± 0.2 18

Pairwise

U-E z = 1.7, p = 0.21 z = 1.6, p = 0.24

U-B z = 2.3, p = 0.05 z = 2.8, p = 0.02

E-B z = 3.0, p = 0.008 z = 3.3, p = 0.003

Recruitment to artificial reefs

A 2.5x higher abundance (p = 0.01) and 2.4x higher species richness (p = 0.007) of reef fish

recruited to artificial reefs stocked with kelp relative to those with turfing algae only (Table

115

5.1). There was no evidence for differential recruitment between reefs stocked with either

kelp (p = 0.21), despite the Undaria treatment having less canopy cover (42 vs. 73 %) (Table

5.1). The canopy cover covariate positively predicted recruit abundance (p = 0.05), indicating

that the amount of cover may be more important than the macrophyte species providing the

cover. Canopy cover did not affect species richness (p = 0.36).

Prediction 2: Native fish will be more abundant and diverse in invasive kelp than urchin

barrens

Visual census survey

We recorded 19 fish species across 25 Undaria and 26 barren plots during in-season surveys

(September-October 2016) (Appendix 5.4) and found a significant effect of Habitat (Undaria or

Barren), Location (west or north Port Phillip Bay), and Habitat×Location on dissimilarity

between plots (PERMANOVA: Appendix 5.5). Canonical analysis of principle coordinates (CAP)

plots revealed visual separation of sites with and without Undaria, driven primarily by a

greater abundance of Heteroclinus spp. and Diodon nicthemerus in Undaria plots (Fig. 5.5).

SIMPER analysis determined that weedfish (Heteroclinus spp.: 28 %), Clarke’s threefin

(Trinorfolkia clarkei: 23 %), Tasmanian blenny (Parablennius tasmanianus: 10 %) and globefish

(D. nichthemerus: 7 %) contributed most to the observed dissimilarity between habitats

(weedfish and globefish in Undaria plots, threefins and blennies in Barren plots). Gross

community metrics also differed; species richness and total abundance were both higher on

Undaria plots (2.0x and 1.9x higher, respectively: Fig. 5.6A, Appendix 5.6). Barren and Undaria

plots were placed at comparable depths (mean ± SD: 2.6 ± 0.5 m cf. 2.7 ± 0.4 m, respectively).

Barren plots contained 6 ± 2 % macroalgal cover, mostly sea lettuce (Ulva spp). Undaria plots

contained 72 ± 3 % macroalgal cover, dominated by Undaria (58 %) with some secondary cover

from other ephemeral macroalgae, including Ulva spp. and Gracilaria spp. Undaria plots did

not contain any other brown algal species such as Ecklonia radiata or Sargassum spp. Where

Ulva or Gracilaria were present on Undaria plots, clinids (Heteroclinus spp. and Cristiceps

australis) were generally found in Undaria microhabitat (12/16 fish).

We resurveyed 18 plots (9 Undaria, 9 Barren) after the Undaria had completely regressed

(April-May 2017). Sparse early stage Undaria recruits (<10 cm) were present in May, but

116

provided a negligible amount of macroalgal cover. The effect of the Habitat factor on overall

fish community structure was significant (PERMANOVA: Appendix 5.5), but neither species

richness nor abundance significantly differed (Appendix 5.6).

Analysing in-season and off-season data together revealed significant overall effects of

Habitat, Season and Location on fish community structure, as well as a Habitat×Location

interaction (PERMANOVA: Appendix 5.5). Off-season Undaria and Barren plots both contained

abundant non-canopy-forming ephemeral macroalgae, dominated by Ulva spp., Gracilaria spp.

and Caulerpa spp., with greater mean coverage (51 ± 12 %) on Undaria plots relative to Barren

plots (32 ± 12 %).

Overall, macroalgal cover negatively correlated with urchin density (r = -0.42, t67 = -3.7, p =

0.0004). We found no evidence that off-season macroalgal cover on resurveyed plots was

predicted by in-season cover of either Undaria (F16 = 0.5, p = 0.5) or all macroalgae (F16 = 1.5, p

= 0.24).

Baited video survey

We recorded 28 fish species across 14 Undaria and 15 barren deployments (Appendix 5.4). Fish

communities did not differ between habitats or locations (PERMANOVA: Appendix 5.6), and

nor did species richness or combined MaxN (Appendix 5.6). BRUV data included some non-

reef-associated species, but restricting the analysis to reef-associated species did not alter our

interpretation.

Diver catch per unit effort

Collection efforts provided 45 Neoodax balteatus, 24 Heteroclinus perspicillatus and 18

Heteroclinus heptaeolus from 12 dives (total 465 min) in Undaria and 13 dives in Ecklonia (total

690 min) habitats over 9 days between 24 October and 28 November 2016. Catch per unit

effort (CPUE) of N. balteatus was 2.6x higher in Ecklonia than Undaria habitat (p = 0.04), but

CPUE of Heteroclinus spp. did not differ (Table 5.2). CPUE of N. balteatus also increased

throughout spring, coinciding with warmer water and greater fish activity (Date covariate: R2 =

0.26, p = 0.007; Table 5.2). There was no evidence that CPUE of Heteroclinus spp. increased

over time (p = 0.6; Table 5.2).

117

Table 5.2. Comparison of reef fish relative abundance, estimated by diver catch per unit effort

(CPUE), in Undaria and Ecklonia habitats, and over time, at Williamstown. Number of

collection efforts (N dives) are given as ‘Undaria, Ecklonia’. N dives differ between species as

not all dives targeted all species. Terms are tested by maximum likelihood ratio comparison of

fitted and null models (X2). Positive Cohen’s d effect sizes indicate metrics were higher in

Undaria habitat.

Habitat effect Undaria Ecklonia N dives X2 p Cohen’s d

Heteroclinus perspicillatus 0.9 ± 0.3 2.0 ± 0.6 13, 10 1.4 0.12 -0.74

Heteroclinus heptaeolus 0.7 ± 0.2 1.0 ± 0.4 13, 12 0.04 0.91 -0.30

Heteroclinus spp. 1.5 ± 0.4 2.7 ± 0.8 13, 12 0.32 0.73 -0.48

Neoodax balteatus 1.6 ± 0.4 4.1 ± 1.0 9, 10 2.1 0.04 -1.08

Temporal effect R2 N dives X2 p

Heteroclinus perspicillatus -0.02 13, 10 0.74 0.33

Heteroclinus heptaeolus 0.10 13, 12 0.77 0.19

Heteroclinus spp. 0.03 13, 12 0.14 0.87

Neoodax balteatus 0.26 9, 10 3.3 0.007

CPUE was not formally assessed in barren habitats, as both Heteroclinus spp. and N. balteatus

are rare in areas without macroalgae. In the course of surveying 26 barren UVC plots during

the Undaria growing season in the northern part of the Bay, we only observed one clinid—a H.

perspicillatus individual inhabiting a macroalgal microhabitat on an otherwise barren plot—

and no N balteatus (Appendix 5.4). Accordingly, CPUE in the barrens would likely be close to

zero.

Prediction 3: Measures of fitness will be higher for fish in native than invasive kelp habitat

Fish collected from natural reefs

N. balteatus individuals collected from Ecklonia habitat were 1.3x larger (p = 0.007) and 2.4x

heavier (p = 0.01) than those in seasonal Undaria habitat (Table 5.3), and a higher proportion

were male (Ecklonia: 7/23; Undaria: 1/22). We found no evidence that body condition or

reproductive fitness metrics differed between Ecklonia and Undaria habitats, although the

direction of effect for reproductive metrics was generally positive in Undaria habitats (Table

5.3).

118

We also found little evidence that H. perspicillatus or H. heptaeolus collected from Ecklonia or

Undaria habitats experience differential habitat quality, with no difference in size, body

condition or reproductive investment between habitat types (Table 5.3). However, the

proportion of fertilised eggs was 1.4x higher in weedfish living in Ecklonia habitats (p = 0.003;

Table 5.3).

Table 5.3. Comparison of body condition and reproductive condition metrics in reef fishes

collected from Undaria and Ecklonia habitats at Williamstown (natural reefs) and Half Moon

Bay (artificial reefs). N = ‘Undaria, Ecklonia’. Habitat effect is tested by comparing fit of models

with and without Habitat term. Positive Cohen’s d effect sizes indicate metrics were higher in

Undaria habitat.

Undaria Ecklonia N Stat p d

Neoodax balteatus

Natural reefs F

Length (mm) 73 ± 4 96 ± 6 22, 23 8.2 0.007 -0.97

Gutted weight (g) 3.9 ± 0.9 9.5 ± 1.7 22, 23 7.2 0.01 -0.87

Wrm 101 ± 1 99 ± 1 22, 23 4.0 0.05 +0.50

Hepatosomatic index (*1000) 30 ± 2 28 ± 2 22, 23 2.1 0.15 +0.20

Stomach index (*1000) 25 ± 2 27 ± 2 22, 23 1.0 0.32 -0.19

Gonadosomatic index (*1000) 51 ± 8 48 ± 10 22, 23 0.2 0.67 +0.07

Mature eggs 789 ± 245 516 ± 213 17, 10 0.5 0.49 +0.30

Eggs 3162 ± 936 2789 ± 850 17, 10 0.5 0.49 +0.11

Mature eggs gutted weight-1 198 ± 39 116 ± 29 17, 10 0.6 0.63 +0.59

Eggs gutted weight-1 742 ± 128 568 ± 176 17, 10 0.3 0.34 +0.32

Egg maturity (%) 26 ± 4 24 ± 8 17, 10 <0.1 0.93 +0.06

Heteroclinus spp.

Natural reefs F

Length (mm) 74 ± 5 65 ± 2 20, 22 1.3 0.26 +0.52

Gutted weight (g) 4.1 ± 0.9 2.6 ± 0.3 20, 22 1.6 0.21 +0.50

Wrm 102 ± 3 100 ± 3 20, 22 0.2 0.63 +0.12

Hepatosomatic index (*1000) 19 ± 1 21 ± 3 20, 22 <0.1 0.98 -0.09

Stomach index (*1000) 24 ± 3 42 ± 17 20, 22 2.0 0.16 -0.31

Gonadosomatic index (*100) 12 ± 2 11 ± 3 20, 22 0.6 0.45 +0.10

Embryos 256 ± 35 198 ± 33 13, 12 0.8 0.39 +0.48

Eggs 516 ± 102 274 ± 66 13, 12 2.9 0.10 +0.81

Embryos gutted weight-1 68 ± 9 92 ± 16 13, 12 2.4 0.14 -0.52

Eggs gutted weight-1 124 ± 21 126 ± 28 13, 12 0.3 0.58 -0.02

Egg fertilisation (%) 60 ± 6 81 ± 7 13, 12 11 0.003 -0.90

Artificial reefs: X2

Length (mm) 51 ± 4 62 ± 4 11, 20 0.6 0.45 -0.71

Gutted weight (g) 1.3 ± 0.3 2.3 ± 0.4 11, 20 0.9 0.35 -0.70

Wrm 1.3 ± 0.1 1.3 ± 0.1 11, 20 0.3 0.58 +0.01

Hepatosomatic index (*1000) 43 ± 17 21 ± 3 11, 20 5.9 0.01 +0.53

Stomach index (*1000) 16 ± 4 15 ± 3 11, 20 <0.1 0.99 +0.08

Gonadosomatic index (*1000) 12 ± 5 55 ± 19 11, 20 0.8 0.37 -2.39

119

Dietary assessment revealed that almost all clinids had consumed mysid shrimp, both in

Undaria and Ecklonia habitats. Some individuals had also consumed amphipods (Undaria and

Ecklonia), isopods (Undaria only) or decapods (Ecklonia only). Amphipods were the most

frequently identifiable prey items for N. balteatus (Ecklonia only). Animal prey items from N.

balteatus in Undaria habitats were not identifiable, although there was no significant

difference in the weight of stomach contents (Table 5.3).

Fish collected from artificial reefs

Hepatosomatic index was 2.0x higher in H. perspicillatus individuals collected from reefs

stocked with Undaria canopy versus Ecklonia canopy (p = 0.01), while other fitness metrics

were unaffected by kelp canopy treatment (Table 5.3).

DISCUSSION

Some results indicate that the presence of the invasive habitat-forming ecosystem engineer

does not create an ecological trap for native fish. Contrary to Prediction 1, that native fish will

prefer native to invasive kelp, the laboratory habitat choice experiment revealed that two

macroalgal-associated reef fishes were equally willing to utilise shelter provided by native and

invasive kelp. Similarly, fish recruits did not distinguish between artificial reefs stocked with

invasive or native kelp, but preferred both to barren reefs. Our population survey data broadly

supported Prediction 2, that fish communities will be more abundant and diverse in native

kelp than invasive kelp than urchin barrens, with some evidence for higher population

densities in native kelp beds, but invasive kelp patches greatly improved fish abundance and

diversity on heavily urchin-grazed reefs where native canopy-forming macroalgae are absent.

Body condition and reproductive investment metrics indicate that fish inhabiting these

invasive kelp patches have similar or better body condition to those in adjacent native kelp

beds, contradicting Prediction 3 (that fish in native kelp habitats will have higher fitness

metrics than those in invasive kelp habitats), although we did not make any direct comparisons

of mortality rates between habitats. Overall, this invasive habitat-forming kelp appears to

provide valuable habitat for native fishes on urchin-grazed urban-impacted rocky reefs.

120

Invasive habitat-forming species can deleteriously affect faunal populations in invaded

landscapes whether the invasive habitat is avoided, leading to habitat loss (Trammell & Butler

1995, Valentine et al. 2007), or occupied, leading to poor fitness outcomes (Remeš 2003, Lloyd

& Martin 2005, Rodewald et al. 2010). However, where the invader adds physical structure or

a novel food source, there can be beneficial effects for some native taxa and increases in local

biodiversity (Baldwin & Lovvorn 1994, Crooks 2002, Castilla et al. 2004, Byers et al. 2012,

Wright et al. 2014). Our study highlights the potential benefits of a weakly-competitive

invasive habitat-forming species in mitigating the impacts of habitat degradation on native

biodiversity.

Figure 5.5. Canonical analysis of principle coordinates (CAP) showing variation in fish

communities across underwater visual census (UVC) plots with and without Undaria

pinnatifida canopy (Undaria: grey; Barren: black), during and after the Undaria growing

season. Radiating lines indicate direction and strength of influence of fish species on the

observed variation between UVC plots. Only species with Pearson’s correlation >0.40 are

shown. Full species names: Brachaluteres jacksonianus, Diodon nicthemerus, Foetorepus

calauropomus, Heteroclinus perspicillatus, Ophiclinus ningulus, Parablennius tasmanianus,

Trinorfolkia clarkei, Trachinops caudimaculatus.

121

In this study system, the invasive habitat-former is highly seasonal, and may influence the

reproductive success of local reef fishes to the extent that it coincides with reproductive

provisioning and larval settlement phases (conceptual model: Appendix 5.7). Common

macroalgae-associated reef fishes such as Heteroclinus spp. and N. balteatus inhabit Undaria

patches throughout vitellogenesis and into the spawning period that occurs from late winter to

late spring (Gunn & Thresher 1991; Neira & Sporcic 2002; this study). Our data on stomach

contents, body condition and reproductive investment indicate that adults residing in Undaria

habitat experience comparable food availability to those in native Ecklonia habitats – perhaps

as a result of lower population density and therefore less competition in Undaria patches. This

would be consistent with the ideal free distribution theory (Fretwell & Lucas 1969). Regardless,

it is unlikely that individuals that choose to reside in Undaria habitats during vitellogenesis are

falling into a condition-driven ecological trap. We did find evidence that fertilisation rates were

lower in the ovaries of H. perspicillatus and H. heptaeolus collected from Undaria patches. This

may reflect mate-finding difficulty in small habitat patches that contain few if any conspecifics

and are separated from adjacent patches by urchin barrens. Undaria patches on these

degraded reefs are increasing in size with successive seasons (pers. obs.), and we expect that

this possible Allee effect will ameliorate over time.

We focus on juvenile and adult fish, but larval settlement is also important in assessing the role

of the invasive habitat-former. At present, no literature exists on the settlement preferences

of larval Heteroclinus spp. or Neoodax balteatus. Clinid larvae are at peak densities in the

water column in spring, with settlement occurring during spring-summer (Gunn & Thresher

1991, Neira & Sporcic 2002). This coincides with peak densities of Undaria canopy, and it is

likely that if larvae have similar habitat preferences to adults, large numbers may recruit to

these habitats shortly before the summer die-off of Undaria canopy (Appendix 5.7). Such a loss

of cover on otherwise barren reefs warrants investigation as a potential temporal ecological

trap for clinid recruits. Such a trap may occur if the benefits of inhabiting the invasive kelp are

outweighed by the eventual loss of cover. However, the rapid appearance of late stage

juvenile and adult fish on our plots and artificial reefs in summer—together with anecdotal

reports of clinids ‘rafting’ in unattached macroalgae—suggests that migration is common and

perhaps does not greatly increase mortality rates. As quantifying movement and mortality of

cryptobenthic fishes settling in invaded and non-invaded habitats remains a considerable

challenge, future studies would benefit from developing novel approaches to discriminate

between these two key demographic processes. N. balteatus settlement peaks later in summer

(Neira & Sporcic 2002), by which time Undaria thalli have regressed to sporophyll and

122

holdfasts and may no longer offer attractive habitat for recruits. Such species likely disperse

into the invasive habitat as adults, perhaps through spillover from areas of high population

density in remnant native kelp habitats.

Figure 5.6. Fish community metrics from underwater visual census plots on urchin-grazed reefs

(A) during and (B) after the Undaria pinnatifida growing season.

Implications for management

In this study we demonstrated the application of the ecological trap theory in assessing the

impacts of novel habitats on native fauna. In this case, rather than revealing the existence of

an ecological trap, our findings provide evidence that an invasive habitat-forming species can

have value for faunal communities when it fills a niche left vacant by native habitat-formers.

Previous work has shown that in contrast to many invasive ecosystem engineers, this invader

appears to be a passenger rather than a driver of ecological change (Valentine & Johnson

2003, Edgar et al. 2004, South & Thomsen 2016), and that it supports diverse invertebrate

communities (Howland 2012). Together with our findings of benign effects on at least some

key native fish taxa and its functional replacement of native macroalgae in large areas, this

leads us to recommend that removal of Undaria should not be a high management priority

within the Bay. Eradication is certainly impossible here, and where an invasive species has

functionally replaced a native species (as is the case with Undaria on urchin-grazed reefs),

removal may not be sufficient to restore the natural ecosystem (Reid et al. 2009), and

furthermore, is likely to drive additional biodiversity loss where native fauna depend on the

invasive habitat for food or shelter (Zavaleta et al. 2001). Instead, we recommend that

managers (a) focus on measures to restore the native kelp canopy, and (b) consider the

123

potential for beneficial effects of introduced habitat-formers, particularly in degraded

environments, and when possible, focus control or eradication efforts on invaders that are

highly competitive in the invaded environment and/or drive declines in biodiversity of native

species.

ACKNOWLEDGEMENTS

We wish to thank Dean Chamberlain, Seann Chia, Ben Cleveland, Emily Fobert, Molly Fredle,

Akiva Gebler, Kevin Jensen, Valeriya Komyakova, Nina Kriegisch, Kevin Menzies, Jack O’Connor,

Simon Reeves, Juan Manuel Valero Rodriguez, Kyler Tan, Chris Taylor and João Teixiera for

their assistance with fieldwork. Members of the SALTT and REEF labs provided useful

comments on the manuscript. All work was conducted in accordance with permits from the

Victorian Fisheries Authority (RP919) and the University of Melbourne’s Faculty of Science

Animal Ethics Committee (1413193 and 1413133). This research was funded by grants to LB

from the Holsworth Wildlife Research Endowment, the PADI Foundation, and the Victorian

Environmental Assessment Council. The authors have no competing interests to declare.

REFERENCES

Anderson M, Gorley R, Clarke R (2008) Permanova+ for Primer: guide to software and

statistical methods. Plymouth.

Baldwin JR, Lovvorn JR (1994) Expansion of seagrass habitat by the exotic Zostera japonica, and

its use by dabbling ducks and brant in Boundary Bay, British Columbia. Mar Ecol Prog Ser

103:119–127

Bates D, Maechler M, Bolker B, Walker S (2015) Fitting linear mixed effects models using lme4.

J Stat Softwar 67:1–48

Battin J (2004) When good animals love bad habitats: ecological traps and the conservation of

animal populations. Conserv Biol 18:1482–1491

Byers JE, Gribben PE, Yeager C, Sotka EE (2012) Impacts of an abundant introduced ecosystem

engineer within mudflats of the southeastern US coast. Biol Invasions 14:2587–2600

Campbell SJ, Burridge TR (1998) Occurrence of Undaria pinnatifida (Phaeophyta: Laminariales)

in Port Phillip Bay, Victoria, Australia. Mar Freshw Res 49:379–381

124

Castilla JC, Lagos NA, Cerda M (2004) Marine ecosystem engineering by the alien ascidian

Pyura praeputialis on a mid-intertidal rocky shore. Mar Ecol Prog Ser 268:119–130

Cren ED Le (1951) The length-weight relationship and seasonal cycle in gonad weight and

condition in the perch (Perca fluviatilis). J Anim Ecol 20:201–219

Crooks JA (2002) Characterizing ecosystem-level consequences of biological invasions: the role

of ecosystem engineers. Oikos 97:153–166

Davidson AD, Campbell ML, Hewitt CL, Schaffelke B (2015) Assessing the impacts of

nonindigenous marine macroalgae: an update of current knowledge. Bot Mar 58:55–79

Dijkstra JA, Harris LG, Mello K, Litterer A, Wells C, Ware C (2017) Invasive seaweeds transform

habitat structure and increase biodiversity of associated species. J Ecol 105:1668–1678

Edgar GJ, Barrett NS, Morton AJ, Samson CR (2004) Effects of algal canopy clearance on plant,

fish and macroinvertebrate communities on eastern Tasmanian reefs. J Exp Mar Biol Ecol

312:67–87

Farrell P, Fletcher RL (2006) An investigation of dispersal of the introduced brown alga Undaria

pinnatifida (Harvey) Suringar and its competition with some species on the man-made

structures of Torquay Marina (Devon, UK). J Exp Mar Biol Ecol 334:236–243

Filbee-Dexter K, Wernberg T (2018) Rise of turfs: a new battlefront for globally declining kelp

forests. Bioscience 68:64–76

Fretwell SD, Lucas HL (1969) On territorial behavior and other factors influencing habitat

distribution in birds. Acta Biotheor 19:16–36

Gribben PE, Wright JT (2006) Invasive seaweed enhances recruitment of a native bivalve: roles

of refuge from predation and the habitat choice of recruits. Mar Ecol Prog Ser 318:177–

185

Gunn JS, Thresher RE (1991) Viviparity and the reproductive ecology of clinid fishes (Clinidae)

from temperate Australian waters. Environ Biol Fishes 31:323–344

Hale R, Swearer SE (2016) Ecological traps: current evidence and future directions. Proc R Soc B

Biol Sci 283:494–499

Hale R, Treml EA, Swearer SE (2015) Evaluating the metapopulation consequences of

ecological traps. Proc R Soc London B Biol Sci 282:20142930

Harrison XA (2014) Using observation-level random effects to model overdispersion in count

data in ecology and evolution. PeerJ 2:e616

Horne B van (1983) Density as a misleading indicator of habitat quality. J Wildl Manage

47:893–901

Howland S (2012) Variation and impacts on biodiversity in kelp holdfast assemblages: a

125

comparison of the introduced Undaria pinnatifida and the indigenous Ecklonia radiata.

University of Melbourne

Irigoyen AJ, Eyras C, Parma AM (2011) Alien algae Undaria pinnatifida causes habitat loss for

rocky reef fishes in north Patagonia. Biol Invasions 13:17–24

Jung CA, Dwyer PD, Minnegal M, Swearer SE (2011) Perceptions of environmental change over

more than six decades in two groups of people interacting with the environment of Port

Phillip Bay, Australia. Ocean Coast Manag 54:93–99

Kokko H, Sutherland WJ (2001) Ecological traps in changing environments: ecological and

evolutionary consequences of a behaviourally mediated Allee effect. Evol Ecol Res 3:537–

551

Leij R de, Epstein G, Brown MP, Smale DA (2017) The influence of native macroalgal canopies

on the distribution and abundance of the non-native kelp Undaria pinnatifida in natural

reef habitats. Mar Biol 164:156

Lenth R V (2016) Least-squares means: the R package lsmeans. J Stat Softw 69:1–33

Ling SD, Ibbott S, Sanderson JC (2010) Recovery of canopy-forming macroalgae following

removal of the enigmatic grazing sea urchin Heliocidaris erythrogramma. J Exp Mar Biol

Ecol 395:135–146

Lloyd JD, Martin TE (2005) Reproductive success of chestnut-collared longspurs in native and

exotic grassland. Condor 107:363–374

MacDougall AS, Turkington R (2005) Are invasive species the drivers of change or passengers

of change in degraded ecosystems? Ecology 86:42–55

McDermott CJ, Shima JS (2006) Ontogenetic shifts in microhabitat preference of the temperate

reef fish Forsterygion lapillum: implications for population limitation. Mar Ecol Prog Ser

320:259–266

Neira FJ, Sporcic MI (2002) Use of ichthyoplankton ecology to evaluate ecosystem changes: a

case study in a large, semi-enclosed Australian bay. Mar Freshw Res 53:339

Patten MA, Kelly JF (2010) Habitat selection and the perceptual trap. Ecol Appl 20:2148–2156

Primo C, Hewitt CL, Campbell ML (2010) Reproductive phenology of the introduced kelp

Undaria pinnatifida (Phaeophyceae, Laminariales) in Port Phillip Bay (Victoria, Australia).

Biol Invasions 12:3081–3092

Pulliam HR (1988) Sources, sinks, and population regulation. Am Nat 132:652–661

Pyšek P, Jarošík V, Hulme PE, Pergl J, Hejda M, Schaffner U, Vilà M (2012) A global assessment

of invasive plant impacts on resident species, communities and ecosystems: the

interaction of impact measures, invading species’ traits and environment. Glob Chang

126

Biol 18:1725–1737

R Core Team (2017) R: A language and environment for statistical computing.

Raffo MP, Eyras MC, Iribarne OO (2009) The invasion of Undaria pinnatifida to a Macrocystis

pyrifera kelp in Patagonia (Argentina, south-west Atlantic). J Mar Biol Assoc UK 89:1571–

1580

Raffo MP, Russo V Lo, Schwindt E (2014) Introduced and native species on rocky shore

macroalgal assemblages: Zonation patterns, composition and diversity. Aquat Bot

112:57–65

Reid AM, Morin L, Downey PO, French K, Virtue JG (2009) Does invasive plant management aid

the restoration of natural ecosystems? Biol Conserv 142:2342–2349

Remeš V (2003) Effects of exotic habitat on nesting success, territory density, and settlement

patterns in the blackcap (Sylvia atricapilla). Conserv Biol 17:1127–1133

Robertson BA, Hutto RL (2006) A framework for understanding ecological traps and an

evaluation of existing evidence. Ecology 87:1075–1085

Robertson BA, Rehage JS, Sih A (2013) Ecological novelty and the emergence of evolutionary

traps. Trends Ecol Evol 28:552–560

Rodewald AD, Shustack DP, Hitchcock LE (2010) Exotic shrubs as ephemeral ecological traps for

nesting birds. Biol Invasions 12:33–39

Sampson J, Easton A, Singh M (2014) Port Phillip Bay. In: Wolanski E (ed) Estuaries of Australia

in 2050 and Beyond. Springer, p 49–68

Schaffelke B, Hewitt CL (2007) Impacts of introduced seaweeds. Bot Mar 50:397–417

Sih A, Ferrari MCO, Harris DJ (2011) Evolution and behavioural responses to human-induced

rapid environmental change. Evol Appl 4:367–387

South PM, Thomsen MS (2016) The ecological role of invading Undaria pinnatifida: an

experimental test of the driver–passenger models. Mar Biol 163:175

Thomsen MS, Wernberg T, Tuya F, Silliman BR (2009) Evidence for impacts of nonindigenous

macroalgae: a meta-analysis of experimental field studies. J Phycol 45:812–819

Trammell MA, Butler JL (1995) Effects of exotic plants on native ungulate use of habitat. J Wildl

Manage 59:808–816

Valentine JP, Johnson CR (2003) Establishment of the introduced kelp Undaria pinnatifida in

Tasmania depends on disturbance to native algal assemblages. J Exp Mar Biol Ecol

295:63–90

Valentine LE, Roberts B, Schwarzkopf LIN (2007) Mechanisms driving avoidance of non-native

plants by lizards. J Appl Ecol 44:228–237

127

Wickham H (2009) ggplot2: elegant graphics for data analysis. Springer-Verlag New York

Willis T, Anderson M (2003) Structure of cryptic reef fish assemblages: relationships with

habitat characteristics and predator density. Mar Ecol Prog Ser 257:209–221

Willis TJ, Babcock RC (2000) A baited underwater video system for the determination of

relative density of carnivorous reef fish. Mar Freshw Res 51:755–763

Wong BBM, Candolin U (2014) Behavioral responses to changing environments. Behav Ecol

26:665–673

Wright JT, Byers JE, DeVore JL, Sotka EE (2014) Engineering or food? mechanisms of facilitation

by a habitat‐forming invasive seaweed. Ecology 95:2699–2706

Zavaleta ES, Hobbs RJ, Mooney HA (2001) Viewing invasive species removal in a whole-

ecosystem context. Trends Ecol Evol 16:454–459

128

CHAPTER SIX: GENERAL DISCUSSION

Incorporating data on individual-level responses to environmental change can fundamentally

improve our understanding of associations between novel habitats and wild animal

populations, and especially the role of novel habitats for population persistence in a source–

sink metapopulation framework. The ecological trap literature highlights the importance of

combining individual- and population-level data to distinguish between attractive population

sources, unattractive population sources, attractive population sinks and unattractive

population sinks.

HOW PREVALENT ARE ECOLOGICAL TRAPS IN THE MARINE ENVIRONMENT?

Ecological traps are probably more common than the current evidence indicates. Clear

demonstrations of traps are rare, but it would be surprising if this reflects consistently

adaptive habitat selection by animals in novel and degraded environments (Sih et al. 2011,

Robertson et al. 2013, Sih 2013). More likely it reflects a lack of research effort on the topic,

with the vast majority of habitat association studies focusing on community-level survey data.

As a result, most researchers are not obtaining suitable estimates of habitat preference, nor

studying the fitness consequences of those habitat selection decisions. Furthermore, Battin

(2004) noted that researchers in terrestrial systems frequently presented findings that were

consistent with the existence of an ecological trap but did not interpret the findings within the

ecological trap framework. Similar cases exist within the marine literature: several of the

examples I cite in this thesis as evidence for marine ecological traps were not interpreted as

such by the authors, including flatfish and crabs selecting oil-contaminated sediment (Moles et

al. 1994, Moles & Norcross 1998, Moles & Stone 2002), bivalves selecting sediment modified

by Caulerpa taxifolia (Wright & Gribben 2008, Gribben et al. 2009a, 2009b, Byers et al. 2010),

and fish recruits avoiding reefs with resident native predators but not those with invasive

lionfish (Benkwitt 2017). In some cases, authors referred to probable ecological trap habitats

as ‘population sinks’ (e.g. Gribben et al. 2009b), consistent with greater awareness of the

source-sink framework, or otherwise noted likely deleterious population effects. However, as

ecological trap models demonstrate, an attractive population sink (ecological trap) has more

severe implications for population persistence than a ‘traditional’ non-preferred population

129

sink as it actively draws in individuals from surrounding higher quality areas (Battin 2004,

Abrams et al. 2012, Hale et al. 2015). In other words, non-preferred population sinks are likely

to conform to the ideal free distribution (with each individual optimising its fitness within

available habitats: Fretwell & Lucas 1969), while ecological traps do not, causing fauna

populations to underperform relative to expectations.

Additional evidence for the likely prevalence of ecological traps comes from the rate at which

researchers looking for traps find them. While it may be impossible to account for bias in

research effort (i.e. researchers selecting systems that are likely to be vulnerable to ecological

traps, or setting aside the ecological trap hypothesis when no evidence is found), there are

notably few cases of researchers looking for traps and not finding them (but see Dempster et

al. 2011; Reubens et al. 2013). In Chapters 3-5, I investigated three novel marine systems that I

considered to be vulnerable to the formation of ecological traps. In Chapter 3 I found some

limited evidence that offspring of Atlantic cod from an area of high salmon farming density

produced smaller eggs and larvae, potentially leading to reduced fitness. In Chapter 4, an

absence of small Asterias seastars within mussel farms was consistent with either high

mortality or emigration in this habitat, while adult Asterias may be avoiding a habitat rich in

food but also potential predators. In Chapter 5, I concluded that juvenile and adult weedfish

and weed whiting are probably making broadly adaptive habitat-selection decisions in Undaria

habitats (consistent with ideal free distribution), but that there are possible Allee effects in

small patches and that loss of canopy during summer senescence might be costly for weedfish

larvae that recruit to the Undaria canopy in spring if the value of the Undaria canopy is

outweighed by its eventual disappearance.

Intriguingly, in all cases, my findings indicated that fauna associated with novel marine habitats

were most likely either making adaptive habitat selection decisions or falling into ecological

traps. I found no evidence that any of these novel habitats were likely to function as

perceptual traps (Fig. 6.1). This is consistent with the current body of literature, in which novel

habitats are more commonly found to be ecological traps than perceptual traps (but see

Patten and Kelly 2010). I see three possible reasons for this trend: (1) novel habitats may, on

average, offer poorer fitness outcomes than natural habitats for native species due to a lack of

evolutionary adaptation to the novel habitat and likely disadvantages relative to non-native

competitors or predators (Byers 2002, Crooks et al. 2011); (2) if a novel habitat appears similar

to an evolutionarily-familiar habitat, individuals will likely attempt to utilise it; and (3) an

abundance of a native species in a novel habitat is more likely to arouse the interest of

130

ecologists than an absence, as we are accustomed to native taxa avoiding novel habitats and

implicitly assume this to be a broadly adaptive response.

Figure 6.1. Conceptual diagram showing speculative positioning of responses of focal species

to novel habitats, indicating potential differences in responses and fitness outcomes across life

stages (Asterias and Coscinasterias at shellfish farms; Heteroclinus and Neoodax in Undaria

canopy; Atlantic cod at salmon farms) placed within the ecological trap framework. Species

placed within the green area exhibit adaptive habitat selection responses. Juvenile

Coscinasterias were not observed inside the shellfish farm, and it is unclear whether this

reflects habitat preferences or high mortality.

HOW DOES THE ECOLOGICAL TRAP CONCEPT RELATE TO THE ATTRACTION–PRODUCTION

DEBATE?

Ultimately, the ecological trap and attraction–production concepts are directed toward the

same objective: understanding the role of a given habitat for population growth, with a focus

on ecologically-novel habitats. The ecological trap concept pre-dates both source–sink models

and the attraction–production debate (Dwernychuk & Boag 1972, Gates & Gysel 1978), but the

source–sink model experienced more rapid uptake among ecologists, while the attraction–

production debate arose in response to questions about the role of artificial reefs for

production of targeted fish species during the 1980s (Lindberg 1997).

131

Uptake of the ecological trap concept has been slow in the fisheries management literature,

but in my view, it offers something that the attraction–production distinction does not. The

typical application of the attraction–production models looks at two questions: are fish

attracted to the given habitat, and does the existence of this habitat lead to more fish being

produced? Under the attraction hypothesis, fish move from surrounding areas to the new

habitat, with no net increase in production. Under the production hypothesis, the new habitat

increases the carrying capacity of a habitat-limited environment and leads to increased fish

production. These two hypothesis are not mutually exclusive and likely co-occur in many cases,

but the limitation of the attraction–production framework is that researchers are liable to miss

a third possibility – that poor individual fitness outcomes in an attractive novel habitat will

create an ecological trap that leads to serious negative effects on fish production, such that it

would be better for the habitat not to exist at all (Reubens et al. 2014). The ecological trap

assessment framework (see Fig. 1.1) accounts for this possibility, and its corollary, an

unattractive but high quality habitat that is underutilised (termed a perceptual trap by Patten

and Kelly 2010). Gutzler et al. (2015) considered both frameworks in their assessment of the

role of casitas for targeted lobsters as individuals and as a population, while Reubens et al.

(2014) proposed a hybrid ‘attraction–ecological trap–production’ framework for assessing the

role of artificial reefs for fish populations, within which researchers ask the following

questions:

(1) Does attraction towards the artificial reef occur?

(2) If there is attraction, is it age-group specific?

(3) Which mechanisms or processes influence production in the ecosystem investigated,

and are these mechanisms/processes affected by the artificial reef?

(4) What is the species-specific behavioural ecology in this ecosystem?

(5) If there is production, is it sufficient to offset associated fishing mortality?

This integration of two assessment frameworks, with two largely independent bodies of

research focusing on terrestrial birds (ecological traps) and artificial reefs (attraction–

production), can also be expressed as a conceptual diagram (Fig. 6.2). The four scenarios

outlined in the diagram are extreme outcomes, and intermediate outcomes are possible. For

example, newly-created, unattractive, low quality habitats may have a small positive effect on

production in habitat-limited systems if they offer a sufficient fitness value for individuals

spilling over from densely-populated preferred habitats (adaptively following the ideal free

distribution).

132

Relative fitness outcomes for individuals

Good Poor

Relative habitat attractiveness

Attractive Source ↑ Ecological trap ↓

Unattractive Perceptual trap — Sink —

Figure 6.2. Conceptual representation of the ecological trap framework applied to an

assessment of the role of artificial reefs for fish production. “↑” indicates an increase in fish

production (e.g. fish utilise a viable new habitat in a habitat-limited system), “↓” indicates a

decrease in fish production (e.g. fish are attracted to a habitat that offers low availability of

food or mates, or confers high mortality), and “—" indicates a neutral effect on fish production

(as the new habitat is not utilised).

BETTER LEFT ALONE? LESSONS FOR THE MANAGEMENT OF HABITAT-FORMING INVASIVE

SPECIES

Most human-mediated species introductions do not establish, but of those that do, a few will

become invasive (Richardson & Pyšek 2012). Such invasions may be an important factor in

native species extinctions worldwide (Clavero & García-Berthou 2005), and some particularly

influential invasive ecosystem engineers have driven phase shifts in natural ecosystems

(Crooks 2002). Management of non-native arrivals often begins with immediate eradication

attempts, but once the window for effective eradication has passed, management strategies

must be reassessed. In cases where the invader is an opportunistic passenger rather than a

driver of ecological change, control measures may be neither effective nor desirable (Zavaleta

et al. 2001, Didham et al. 2005, MacDougall & Turkington 2005). This may be especially true

where invasive habitat-forming species replace native habitat-formers and native animals

come to rely on the invasive habitat (Zavaleta et al. 2001). However, a major limitation in our

understanding of the impacts of invasive habitat-formers is a scarcity of data on habitat

preferences and fitness outcomes for native fauna inhabiting invasive habitats.

133

Figure 6.3. Current distribution of Undaria pinnatifida within Port Phillip Bay in 2017. Red

markers denote documented established populations on rocky reefs or marine infrastructure;

blue markers denote recent infestations. A satellite population has also established at Apollo

Bay (97 km SW) and another was discovered and eradicated at Flinders (45 km SE). Both

expansions were likely the result of boat transport.

The invasive kelp Undaria pinnatifida, a habitat-forming ecosystem engineer, provides an

illustrative case study. Despite being one of only two macroalgal taxa included on the IUCN’s

list of 100 of the world’s worst invasive alien species (Lowe et al. 2000), the subsequent

literature on invasive Undaria populations in Australia, New Zealand and Europe indicates that

this species is a passenger of general ecological degradation, taking advantage of rocky reef

substrate left vacant by declining native kelp canopy but only persisting at low population

densities in intact native macroalgal habitats (Valentine & Johnson 2003, Edgar et al. 2004,

South & Thomsen 2016, Kriegisch et al. 2016, de Leij et al. 2017, South et al. 2017). Effects of

the invasive canopy on native faunal biodiversity are less well known, but in South America,

invasive Undaria cover was associated with elevated macroinvertebrate species richness,

diversity and abundance compared to plots where early stage Undaria recruits were removed

to prevent canopy formation (Irigoyen et al. 2011). Similarly, work in Australia found that

invasive Undaria and native Ecklonia radiata holdfasts contain comparable abundance and

species richness of epibenthic invertebrates (Howland 2012), indicating that Undaria is likely to

have neutral or positive effects on biodiversity in degraded areas. In Chapter Five, I

134

demonstrated that the presence of Undaria on degraded reefs is associated with elevated fish

biodiversity and provides viable habitat for a variety of endemic fish species in the months

leading up to spring spawning. This work is a rare assessment of outcomes for native species

utilising invasive habitats.

Despite this growing body of evidence that Undaria canopy is utilised by native fish and

invertebrates, coupled with broadly benign effects on native macroalgal communities,

authorities are undertaking ongoing eradication efforts at local sites. Such efforts are unlikely

to be effective in the long term (e.g. Hewitt et al. 2005), and as they necessarily occur during

the Undaria growing season in spring and summer, are likely to displace resident macroalgae-

associated fish in spawning condition. Accordingly, I recommend that eradication efforts cease

in locations where Undaria is established (see Fig. 6.3 for distribution in Port Phillip Bay), and

instead focus on eradicating satellite populations using approaches modelled on successful

removals of localised infestations at Flinders (Parry & Cohen 2001) and the Chatham Islands

(Wotton et al. 2004). Where the invader is established, management efforts may be better

directed toward promoting the recovery of Ecklonia and other native macrophytes that

compete with Undaria (Ling et al. 2009, Carnell & Keough 2016, Kriegisch et al. 2016), for

example by investigating methods to reduce sea urchin grazing pressure (e.g. Pert et al. 2018)

and other stressors.

Farming of Undaria may also be considered in areas where the invader is already abundant

and where farming is unlikely to increase the severity of any ecological impacts. Accordingly,

the New Zealand government allowed farming of Undaria in several heavily infested areas in

2012, with interest along similar lines in Europe (Peteiro et al. 2016). Farming would greatly

increase propagule pressure locally if the reproductive plants reach maturity before

harvesting, so the potential for increased local impacts as well as the risk of accelerated range

expansion should be considered carefully. Undaria spores do not naturally disperse large

distances (Forrest et al. 2000), but caution is required given the likelihood of detached mature

plants drifting large distances (Russell et al. 2008), and a history of long distance dispersal

mediated by maritime transport.

Undaria may not be alone in being more benign than initially feared. A recent global review of

non-native seaweeds found that introduced seaweeds generally had negative effects on

macroalgal diversity, but neutral or positive effects on faunal communities (Thomsen et al.

2016). In many cases, there is evidence that the spread of non-native taxa is facilitated by prior

declines in native cover (Ceccherelli et al. 2014, Gennaro & Piazzi 2014, Luigi & Giulia 2017).

135

More research is needed to understand the value of non-native habitats for native fauna, and

to avoid conflating effects of non-native habitat-formers with negative effects of other forms

of HIREC.

WHERE TO NEXT FOR THIS WORK?

The meta-analysis in Chapter Two revealed substantial gaps in our understanding of the

impacts of marine and freshwater aquaculture on wildlife. In particular, studies that rigorously

demonstrate attractive or repulsive effects of farms are rare, as are studies quantifying direct

or indirect fitness effects on wild taxa. Chapter Three addressed one of these key knowledge

gaps—effects of farm-association on reproductive fitness of wild fish—and revealed some

evidence for negative effects, but the study was nonetheless limited in spatial replication,

experimental power and duration of larval rearing. Ideally, the geographical extent of this work

should be expanded, with additional resources to support larger sample sizes, and/or raise

larvae to juvenile or adult stages to test for delayed effects of farm-association on offspring.

Adult mortality is also important in assessing the impacts of aquaculture on wild fish

populations, especially where farms bring elevated infection, predation and fishing mortality

rates. Long term acoustic tracking of adult fish within farm-affected or non-affected areas,

replicated across multiple locations, may be the most viable approach for quantifying mortality

and emigration in semi-enclosed systems such as fjords (Olsen & Moland 2011, Olsen et al.

2012, Fernández-Chacón et al. 2015).

In Chapter Four, I showed that despite an underlying attraction to shellfish prey and benefits

of high food availability for individuals residing inside shellfish farms, the invasive seastar

Asterias amurensis was not more abundant inside farms. Predator avoidance experiments

together with a truncated size distribution indicated that the threat of predation by the native

seastar Coscinasterias muricata may be preventing the invader from exploiting the food-rich

resource. However, I was unable to clearly partition the effects of behavioural predator

avoidance and mortality. A follow-up study quantifying relative predation risk for tethered

seastars, as well as mass mark-recapture of seastars at locations inside and outside farms

(Loosanoff 1937, Lamare et al. 2009, Barahona & Navarrete 2010) may shed light on the

relative importance of these processes.

136

In Chapter Five, I provided estimates of habitat preference (via a controlled habitat choice

experiment and recruitment patterns to mesocosm reefs) and relative fitness (via condition

and fecundity metrics) for native reef fish inhabiting native and invasive kelp canopy, but

further work is required to quantify patterns of movement, site fidelity and mortality,

especially during the summer senescence of the Undaria canopy. Acoustic tags can be

appropriate for tracking movement of large-bodied taxa (Parsons et al. 2003), while site

fidelity of smaller site-attached taxa may be estimated by a mark-recapture study using visible

implant elastomer tags (Willis et al. 2001). However, a pilot mark-recapture study of the

common weedfish Heteroclinus perspicillatus translocated to mesocosm reefs in Chapter Five

indicated low retention rates, with simultaneous high immigration of new juveniles and adults

indicating that this species is highly transient, consistent with previous work tracking temporal

trends in fish communities on artificial seaweed beds (Jenkins & Sutherland 1997). Novel

techniques may be required to track movement and mortality of small cryptobenthic fishes

during and after the Undaria growing season. The little weed whiting Neoodax balteatus may

be a better focal species for a study of survivorship rates in native and invasive kelp habitats,

as dominant terminal phase males maintain harems and appear to be more strongly site-

attached (pers. obs.), such that it may be reasonable to infer mortality rates from the loss of

dominant individuals, although a full assessment of the ecological role of Undaria patches

should consider a wider suite of associated vertebrate and invertebrate fauna.

CONCLUSION

A reliance on population- and community-level metrics can result in a failure to detect

important individual-level processes that influence population growth and persistence, such as

habitat selection, survivorship and reproductive output. Wherever possible, researchers

assessing the role of novel or degraded habitats as population sources or sinks should

complement population- and community-level metrics with direct or indirect measures of

habitat preference and individual fitness. Applying the ecological trap assessment framework

can also improve our understanding of source–sink dynamics by accounting for maladaptive

habitat selection decisions that can lead to attractive population sinks or underutilised

population sources. Correctly characterising the value of specific habitats for fauna will assist

in targeting of management measures, for example by: (1) protecting attractive, highly

productive habitats, (2) disarming ecological traps by removing attractive cues or drivers of

137

poor fitness outcomes, (3) investigating ecological engineering approaches to increase

utilisation of perceptual traps by native fauna, or (4) leveraging ecological traps to assist in the

control of non-native fauna.

REFERENCES

Abrams PA, Ruokolainen L, Shuter BJ, McCann KS (2012) Harvesting creates ecological traps:

consequences of invisible mortality risks in predator-prey metacommunities. Ecology

93:281–293

Barahona M, Navarrete SA (2010) Movement patterns of the seastar Heliaster helianthus in

central Chile: relationship with environmental conditions and prey availability. Mar Biol

157:647–661

Battin J (2004) When good animals love bad habitats: ecological traps and the conservation of

animal populations. Conserv Biol 18:1482–1491

Benkwitt CE (2017) Predator effects on reef fish settlement depend on predator origin and

recruit density. Ecology 98:896–902

Byers JE (2002) Impact of non-indigenous species on natives enhanced by anthropogenic

alteration of selection regimes. Oikos 97:449–458

Byers JE, Wright JT, Gribben PE (2010) Variable direct and indirect effects of a habitat-

modifying invasive species on mortality of native fauna. Ecology 91:1787–1798

Carnell PE, Keough MJ (2016) The influence of herbivores on primary producers can vary

spatially and interact with disturbance. Oikos 125:1273–1283

Ceccherelli G, Pinna S, Cusseddu V, Bulleri F (2014) The role of disturbance in promoting the

spread of the invasive seaweed Caulerpa racemosa in seagrass meadows. Biol Invasions

16:2737–2745

Clavero M, García-Berthou E (2005) Invasive species are a leading cause of animal extinctions.

Trends Ecol Evol 20:110

Crooks JA (2002) Characterizing ecosystem-level consequences of biological invasions: the role

of ecosystem engineers. Oikos 97:153–166

Crooks JA, Chang AL, Ruiz GM (2011) Aquatic pollution increases the relative success of

invasive species. Biol Invasions 13:165–176

Dempster T, Sanchez-Jerez P, Fernandez-Jover D, Bayle-Sempere JT, Nilsen R, Bjørn P-A, Uglem

I (2011) Proxy measures of fitness suggest coastal fish farms can act as population

138

sources and not ecological traps for wild gadoid fish. PLoS One 6:e15646

Didham RK, Tylianakis JM, Hutchison MA, Ewers RM, Gemmell NJ (2005) Are invasive species

the drivers of ecological change? Trends Ecol Evol 20:470–474

Dwernychuk LW, Boag DA (1972) Ducks nesting in association with gulls - an ecological trap?

Can J Zool 50:559–563

Edgar GJ, Barrett NS, Morton AJ, Samson CR (2004) Effects of algal canopy clearance on plant,

fish and macroinvertebrate communities on eastern Tasmanian reefs. J Exp Mar Bio Ecol

312:67–87

Fernández-Chacón A, Moland E, Espeland SH, Olsen EM (2015) Demographic effects of full vs.

partial protection from harvesting: inference from an empirical before-after control-

impact study on Atlantic cod. J Appl Ecol 52:1206–1215

Forrest BM, Brown SN, Taylor MD, Hurd CL, Hay CH (2000) The role of natural dispersal

mechanisms in the spread of Undaria pinnatifida (Laminariales, Phaeophyceae).

Phycologia 39:547–553

Fretwell SD, Lucas HL (1969) On territorial behavior and other factors influencing habitat

distribution in birds. Acta Biotheor 19:16–36

Gates JE, Gysel LW (1978) Avian nest dispersion and fledging success in field-forest ecotones.

Ecology 59:871–883

Gennaro P, Piazzi L (2014) The indirect role of nutrients in enhancing the invasion of Caulerpa

racemosa var cylindracea. Biol Invasions 16:1709–1717

Gribben PE, Wright JT, O’Connor WA, Doblin MA, Eyre B, Steinberg PD (2009a) Reduced

performance of native infauna following recruitment to a habitat-forming invasive

marine alga. Oecologia 158:733–745

Gribben PE, Wright JT, O’Connor WA, Steinberg P (2009b) Larval settlement preference of a

native bivalve: the influence of an invasive alga versus native substrata. Aquat Biol 7:217–

227

Gutzler BC, Butler MJ IV, Behringer DC (2015) Casitas: a location-dependent ecological trap for

juvenile Caribbean spiny lobsters, Panulirus argus. ICES J Mar Sci 72:i177–i184.

Hale R, Treml EA, Swearer SE (2015) Evaluating the metapopulation consequences of

ecological traps. Proc R Soc London B Biol Sci 282:20142930

Hewitt CL, Campbell ML, McEnnulty F, Moore KM, Murfet NB, Robertson B, Schaffelke B (2005)

Efficacy of physical removal of a marine pest: the introduced kelp Undaria pinnatifida in a

Tasmanian Marine Reserve. Biol Invasions 7:251–263

Howland S (2012) Variation and impacts on biodiversity in kelp holdfast assemblages: a

139

comparison of the introduced Undaria pinnatifida and the indigenous Ecklonia radiata.

University of Melbourne

Irigoyen AJ, Trobbiani G, Sgarlatta MP, Raffo MP (2011) Effects of the alien algae Undaria

pinnatifida (Phaeophyceae, Laminariales) on the diversity and abundance of benthic

macrofauna in Golfo Nuevo (Patagonia, Argentina): potential implications for local food

webs. Biol Invasions 13:1521–1532

Jenkins G., Sutherland C. (1997) The influence of habitat structure on nearshore fish

assemblages in a southern Australian embayment: colonisation and turnover rate of

fishes associated with artificial macrophyte beds of varying physical structure. J Exp Mar

Bio Ecol 218:103–125

Kriegisch N, Reeves S, Johnson CR, Ling SD, Jernakoff P (2016) Phase-shift dynamics of sea

urchin overgrazing on nutrified reefs (BD Russell, Ed.). PLoS One 11:e0168333

Lamare MD, Channon T, Cornelisen C, Clarke M (2009) Archival electronic tagging of a

predatory sea star - testing a new technique to study movement at the individual level. J

Exp Mar Bio Ecol 373:1–10

Leij R de, Epstein G, Brown MP, Smale DA (2017) The influence of native macroalgal canopies

on the distribution and abundance of the non-native kelp Undaria pinnatifida in natural

reef habitats. Mar Biol 164:156

Lindberg WJ (1997) Can science resolve the attraction-production issue? Fisheries 22:10–13

Ling SD, Johnson CR, Frusher SD, Ridgway KR (2009) Overfishing reduces resilience of kelp beds

to climate-driven catastrophic phase shift. Proc Natl Acad Sci USA 106:22341–22345

Loosanoff VL (1937) Use of nile blue sulfate in marking starfish. Science 85:412

Lowe S, Browne M, Boudjelas S, Poorter M De (2000) 100 of the world’s worst invasive alien

species: a selection from the global invasive species database. Auckland

Luigi P, Giulia C (2017) Eutrophication affects the resistance of fucoids to an introduced alga

spread. Mar Environ Res 129:189–194

MacDougall AS, Turkington R (2005) Are invasive species the drivers of change or passengers

of change in degraded ecosystems? Ecology 86:42–55

Moles A, Norcross BL (1998) Effects of oil-laden sediments on growth and health of juvenile

flatfishes. Can J Fish Aquat Sci 55:605–610

Moles A, Rice S, Norcross BL (1994) Non-avoidance of hydrocarbon laden sediments by

juvenile flatfishes. Netherlands J Sea Res 32:361–367

Moles A, Stone RP (2002) Habitat preferences of juvenile tanner and red king crabs: substrate

and crude oil. In: Paul AJ, Dawe EG, Elner R, Jamieson GS, Kruse GH, Otto RS, Sainte-Marie

140

B, Shirley TC, Woodby D (eds) Crabs in Cold Water Regions: Biology, Management, and

Economics. University of Alaska Sea Grant, p 631–644

Olsen EM, Heupel MR, Simpfendorfer CA, Moland E (2012) Harvest selection on Atlantic cod

behavioral traits: implications for spatial management. Ecol Evol 2:1549–1562

Olsen EM, Moland E (2011) Fitness landscape of Atlantic cod shaped by harvest selection and

natural selection. Evol Ecol 25:695–710

Parry GD, Cohen BF (2001) Exotic species established in Western Port, including an assessment

of the status of the exotic species Corbula gibba, Alexandrium spp, Gymnodinium spp and

Undaria pinnatifida. Queenscliff

Parsons DM, Babcock RC, Hankin RKS, Willis TJ, Aitken JP, O’dor RK, Jackson GD (2003) Snapper

Pagrus auratus (Sparidae) home range dynamics: acoustic tagging studies in a marine

reserve. Mar Ecol Prog Ser 262:253–265

Patten MA, Kelly JF (2010) Habitat selection and the perceptual trap. Ecol Appl 20:2148–2156

Pert CG, Swearer SE, Dworjanyn SA, Turchini GM, Francis DS and Dempster T (2018) Barrens of

gold: gonad conditioning of an overabundant sea urchin. Aquac Environ Interact 10:345–

361

Peteiro C, Sánchez N, Martínez B (2016) Mariculture of the Asian kelp Undaria pinnatifida and

the native kelp Saccharina latissima along the Atlantic coast of southern Europe: an

overview. Algal Res 15:9–23

Reubens JT, Degraer S, Vincx M (2014) The ecology of benthopelagic fishes at offshore wind

farms: a synthesis of 4 years of research. Hydrobiologia 727:121–136

Reubens JT, Vandendriessche S, Zenner AN, Degraer S, Vincx M (2013) Offshore wind farms as

productive sites or ecological traps for gadoid fishes? Impact on growth, condition index

and diet composition. Mar Environ Res 90:66–74

Richardson DM, Pyšek P (2012) Naturalization of introduced plants: ecological drivers of

biogeographical patterns. New Phytol 196:383–396

Robertson BA, Rehage JS, Sih A (2013) Ecological novelty and the emergence of evolutionary

traps. Trends Ecol Evol 28:552–560

Russell LK, Hepburn CD, Hurd CL, Stuart MD (2008) The expanding range of Undaria pinnatifida

in southern New Zealand: distribution, dispersal mechanisms and the invasion of wave-

exposed environments. Biol Invasions 10:103–115

Sih A (2013) Understanding variation in behavioural responses to human-induced rapid

environmental change: a conceptual overview. Anim Behav 85:1077–1088

Sih A, Ferrari MCO, Harris DJ (2011) Evolution and behavioural responses to human-induced

141

rapid environmental change. Evol Appl 4:367–387

South PM, Floerl O, Forrest BM, Thomsen MS (2017) A review of three decades of research on

the invasive kelp Undaria pinnatifida in Australasia: an assessment of its success, impacts

and status as one of the world’s worst invaders. Mar Environ Res 131: 243–257

South PM, Thomsen MS (2016) The ecological role of invading Undaria pinnatifida: an

experimental test of the driver–passenger models. Mar Biol 163:175

Thomsen MS, Wernberg T, South PM, Schiel DR (2016) Non-native seaweeds drive changes in

marine coastal communities around the world. In: Hu Z-M, Fraser C (eds) Seaweed

Phylogeography: Adaptation and Evolution of Seaweeds under Environmental Change.

Springer Netherlands, Dordrecht, p 147–185

Valentine JP, Johnson CR (2003) Establishment of the introduced kelp Undaria pinnatifida in

Tasmania depends on disturbance to native algal assemblages. J Exp Mar Bio Ecol

295:63–90

Willis TJ, Parsons DM, Babcock RC (2001) Evidence for long‐term site fidelity of snapper

(Pagrus auratus) within a marine reserve. New Zeal J Mar Freshw Res 35:581–590

Wotton DM, O’Brien C, Stuart MD, Fergus DJ (2004) Eradication success down under: heat

treatment of a sunken trawler to kill the invasive seaweed Undaria pinnatifida. Mar

Pollut Bull 49:844–9

Wright JI, Gribben PE (2008) Predicting the impact of an invasive seaweed on the fitness of

native fauna. J Appl Ecol 45:1540–1549

Zavaleta ES, Hobbs RJ, Mooney HA (2001) Viewing invasive species removal in a whole-

ecosystem context. Trends Ecol Evol 16:454–459

142

APPENDICES

Appendix 2.1. List of 191 articles included in systematic review and meta-analysis.

Abaad, M., Tuset, V.M., Montero, D., Lombarte, A., Otero-Ferrer, J.L. and Haroun, R. (2016) Phenotypic

plasticity in wild marine fishes associated with fish-cage aquaculture. Hydrobiologia 765, 343–358.

Adámek, Z., Kortan, D., Lepič, P. and Andreji, J. (2003) Impacts of otter (Lutra lutra L.) predation on

fishponds: A study of fish remains at ponds in the Czech Republic. Aquaculture International 11,

389–396.

Anyango, J.O., Mlewa, C.M. and Mwaluma, J. (2017) Abundance, diversity and trophic status of wild fish

around seaweed farms in Kibuyuni, South Coast Kenya. International Journal of Fisheries and

Aquatic Studies 5, 440–446.

Arechavala-Lopez, P., Izquierdo-Gomez, D., Uglem, I. and Sanchez-Jerez, P. (2015a) Aggregations of

bluefish Pomatomus saltatrix (L.) at Mediterranean coastal fish farms: seasonal presence, daily

patterns and influence of farming activity. Environmental Biology of Fishes 98, 499–510.

Arechavala-Lopez, P., Sæther, B.-S., Marhuenda-Egea, F., Sanchez-Jerez, P. and Uglem, I. (2015b)

Assessing the influence of salmon farming through total lipids, fatty acids, and trace elements in

the liver and muscle of wild saithe Pollachius virens. Marine and Coastal Fisheries 7, 59–67.

Arechavala-Lopez, P., Sanchez-Jerez, P., Bayle-Sempere, J., Fernandez-Jover, D., Martinez-Rubio, L., Lopez-

Jimenez, J.A. and Martinez-Lopez, F.J. (2011) Direct interaction between wild fish aggregations at

fish farms and fisheries activity at fishing grounds: a case study with Boops boops. Aquaculture

Research 42, 996–1010.

Arechavala-Lopez, P., Uglem, I., Sanchez-Jerez, P., Fernandez-Jover, D., Bayle-Sempere, J.T. and Nilsen, R.

(2010) Movements of grey mullet Liza aurata and Chelon labrosus associated with coastal fish

farms in the western Mediterranean Sea. Aquaculture Environment Interactions 1, 127–136.

Avery, M.L., Eiselman, D.S., Young, M.K., Humphrey, J.S. and Decker, D.G. (1999) Wading bird predation

at tropical aquaculture facilities in central Florida. North American Journal of Aquaculture 61, 64–

69.

Bacher, K. and Gordoa, A. (2016) Does marine fish farming affect local small-scale fishery catches? A case

study in the NW Mediterranean Sea. Aquaculture Research 47, 2444–2454.

Bacher, K., Gordoa, A. and Sague, O. (2012) Spatial and temporal extension of wild fish aggregations at

Sparus aurata and Thunnus thynnus farms in the north-western Mediterranean. Aquaculture

Environment Interactions 2, 239–252.

Bacher, K., Gordoa, A. and Sagué, O. (2015) Feeding activity strongly affects the variability of wild fish

aggregations within fish farms: a sea bream farm as a case study. Aquaculture Research 46, 552–

564.

Bagdonas, K., Humborstad, O.-B. and Løkkeborg, S. (2012) Capture of wild saithe (Pollachius virens) and

cod (Gadus morhua) in the vicinity of salmon farms: three pot types compared. Fisheries Research

134–136, 1–5.

Ballester-Moltó, M., Sanchez-Jerez, P. and Aguado-Giménez, F. (2017) Consumption of particulate wastes

derived from cage fish farming by aggregated wild fish. An experimental approach. Marine

Environmental Research 130, 166–173.

Ballester-Moltó, M., Sanchez-Jerez, P., Garcia-Garcia, B. and Aguado-Giménez, F. (2015) Husbandry and

environmental conditions explain temporal variability of wild fish assemblages aggregated around

a Mediterranean fish farm. Aquaculture Environment Interactions 7, 193–203.

Baltrūnaitė, L. (2009) Diet of otters in fish farms in Lithuania. Acta Zoologica Lituanica 19, 182–187.

Barlow, C.G. and Bock, K. (1984) Predation of fish in farm dams by cormorants, Phalacrocorax spp.

Australian Wildlife Research 11, 559–566.

Bechard, M.J. and Márquez-Reyes, C. (2003) Mortality of wintering ospreys and other birds at aquaculture

facilities in Colombia. Journal of Raptor Research 37, 292–298.

143

Becker, B.H., Press, D.T. and Allen, S.G. (2011) Evidence for long-term spatial displacement of breeding

and pupping harbour seals by shellfish aquaculture over three decades. Aquatic Conservation-

Marine and Freshwater Ecosystems 21, 247–260.

Belant, J.L., Tyson, L.A. and Mastrangelo, P.A. (2000) Effects of lethal control at aquaculture facilities on

populations of piscivorous birds. Wildlife Society Bulletin 28, 379–384.

Bergman, K.C., Svensson, S. and Ohman, M.C. (2001) Influence of algal farming on fish assemblages.

Marine Pollution Bulletin 42, 1379–1389.

Beynon, J.L., Hutchins, D.L., Rubino, A.J., Lawrence, A.L. and Chapman, B.R. (1981) Nocturnal activity of

birds on shrimp mariculture ponds. Journal of the World Mariculture Society 12, 63–70.

Bjordal, Å. and Johnstone, A.D.F. (1993) Local movements of saithe (Pollachius virens L.) in the vicinity of

fish farm cages. In: ICES Mar Sci Symp, Vol. 196. pp 143–146.

Björklund, H., Bondestam, J. and Bylund, G. (1990) Residues of oxytetracycline in wild fish and sediments

from fish farms. Aquaculture 86, 359–367.

Bjørn, P.-A., Uglem, I., Kerwath, S., Sæther, B.-S. and Nilsen, R. (2009) Spatiotemporal distribution of

Atlantic cod (Gadus morhua L.) with intact and blocked olfactory sense during the spawning season

in a Norwegian fjord with intensive salmon farming. Aquaculture 286, 36–44.

Bjørn, P.A., Finstad, B. and Kristoffersen, R. (2001) Salmon lice infection of wild sea trout and Arctic char

in marine and freshwaters: the effects of salmon farms. Aquaculture Research 32, 947–962.

Blackwell, B.F., Dolbeer, R.A. and Tyson, L.A. (2000) Lethal control of piscivorous birds at aquaculture

facilities in the northeast United States: effects on populations. North American Journal of

Aquaculture 62, 300–307.

Bogdanović, T., Šimat, V., Frka-Roić, A. and Marković, K. (2012) Development and application of quality

index method scheme in a shelf-life study of wild and fish farm affected bogue (Boops boops, L.).

Journal of Food Science 77, S99–S106.

Bonizzoni, S., Furey, N.B., Pirotta, E., Valavanis, V.D., Wuersig, B. and Bearzi, G. (2014) Fish farming and its

appeal to common bottlenose dolphins: modelling habitat use in a Mediterranean embayment.

Aquatic Conservation-Marine and Freshwater Ecosystems 24, 696–711.

Boyra, A., Sanchez-Jerez, P., Tuya, F., Espino, F. and Haroun, R. (2004) Attraction of wild coastal fishes to

an Atlantic subtropical cage fish farms, Gran Canaria, Canary Islands. Environmental Biology of

Fishes 70, 393–401.

Brandt, M.J., Höschle, C., Diederichs, A., Betke, K., Matuschek, R., Witte, S. and Nehls, G. (2013) Far‐

reaching effects of a seal scarer on harbour porpoises, Phocoena phocoena. Aquatic Conservation:

Marine and Freshwater Ecosystems 23, 222–232.

Brehmer, P., Gerlotto, F., Guillard, J., Sanguinède, F., Guénnegan, Y. and Buestel, D. (2003) New

applications of hydroacoustic methods for monitoring shallow water aquatic ecosystems: the case

of mussel culture grounds. Aquatic Living Resources 16, 333–338.

Bridgman, G.K., Rave, E.H. and Rafferty, J.M. (2000) Piscivorous bird depredation at northern Minnesota

aquaculture facilities. Prairie Naturalist 32, 17–28.

Broyer, J. and Calenge, C. (2010) Influence of fish-farming management on duck breeding in French fish

pond systems. Hydrobiologia 637, 173–185.

Broyer, J., Chavas, G. and Chazal, R. (2017) The effects of cessation of fish farming on duck breeding in

French fishpond systems. Hydrobiologia 788, 47–53.

Bustnes, J.O., Lie, E., Herzke, D., Dempster, T., Bjørn, P.A., Nygård, T. and Uglem, I. (2010) Salmon farms

as a source of organohalogenated contaminants in wild fish. Environmental Science & Technology

44, 8736–43.

Bustnes, J.O., Nygard, T., Dempster, T., Ciesielski, T., Jenssen, B.M., Bjorn, P.A. and Uglem, I. (2011) Do

salmon farms increase the concentrations of mercury and other elements in wild fish? Journal of

Environmental Monitoring 13, 1687–1694.

Caldow, R., Beadman, H., McGrorty, S., Kaiser, M., Goss-Custard, J., Mould, K. and Wilson, A. (2003) Effects

of intertidal mussel cultivation on bird assemblages. Marine Ecology Progress Series 259, 173–183.

Carss, D.N. (1990) Concentrations of wild and escaped fishes immediately adjacent to fish farm cages.

Aquaculture 90, 29–40.

144

Carss, D.N. (1993a) Cormorants Phalacrocorax carbo at cage fish farms in Argyll, western Scotland. Seabird

15, 38–44.

Carss, D.N. (1993b) Grey heron, Ardea cinerea L., predation at cage fish farms in Argyll, western Scotland.

Aquaculture and Fisheries Management 24, 29–45.

Carss, D.N. (1994) Killing of piscivorous birds at Scottish fin fish farms, 1984-87. Biological Conservation

68, 181–188.

Cartier, L.E. and Carpenter, K.E. (2014) The influence of pearl oyster farming on reef fish abundance and

diversity in Ahe, French Polynesia. Marine Pollution Bulletin 78, 43–50.

de Carvalho, L.L., de Souza, E.G.A., da Mata Júnior, M.R. and Villaça, R.C. (2017) Assessment of rocky reef

fish assemblages close to seaweed farming. Aquaculture Research 48, 481–493.

Cervencl, A., Troost, K., Dijkman, E., de Jong, M., Smit, C.J., Leopold, M.F. and Ens, B.J. (2015) Distribution

of wintering common eider Somateria mollissima in the Dutch Wadden Sea in relation to available

food stocks. Marine Biology 162, 153–168.

Chesney, E.J. and Iglesias, J. (1979) Seasonal distribution, abundance and diversity of demersal fishes in

the inner Ria de Arosa, Northwest Spain. Estuarine and Coastal Marine Science 8, 227–239.

Clynick, B.G., McKindsey, C.W. and Archambault, P. (2008) Distribution and productivity of fish and

macroinvertebrates in mussel aquaculture sites in the Magdalen islands (Québec, Canada).

Colorni, A., Diamant, A., Eldar, A., Kvitt, H. and Zlotkin, A. (2002) Streptococcus iniae infections in Red Sea

cage-cultured and wild fishes. Diseases of Aquatic Organisms 49, 165–170.

Connolly, L.M. and Colwell, M.A. (2005) Comparative use of longline oysterbeds and adjacent tidal flats

by waterbirds. Bird Conservation International 15, 237–255.

D’Amours, O., Archambault, P., McKindsey, C.W. and Johnson, L.E. (2008) Local enhancement of

epibenthic macrofauna by aquaculture activities. Marine Ecology Progress Series 371, 73–84.

Dealteris, J.T., Kilpatrick, B.D. and Rheault, R.B. (2004) A comparative evaluation of the habitat value of

shellfish aquaculture gear, submerged aquatic vegetation and a non-vegetated seabed. Journal of

Shellfish Research 23, 867–874.

DeBruyn, A.M.H., Trudel, M., Eyding, N., et al. (2006) Ecosystemic effects of salmon farming increase

mercury contamination in wild fish. Environmental Science & Technology 40, 3489–3493.

Demétrio, J.A., Gomes, L.C., Latini, J.D. and Agostinho, A.A. (2012) Influence of net cage farming on the

diet of associated wild fish in a Neotropical reservoir. Aquaculture 330, 172–178.

Dempster, T., Fernandez-Jover, D., Sanchez-Jerez, P., Tuya, F., Bayle-Sempere, J., Boyra, A. and Haroun,

R.J. (2005) Vertical variability of wild fish assemblages around sea-cage fish farms: implications for

management. Marine Ecology Progress Series 304, 15–29.

Dempster, T., Sanchez-Jerez, P., Bayle-Sempere, J.T. and Kingsford, M. (2004) Extensive aggregations of

wild fish at coastal sea-cage fish farms. Hydrobiologia 525, 245–248.

Dempster, T., Sanchez-Jerez, P., Fernandez-Jover, D., Bayle-Sempere, J.T., Nilsen, R., Bjørn, P.-A. and

Uglem, I. (2011) Proxy measures of fitness suggest coastal fish farms can act as population sources

and not ecological traps for wild gadoid fish. PLoS ONE 6, e15646.

Dempster, T., Sanchez-Jerez, P., Uglem, I. and Bjørn, P.-A. (2010) Species-specific patterns of aggregation

of wild fish around fish farms. Estuarine, Coastal and Shelf Science 86, 271–275.

Dempster, T., Uglem, I., Sanchez-Jerez, P., Fernandez-Jover, D., Bayle-Sempere, J. and Nilsen, R. (2009)

Coastal salmon farms attract large and persistent aggregations of wild fish: an ecosystem effect.

Marine Ecology Progress Series 385, 1–14.

Diamant, A., Banet, A., Ucko, M., Colorni, A., Knibb, W. and Kvitt, H. (2000) Mycobacteriosis in wild

rabbitfish Siganus rivulatus associated with cage farming in the Gulf of Eilat, Red Sea. Diseases of

Aquatic Organisms 39, 211–219.

Diaz López, B. (2006) Bottlenose dolphin (Tursiops truncatus) predation on a marine fin fish farm: some

underwater observations. Aquatic Mammals 32, 305–310.

Diaz López, B. (2012) Bottlenose dolphins and aquaculture: interaction and site fidelity on the north-

eastern coast of Sardinia (Italy). Marine Biology 159, 2161–2172.

Diaz López, B. (2017) Temporal variability in predator presence around a fin fish farm in the Northwestern

Mediterranean Sea. Marine Ecology 38, e12378.

145

Diaz López, B. (2009) The bottlenose dolphin Tursiops truncatus foraging around a fish farm: effects of

prey abundance on dolphins’ behavior. Current Zoology 55, e11221–e11221.

Diaz López, B. and Bernal-Shirai, J.A. (2007) Bottlenose dolphin (Tursiops truncatus) presence and

incidental capture in a marine fish farm on the north-eastern coast of Sardinia (Italy). Journal of

the Marine Biological Association of the United Kingdom 87, 113–117.

Diaz López, B. and Bernal-Shirai, J.A. (2008) Marine aquaculture and bottlenose dolphins’ (Tursiops

truncatus) social structure. Behavioral Ecology and Sociobiology 62, 887–894.

Dorr, B.S., Burger, L.W., Barras, S.C. and Godwin, K.C. (2012) Economic impact of double-crested

cormorant, Phalacrocorax auritus, depredation on channel catfish, Ictalurus punctatus,

aquaculture in Mississippi, USA. Journal of the World Aquaculture Society 43, 502–513.

Draulans, D. and van Vessem, J. (1985) The effect of disturbance on nocturnal abundance and behaviour

of grey herons (Ardea cinerea) at a fish-farm in winter. Journal of applied ecology, 19–27.

Dunthorn, A.A. (1971) The predation of cultivated mussels by eiders. Bird Study 18, 107–11.

Eklöf, J.S., de la Torre-Castro, M., Nilsson, C. and Rönnbäck, P. (2006) How do seaweed farms influence

local fishery catches in a seagrass-dominated setting in Chwaka Bay, Zanzibar? Aquatic Living

Resources 19, 137–147.

Erbland, P.J. and Ozbay, G. (2008) A comparison of the macrofaunal communities inhabiting a Crassostrea

virginica oyster reef and oyster aquaculture gear in Indian River Bay, Delaware. Journal of Shellfish

Research 27, 757–768.

Ervik, A., Thorsen, B., Eriksen, V., Lunestad, B.T. and Samuelsen, O.B. (1994) Impact of administering

antibacterial agents on wild fish and blue mussels Mytilus edulis in the vicinity of fish farms.

Diseases of Aquatic Organisms 18, 45–51.

Fernandez-Jover, D., Faliex, E., Sanchez-Jerez, P., Sasal, P. and Bayle-Sempere, J.T. (2010) Coastal fish

farming does not affect the total parasite communities of wild fish in SW Mediterranean.

Aquaculture 300, 10–16.

Fernandez-Jover, D., Jimenez, J.A.L., Sanchez-Jerez, P., Bayle-Sempere, J.T., Giménez-Casalduero, F.,

Lopez, F.J.M. and Dempster, T. (2007a) Changes in body condition and fatty acid composition of

wild Mediterranean horse mackerel (Trachurus mediterraneus, Steindachner, 1868) associated to

sea cage fish farms. Marine Environmental Research 63, 1–18.

Fernandez-Jover, D., Martinez-Rubio, L., Sanchez-Jerez, P., et al. (2011) Waste feed from coastal fish

farms: a trophic subsidy with compositional side-effects for wild gadoids. Estuarine, Coastal and

Shelf Science 91, 559–568.

Fernandez-Jover, D. and Sanchez-Jerez, P. (2015) Comparison of diet and otolith growth of juvenile wild

fish communities at fish farms and natural habitats. ICES Journal of Marine Science 72, 916–929.

Fernandez-Jover, D., Sanchez-Jerez, P., Bayle-Sempere, J., Carratala, A. and Leon, V.M. (2007b) Addition

of dissolved nitrogen and dissolved organic carbon from wild fish faeces and food around

Mediterranean fish farms: Implications for waste-dispersal models. Journal of Experimental Marine

Biology and Ecology 340, 160–168.

Fernandez-Jover, D., Sanchez-Jerez, P., Bayle-Sempere, J.T., Arechavala-Lopez, P., Martinez-Rubio, L.,

Lopez Jimenez, J.A. and Martinez Lopez, F.J. (2009) Coastal fish farms are settlement sites for

juvenile fish. Marine environmental research 68, 89–96.

Fernandez-Jover, D., Sanchez-Jerez, P., Bayle-Sempere, J.T., Valle, C. and Dempster, T. (2008) Seasonal

patterns and diets of wild fish assemblages associated with Mediterranean coastal fish farms. ICES

Journal of Marine Science 65, 1153–1160.

Fleury, B.E. and Sherry, T.W. (1995) Long-term population trends of colonial wading birds in the southern

United States: The impact of crayfish aquaculture on Louisiana populations. Auk 112, 613–632.

Freitas, D., Gomes, J., Luis, T.S., et al. (2007) Otters and fish farms in the Sado estuary: ecological and

socio-economic basis of a conflict. Hydrobiologia 587, 51–62.

Gabrielsen, S.-E. (1999) Effects of fish-farm activity on the limnetic community structure of brown trout,

Salmo trutta, and Arctic charr, Salvelinus alpinus. Environmental biology of fishes 55, 321–332.

Gaitán‐Espitia, J.D., Gómez, D., Hobday, A.J., Daley, R., Lamilla, J. and Cárdenas, L. (2017) Spatial overlap

of shark nursery areas and the salmon farming industry influences the trophic ecology of Squalus

acanthias on the southern coast of Chile. Ecology and evolution 7, 3773–3783.

146

Giannoulaki, M., Machias, A., Somarakis, S. and Karakassis, I. (2005) Wild fish spatial structure in response

to presence of fish farms. Journal of the Marine Biological Association of the United Kingdom 85,

1271–1277.

Glahn, J.F., Rasmussen, E.S., Tomsa, T. and Preusser, K.J. (1999a) Distribution and relative impact of avian

predators at aquaculture facilities in the northeastern United States. North American Journal of

Aquaculture 61, 340–348.

Glahn, J.F., Tomsa, T. and Preusser, K.J. (1999b) Impact of great blue heron predation at trout-rearing

facilities in the northeastern United States. North American Journal of Aquaculture 61, 349–354.

Glover, K.A., Sørvik, A.G.E., Karlsbakk, E., Zhang, Z. and Skaala, Ø. (2013) Molecular genetic analysis of

stomach contents reveals wild Atlantic cod feeding on piscine reovirus (PRV) infected Atlantic

salmon originating from a commercial fish farm. PLoS ONE 8, e60924.

Godet, L., Toupoint, N., Fournier, J., Le Mao, P., Retiere, C. and Olivier, F. (2009) Clam farmers and

oystercatchers: effects of the degradation of Lanice conchilega beds by shellfish farming on the

spatial distribution of shorebirds. Marine Pollution Bulletin 58, 589–595.

Goodbrand, L., Abrahams, M. V and Rose, G.A. (2013) Sea cage aquaculture affects distribution of wild

fish at large spatial scales. Canadian Journal of Fisheries and Aquatic Sciences 70, 1289–1295.

Gregory, P.T. and Nelson, K.J. (1991) Predation on fish and intersite variation in the diet of common garter

snakes, Thamnophis sirtalis, on Vancouver Island. Canadian Journal of Zoology 69, 988–994.

Güçlüsoy, H. and Savas, Y. (2003) Interaction between monk seals Monachus monachus (Hermann, 1779)

and marine fish farms in the Turkish Aegean and management of the problem. Aquaculture

Research 34, 777–783.

Hayward, C.J., Svane, I., Lachimpadi, S.K., Itoh, N., Bott, N.J. and Nowak, B.F. (2011) Sea lice infections of

wild fishes near ranched southern bluefin tuna (Thunnus maccoyii) in South Australia. Aquaculture

320, 178–182.

Hehre, E.J. and Meeuwig, J.J. (2016) A global analysis of the relationship between farmed seaweed

production and herbivorous fish catch. PLoS ONE 11, e0148250.

Hume, F., Pemberton, D., Gales, R., Brothers, N. and Greenwood, M. (2002) Trapping and relocating seals

from salmonid fish farms in Tasmania, 1990-2000: was it a success? Papers and Proceedings of the

Royal Society of Tasmania 136, 1–6.

Iglesias, J. (1981) Spatial and temporal changes in the demersal fish community of the Ria de Arosa (NW

Spain). Marine Biology 65, 199–208.

Hilgerloh, G., Halloran, J.O., Kelly, T.C. and Burnell, G.M. (2001) A preliminary study on the effects of oyster

culturing structures on birds in a sheltered Irish estuary. Hydrobiologia 465, 175–180.

Jacobs, S.R. and Terhune, J.M. (2000) Harbor seal (Phoca vitulina) numbers along the New Brunswick coast

of the Bay of Fundy in autumn in relation to aquaculture. Northeastern Naturalist 7, 289–296.

Jimenez, J.E., Arriagada, A.M., Fonturbel, F.E., Camus, P.A. and Avila-Thieme, M.I. (2013) Effects of exotic

fish farms on bird communities in lake and marine ecosystems. Naturwissenschaften 100, 779–

787.

Johansson, T., Hakanson, L., Borum, K. and Persson, J. (1998) Direct flows of phosphorus and suspended

matter from a fish farm to wild fish in lake southern Bullaren, Sweden. Aquacultural Engineering

17, 111–137.

Johnston, T.A., Keir, M. and Power, M. (2010) Response of native and naturalized fish to salmonid cage

culture farms in northern Lake Huron, Canada. Transactions of the American Fisheries Society 139,

660–670.

Kelly, J.P., Evens, J.G. and Stallcup, R.W. (1996) Effects of aquaculture on habitat use by wintering

shorebirds in Tomales Bay, California. California Fish and Game 82, 160–174.

Kemper, C.M. and Gibbs, S.E. (2001) Dolphin interactions with tuna feedlots at Port Lincoln, South

Australia and recommendations for minimising entanglements. Journal of Cetacean Research and

Management 3, 283–292.

Kilambi, R. V., Adams, J.C. and Wickizer, W.A. (1978) Effects of cage culture on growth, abundance, and

survival of resident largemouth bass (Micropterus salmoides). Journal of the Fisheries Research

Board of Canada 35, 157–160.

147

Kloskowski, J. (2010) Fish farms as amphibian habitats: factors affecting amphibian species richness and

community structure at carp ponds in Poland. Environmental Conservation 37, 187–194.

Kloskowski, J. (2012) Fish stocking creates an ecological trap for an avian predator via effects on prey

availability. Oikos 121, 1567–1576.

Kloskowski, J. (2005) Otter Lutra lutra damage at farmed fisheries in southeastern Poland, II: exploitation

of common carp Cyprinus carpio. Wildlife Biology 11, 257–261.

Kloskowski, J. (2009) Size-structured effects of common carp on reproduction of pond-breeding

amphibians. Hydrobiologia 635, 205–213.

Kloskowski, J., Grendel, A. and Wronka, M. (2000) The use of fish bones of three farm fish species in diet

analysis of the Eurasian otter, Lutra lutra. Folia Zoologica 49, 183–190.

Kloskowski, J., Kaczanowska, E., Krogulec, J. and Grela, P. (2017) Hematological indicators of habitat

quality: erythrocyte parameters reflect greater parental effort of red-necked grebes under

ecological trap conditions. The Condor 119, 239–250.

Kloskowski, J., Nieoczym, M., Polak, M. and Pitucha, P. (2010) Habitat selection by breeding waterbirds at

ponds with size-structured fish populations. Naturwissenschaften 97, 673–682.

Kortan, D., Adámek, Z. and Poláková, S. (2007) Winter predation by otter, Lutra lutra on carp pond systems

in South Bohemia (Czech Republic). Folia Zoologica 56, 416.

Krkosek, M., Lewis, M.A., Morton, A., Frazer, L.N. and Volpe, J.P. (2006) Epizootics of wild fish induced by

farm fish. Proceedings of the National Academy of Sciences of the United States of America 103,

15506–15510.

Krkosek, M., Lewis, M.A. and Volpe, J.P. (2005) Transmission dynamics of parasitic sea lice from farm to

wild salmon. Proceedings. Biological sciences 272, 689–96.

Laffargue, P., Begout, M. and Lagardere, F. (2006) Testing the potential effects of shellfish farming on

swimming activity and spatial distribution of sole (Solea solea) in a mesocosm. ICES Journal of

Marine Science 63, 1014–1028.

Lanszki, J., Pallos, Z.S., Nagy, D. and Yoxon, G. (2007) Diet and fish choice of Eurasian otters (Lutra lutra L.)

in fish wintering ponds in Hungary. Aquaculture international 15, 393–402.

Lin, H.-J., Shao, K.-T., Hsieh, H.-L., Lo, W.-T. and Dai, X.-X. (2009) The effects of system-scale removal of

oyster-culture racks from Tapong Bay, southwestern Taiwan: model exploration and comparison

with field observations. ICES Journal of Marine Science 66, 797–810.

Loiseau, N., Kiszka, J.J., Bouveroux, T., Heithaus, M.R., Soria, M. and Chabanet, P. (2016) Using an unbaited

stationary video system to investigate the behaviour and interactions of bull sharks Carcharhinus

leucas under an aquaculture farm. African Journal of Marine Science 38, 73–79.

Ludwig, G.X., Hokka, V., Sulkava, R. and Ylönen, H. (2002) Otter Lutra lutra predation on farmed and free-

living salmonids in. Wildlife Biology 8.

Machias, A., Karakassis, I., Giannoulaki, M., Papadopoulou, K.N., Smith, C.J. and Somarakis, S. (2005)

Response of demersal fish communities to the presence of fish farms. Marine Ecology Progress

Series 288, 241–250.

Machias, A., Karakassis, I., Labropoulou, M., Somarakis, S., Papadopoulou, K.N. and Papaconstantinou, C.

(2004) Changes in wild fish assemblages after the establishment of a fish farming zone in an

oligotrophic marine ecosystem. Estuarine Coastal and Shelf Science 60, 771–779.

MacKenzie, K., Longshaw, M., Begg, G.S. and McVicar, A.H. (1998) Sea lice (Copepoda: Caligidae) on wild

sea trout (Salmo trutta L.) in Scotland. ICES Journal of Marine Science 55, 151–162.

Manikowska-Ślepowrońska, B., Szydzik, B. and Jakubas, D. (2016) Determinants of the presence of conflict

bird and mammal species at pond fisheries in western Poland. Aquatic Ecology 50, 87–95.

Marenghi, F.P. and Ozbay, G. (2010) Floating oyster, Crassostrea virginica Gmelin 1791, aquaculture as

habitat for fishes and macroinvertebrates in Delaware Inland Bays: the comparative value of oyster

clusters and loose shell. Journal of Shellfish Research 29, 889–904.

Markowitz, T.M., Harlin, A.D., Wursig, B. and Mcfadden, C.J. (2004) Dusky dolphin foraging habitat:

overlap with aquaculture in New Zealand. Aquatic Conservation-marine and Freshwater

Ecosystems 14, 133–149.

Marques, C., Rosalino, L.M. and Santos-Reis, M. (2007) Otter predation in a trout fish farm of central-east

Portugal: preference for “fast-food”? River Research and Applications 23, 1147–1153.

148

McAllister, P.E. and Owens, W.J. (1992) Recovery of infectious pancreatic necrosis virus from the faeces

of wild piscivorous birds. Aquaculture 106, 227–232.

McConnell, A., Routledge, R. and Connors, B. (2010) Effect of artificial light on marine invertebrate and

fish abundance in an area of salmon farming. Marine Ecology Progress Series 419, 147–156.

McPeek, K.C., McDonald, P.S. and VanBlaricom, G.R. (2015) Aquaculture disturbance impacts the diet but

not ecological linkages of a ubiquitous predatory fish. Estuaries and Coasts 38, 1520–1534.

Middlemas, S.J., Fryer, R.J., Tulett, D. and Armstrong, J.D. (2013) Relationship between sea lice levels on

sea trout and fish farm activity in western Scotland. Fisheries Management and Ecology 20, 68–74.

Mladineo, I., Segvic, T. and Grubisic, L. (2009) Molecular evidence for the lack of transmission of the

monogenean Sparicotyle chrysophrii (Monogenea, Polyopisthocotylea) and isopod Ceratothoa

oestroides (Crustacea, Cymothoidae) between wild bogue (Boops boops) and cage-reared sea

bream (Sparus aurata) . Aquaculture 295, 160–167.

Morrisey, D.J., Cole, R.G., Davey, N.K., Handley, S.J., Bradley, A., Brown, S.N. and Madarasz, A.L. (2006)

Abundance and diversity of fish on mussel farms in New Zealand. Aquaculture 252, 277–288.

Morrison, S.S. and Vogel, P. (2009) Aquaculture ponds, a Jamaican study: The impact of birds on fish

production. African Journal of Agricultural Research 4, 1447–1454.

Morton, A., Routledge, R. and Krkosek, M. (2008) Sea louse infestation in wild juvenile salmon and Pacific

herring associated with fish farms off the east-central coast of Vancouver Island, British Columbia.

North American Journal of Fisheries Management 28, 523–532.

Morton, A., Routledge, R., McConnell, A. and Krkosek, M. (2011) Sea lice dispersion and salmon survival

in relation to salmon farm activity in the Broughton Archipelago. ICES Journal of Marine Science

68, 144–156.

Morton, A., Routledge, R., Peet, C. and Ladwig, A. (2004) Sea lice (Lepeophtheirus salmonis) infection rates

on juvenile pink (Oncorhynchus gorbuscha) and chum (Oncorhynchus keta) salmon in the

nearshore marine environment of British Columbia, Canada. Canadian Journal of Fisheries and

Aquatic Sciences 61, 147–157.

Morton, A., Routledge, R.D. and Williams, R. (2005) Temporal patterns of sea louse infestation on wild

Pacific salmon in relation to the fallowing of Atlantic salmon farms. North American Journal of

Fisheries Management 25, 811–821.

Mott, D.E. and Flynt, R.D. (1995) Evaluation of an electric fence system for excluding wading birds at

catfish ponds. The Progressive Fish-Culturist 57, 88–90.

Nelson, K.J. and Gregory, P.T. (2000) Activity patterns of garter snakes, Thamnophis sirtalis, in relation to

weather conditions at a fish hatchery on Vancouver Island, British Columbia. Journal of

Herpetology 34, 32.

Neofitou, N. (2016) Waste feed from fish farms of the eastern Mediterranean and attraction of wild fish.

Universal Journal of Geoscience 4, 112–115.

Oakes, C.T. and Pondella, D.J. (2009) The value of a net‐cage as a fish aggregating device in southern

California. Journal of the World Aquaculture Society 40, 1–21.

Otterå, H. and Skilbrei, O.T. (2014) Possible influence of salmon farming on long-term resident behaviour

of wild saithe (Pollachius virens L.). ICES Journal of Marine Science 71, 2484–2493.

Özgül, A. and Angel, D. (2013) Wild fish aggregations around fish farms in the Gulf of Aqaba, Red Sea:

implications for fisheries management and conservation. Aquaculture Environment Interactions 4,

135–145.

Parkhurst, J.A., Brooks, R.P. and Arnold, D.E. (1992) Assessment of predation at trout hatcheries in central

Pennsylvania. Wildlife Society Bulletin (1973-2006) 20, 411–419.

Pemberton, D., Brothers, N. and Copson, G. (1991) Predators on marine fish farms in Tasmania. Papers

and Proceedings of the Royal Society of Tasmania 125, 33–35.

Pemberton, D. and Shaughnessy, P.D. (1993) Interaction between seals and marine fish-farms in

Tasmania, and management of the problem. Aquatic Conservation-marine and Freshwater

Ecosystems 3, 149–158.

Pitt, W.C. and Conover, M.R. (1996) Predation at intermountain west fish hatcheries. The Journal of

Wildlife Management 60, 616–624.

149

Plummer, M. V. and Goy, J.M. (1984) Ontogenetic dietary shift of water snakes (Nerodia rhombifera) in a

fish hatchery. Copeia 1984, 550.

Powers, M., Peterson, C., Summerson, H. and Powers, S. (2007) Macroalgal growth on bivalve aquaculture

netting enhances nursery habitat for mobile invertebrates and juvenile fishes. Marine Ecology

Progress Series 339, 109–122.

Quick, N.J., Middlemas, S.J. and Armstrong, J.D. (2004) A survey of antipredator controls at marine salmon

farms in Scotland. Aquaculture 230, 169–180.

Radomski, A.A. and Zimba, P. V (2010) Does pond water reflectance influence double‐crested cormorant

selection of aquaculture ponds? Journal of the World Aquaculture Society 41, 430–437.

Ramos, I.P., Brandão, H., Zanatta, A.S., Zica, E. de O.P., da Silva, R.J., de Rezende-Ayroza, D.M.M. and

Carvalho, E.D. (2013) Interference of cage fish farm on diet, condition factor and numeric

abundance on wild fish in a Neotropical reservoir. Aquaculture 414, 56–62.

Ribeiro, S., Viddi, F.A., Cordeiro, J.K. and Freitas, T.R.O. (2007) Fine-scale habitat selection of Chilean

dolphins (Cephalorhynchus eutropia): interactions with aquaculture activities in southern Chiloe

Island, Chile. Journal of the Marine Biological Association of the United Kingdom 87, 119–128.

Riera, R., Tuset, V.M., Rodríguez, M., Monterroso, Ó. and Lombarte, A. (2017) Analyzing functional

diversity to determine the effects of fish cages in insular coastal wild fish assemblages. Aquaculture

479, 384–395.

Robinson, S., Gales, R., Terauds, A. and Greenwood, M. (2008a) Movements of fur seals following

relocation from fish farms. Aquatic Conservation-marine and Freshwater Ecosystems 18, 1189–

1199.

Robinson, S., Terauds, A., Gales, R. and Greenwood, M. (2008b) Mitigating fur seal interactions: relocation

from Tasmanian aquaculture farms. Aquatic Conservation-marine and Freshwater Ecosystems 18,

1180–1188.

Ross, B., Lien, J. and Furness, R.W. (2001) Use of underwater playback to reduce the impact of eiders on

mussel farms. ICES Journal of Marine Science 58, 517–524.

Roycroft, D., Kelly, T.C. and Lewis, L.J. (2004) Birds, seals and the suspension culture of mussels in Bantry

Bay, a non-seaduck area in Southwest Ireland. Estuarine, Coastal and Shelf Science 61, 703–712.

Roycroft, D., Kelly, T.C. and Lewis, L.J. (2007) Behavioural interactions of seabirds with suspended mussel

longlines. Aquaculture International 15, 25–36.

Sales-Luis, T., Freitas, D. and Santos-Reis, M. (2009) Key landscape factors for Eurasian otter Lutra lutra

visiting rates and fish loss in estuarine fish farms. European Journal of Wildlife Research 55, 345–

355.

Samuelsen, O.B., Lunestad, B.T., Husevag, B., Holleland, T. and Ervik, A. (1992) Residues of oxolinic acid in

wild fauna following medication in fish farms. Diseases of Aquatic Organisms 12, 111–119.

Sandilyan, S. (2017) A preliminary assessment on the role of abandoned shrimp farms on supporting

waterbirds in Pichavaram mangrove, Tamilnadu, Southern India. Journal of Coastal Conservation

21, 255–263.

Sanz-Lázaro, C., Belando, M.D., Navarrete-Mier, F. and Marín, A. (2011) Effects of wild fish and motile

epibenthic invertebrates on the benthos below an open water fish farm. Estuarine, Coastal and

Shelf Science 91, 216–223.

Schramm JR, H.L., French, B. and Ednoff, M. (1984) Depredation of channel catfish by Florida double-

crested cormorants. The Progressive Fish-Culturist 46, 41–43.

Segvic-Bubic, T., Grubisic, L., Karaman, N., Ticina, V., Jelavic, K.M. and Katavic, I. (2011) Damages on mussel

farms potentially caused by fish predation - self service on the ropes? Aquaculture 319, 497–504.

Segvić Bubić, T., Grubišić, L., Tičina, V. and Katavić, I. (2011) Temporal and spatial variability of pelagic wild

fish assemblages around Atlantic bluefin tuna Thunnus thynnus farms in the eastern Adriatic Sea.

Journal of Fish Biology 78, 78–97.

Sepúlveda, F., Marin, S.L. and Carvajal, J. (2004) Metazoan parasites in wild fish and farmed salmon from

aquaculture sites in southern Chile. Aquaculture 235, 89–100.

Sepúlveda, M., Newsome, S.D., Pavez, G., Oliva, D., Costa, D.P. and Hueckstaedt, L.A. (2015) Using satellite

tracking and isotopic information to characterize the impact of South American sea lions on

salmonid aquaculture in southern Chile. PLoS ONE 10, e0134926–e0134926.

150

Sepúlveda, M. and Oliva, D. (2005) Interactions between South American sea lions Otaria flavescens

(Shaw) and salmon farms in southern Chile. Aquaculture Research 36, 1062–1068.

Sepúlveda, M., Pavez, G., Santos‐Carvallo, M., Balbontín, C., Pequeño, G. and Newsome, S.D. (2017)

Spatial, temporal, age, and sex related variation in the diet of South American sea lions in southern

Chile. Marine Mammal Science 33, 480–495.

Skilbrei, O.T. and Otterå, H. (2016) Vertical distribution of saithe (Pollachius virens) aggregating around

fish farms. ICES Journal of Marine Science: Journal du Conseil 73, 1186–1195.

Skog, T.-E., Hylland, K., Torstensen, B.E. and Berntssen, M.H.G. (2003) Salmon farming affects the fatty

acid composition and taste of wild saithe Pollachius virens L. Aquaculture Research 34, 999–1007.

Snow, M., Black, J., Matejusova, I., McIntosh, R., Baretto, E., Wallace, I.S. and Bruno, D.W. (2010)

Detection of salmonid alphavirus RNA in wild marine fish: implications for the origins of salmon

pancreas disease in aquaculture. Diseases of Aquatic Organisms 91, 177–188.

Stickley, A.R., Glahn, J.F., King, J.O. and King, D.T. (1995) Impact of great blue heron depredations on

channel catfish farms. Journal of the World Aquaculture Society 26, 194–199.

Stickley, A.R., Warrick, G.L. and Glahn, J.F. (1992) Impact of double‐crested cormorant depredations on

channel catfish farms. Journal of the World Aquaculture Society 23, 192–198.

Strictar-Pereira, L., Agostinho, A.A. and Gomes, L.C. (2010) Cage culture with tilapia induces alteration in

the diet of natural fish populations: the case of Auchenipterus osteomystax. Brazilian Journal of

Biology 70, 1021–1030.

Sudirman, H.H., Jompa, J., Zulfikar, Iswahyudin and McKinnon, A.D. (2009) Wild fish associated with

tropical sea cage aquaculture in South Sulawesi, Indonesia. Aquaculture 286, 233–239.

Sun, Y.H., Wu, H.J. and Wang, Y. (2004) Tawny fish-owl predation at fish farms in Taiwan. Journal of Raptor

Research 38, 326–333.

Tallman, J.C. and Forrester, G.E. (2007) Oyster grow-out cages function as artificial reefs for temperate

fishes. Transactions of the American Fisheries Society 136, 790–799.

Tsuyuki, A. and Umino, T. (2017) Spatial movement of black sea bream Acanthopagrus schlegelii around

the oyster farming area in Hiroshima Bay, Japan. Fisheries Science 83, 235–244.

Tuckett, Q.M., Ritch, J.L., Lawson, K.M. and Hill, J.E. (2017) Landscape-scale survey of non-native fishes

near ornamental aquaculture facilities in Florida, USA. Biological Invasions 19, 223–237.

Tully, O., Gargan, P., Poole, W.R. and Whelan, K.F. (1999) Spatial and temporal variation in the infestation

of sea trout (Salmo trutta L.) by the caligid copepod Lepeophtheirus salmonis (Krøyer) in relation

to sources of infection in Ireland. Parasitology 119, 41–51.

Tuya, F., Sanchez-Jerez, P., Dempster, T., Boyra, A. and Haroun, R.J. (2006) Changes in demersal wild fish

aggregations beneath a sea-cage fish farm after the cessation of farming. Journal of Fish Biology

69, 682–697.

Uglem, I., Dempster, T., Bjørn, P.-A., Sanchez-Jerez, P. and Økland, F. (2009) High connectivity of salmon

farms revealed by aggregation, residence and repeated movements of wild fish among farms.

Marine Ecology Progress Series 384, 251–260.

Ulenaers, P. and Dhondt, A.A. (1994) Great crested grebe Podiceps cristatus chick mortality in relation to

parental fishing. Bird Study 41, 211–220.

Ulenaers, P. and van Vessem, J. (1994) Impact of great crested grebes (Podiceps cristatus L.) on fish ponds.

Hydrobiologia 279, 353–366.

Valle, C., Bayle-Sempere, J.T., Dempster, T., Sanchez-Jerez, P. and Gimenez-Casalduero, F. (2007)

Temporal variability of wild fish assemblages associated with a sea-cage fish farm in the south-

western Mediterranean Sea. Estuarine Coastal and Shelf Science 72, 299–307.

Valtonen, E.T. and Koskivaara, M. (1994) Relationships between the parasites of some wild and cultured

fishes in two lakes and a fish farm in central Finland. International Journal for Parasitology 24, 109–

118.

Vita, R., Marin, A., Madrid, J.A., Jimenez-Brinquis, B., Cesar, A. and Marin-Guirao, L. (2004) Effects of wild

fishes on waste exportation from a Mediterranean fish farm. Marine Ecology Progress Series 277,

253–261.

151

Wallace, I.S., Donald, K., Munro, L.A., et al. (2015) A survey of wild marine fish identifies a potential origin

of an outbreak of viral haemorrhagic septicaemia in wrasse, Labridae, used as cleaner fish on

marine Atlantic salmon, Salmo salar L., farms. Journal of Fish Diseases 38, 515–521.

Wallace, I.S., Gregory, A., Murray, A.G., Munro, E.S. and Raynard, R.S. (2008) Distribution of infectious

pancreatic necrosis virus (IPNV) in wild marine fish from Scottish waters with respect to clinically

infected aquaculture sites producing Atlantic salmon, Salmo salar L. Journal of Fish Diseases 31,

177–186.

Watson-Capps, J.J. and Mann, J. (2005) The effects of aquaculture on bottlenose dolphin (Tursiops sp.)

ranging in Shark Bay, Western Australia. Biological Conservation 124, 519–526.

Wisniowska, L. (2007) Otter (Lutra lutra L.) damage in commercial carp ponds of southern Poland. Hystrix,

the Italian Journal of Mammalogy 17.

Yasue, M. and Dearden, P. (2009) The importance of supratidal habitats for wintering shorebirds and the

potential impacts of shrimp aquaculture. Environmental Management 43, 1108–1121.

Zlotkin, A., Hershko, H. and Eldar, A. (1998) Possible transmission of Streptococcus iniae from wild fish to

cultured marine fish. Applied and Environmental Microbiology 64, 4065–4067.

Zydelis, R., Esler, D., Kirk, M. and Sean Boyd, W. (2009) Effects of off-bottom shellfish aquaculture on

winter habitat use by molluscivorous sea ducks. Aquatic Conservation: Marine and Freshwater

Ecosystems 19, 34–42.

152

Appendix 2.2. Summary of results for zero-inflated Poisson model comparing research effort on

interactions between wildlife (number of articles included in our systematic review) and

aquaculture between nations according to both domestic aquaculture production (t) in 2015

and Human Development Index (HDI) in 2015. N = 188 nations.

Model term Estimate SE z p

HDI 9.4 2.3 4.1 <0.0001 Production -0.01 0.006 -2.2 0.02 HDI × Production 0.02 0.01 2.2 0.03 HDI > 0.8, N = 55 Production 0.008 0.002 5.6 <0.0001

153

Appendix 2.3. List of the six most parsimonious linear models predicting log response ratios for

wildlife abundance (a) and species richness (b) at aquaculture sites relative to reference sites. In

each case, the full model was specified using the following terms: Response ~ Year + Country +

Region + Environment + Culture System + Culture Organism + Wild Taxa + Reference Habitat. A

model selection process using the MuMIn package for R (Barton 2016) fitted every possible

combination of these terms using maximum likelihood estimation and ranked them by AICC

score. ‘+’ indicates the corresponding term was included in the given model.

Model ID

Cultured Taxa

Env Ref Hab Wild Taxa

Year logLik AICC ΔAICC Wt

27 + + + -115.6 247.2 0 0.33 91 + + + + -111.7 247.6 0.3 0.28

155 + + + + -115.3 249.2 2.0 0.12 75 + + + -113.9 249.2 2.0 0.12 83 + + + -114.3 250.0 2.8 0.08

219 + + + + + -111.6 250.3 3.1 0.07

Model ID

Cultured Taxa

Env Ref Hab Wild Taxa

Year logLik AICC ΔAICC Wt

81 + + -24.8 64.4 0 0.34 65 + -26.5 65.0 0.6 0.26

209 + + + -24.1 66.0 1.7 0.15 89 + + + -24.4 66.7 2.4 0.11 73 + + -26.2 67.2 2.9 0.08

193 + + -26.5 67.8 3.4 0.06

154

Appendix 2.4. ANOVA table and Tukey’s post-hoc test results (multiple comparisons of means

with 95% family-wise confidence level) for best fitting linear model for factors predicting log

response ratios (RR) for wildlife abundance at aquaculture sites. The best fitting model was

selected from a full model containing the following factors: Year, Country, Continent,

Environment (Marine, Freshwater), Culture System (Cage, Pond, Longline, Rack, Bed), Cultured

Taxa (Fish, Shellfish, Crustacean, Alga), Wild Taxa (Fish, Bird, Mammal, Reptile, Amphibian), and

Reference Habitat (Structured, Unstructured). Post-hoc testing compared all pairwise

combinations of levels of significant model terms (Cultured Taxa and Reference Habitat).

Positive values for ‘Difference’ indicate that the former level has a higher response to farms than

the latter level in the pairwise comparison.

ANOVA Model term df SS MS F p

Environment 1 6.87 6.87 2.7 0.11 Cultured Taxa 3 62.0 20.7 8.1 0.0001 Reference Habitat 1 24.9 24.9 9.8 0.003 Residuals 57 144 2.54 POST HOC Significant model term Difference Lower Upper Adjusted p

Cultured Taxa Crustacean-Alga -2.33 -6.95 2.29 0.54 Fish-Alga 1.77 -0.24 3.78 0.10 Shellfish-Alga 0.09 -2.01 2.19 0.99 Fish-Crustacean 4.11 -0.17 8.38 0.06 Shellfish-Crustacean 2.42 -1.89 6.74 0.45 Shellfish-Fish -1.68 -2.84 -0.52 0.002 Reference Habitat Unstructured-Structured 1.17 0.37 1.97 0.005

155

Appendix 2.5. ANOVA table and Tukey’s post-hoc test results for best fitting linear model for

factors predicting log response ratios (RR) for wildlife species richness at aquaculture sites. The

best fitting model was selected from a full model containing the following factors: Year, Country,

Continent, Environment (Marine, Freshwater), Culture System (Cage, Pond, Longline, Rack, Bed),

Cultured Taxa (Fish, Shellfish, Crustacean, Alga), Wild Taxa (Fish, Bird, Mammal, Reptile,

Amphibian), and Reference Habitat (Structured, Unstructured). Post-hoc testing compared all

pairwise combinations of wild taxonomic groups represented in the dataset. Positive values for

‘Difference’ indicate that the former level has a higher response to farms than the latter level in

the pairwise comparison.

ANOVA Model term df SS MS F p

Wild Taxa 3 3.63 1.21 4.4 0.01 Reference Habitat 1 0.81 0.81 3.1 0.09 Residuals 32 8.27 0.26 POST HOC Significant model term Difference Lower Upper Adjusted p

Wild Taxa Bird-Amphibian 1.22 -0.25 2.70 0.13 Fish-Amphibian 1.53 0.13 2.93 0.03 Mammal-Amphibian 0.41 -1.54 2.35 0.94 Fish-Bird 0.30 -0.28 0.89 0.50 Mammal-Bird -0.82 -2.29 0.65 0.44 Mammal-Fish -1.12 -2.53 0.28 0.15

156

Appendix 3.1. Ovarian fatty acid profiles in cod (Gadus morhua) collected from areas of high and

low salmon farming density. Fatty acids that represent less than 0.1 % of the total fatty acids are

omitted. Data are presented as percentage of total fatty acids (mean ± SD, n = 10 per group).

Statistical comparisons are from univariate linear analyses of variance with 1 on 18 df.

Fatty acids Low farm density High farm density F p

14:0 1.63 ± 0.36 1.53 ± 0.39 Iso 15:0 0.25 ± 0.12 0.20 ± 0.04 Iso 16:0 0.17 ± 0.06 0.15 ± 0.04 Iso 17:0 0.46 ± 0.09 0.46 ± 0.07

Antiso 17:0 0.23 ± 0.09 0.21 ± 0.05 17:0 0.37 ± 0.06 0.41 ± 0.08

iso 18:0 0.18 ± 0.05 0.19 ± 0.04 18:0 2.44 ± 0.40 2.55 ± 0.54

∑SFA 23.87 ± 0.71 24.00 ± 1.05 0.10 0.76

16:1 (n-11) 0.14 ± 0.04 0.13 ± 0.05 16:1 (n-9) 1.27 ± 0.13 1.30 ± 0.28 18:1 (n-9) 12.77 ± 0.71 13.57 ± 1.67 1.9 0.18 18:1 (n-7) 4.25 ± 0.89 3.98 ± 0.47 18:1 (n-5) 0.28 ± 0.04 0.31 ± 0.07

20:1 (n-11) 0.33 ± 0.23 0.20 ± 0.08 20:1 (n-9) 1.02 ± 0.56 0.61 ± 0.23

22:1 (n-11) 0.38 ± 0.32 0.21 ± 0.10 22:1 (n-9) 0.08 ± 0.04 0.07 ± 0.01 24:1 (n-9) 0.94 ± 0.20 1.02 ± 0.34 24:1 (n-7) 0.24 ± 0.05 0.32 ± 0.12

∑MUFA 26.16 ± 3.10 26.12 ± 1.51 0.00 0.97

18:2 (n-6) 1.06 ± 0.53 1.26 ± 0.89 0.34 0.57 18:3 (n-6) 0.18 ± 0.05 0.17 ± 0.04 20:2 (n-6) 0.32 ± 0.15 0.25 ± 0.07 20:3 (n-6) 0.10 ± 0.01 0.09 ± 0.03 20:4 (n-6) 3.90 ± 0.69 4.11 ± 1.72 22:4 (n-6) 0.68 ± 0.20 0.50 ± 0.20 22:5 (n-6) 0.36 ± 0.04 0.40 ± 0.05 18:3 (n-3) 0.39 ± 0.07 0.52 ± 0.25 18:4 (n-3) 0.44 ± 0.13 0.68 ± 0.59 20:3 (n-3) 0.14 ± 0.02 0.14 ± 0.03 20:4 (n-3) 0.45 ± 0.11 0.50 ± 0.19 20:5 (n-3) 11.14 ± 1.00 10.34 ± 0.94 21:5 (n-3) 0.19 ± 0.04 0.17 ± 0.03 22:5 (n-3) 1.54 ± 0.47 1.35 ± 0.45 22:6 (n-3) 26.64 ± 3.43 26.87 ± 2.08

∑PUFA 47.53 ± 2.78 47.33 ± 1.03 0.04 0.84

∑PUFA (n-3) 40.92 ± 3.26 40.57 ± 2.34 0.08 0.78

∑PUFA (n-6) 6.61 ± 0.84 6.77 ± 1.78 0.07 0.80

(n-3)/(n-6) 6.32 ± 1.20 6.47 ± 2.12 0.03 0.86

Fatty acids relative to sample wet weight (%)

2.12 ± 0.47 2.27 ± 0.73 0.32 0.58

Cholesterol relative to sample wet weight (%)

0.21 ± 0.06 0.20 ± 0.03 0.17 0.68

157

Appendix 3.2. Model summaries for linear regression of log-transformed weight-at-length

relationships in female and male Atlantic cod from areas of high and low salmon farming density.

FEMALES

Model term Estimate SE t p

Intercept 1.63 0.10 17 <0.0001

log(W) 0.32 0.01 26 <0.0001

Treatment(LFD) -0.02 0.01 -2.5 0.014

R2 = 0.88 Residual df = 105

MALES

Model term Estimate SE t p

Intercept 1.56 0.96 19 <0.0001

log(W) 0.33 0.00 31 <0.0001

Treatment(LFD) 0.02 0.00 2.0 0.06

R2 = 0.95 Residual df = 45

158

Appendix 3.3. Model summaries for effects of farm density on egg production, egg quality and

larval quality of wild Atlantic cod. Model terms are Farm density group (LFD and HFD), Day, Time

(early or late season sampling), total female length (TotalFL) and mean female length (MeanFL).

Model summaries for X2 likelihood ratio tests of model terms are omitted, as are model

summaries for analyses within farm density groups.

Daily egg production (DEP) Model specification: DEP ~ Group + Day + TotalFL + (1|Group/Tank) Model term Estimate SE z p

Intercept 1.55 0.48 3.2 0.001 Group (LFD) 0.53 0.28 1.9 0.056 Day -0.017 0.008 -2.0 0.042 TotalFL 0.008 0.006 14 <0.0001 Observations 420 Residual df 413

Relative daily egg production (RDEP) Model specification: RDEP ~ Group + Day + (1|Group/Tank) Model term Estimate SE z p

Intercept 5.32 0.19 27 <0.0001 Group (LFD) 0.12 0.18 0.7 0.52 Day -0.07 0.007 -10 <0.0001 Observations 420 Residual df 414

Egg viability rate (ViablePr) Model specification: ViablePr ~ Group + Time + meanFL + (1|Group/Tank) Model term Estimate SE z p

Intercept 3.7 1.48 2.5 0.012 Group (LFD) -0.09 0.18 -0.5 0.61 Time (late season) -0.81 0.03 -6.3 <0.0001 MeanFL -0.03 0.02 -1.4 0.18 Observations 348 Residual df 341

Egg diameter (EggSize) Model specification: EggSize ~ Group + Time + meanFL + (1|Group/Tank) Model term Estimate SE z p

Intercept 0.82 0.16 5.6 <0.0001 Group (LFD) 0.07 0.02 3.9 0.0001 Time (late season) 0.08 0.004 20 <0.0001 MeanFL 0.005 0.002 2.1 0.045 Observations 1038 Residual df 1031

Egg fertilisation rate (FertPr) Model specification*: FertPr ~ Group + Time + (1|Group/Tank) *MeanFL omitted due to poor model fit Model term Estimate SE z p

159

Intercept 0.96 0.22 4.4 <0.0001 Group (LFD) -0.16 0.25 -0.6 0.52 Time (late season) -0.64 0.25 -2.6 0.009 Observations 72 Residual df 66

Egg symmetry rate (SymPr) Model specification: SymPr ~ Group + Time + (1|Group/Tank) Model term Estimate SE z p

Intercept -2.0 2.7 -0.7 0.50 Group (LFD) 0.15 0.35 0.4 0.67 Time (late season) 0.36 0.26 1.4 0.18 MeanFL 0.05 0.04 1.3 0.20 Observations 72 Residual df 65

Egg survival rate during incubation (SurvPr) Model specification: SurvPr ~ Group + Time + MeanFL + (1|Group/Tank) Model term Estimate SE z p

Intercept -2.2 3.6 -0.6 0.54 Group (LFD) 0.03 0.41 0.07 0.95 Time (late season) -1.0 0.39 -2.7 0.007 MeanFL 0.06 0.06 1.0 0.31 Observations 23 Residual df 16

Egg hatching rate (HatchPr) Model specification: HatchPr ~ Group + Time + MeanFL + (1|Group/Tank) Model term Estimate SE z p

Intercept 3.1 2.6 1.2 0.24 Group (LFD) 0.13 0.30 0.43 0.67 Time (late season) 0.57 0.26 2.2 0.030 MeanFL -0.02 0.04 -0.54 0.59 Observations 45 Residual df 38

Larval length (LarvL) Model specification: LarvL ~ Group + MeanFL + (1|Group/Tank) Model term Estimate SE z p

Intercept 6.4 1.7 3.6 0.0002 Group 0.46 0.21 2.2 0.029 MeanFL -0.006 0.03 -0.2 0.83 Observations 356 Residual df 350

Maximum larval length (MaxLarvL) Model specification: MaxLarvL ~ Group + MeanFL Model term df SS F p

Group 1 0.39 1.6 0.24 MeanFL 1 0.01 0.04 0.84 Residuals 9 2.2

160

Observations 12

Larval deformity rate (DeformPr) Model specification: DeformPr ~ Group + MeanFL Model term Estimate SE z p

Intercept 0.88 3.1 0.28 0.77 Group -0.07 0.38 -0.18 0.85 MeanFL -0.02 0.05 -0.31 0.75 Observations 12 Residual df 8

Larval phototaxis rate (PhotoPr) Model specification: PhotoPr ~ Group + MeanFL Model term Estimate SE z p

Intercept 0.07 3.3 0.02 0.98 Group 0.30 0.41 0.74 0.46 MeanFL 0.01 0.05 0.19 0.85 Observations 12 Residual df 8

161

Appendix 4.1. (Left) Large Asterias amurensis individuals (>20 cm) feeding on fallen mussel

clumps with Coscinasterias muricata. (Right) Attempted predation of Asterias amurensis by

Coscinasterias muricata. Image credit: Emily Fobert.

162

Appendix 4.2. Model summaries for effects of shellfish farms on seastar population metrics.

Model summaries for X2 likelihood ratio tests of model terms are omitted, as are model

summaries for analyses within locations.

Asterias population density (Density) Model specification: Density ~ Habitat + Location + (1|Date) Model term Estimate SE z p

Intercept 7.6 0.96 8.1 <0.0001 Habitat (Out) -0.81 0.60 -1.4 0.18 Location (GP) -10.5 1.5 -6.8 <0.0001 Observations 76 Residual df 71

Asterias mean transect arm span at Clifton Springs (Size) Model specification: Size ~ Habitat + (1|Date) Model term Estimate SE z p

Intercept 3.0 0.05 55 <0.0001 Habitat (Out) -0.2 0.04 -5.5 <0.0001 Observations 38 Residual df 34

Asterias gutted weight at Clifton Springs (GuttedW) Model specification: GuttedW ~ Habitat + (1|Date) + (1|Sex) Model term Estimate SE z p

Intercept 9.2 1.1 8.4 <0.0001 Habitat (Out) -4.3 0.46 -9.4 <0.0001 Observations 192 Residual df 187

Asterias drained weight at Clifton Springs (DrainedW) Model specification: DrainedW ~ Habitat + (1|Date) + (1|Sex) Model term Estimate SE z p

Intercept 16 2.0 8.1 <0.0001 Habitat (Out) -6.5 0.78 -8.3 <0.0001 Observations 192 Residual df 187

Asterias gonad weight at Clifton Springs (GonadW) Model specification: GonadW ~ Habitat + (1|Date) + (1|Sex) Model term Estimate SE z p

Intercept 4.2 1.8 2.3 0.022 Habitat (Out) -1.1 0.33 -3.5 0.0004 Observations 192 Residual df 187

Asterias gonadosomatic index at Clifton Springs (GSI) Model specification: GSI ~ Habitat + (1|Date) + (1|Sex) Model term Estimate SE z p

Intercept 0.42 0.15 2.8 0.005 Habitat (Out) -0.08 0.02 -4.9 <0.0001

163

Observations 192 Residual df 187

Asterias pyloric caeca index at Clifton Springs (PCI) Model specification: PCI ~ Habitat + (1|Date) + (1|Sex) Model term Estimate SE z p

Intercept 0.69 0.02 33 <0.0001 Habitat (Out) -0.04 0.01 -4.1 <0.0001 Observations 192 Residual df 187

Coscinasterias population density (Density) Model specification: Density ~ Habitat + Location + (1|Date) Model term Estimate SE z p

Intercept 7.1 0.31 22.7 <0.0001 Habitat (Out) -6.5 1.1 -5.8 <0.0001 Location (GP) 4.2 1.1 3.7 0.0002 Observations 76 Residual df 71

Coscinasterias mean transect arm span at Grassy Point (Size) Model specification: Size ~ Habitat + (1|Date) Model term Estimate SE z p

Intercept 31 0.56 55 <0.0001 Habitat (Out) -2.1 0.66 -3.2 0.001 Observations 32 Residual df 28

Coscinasterias gutted weight at Clifton Springs (GuttedW) Model specification: GuttedW ~ Habitat + (1|Date) + (1|Sex) Model term Estimate SE z p

Intercept 16.4 1.5 11 <0.0001 Habitat (Out) -2.0 0.53 -3.7 0.0002 Observations 158 Residual df 153

Coscinasterias drained weight at Clifton Springs (DrainedW) Model specification: DrainedW ~ Habitat + (1|Date) + (1|Sex) Model term Estimate SE z p

Intercept 21 2.3 9.1 <0.0001 Habitat (Out) -2.7 0.76 -3.6 0.0004 Observations 158 Residual df 153

Coscinasterias gonad weight at Clifton Springs (GonadW) Model specification: GonadW ~ Habitat + (1|Date) + (1|Sex) Model term Estimate SE z p

Intercept 1.5 0.56 2.6 0.009 Habitat (Out) -0.01 0.23 -0.04 0.96

164

Observations 158 Residual df 153

Coscinasterias gonadosomatic index at Clifton Springs (GSI) Model specification: GSI ~ Habitat + (1|Date) + (1|Sex) Model term Estimate SE z p

Intercept 0.41 0.12 3.4 0.0008 Habitat (Out) -0.005 0.02 -2.8 0.78 Observations 158 Residual df 153

Coscinasterias pyloric caeca index at Clifton Springs (PCI) Model specification: PCI ~ Habitat + (1|Date) + (1|Sex) Model term Estimate SE z p

Intercept 0.66 0.01 50 <0.0001 Habitat (Out) -0.02 0.009 -2.6 0.010 Observations 158 Residual df 153

165

Appendix 5.1. Study site characteristics for Chapter Five.

We conducted surveys of reef fish communities at Kirk Point, Point Cook, Altona Reef,

Williamstown and Half Moon Bay (Fig. 5.2). These locations had dense Undaria patches adjacent

to reference habitats of turfing algae or urchin barrens. Kirk Point is primarily urchin barrens

with areas of ephemeral macroalgae, including Undaria, Ulva and Gracilaria. Point Cook is

adjacent to a no-take marine reserve and contains a variety of rocky reef habitats, including

large areas of urchin barrens adjacent to mixed Ecklonia and Sargassum patches and areas of

Undaria, Ulva, Gracilaria and Caulerpa. Altona Reef is largely barren, with patchy, ephemeral

growth of Undaria, Ulva and Gracilaria. Williamstown contain large remnant beds of Ecklonia,

alongside habitats such as urchin barrens, turf, mixed Sargassum/Cystophora forests and diverse

ephemeral macroalgae including dense Undaria patches. Half Moon Bay sites were

characterised by extensive urchin barrens, mussel beds, and diverse macroalgae including

Sargassum and seasonal growth of Undaria and Ulva.

166

Appendix 5.2. Diagram of the tank used for the habitat choice experiment.

167

Appendix 5.3. Results of habitat choice experiment. Preference was tested using X2 test of

proportions, using equal preference for habitat options as the null hypothesis. Sample sizes are

lower for comparisons between macroalgae species as fish that selected barren rock were

omitted from this comparison.

N df X2 p

Heteroclinus perspicillatus

Initial choice

Overall 48 3 20 0.0002

Macroalgae vs. barren 48 1 36 <0.0001

Between macroalgae 45 2 8.4 0.02

20 min choice

Overall 48 3 7.3 0.06

Macroalgae vs. barren 48 1

121

33 <0.0001

Between macroalgae 44 2 0.2 0.9

Neoodax balteatus

Initial choice

Overall 23 3 6.7 0.08

Macroalgae vs. barren 23 1 12 0.0004

Between macroalgae 20 2 4.3 0.12

20 min choice

Overall 23 3 16 0.13

Macroalgae vs. barren 23 1

121

16 <0.0001

Between macroalgae 21 2 2.0 0.4

168

Appendix 5.4. Species counts from underwater visual census (UVC) and baited remote underwater video (BRUV) surveys. Number of UVC plots or BRUV

deployments given in parentheses. Focal species for this study are highlighted in bold type.

Family Scientific name Common name

In-season UVC Off-season UVC In-season BRUV

Undaria Barren Undaria Barren Undaria Barren Ecklonia

(25) (26) (9) (9) (14) (15) (5)

Apogonidae Vincentia conspersa Southern cardinalfish 2 2 1 1 2

Apogonidae Siphamia cephalotes Wood's siphonfish 1

Atherinidae Atherinosoma microstoma Smallmouth hardyhead 1

Blenniidae Parablennius tasmanianus Tasmanian blenny 3 8 2 8 2

Callionymidae Foetorepus calauropomus Common stinkfish 1 4 1 4 1

Centrolophidae Seriolella brama Blue warehou 16 3

Clinidae Cristiceps australis Crested weedfish 2 1 1

Clinidae Heteroclinus tristis Longnose weedfish 2

Clinidae Heteroclinus spp. Common weedfish and allies 38 1 8 3

Clinidae Ophiclinus ningulus Variable snakeblenny 1

Dinolestidae Dinolestes lewini Longfin pike 1

Diodontidae Diodon nicthemerus Globefish 5 1 1 4 3

Engraulidae Engraulis australis Australian anchovy 65

Gobiidae Nesogobius pulchellus Sailfin goby 2 1 4

Gobiidae Unidentified Goby 1

Hemiramphidae Hyporhamphus melanochir Southern garfish 11 65

Heterodontidae Heterodontus portusjacksoni Port Jackson shark 1 1

Kyphosidae Girella zebra Zebrafish 5

Kyphosidae Tilodon sexfasciatus Moonlighter 1

Labridae Notolabrus tetricus Bluethroat wrasse 1 3

Labridae Neoodax balteatus Little weed whiting 3 9 4 7

Labridae Pictilabrus laticlavius Senator wrasse 1

Loliginidae Sepioteuthis australis Southern calamari squid 1 1

Monacanthidae Brachaluteres jacksonianus Southern pygmy leatherjacket 1 2 4 1

Monacanthidae Meuschenia freycineti Sixspine leatherjacket 1 1 2 1 1

Monacanthidae Scobinichthys granulatus Rough leatherjacket 1 1 2

Monacanthidae Acanthaluteres spilomelanurus Bridled leatherjacket 1

Monacanthidae Eubalichthys mosaicus Mosaic leatherjacket 1

Monacanthidae Meuschenia hippocrepis Horseshoe leatherjacket 4 5

Mullidae Upeneichthys vlamingii Bluespotted goatfish 1 1

Platycephalidae Platycephalus laevigatus Rock flathead 1 1

Plesiopidae Trachinops caudimaculatus Southern hulafish 5 3 9 34 6

169

Pomacentridae Parma victoriae Scalyfin 1

Rhinobatidae Trygonorrhina dumerilii Southern fiddler ray 3 1

Sillaginidae Sillaginodes punctata King George whiting 1

Sphyraenidae Sphyraena novaehollandiae Snook 5

Tetraodontidae Tetractenos glaber Smooth toadfish 5 4

Antennariidae Trichophryne mitchellii Spinycoat anglerfish 1

Tripterygiidae Trinorfolkia clarkei Clark's threefin 20 19 15 25 1

Urolophidae Urolophus paucimaculatus Sparsely-spotted stingaree 2 1

Unidentified Unidentified Unidentified 31

Species richness 14 9 13 11 19 18 11

Combined MaxN 81 43 39 61 169 100 54

170

Appendix 5.5. Results of permutational ANOVA comparing fish communities surveyed by underwater

visual census (UVC) and baited remote underwater video (BRUV). Habitats are Undaria and Barren;

Locations are pooled into western or northern Port Phillip Bay.

df pseudo-F p

UVC IN-SEASON

Habitat 1 10 0.0002

Location 1 4.4 0.002

Habitat x Location 1 4.9 0.002

Residual 47

UVC OFF-SEASON

Habitat 1 2.5 0.04

Location 1 1.9 0.10

Habitat x Location 1 1.1 0.41

Residual 14

UVC OVERALL

Habitat 1 7.6 <0.0001

00 Season 1 6.9 <0.0001

Location 1 4.0 0.002

Habitat x Season 1 1.3 0.26

Habitat x Location 1 3.0 0.01

Season x Location 1 1.7 0.14

Habitat x Season x Location 1 1.0 0.43

Residual 61

BRUV IN-SEASON

Habitat 1 1.2 0.28

Location 1 1.9 0.06

Habitat x Location 1 1.6 0.11

Residual 25

171

Appendix 5.6. Fish population metrics from baited remote underwater video (BRUV) and underwater

visual census (UVC) surveys conducted on Undaria and Barren plots inside and outside the Undaria

growing season. Positive Cohen’s d effect sizes indicate metrics were higher in Undaria plots. Metrics

were compared using generalized linear models with a quasi-Poisson error distribution. Five BRUV

deployments in the few remaining Ecklonia beds at Williamstown (northern bay) and St Leonards

(southern bay) were suggestive of still higher species richness, but similar MaxN to Undaria

deployments (S = 4.2 ± 1.2, MaxN = 11 ± 5.1).

Undaria (µ ± SE) Barren (µ ± SE) Statmod df, res df p Cohen’s d

UVC IN-SEASON

Species richness 2.2 ± 0.2 1.1 ± 0.1 t1,49 = 4.4 <0.0001 +1.19

Abundance 3.3 ± 0.5 1.7 ± 0.2 t1,49 = 3.2 0.002 +0.87

UVC OFF-SEASON

Species richness 3.0 ± 0.4 3.1 ± 0.5 t1,16 = 0.17 0.87 -0.08

Abundance 4.3 ± 0.8 6.7 ± 1.0 t1,16 = 1.8 0.09 -0.85

BRUV IN-SEASON

Species richness 2.8 ± 0.6 2.0 ± 0.4 t1,31 = 1.1 0.27 +0.39

Combined MaxN 12.6 ± 4.3 6.7 ± 4.3 t1,31 = 1.0 0.34 +0.37

172

Appendix 5.7. Conceptual model outlining the temporal relationship between Undaria pinnatifida and

reproductive phases of macroalgae-associated reef fishes. Undaria habitat quality is most likely to

influence reproductive success when seasonal Undaria canopy coincides with periods of reproductive

provisioning, larval settlement, or both. The kernel area where Undaria density overlaps with a

reproductive phase is proportional to the likelihood of an individual being influenced by Undaria

habitat during that phase. Plots represent (A) minimal temporal overlap between Undaria canopy and

reproductive provisioning and no overlap with settlement, (B) moderate coincidence during

reproductive provisioning but settlement largely decoupled (e.g. Neoodax balteatus), and (C) strong

coincidence with both reproductive provisioning and settlement (e.g. Heteroclinus spp.).

Minerva Access is the Institutional Repository of The University of Melbourne

Author/s:

Barrett, Luke

Title:

Habitat preferences and fitness consequences for fauna associated with novel marine

environments

Date:

2017

Persistent Link:

http://hdl.handle.net/11343/216400

File Description:

Complete thesis

Terms and Conditions:

Terms and Conditions: Copyright in works deposited in Minerva Access is retained by the

copyright owner. The work may not be altered without permission from the copyright owner.

Readers may only download, print and save electronic copies of whole works for their own

personal non-commercial use. Any use that exceeds these limits requires permission from

the copyright owner. Attribution is essential when quoting or paraphrasing from these works.