Assessing and Understanding Ecological Changes to Fish Communities in Highly Disturbed Estuaries

ASSESSING AND UNDERSTANDING ECOLOGICAL CHANGES TO FISH COMMUNITIES IN HIGHLY DISTURBED ESTUARIES

Andrew McKinley

MA.Sc., B.A. (Hons)

Evolution and Ecology Research Center

School of Biological, Earth and Environmental Sciences

University of New South Wales

Sydney, Australia

A thesis submitted in fulfillment of the requirements for the degree of

Doctor of Philosophy within the University of New South Wales

August 2012 Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Thesis Statements

‘I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.'

Signed

Date August 13, 2012

Authenticity Statement

‘I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.’

Signed

Date August 13, 2012

Originality Statement

‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.’

Signed

Date August 13, 2012

i Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Thesis Dissertation Statement

THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet

Surname or Family name: McKinley

First name: Andrew Other name/s:

Abbreviation for degree as given in the University calendar: PhD

School: Biological, Earth, and Environmental Science Faculty: Science

Title: Assessing and Understanding Ecological Changes to Fish Communities in Highly Disturbed Estuaries

Abstract

Marine fish communities are threatened by the interrelated effects of anthropogenic disturbance, habitat modification, and contamination. This thesis examines the ways in which these stressors affect the ecology of fish communities living in estuaries. Using a variety of methods I documented large scale impacts among fish communities in heavily modified vs. relatively unmodified estuaries. First, a comprehensive meta-analysis identified trends and knowledge gaps. Studies on adult fish have generally shown weakly negative impacts from contamination or largely positive impacts where enriching contaminants are present. I conducted two field studies of adult fish, the results of which were broadly consistent with the literature. Surveys of large bentho-pelagic fish using underwater video showed increased abundance of targeted species in heavily modified/nutrient enriched estuaries. However, beach seine surveys of small bodied species indicated little impact on the beach fish community, even where high levels of modification and contamination were detected. Instead, beach fish were highly correlated to physico- chemical gradients. This suggests that impacts are highly variable among adult fish, and that both ecological characteristics and habitat preferences play an important role. Prior to this study nearly no research had been published assessing the impacts of stressors on marine larvae. Substantial differences in larval communities were detected in heavily modified areas, including increased abundance and diversity, large shifts in the occurrence of species, and changes to the overall composition of the community. These trends were highly correlated to contamination of trace metals in the sediment and loss of seagrass cover. Impacts on larval fish were greatest among fully estuarine taxa and those with benthic eggs. Lastly, a final project indicated strong impacts from the accumulation of some trace metals (Cu, Zn, and Se) in muscle tissue of Psuedorhombus jenynsii on the relative body size of the species. In combination these results suggest that stressors have their greatest impact at the early stages of a fishes’ life cycle and that many positive impacts exist in heavily modified estuaries (e.g. increased abundance/diversity). Also, my results suggest that ecological characteristics and habitat mediation may play an important role in determining the relative sensitivity of taxa.

Declaration relating to disposition of project thesis/dissertation

I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).

August 13, 2012

…………………………………………………………… ……………………………………..……………… ……….……………………...…….… Signature Witness Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

THIS SHEET IS TO BE GLUED TO THE INSIDE FRONT COVER OF THE THESIS

ii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Acknowledgments

Firstly, I would like to thank my supervisors Dr. Emma Johnston and Dr. Matthew Taylor. Thank you so much for guiding me through an amazing thesis experience and for enriching my time in Australia immensely. They say that having a smart, attentive, and dedicated supervisor is essential for doing a good PhD, and I was lucky enough to have two!

I would also like to thank all of my other collaborators and co-authors. All of you made huge contributions to my work and the papers that we did together. Without your help and guidance, we never would have achieved so much; Tony Miskiewicz, Melinda Coleman, Nathan Knott, Katherine Dafforn, and Graeme Clark.

Many thanks to all of the advisors who helped to shape this project with their expert knowledge and advice; Iain Suthers, Richard Kingsford, Stuart Simpson, Jerom Stocks and Bob Crease. I would also like to thank the late Tony Roach for your advice in planning my projects and with the toadfish work. You will be sorely missed.

I am really grateful to everyone in the Subtidal Ecology and Ecotoxicology Group for their help and support. In particular, I would like to thank all my friends and field work helpers: Shinjiro Ushiama, Cian Foster-Thorpe, Valeriya Komyakova, David Day, Joey Baba, Laura Ryan, Anthousa Harris, Katelyn Edge, Andrew Johnson, and Natalie Rivero. Every one of you put up with exhausting work in difficult conditions (day and night, rain or shine!) to help me with my project. All of you made a tremendous contribution to my PhD and I am forever grateful for the good times we had together.

I would also like to thank the organizations that helped make this project possible including: Bluescope Steel, UNSW, the EERC & BEES, the National Science and Engineering Research Council of Canada (NSERC), the Sydney Aquarium/AMSA Conservation Fund, I&I NSW, Marine Parks NSW, CSIRO, ASFB and DPI Fisheries. Thanks to all of you for making this thesis possible.

Lastly, I would like to thank my family for all of their love and support. Thank you to my Dad, Garth, Evan, Laura, Daisy, and Grandmum for caring for me from the other side of the world and for the visits. I’ve missed you all dearly and your support and love have been essential to my happiness. Finally, I would like to thank my wife Anthousa, who traveled with me all the way to Australia to support me in this endeavor. We’ve had a wonderful adventure together and your support and love have always been the foundation of my life. Without you none of it would ever have been possible….

iii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Table of Contents

Thesis Statements i

Thesis Dissertation Statement ii

Acknowledgments iii

Table of Contents iv

Abstract v

Acknowledgments of Contributions by Co-Authors vi

Introduction ix

Chapter 1: Impacts of contaminants on fish abundance and species richness: A review and meta-analysis of evidence from the field pp. 175-191

Chapter 2: High levels of sediment contamination have little influence on estuarine beach fish communities pp. 1-15

Chapter 3: Putting marine sanctuaries into context: A comparison of estuary fish assemblages over multiple levels of protection and disturbance pp. 1-13

Chapter 4: Strong links between metal contamination, habitat modification and estuarine larval fish distributions pp. 1499-1509

Chapter 5: Anthropogenic activities differentially impact fish guilds: The importance of understanding life history characteristics pp. 1-43

Chapter 6: Relationships between body burdens of trace metals (As, Cu, Fe, Hg, Mn, Se, Zn) and the relative body size of small tooth flounder (Pseudorhombus jenynsii) pp. 84-94

Discussion and Conclusions xv

Future Considerations xxxiii

Works Cited (Introduction and Discussion) xxxv

iv Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Abstract

Marine fish communities are threatened by the interrelated effects of a variety of forms of anthropogenic disturbance, including habitat modification and contamination. This thesis examines the ways in which these stressors affect the ecology of fish communities living in estuaries. Using a variety of methods I documented patterns of large scale impacts among fish communities in heavily modified vs. relatively unmodified estuaries. First, a comprehensive meta-analysis identified trends and knowledge gaps. Studies on adult fish have generally indicated weak negative impacts from contamination or largely positive impacts where enriching contaminants are present. I conducted two field studies of adult fish, the results of which were broadly consistent with the literature. Surveys of large bentho-pelagic fish using underwater video showed increased abundance of targeted species in heavily modified/nutrient enriched estuaries. However, beach seine surveys of small bodied species indicated little impact on the beach fish community, even where high levels of modification and contamination were detected. Instead, beach fish were highly correlated to physico-chemical gradients. This suggests that impacts are highly variable among adult fish, and that both ecological characteristics and habitat preferences play an important role in mediating fish responses. Prior to this study very little research had been published assessing the impacts of stressors on marine larvae. Substantial differences in larval communities were detected in heavily modified areas, including increased abundance and diversity, large shifts in the occurrence of species, and changes to the overall composition of the community. These trends were highly correlated to contamination of trace metals in the sediment and loss of seagrass cover. The strongest patterns were observed for fully estuarine taxa and those with benthic eggs. Lastly, a final project indicated strong impacts from the accumulation of some trace metals (Cu, Zn, and Se) in muscle tissue of Psuedorhombus jenynsii on the relative body size of the species. In combination these results suggest that stressors have their greatest impact at the early stages of a fishes’ life cycle and that many positive impacts exist in heavily modified estuaries (e.g. increased abundance/diversity) possibly as a result of nutrient enrichment. Also, my results suggest that ecological characteristics and habitat mediation play an important role in determining the relative sensitivity of taxa.

v Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Acknowledgments of Contributions by Co-Authors

At the time of submission, all six data chapters are published, in press, or in review in high impact peer reviewed scientific journals. As such, this thesis represents a submission as a series of publications. I am the primary author of all six papers and the majority of the data which comprises these works is original data which I gathered. I have also conducted the primary analysis of all data, the writing of the manuscripts, and undertaken the submission/revisions process for all of these publications. As such, I have conducted the majority of the work in these publications.

However, I have had significant contributions from co-authors and collaborators, including data contributions. In most cases I have used data from my co-authors as covariates in the analysis. In this way, the primary biological dataset (fish data) has been my own while I have used outside data to provide context in the analysis. In all cases these contributions have been acknowledged in the methods section of the manuscript or through the inclusion of the collaborator as a co- author.

The specific contributions to each chapter are detailed here (initials as in citation):

Chapter 1:

McKinley, A.C. & Johnston, E.L. (2010) Impacts of contaminants on fish abundance and species richness: A review and meta-analysis of evidence from the field. Marine Ecology Progress Series, 420: 175-191.

x ACM - All primary data gathering, primary analysis, manuscript writing, submission, and revisions. x ELJ – Concept and guidance, manuscript review and revisions. Supervision and funding.

Chapter 2:

McKinley, A.C. Dafforn K.A., Taylor, M.D. & E.L. Johnston (2011) High levels of sediment contamination have little influence on estuarine beach fish communities. PLoS One, 6(10).

x ACM - All biological and physico-chemical covariates data gathering. Primary analysis, manuscript writing, submission, and revisions. x KAD – Provided sediment metals and PAH data. x ELJ/MDT/KAD – Concept, design and guidance, review and revisions of manuscript. Supervision, funding and equipment.

vi Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Chapter 3:

x ACM - All field work for biological data. Processing of BRUV tapes for Sydney Harbor/Port Hacking. Primary analysis, manuscript writing, submission, and revisions. x LR - Processing of BRUV tapes for Jervis Bay/Clyde River, review of manuscript. x GC – Statistical design and support, supervision of LR, review of manuscript. x MAC/NAK – Field work support, BRUV equipment, study design, review of manuscript. x ACM/MAC – Revisions for resubmission to Aquatic Conservation (following rejection from ECSS). x MDT/ELJ – Concept and guidance, review and revisions to manuscript. Supervision, funding and equipment.

Chapter 4:

x ACM - All field and lab work for biological and physico-chemical data. Primary analysis, manuscript writing, submission, and revisions. x AM – Taxonomic verification of larvae. Guidance in larval methodology and taxonomy. Concept and review of manuscript. x MDT/ELJ – Concept, design and guidance, review and revisions to manuscript. Supervision, funding and equipment. o Sediment metals contamination data provided by Katherine A. Dafforn (acknowledged in methods).

Chapter 5:

McKinley, A.C., Foster-Thorpe, C., Miskiewicz, A., Taylor, M.D. & E.L. Johnston. (2012) Anthropogenic activities differentially impact fish guilds: The importance of understanding life history characteristics.

x ACM - Primary analysis, manuscript writing, submission, and revisions. All literature review and data gathering for ecological dataset (guild classifications) including for benthic and surface larvae. x ACM/CFT – All fieldwork for biological and physico-chemical datasets. x ACM – Laboratory identification and sorting of all benthic larvae - 73% of dataset. x CFT – Laboratory identification and sorting of all surface larvae – 27% of dataset. Review of manuscript.

vii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

o Note; Surface larvae data was needed in this study in order to increase the diversity and replication of estuarine opportunists and marine straggler taxa in order to facilitate the guild analysis. In isolation, the benthic larval data did not have enough taxa from these guilds for a robust analysis. x AM – Taxonomic verification of larvae. Guidance in larval methodology and taxonomy. Concept and review of manuscript. x MDT/ELJ – Concept, design and guidance, review and revisions to manuscript. Supervision, funding and equipment. o Sediment metals contamination data provided by Katherine A. Dafforn (acknowledged in methods).

Chapter 6:

McKinley, A.C., Taylor, M.D. & E.L. Johnston (2012) Relationships between body burdens of trace metals (As, Cu, Fe, Hg, Mn, Se, Zn) and the relative body size of small tooth flounder (Pseudorhombus jenynsii). Science of the Total Environment, 423: 84-94.

x ACM - All flounder data, benthic fish abundance data, and physico-chemical covariates data gathering. Primary analysis, manuscript writing, submission, and revisions. x ACM – Flounder dissections, otolith extraction, processing, and counting, tissue sample preparation and freeze drying. o Prepared tissue samples were analyzed for metals content by Stuart Simpson and Chad Jaromilek at CSIRO (acknowledged in methods). o Sediment metals contamination and sediment grain size data provided by Katherine A. Dafforn (acknowledged in methods). x MDT/ELJ – Concept, design and guidance, review and revisions to manuscript. Supervision, funding and equipment.

viii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Introduction

A variety of anthropogenic activities contribute to widespread pollution and contamination in the marine environment, which influences the composition and health of ecological communities (Johnston & Roberts 2009). Estuaries are generally believed to be exposed to the highest levels of contamination of any marine environment due to their proximity to human settlements and their position directly downstream of agricultural and industrial activities (Kennish 2002, Lotze et al. 2006). Similarly, habitat modification in estuarine systems is widespread, and many estuaries around the world have experienced losses of seagrass, mangrove, saltmarsh and other vegetated habitats (Lotze et al. 2006, Duke et al. 2007, Waycott et al. 2009). Many of these complex estuarine habitats provide a ‘nursery’ function for ecologically and economically important species of fish and are essential to their life cycle (Boesch & Turner 1984, Robertson & Duke 1987, Beck et al. 2001, Dorenbosch et al. 2004, Taylor et al. 2005). Assessing and understanding the ecological impact of these stressors is critical to managing and conserving native biodiversity in these systems. Many marine fish species have experienced both rapid and long term population declines attributable to a variety of anthropogenic stressors (Pauly et al. 2002). Anthropogenic impacts on fish abundance and diversity are a significant management concern, and it is likely that shifting fish populations are driving further changes to marine ecosystems (Frank et al. 2005, Jiao 2009). The majority of scientific literature addressing declining stocks focuses on the impacts of commercial fishing on a few economically important species. Anthropogenic factors such as climate change, invasive species, habitat modification, and marine contamination are little studied within the context of changing marine fish assemblages (Rose 2000, Perry et al. 2005). However, these stressors may play a large and poorly understood role in structuring fish communities and may act synergistically with commercial fishing pressures in some contexts (Micheli 1999, Islam & Tanaka 2004, Hylland 2006b, Breitburg et al. 2009). This thesis focuses primarily on the impacts of contamination, habitat modification, and to a lesser extent fishing pressure within the context of large scale estuarine disturbance. A great deal of research examines the physiological effects of contaminants, including their presence, biomagnification, toxicology, and biomarker induction in marine fish populations (Costello et al. 1994, Wirgin et al. 1998, Austin 1999, van der Oost et al. 2003, Hylland 2006b) It is well documented that toxic contaminants such as metals are found in fish at various stages

ix Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

of their life cycle, often at levels that may potentially reduce growth or survivorship (Miskiewicz & Gibbs 1994, Alquezar et al. 2006, Isosaari et al. 2006, Kojadinovic et al. 2007, Guo et al. 2008). The less toxic enriching contaminants (such as nutrients or sewage) may have either a weakly negative or largely positive effect on abundance and diversity of adult fish (see Chapter 1). The majority of this research addresses the issue at the genetic, cellular, chemical, or organism level, while comparatively little research has been done linking this information to changes in wild populations and assemblages (Rose 2000, Clements & Rohr 2009). As a result, ecologically pertinent information regarding the impacts of contaminants on wild fish abundance, assemblage structure, ecological function, and biodiversity is lacking. This contrasts with other aquatic organisms, such as infaunal invertebrates, which are comparatively well studied within this context (Johnston & Roberts 2009). Modification to marine habitats represents another potential stressor of fish communities. Habitat degradation is the largest source of ecological modification globally and the greatest threat to biodiversity (Tilman et al. 1994). Australian estuaries have experienced widespread changes to vegetative habitats over the last century, with well documented losses of mangrove (Valiela et al. 2001), salt marsh (Saintilan & Williams 1999), and seagrass habitats (Walker & McComb 1992) in south-eastern Australia. The reduced extent of these habitats within heavily modified estuaries arises due to a variety of anthropogenic activities including dredging, increased siltation, nutrient enrichment, contamination, clearing for coastal development, and alterations to natural tidal or fluvial patterns (Walker & McComb 1992, Saintilan & Williams 1999, Valiela et al. 2001). Degradation of estuarine macrophyte communities is likely to lead to changes to fish communities, as many species require these habitats for food, shelter, and reproduction (Boesch & Turner 1984, Robertson & Duke 1987, Beck et al. 2001, Dorenbosch et al. 2004). Despite the potential importance of these stressors, the effects of contamination and habitat modification on fish ecology, abundance, and biodiversity have received relatively little attention in the scientific literature (Murphy et al. 2008) (see Chapter 1). While toxicants and habitat modification are thought to have significant impacts on wild fish communities, natural variation in physico-chemical conditions such as changes in turbidity, salinity, temperature and pH, as well as geomorphological and oceanographic factors, have consistently been shown to have a large influence on the composition and species richness of fish assemblages (Potter & Hyndes 1999). Spatial variation in physico-chemical factors can manifest

x Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

as gradients within estuarine systems, influencing the distribution of fish species along the length of an estuary (Rakocinski et al. 1992). Similarly, seasonal and temporal variability in physico- chemical factors can influence fish communities (Taylor et al. 2006). Estuary geomorphology and physical structure will also affect estuarine ecology through variation in entrance conditions (e.g. permanently or intermittently open estuaries), the relative size of the fluvial and tidal deltas, and the evolutionary maturity (stage of sediment filling) of the estuarine system (Roy & Williams 2001). The way in which these factors influence the fish assemblage is increasingly juxtaposed against the effects of anthropogenic modification to estuaries. As such, identifying the major drivers of fish distribution is likely to be more complicated in modified habitats, and any assessment of the impacts of anthropogenic stressors on fish communities must be conducted within the context of physico-chemical and geomorphological variability. The way in which anthropogenic stressors affect fish communities is complex and it is likely that a variety of factors will influence or mediate ecological effects (discussed fully in Chapter 1). Some of these factors have been examined in detail in this thesis. This includes chapters focused on assessing the role of physico-chemical variability (Chapter 2), fishing pressure and conservation initiatives (sanctuary zones) relative to large scale disturbance (Chapter 3), increased sensitivity at early life stages (Chapter 4&5), ecological characteristics in determining relative sensitivity (Chapter 5), and physiological impacts (Chapter 6). In this thesis the focus is primarily on assessing in situ community level impacts, though Chapter 6 also examines physiological impacts on a model species. This is achieved both through traditional survey techniques and the employment of novel methods for observing real impacts in a field environment.

Terminology Throughout this thesis estuaries have been classified as either heavily modified or relatively unmodified. This classification is made based on a variety of factors. Heavily modified estuaries are all highly anthropogenically disturbed environments near large urban and industrial areas and are subject to intense commercial and recreational boating traffic, historic and ongoing contamination, greater recreational fishing activity, and widespread urbanization of their shoreline and catchment (Birch & Taylor 1999, Henry & Lyle 2003, DPI 2010, Scanes 2010). The results presented in Chapters 2-6 also clearly indicate that these estuaries have greater levels

xi Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

of sediment contamination and decreased cover of vegetative environments (e.g. mangroves and seagrass beds) compared to the relatively unmodified estuaries. Compared to the modified estuaries, the relatively unmodified estuaries have less recreational fishing activity, less boating traffic, less urbanization of the coastline and catchment, and virtually no heavy industry (Birch & Taylor 1999, Henry & Lyle 2003, DPI 2010, Scanes 2010). Both the Clyde River (within Bateman’s Bay Marine Park), Jervis Bay (Jervis Bay Marine Park), and Karuah River (Port Stephens Marine Park) are within marine parks (NSW 1999). Port Hacking is located between the suburbs of southern Sydney and the forested slopes of Royal National Park, which lines the southern border of the estuary. While not strictly within a marine park, Port Hacking’s catchment is largely intact due to its proximity to the Royal National Park and there is no major industrial activity within the estuary (NSWDNR 2010). Previous monitoring indicates that some of the heavily modified estuaries are nutrient enriched while nutrient levels in the relatively unmodified estuaries are less elevated (Scanes 2010). Heavily modified estuaries referred to in Chapters 2-6 include the Hunter River, Port Jackson, Botany Bay, and Port Kembla. Relatively unmodified estuaries include Karuah River, Broken Bay, Port Hacking, Clyde River (Batemans Bay), Jervis Bay, and Wagonga Inlet. Note that not all of these estuaries appear in every chapter. Throughout this thesis trends are described as ‘positive’ or ‘increasing’ effects or conversely as ‘negative’ or ‘decreasing’ effects. These descriptions are usually made in reference to trends in abundance, biomass, species richness, or diversity. For example, “…contaminants were shown to have a positive effect on community abundance.” In this thesis the terms ‘positive’ and ‘negative’ are not meant to indicate a judgement on the benefit or desirability of an observed trend. Instead, these terms are meant to impartially describe the direction of an effect. A positive effect is one where there is an increase, e.g. a positive effect on community abundance is meant to describe an increase in community abundance, while a negative effect would describe a decrease. No evaluation or judgement is made on the desirability of these effects.

xii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

In summary, this thesis describes fish assemblages in relation to contamination, habitat modification, and fishing pressure. The initial goals of this study were to: 1. Provide a better understanding of the impacts of contamination and habitat modification on estuarine fish assemblages. 2. Examine potential impacts primarily at the community level in a field environment. 3. Undertake novel methods for assessing potential impacts in a field situation (Chapter 6). 4. Examine the influence of mediating factors including physico-chemical variability, fishing pressure and conservation initiatives, susceptibility of early life stages, the importance of ecological characteristics in determining relative sensitivity, and direct physiological impacts.

The topics addressed in each chapter include the following: x Chapter 1: Impacts of contaminants on fish abundance and species richness: A review and meta-analysis of evidence from the field. This chapter provides a synthesis of the literature on this subject and provides a conceptual foundation for the remaining chapters. It also summarizes trends in the literature from similar studies through a meta-analysis. x Chapter 2: High levels of sediment contamination have little influence on estuarine beach fish communities. This chapter examines patterns of juvenile estuarine beach fish in relation to levels of disturbance and contamination, within the context of physico- chemical and temporal variability. Sampling was conducted using a beach seine net and focused on juvenile and small bodied fishes living in shallow beach environments. x Chapter 3: Putting marine sanctuaries into context: A comparison of estuary fish assemblages over multiple levels of protection and disturbance. This chapter examines the effectiveness of sanctuary zones and places the impacts of recreational fishing pressure within the context of large scale disturbance patterns. Sampling was conducted using baited underwater video systems targeting medium to large bentho-pelagic and pelagic fishes. x Chapter 4: Strong links between metal contamination, habitat modification and estuarine larval fish distributions. This chapter examines the sensitivity of early life stages (larvae) and the potential impacts of both vegetative habitat loss and sediment contamination. Chapter 4 & 5 demonstrate much stronger patterns of potential impact at early life stages compared to adults (Chapters 2&3). Sampling was conducted using a benthic larval trawl. x Chapter 5: Anthropogenic activities differentially impact fish guilds: The importance of understanding life history characteristics. This chapter utilizes an ecological guild approach to determine the relative sensitivity of different larval guilds. The impacts of vegetative habitat loss, sediment contamination, and estuary morphological factors are addressed. Sampling of larvae was conducted using a benthic trawl and a surface tow. x Chapter 6: Relationships between body burdens of trace metals (As, Cu, Fe, Hg, Mn, Se, Zn) and the relative body size of small tooth flounder (Pseudorhombus jenynsii). This

xiii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

chapter examines trace metal concentrations in a model species, Pseudorhombus jenynsii. Specifically, it examines how the relative body size (as an indirect measure of the relative growth rate) is potentially influenced by trace metals accumulation in muscle tissue. Flounder were sampled using an otter trawl, growth rates measured using otolith ring counting, and tissue accumulation using ICP-AES/ICP-MS.

xiv

Chapter 1

IMPACTS OF CONTAMINANTS ON FISH ABUNDANCE AND SPECIES RICHNESS: A REVIEW AND META-ANALYSIS OF EVIDENCE FROM THE FIELD

Final Version:

Vol. 420: 175–191, 2010 MARINE ECOLOGY PROGRESS SERIES Published December 16 doi: 10.3354/meps08856 Mar Ecol Prog Ser

Impacts of contaminant sources on marine fish abundance and species richness: a review and meta-analysis of evidence from the field

Andrew McKinley*, Emma L. Johnston

Evolution & Ecology Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, New South Wales 2052, Australia

ABSTRACT: We conducted a systematic review and meta-analysis of the in situ effects of contaminant sources on fish abundance and species richness. We discuss these effects and examine the role of contaminant sources, habitats, and study systems. We also highlight the role of fish ecology in determining contaminant impacts, mechanisms of impact, and areas for future research. A total of 45 papers were analyzed in the meta-analysis, which revealed that the average change in abundance at contaminated sites was +103× (fish farms), +40% (sewage studies), –52% (industrial effluent), and –65% (run-off). This analysis suggested that the abundance of fish assemblages in coral reefs was negatively impacted by contaminants and that these reefs are more sensitive than other habitats. Weak trends were observed for species richness, which may suggest that contamination is not having an impact on fish diversity. Some sources of contamination that are also sources of food are having sizeable effects on abundance and richness and are likely to be of commercial and environmental significance. Further research is warranted, particularly where contamination may act as an additional stressor in impacted communities.

KEY WORDS: Marine contamination · Abundance and richness of fish

Resale or republication not permitted without written consent of the publisher

INTRODUCTION with commercial fishing pressures in some contexts (Micheli 1999, Islam & Tanaka 2004, Hylland 2006b, Many marine fish species have experienced both Breitburg et al. 2009). rapid and long-term population declines attributable to The effects of marine contamination on fish ecology, a variety of anthropogenic stressors (Pauly et al. 2002). abundance, and biodiversity have received relatively Anthropogenic impacts on fish abundance and diver- little attention in the scientific literature (Murphy et al. sity are a significant management concern, and it is 2008). A great deal of research examines the physio- likely that shifting fish populations are driving further logical effects of contaminants, including their pres- changes to marine ecosystems (Frank et al. 2005, Jiao ence, biomagnification, toxicology, and biomarker 2009). The majority of scientific literature addressing induction in marine fish populations (Costello & Read declining stocks focuses on the impacts of commercial 1994, Wirgin & Waldman 1998, Austin 1999, van der fishing on a few economically important species. Oost et al. 2003, Hylland 2006b). Contaminants are Anthropogenic factors such as climate change, inva- widespread within fish populations, often at levels that sive species, habitat modification, and marine contam- could reduce growth, fecundity, or survivorship, and ination are little studied within the context of changing which may ultimately have an influence on structuring marine fish assemblages (Rose 2000, Perry et al. 2005). fish communities (Jones & Reynolds 1997, Kingsford et These stressors may play a large and poorly under- al. 1996a, Arkoosh et al. 1998a, Robinet & Feunteun stood role in structuring fish communities, and there is 2002). The majority of this research addresses the issue evidence to suggest that they may act synergistically at the genetic, cellular, chemical, or organism level,

*Email: [email protected] © Inter-Research 2010 · www.int-res.com 176 Mar Ecol Prog Ser 420: 175–191, 2010

while comparatively little research has been done link- shown to have a negative effect at a given concentra- ing this information to changes in wild populations and tion. Thus, we have chosen the term ‘contaminant’, as assemblages (Rose 2000, Clements & Rohr 2009). As a we do not wish to imply that these chemical alterations result, ecologically pertinent information regarding necessarily have a significant, positive, or negative the impacts of contaminants on wild fish abundance, environmental impact. Instead, these terms are simply assemblage structure, ecological function, and biodi- meant to describe the addition of the chemical sub- versity is lacking. This contrasts with other aquatic stances discussed above, irrespective of any normative organisms, such as infaunal invertebrates, which are judgment of environmental impacts. comparatively well studied within this context (John- Because very few studies isolate the effects of a sin- ston & Roberts 2009). gle contaminant class on fish, and because contamina- Review articles addressing contamination effects in tion in the natural environment usually consists of a fish have focused on nitrogen enrichment (Micheli mix of chemical inputs from a range of sources, studies 1999, Nixon & Buckley 2002, Breitburg et al. 2009), are analyzed in the following meta-analysis according sewage outflows (Pastorok & Bilyard 1985, Grigg to broad contaminant source groups. Each of these 1994), and regional impacts of contamination (Dethlef- represents a mixed range of contaminants, some of sen & Tiews 1985, Nagai 2003, Hylland et al. 2006a). which may have positive effects for fish communities. No comprehensive review examining the effects of These include fish farms, sewage outflows, nutrient contamination on fish abundance and diversity cur- run-off, and industrial effluent. These contaminant rently exists. Key questions remain regarding the mag- sources can be broadly characterized based on the typ- nitude of impacts on fish populations and diversity, ical mixture of chemicals that they emit. Typically, fish which families or species are most affected, and which farms emit a mixture of fish faeces, dissolved nitrogen habitats are most vulnerable. Reliable information waste, parasite and chemical treatment baths, anti- addressing these topics could improve management foulants, and excess fish feed (Wu et al. 1994). Fish of fisheries resources and wild fish assemblages. farms are thus both a source of food subsidy/nutrient Many forms of contaminants are present in the enrichment and to a lesser extent a source of tradi- marine environment with varying toxicities, and each tional chemical contaminants. This does not imply that may impact diversity and abundance differently. Con- these farms necessarily have a negative or positive taminants such as metals are present in the natural environmental impact (discussed in detail in ‘Discus- environment at trace levels, while chemicals such as sion’ section ‘Contaminant sources’). some pesticides and persistent organic pollutants Both run-off and sewage outflows are highly vari- (POPs) are entirely artificial compounds that do not able contaminant sources that may simultaneously exist naturally. Metals, polycyclic aromatic hydrocar- provide food subsidies, nutrient enrichment, and high bons (PAHs), pesticides, and POPs are contaminants concentrations of potentially toxic chemicals to marine that display toxicological effects in fish at some con- systems. It is possible that differential responses to centrations (Hylland 2006b). These contaminants can these contaminant sources are a result of different be broadly characterized as ‘potentially toxic contami- chemical mixtures, concentrations of high impact pol- nants’. Other classes of contaminants such as nutrients, lutants, or effluent treatment practices (Bishop et al. organic sewage, food subsidies, and thermal effluents 2006b, Ribeiro et al. 2008). Industrial effluents repre- may not be acutely toxic but instead affect the ecosys- sent another mixed contaminant source; they differ tem by altering food availability or the physio-chemi- from the other contamination sources examined in this cal conditions of the environment. These chemicals study, as they contain comparatively few food subsi- can be broadly characterized as ‘enriching’ or ‘physio- dies and nutrients while potentially toxic contaminants chemical contaminants’. Responses to physio-chemical are often present (such as metals and pesticides) alteration are not always linear, and they are known to (Costello & Read 1994, Austin 1999, Hylland 2006b). enhance the abundance and diversity of fish in some It is likely that fish living in some marine habitats are contexts (Micheli 1999, Nixon & Buckley 2002, Breit- more susceptible to contamination than others. Habi- burg et al. 2009). It is important to note that in this tats such as seagrasses and coral reefs are known to be review we are using the term ‘contaminant’ rather highly sensitive to changing environmental conditions than ‘pollutant’, as it is generally accepted in the eco- and contamination; this sensitivity may mean that fish toxicology literature that ‘contaminants’ describe that are closely associated with these habitats are dis- chemicals added to the environment above their nat- proportionately affected by contaminants (Pastorok ural concentrations that have not necessarily been & Bilyard 1985, Fabricius 2005b, Deegan et al. 2007). demonstrated to have a negative impact at a given Similarly, habitats are subject to varying levels of concentration. By contrast, ‘pollutants’ are generally contaminant exposure, and those that are situated thought of as chemical substances that have been closer to contamination sources are normally exposed McKinley & Johnston: Impacts of contaminant sources on marine fish 177

to higher levels of contaminants than those further search terms was applied in 4 major biological data- away. The amount of water flow or flushing in a habitat bases: Aquatic Sciences and Fisheries Abstracts is also a significant determinant of contamination expo- (1971–present), Biological Abstracts (1969–2003), Cur- sure. Ecosystems in shallow protected waters that are rent Contents (1998–present), and Web of Science poorly flushed are more likely to accumulate and retain (1900–present). The following search terms were used higher concentrations of contaminants than habitats to search these databases, and results were limited to that are well flushed, in deeper water, or offshore English language studies, in peer reviewed journals, (Roberts et al. 2010). In this meta-analysis studies have that were available in full text online: been broadly categorized according to habitat types in- Search 1: marine* AND fish* AND (contamina* OR cluding bare sediment, coral reefs, rocky reef and rocky pollut*); bottom, and vegetated habitats. Bare sediment habitats Search 2: divers* OR biodivers* OR index* OR com- are areas in which vegetation and hard substrate are munity* OR assemblage* OR abundance* OR stock* mostly absent and the bottom is characterized primarily OR population*; by sand, mud, or silt. These habitats may occur in shal- Search 3: hydrocarbon* OR PAH* OR oil* OR metal* low or deep waters; they can be poorly or well flushed OR nutrient* OR sewage* OR solid waste* OR efflu- depending on where they occur. Contaminants prefer- ent* OR estrogen* OR androgen* OR thermal* OR entially bind to fine sediments and are a common envi- pesticide* OR herbicide* OR brine* OR farm* OR ronmental concern in soft sediment habitats from an- aquaculture*. thropogenically modified harbors (Knott et al. 2009). These 3 searches were cross-referenced against one Coral reefs are primarily tropical and sub-tropical envi- another to produce a final selection of studies. We read ronments where habitat-forming colonial polyps have the abstracts of all papers that emerged from this created a hard substrate environment. These areas search (n = >1650). These studies were included in the have high fish diversity and typically occur in shallow, meta-analysis if they contained the following ele- nutrient-poor waters. Rocky reefs and rocky bottom ments: (1) Quantitative data on the species richness/ habitats typically occur in shallow near shore waters abundance of a fish species or assemblage. (2) A com- and are high energy environments that are well parison of this data between contaminated and non- flushed. Lastly, vegetated habitats include seagrass contaminated sites, before and after contamination, or beds, kelp forests, mangrove creeks, and other areas along contamination gradients. The research design containing habitat forming plant species. Most of these must have been appropriate for the investigation of the vegetative communities grow in the shallow photic impacts of contamination on the species richness/ zone and are typically in protected to moderately pro- abundance data. (3) Only in situ studies placed within tected waters. In coral reef and vegetated environ- the context of functional natural habitats were in- ments, habitat-forming organisms may be affected first cluded. Following this criteria, field surveys, field ex- by contaminants, and changes in their communities periments, and fisheries catch studies were included, may impact fish indirectly (discussed in more detail in while laboratory experiments, mesocosms, or model- ‘Discussion’ section ‘Study systems and habitat media- ing studies were excluded if they did not report field tion’) (Deegan et al. 2002, Reopanichkul et al. 2009). data. This distinction was made because this review Here, we present the results of a systematic literature focuses on in situ changes to abundance and species review and meta-analysis in which we address 4 ques- richness. tions about contaminant impacts on fish abundance and We then examined the citation lists of papers se- species richness: (1) In which habitats are researchers lected in the first round and investigated any relevant typically assessing the effects of contaminants on fish studies for inclusion in the meta-analysis. In total, this abundance and species richness? (2) Do marine habi- process produced 45 studies (<3%) that satisfied the tats vary in their susceptibility to the impacts of contam- above criteria for the meta-analysis portion of our study. ination? (3) Do different contaminant sources vary in From these studies we extracted qualitative data on their ability to impact fish abundance and species rich- the contaminant type (where described), contaminant ness? (4) Are some fish families or functional groups source, and habitat type. Due to the limited number of more susceptible to contamination than others? studies, data was clustered into broad categories according to contaminant source (fish farm, sewage, run-off, industrial effluent) and habitat type, as MATERIALS AND METHODS defined by substrate characteristics (fish farms over bare sediment, coral reef, rocky reef and rocky bottom, Search methods. In order to capture a representative vegetated habitat, bare sediment). From each study, sample of the marine contamination literature, a sys- quantitative data on species richness and abundance tematic literature review was conducted. A list of were extracted. We then collated data on the overall 178 Mar Ecol Prog Ser 420: 175–191, 2010

finding of the study (reduced, increased, or no effect measurements for an entire fish community and also on abundance/richness) as concluded by the authors for individual species, families, or trophic categories, and calculated the effect size (see Figs. 1 & 2). Not all the community measurement was always selected over studies presented data on both abundance and species measurements describing smaller species groupings. richness. Studies with only one of these measures were However, information on the abundance response of still included in the review but were partitioned into fish families or individual species was also recorded separate analyses. Some studies did not perform for- for qualitative analysis (discussed in ‘Results of meta- mal statistical analysis of results, but because they pre- analysis). Due to the different ways in which studies sented information on abundance or species richness, were structured, several procedures were created we were still able to extract the required information. to ensure consistent data extraction. Survey studies Where abundance data was presented for individual either examined abundance/richness along gradients fish species (n = 208), those species were characterized or contrasted contaminated and non-contaminated according to their swimming behavior and feeding sites. For gradient studies the effect size was calcu- pattern. These characterizations were undertaken lated by taking data from the sampling locations clos- using basic ecological information available on the est to the contamination source (or the most contami- website www.fishbase.org. In some cases descriptive nated site where contamination data was available) information provided by the authors was used to char- and contrasting it to the reference site that was furthest acterize species. Species were divided into 3 distinct from the contamination source (or the least contami- behavioral groups following the organizational system nated site). For studies that sampled over multiple used by www.fishbase.org: demersal, benthopelagic, periods (with stable contaminant conditions), data was and pelagic-neritic/pelagic/bathypelagic. Fishes cate- averaged over all periods. For studies that sampled gorized as ‘reef associated’ were removed due to an over multiple periods with increasing or decreasing insufficient sample size (n = 2). Due to a lack of suffi- contamination loads, data was extracted from the time cient information, fish could not be fully characterized of highest contamination and compared to the time of by trophic level. Instead, fish were characterized into 3 lowest contamination. One paper (Dempster et al. broad feeding guilds that approximate trophic levels. 2004) is treated throughout our review as 2 indepen- These are herbivores and planktivores, omnivores dent studies, as it presents data separately for 2 dis- (eating a mixture of algae, invertebrates, and some parate geographic regions: Spain and Australia. Lastly, fish), and predatory fishes (eating a mixture of macro- for studies that sampled over multiple sampling sea- invertebrates and fish). Fish farm data was excluded sons but found significant results only in one season, from this analysis, as the overwhelmingly positive data was extracted from the significant season. Due to effect of fish farms would obscure effects from other the limited number of papers available for the meta- forms of contamination. analysis, studies of adult and larval fishes were consid- Meta-analyses. In addition to the qualitative review, ered together. The impact of contaminants at different we performed a quantitative meta-analysis. The meta- stages of the fish life cycle is discussed in more detail analysis was partitioned into 2 distinct analyses: in ‘Discussion’ section ‘Linking contaminant effects to (1) examining any studies that reported a measure- fish ecology’. Studies examining the effects of con- ment of abundance (n = 44), and (2) examining any tamination on larval fishes (ichthyoplankton) included studies which reported species richness (n = 19). (Gray et al. 1992, Gray 1996, 1997). Effect sizes were For these meta-analyses we calculated effect sizes contrasted by habitat and contaminant source using that were attributed to contamination. We defined the separate 1-way analysis of variance (ANOVA) followed effect size as the proportional change in mean abun- by post hoc Tukey’s tests. dance/species richness between control and impacted sites (or experimental and treatment sites for field experiments). The effect size is expressed as a natural RESULTS logarithm, such that ln(effect size) = ln(impacted) – ln(control). All statistical analysis and graphing of Summary statistics results was conducted in the natural logarithm of the response variables (Hedges et al. 1999, Johnston & Of the over 1650 titles and abstracts examined, a Roberts 2009). In some cases, effect size data was total of 45 studies fulfilled the criteria for inclusion in extracted from graphs using a computerized measure- the meta-analysis (Appendix 1). The majority of stud- ment tool; these are considered estimates only. ies was conducted in highly developed nations (87%) In all cases the abundance measure was extracted while a minority was undertaken in developing coun- that represented the greatest number of species within tries (13%). Studies were clustered regionally with the a community; in studies that presented abundance majority of studies occurring in Europe (38%), Aus- McKinley & Johnston: Impacts of contaminant sources on marine fish 179

tralia (26%), and North America (21%). A minority of view, though urban sewage (36%) was most studied studies was conducted in other regions including the followed by industrial effluents (28%), fish farms Caribbean (4%), Middle East (4%) and East Asia/ (21%), and run-off (15%). The vast majority of studies Pacific (6%). The countries with the greatest propor- consisted of field surveys (91%), while fisheries catch tion of studies were Australia (26%), USA (including studies and field experiments were each ~5%. The Hawaii) (19%), Sweden (13%), and Spain (including small number of field experiments is likely due to the Canary Islands) (11%). Limiting search results to stud- comparatively high cost, as well as practical and ethical ies published in English may have resulted in the difficulty of conducting such investigations with fish. exclusion of a significant body of literature published Responses by contaminant source were variable, in other languages. The latitude of sampling was with the majority of fish farm studies (90%) reporting biased towards mid-latitude research sites, with 60% increased abundance and richness, while the majority of studies occurring between 31 and 50° N or S. Studies of industrial effluent (62%) and run-off (67%) studies at high latitude +51° N or S (19%) and low latitude 0 to reported negative abundance responses. Sewage stud- 30° N or S (21%) were less represented in the data set. ies reported negative and positive responses equally The average maximum spatial scale (defined as the for both abundance and richness (Fig. 1). Fish farms largest distance between an impacted and control site) over bare sediment were separated from other bare was 53.9 km, while the largest study was conducted sediment studies due to the large difference in the over a spatial scale of approximately 600 km. The aver- magnitude and direction of reported responses. age minimum sampling depth was 12.5 m, while the Responses by habitat type were similarly variable average maximum sampling depth was 30.7 m. Only 5 (Fig. 1). Of the studies reviewed, a minority reported studies sampled in water >40 m deep, and only 2 sam- no detectable effects of contamination upon abun- pled in water >80 m deep. dance (9%) and richness (11%), suggesting a possible The greatest number of studies were conducted over bias towards publishing research that yields statisti- bare sediment (25%), while vegetated habitats, coral cally significant results. The ecological impacts of reefs, fish farms over bare sediment, and rocky bottom marine contamination may be overestimated due to and rocky reef systems were roughly equal (~15 to 20% the tendency of journals to publish studies which find each). No single contaminant source dominated the re- significant results (Browman 1999).

a) Total abundance response by contaminant source b) Species richness response by contaminant source 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0 Fish farm (10) Industrial Run-off (6) Sewage (15) Fish farm (4) Industrial Run-off (2) Sewage (9) effluent (13) effluent (4)

c) Total abundance response by study system d) Species richness response by study system 1.0 1.0 Proportion of total 0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0 Fish farms Bare Coral reef Rocky Vegetated Fish farms Bare Coral reef Rocky Vegetated over bare sediment (6) bottom & habitats (8) over bare sediment (4) bottom & habitats (3) sediment (12) rocky reef sediment (3) rocky reef (8) (7) (4) (4)

Decreased total abundance No change Increased total abundance Decreased species richness No change Increased species richness Fig. 1. Proportion of studies in the marine contamination literature concluding that contamination has negative, positive, or no effect on abundance and richness. The number of research papers in each category is shown in parentheses 180 Mar Ecol Prog Ser 420: 175–191, 2010

Results of meta-analysis (fish farms), +40% (sewage studies), –52% (industrial effluent), and –65% (run-off). While the variation The meta-analysis supported the qualitative assess- between groups was significant (p = 0.000), Tukey’s ment; abundance and species richness responses to post hoc analysis indicated that this variation was due contamination were variable both for contaminant to the difference between fish farms and the other 3 sources and sampling habitats. Among contaminant groups. Sewage, industrial effluent, and run-off were a sources, the greatest effect sizes for abundance were observed among fish farms, which displayed a signifi- Table 1. Analysis of variance contrasting impacts of contami- cant trend of increased abundance (Fig. 2a, Table 1). nation on abundance and species richness The effect sizes for industrial effluent and run-off displayed slight trends towards decreased abundance, Effect on abundance Effect on richness and sewage appeared to show little change on aver- df MS F pdfMSF p age. While the effect sizes for these categories showed Contam- 3 32.144 11.513 0.000 3 2.638 5.617 0.009 little change on average, this obscures the fact that the inant majority of studies in both categories reported nega- Study 4 23.340 9.386 0.000 4 1.623 3.804 0.029 tive responses to contamination (Fig. 1). The average system change in abundance at contaminated sites was +103×

Fig. 2. (a) Total abundance response and (b) species richness response by contaminant source. (c) Total abundance response and (d) species richness response by study system. *Category is significantly different from all other categories (p < 0.05). x-yz pairings are significantly different (p < 0.05). All other pairings are homogenous. Error bars are ±1 SE. Note that effect sizes are expressed as a natural logarithm. Other details as in Fig. 1 McKinley & Johnston: Impacts of contaminant sources on marine fish 181

homogenous subset (Fig. 2a). For the species richness presented for a total of 208 species individually (as analysis, the greatest effect sizes were again observed opposed to being summarized within a community for fish farms, which displayed a significant trend of measurement). This data was utilized to analyze abun- increased species richness (Fig. 2b, Table 1). Industrial dance responses according to swimming behavior and effluent and run-off displayed slight trends towards feeding pattern groupings. For the swimming behavior decreased species richness and sewage appeared to analysis, fish were classified as demersal (50), bentho- show little change on average. The average change in pelagic (28), and pelagic-neritic/pelagic/bathypelagic species richness at contaminated sites was +56% (fish (16). For the feeding pattern analysis fish were charac- farms), +6% (sewage studies), –29% (industrial efflu- terized as planktivores (20), omnivores (25), and pre- ent), and –8% (run-off). While the variation between dators (57). Abundance responses were not found to be groups was significant (p = 0.009), Tukey’s post hoc statistically different across all swimming behavior analysis indicated that this variation was due to the dif- categories (p = 0.751) and feeding guilds (p = 0.639). ference between fish farms and the other three groups. Sewage, industrial effluent, and run-off were a homogenous subset (Fig. 2b). DISCUSSION Among study systems, the greatest abundance effect sizes were observed for fish farms (over bare sediment) Fish assemblages are responding differently to dis- and coral reef habitats, which displayed positive and tinct sources of marine contamination. Fish farms are a negative responses, respectively (Fig. 2c, Table 1). The substantial source of marine contamination (Wu et al. vegetated habitats, bare sediment, and rocky bottom 1994), but they are also a source of fish food and habi- and rocky reef groupings did not display strong up- tat structure in an otherwise pelagic environment. Fish wards or downwards trends in abundance. The aver- farms were consistently associated with substantial age change in abundance at contaminated sites was increases in fish abundance and moderate increases in +106× (fish farms over bare sediment), +8% (bare species richness. Sewage is another potential source of sediment), +33% (vegetated habitats), +1.3× (rocky food and contamination and was associated with posi- reef and rocky bottom), and –63% (coral reef). While tive and negative responses in equal measure. In con- the variation between groups was significant (p = trast, industrial effluent and run-off were more closely 0.000), Tukey’s post-hoc analysis again indicated that associated with reports of decreased fish abundance this variation was due to the difference between fish and richness, and the average effect sizes for both of farms and the other 4 groups. Bare sediment, coral these contaminant sources was negative. Differences reefs, rocky bottom and rocky reef, and vegetated amongst habitats were again driven by the strong habitats were a homogenous subset (Fig. 2c). For spe- association of fish farms (all located over bare sedi- cies richness, the greatest effect size was observed for ment) with increased fish abundance and species rich- fish farms (Fig. 2d, Table 1). Other categories did not ness. In the absence of fish farms, there were no obvi- show consistent upwards or downwards trends. The ous effects of contaminants in bare sediment habitats. average change in species richness at contaminated Fish abundance in coral reefs suggested a weaker sites was +56% (fish farms over bare sediment), +18% trend towards decreased abundance but nearly no (rocky reef and rocky bottom), –15% (bare sediment), change in species richness. Our results contrast with –13% (vegetated habitats), and –2% (coral reef). While other meta-analyses in marine systems, which have the variation between groups was significant (p = found more consistent negative effects of contaminants 0.029), Tukey’s post-hoc analysis indicated that this in abundance and diversity responses (Micheli 1999, variation was due to the difference between fish farms Breitburg et al. 2009, Johnston & Roberts 2009). over bare sediment and vegetated habitats. All other pairings were non-significant (Fig. 2d). In addition to the data discussed thus far, other qual- Contaminant sources itative information was collected during the review. The 45 studies examined in the meta-analyses re- In our meta-analysis, fish farms displayed the most ported data that summarized the abundance or species consistent and largest responses to contamination richness responses of more than 1300 fish species among study groups, and this is likely due to the spe- (note: exact numbers are not known as not all authors cial nature of the contaminants emitted by these oper- reported the total number of species studied. The ations. Unlike most of the other contaminant sources, degree of overlap between studies is unknown as not fish farms can primarily be viewed as a source of food all authors reported the names of all species investi- subsidies and enriching organic nutrients, while other gated); on average, each study reported data for ~34 chemicals are present in lower concentrations (Wu et species. Of this total, abundance response data was al. 1994). Intuitively it can be expected that these forms 182 Mar Ecol Prog Ser 420: 175–191, 2010

of contamination will have a positive effect on fish of wild fish in just 750 ha of coastal waters. Aggrega- abundance and possibly also diversity. In addition to tion on such a large scale would be likely to have far the papers examined in the meta-analysis, several reaching ecological and fisheries impacts. other studies have documented the aggregation, distri- Run-off and sewage outflows represent more diverse bution, and general increase of fish near aquaculture contaminant sources compared to fish farms. Both run- facilities (Dempster et al. 2005, 2009, 2010). Several off and sewage outflows are highly variable contami- studies have also demonstrated that wild fish are eat- nant sources that may simultaneously provide food ing food or captive fish from these structures, confirm- subsidies, nutrient enrichment, and high concentra- ing that they provide food subsidies to wild populations of potentially toxic chemicals to marine systems. tions (Fernandez-Jover et al. 2007, 2008). For example, It is possible that differential responses to these conta- Fernandez-Jover et al. (2008) conducted a study of the minant sources are a result of different chemical gut contents of 5 dominant taxa around fish farms in mixtures, concentrations of high impact chemicals, or the Mediterranean, and they found that 66 to 89% of effluent treatment practices. In several cases, fish wild fish around these farms were consuming the food abundance was increased by these contamination pellets and that they ate up to 10% of the feed input at sources even where a variety of toxic contaminants these facilities. In the past, it has been argued that the were present at high concentrations, and in some food subsidy and nutrient enrichment effects of fish instances declines in commercially important species farms may be somewhat confounded by the impacts of were observed following the implementation of im- the fish farm structure itself. Tuya et al. (2006) have proved water quality management (Bishop et al. contributed towards clarifying this issue by employing 2006b, Ribeiro et al. 2008). In one such study, Ribeiro et an advanced ‘before–after control impact’ (BACI) al. (2008) investigated the impacts of improvements in research design. In their study the authors investigated sewage treatment on fish assemblages living in coastal the presence of wild fish around farms before and after lagoons in southern Portugal. They found that the the closure of a farm. Throughout the study the fish implementation of improved sewage treatment (and farm structure remained in place, and the authors subsequent declines in organic matter and nutrient demonstrated that, in the absence of feeding, the concentrations) in the lagoons were associated with structure has <2× the fish of control sites. In contrast, significantly decreased fish biomass. In particular, nearby farms that remained active displayed an commercially important Mugilidae species experienced increase in fish abundance of approximately 50× com- a 72% decrease in their abundance. Studies such as pared to control sites. The results of this study suggest this suggest that in some systems the positive effects of that food subsidies and nutrient enrichment play the food subsidies and nutrient enrichment are overpower- dominant role in aggregating wild fish near farms, and ing the potentially negative effects of other contami- that the fish farm structure has a negligible influence nants and that some forms of contamination may be on this phenomenon (Tuya et al. 2006). enhancing wild fish stocks. While most studies of fish farms have examined the Negative responses to sewage outflows were also localized impacts of these operations, there is some reported, chiefly in instances where primary sewage indication that aquaculture is having regional effects treatment was absent and potentially toxic chemical on fish assemblages and fisheries catches. A pioneer- concentrations were comparatively high (Aguilar et al. ing study by Machias et al. (2006) attempted to quan- 2007, Reopanichkul et al. 2009). Several authors also tify the regional effects of aquaculture in a highly reported negative effects of run-off, primarily where developed aquaculture region, the Aegean and Ionian high nutrient concentrations were causing some form seas of the eastern Mediterranean basin. Machias et al. of eutrophication/hypoxia. While eutrophication and (2006) used time series data of fisheries landings, fish associated hypoxia are relatively uncommon in open farm productivity, fishing fleet activity, and environ- marine systems (Breitburg et al. 2009), several studies mental factors over a 17 yr period to analyze the im- in this review reported the development of these pacts of aquaculture development on wild fish abun- negative feedbacks as a result of very high nutrient dance and fisheries landings. The results of their study enrichment in enclosed marine environments (Baden suggest that high concentrations of aquaculture activ- et al. 1990, Nagai 2003). This was demonstrated by ity in an oligotrophic sea could be linked to increased Oczkowski & Nixon (2008), who studied fisheries fisheries landings (Machias et al. 2004, 2006). In an- yields in relation to nutrient enrichment over a forty other large scale study, Dempster et al. (2009) studied year period in coastal lagoons of the Nile delta (Egypt). the aggregation effects of fish farms along the west They found that nutrient enrichment (originating pri- coast of Norway, an area spanning more than 1500 km. marily from agricultural run-off) initially increased the They estimated that the 1200 salmon farms operating abundance of a variety of commercial fish species, over this range are concentrating about 12 000 tonnes resulting in increased fisheries landings. However, a McKinley & Johnston: Impacts of contaminant sources on marine fish 183

threshold enrichment level was eventually reached to be particularly sensitive to contamination sources (~100 μM dissolved inorganic nitrogen), after which (such as sewage outflows), which alter water clarity fish stocks declined exponentially. The authors argued through sedimentation or nutrient induced increases in that this decline was due to nutrient induced eutrophi- algal growth (Pastorok & Bilyard 1985, Fabricius et al. cation and that commercial fishing, pesticide build-up, 2005a, Fabricius 2005b). increased fishing effort, and metals contamination It is interesting to note that the other habitat classifi- played only minor roles in the fishery collapse. It is cations (bare sediment, rocky bottom and rocky reef, interesting to note that even after the development of and vegetated habitats) did not show strong trends for eutrophic conditions, fish stocks remained elevated abundance or species richness response. This could compared to pre-enrichment levels, though they were suggest that contaminant sources or certain types of much reduced compared to their pre-eutrophication chemicals are more important in determining fish ecol- peak (Oczkowski & Nixon 2008). Studies such as this ogy impacts in these ecosystems than the characteris- highlight the importance of quantifying and monitor- tics of the habitat itself. Relatively little research has ing contaminants over long periods, as ecological been undertaken examining how different habitat responses may be non-linear across concentration characteristics could be mediating or affecting contam- gradients. ination impacts in fish assemblages. However, 2 well- While industrial effluents represent another mixed designed studies (Deegan et al. 2002, Reopanichkul et contaminant source, they differed from the other cont- al. 2009) have investigated effects over multiple levels amination sources examined in the present study as of ecological organization and have demonstrated the they contain comparatively few food subsidies and importance of habitat as a mediator of contamination nutrients. The industrial effluents examined consisted impacts. primarily of pulp mill, mine, petrochemical plant, In Reopanichkul et al. (2009), the researchers evalu- power station, and chemical plant effluents, which ated the impacts of a sewage outflow on a coral reef contain a variety of contaminants that have been ecosystem in Thailand. They demonstrated that shown to have toxic effects (e.g. metals and pesticides) sewage outflows were simultaneously associated with (Costello & Read 1994, Austin 1999, Hylland 2006b). increased macro-algal density, decreased hard coral Industrial effluent studies primarily reported a cover, and significant declines in fish abundance decrease in the abundance of fish, which could indi- within the coral reef environment. This study was cate that in the absence of nutrients and food subsidies designed to assess the role of habitat forming organ- other contaminant classes are producing a negative isms in mediating fish declines, and it is one of few to response. Interestingly, where increased abundances demonstrate the cascade of ecological effects over were associated with industrial effluents (Jones et al. multiple levels of biological organization. Deegan et al. 1996, Hoisington & Lowe 2005) water temperature (2002) also investigated the impacts of contamination was also elevated by the effluent. It is possible that in across multiple levels of biological organization, such cases thermal outflows are attracting greater num- though their study was situated within a temperate bers of fish, masking the potentially negative effects of seagrass bed. They demonstrated that contamination contaminants. had a significant impact on the macrophyte community structure, where primary production shifted from eel- grass to macroalgae with increased nutrient loading. Study systems and habitat mediation Changes to the macrophyte community were strongly associated with declines in fish abundance and diver- We examined the responses of fish abundance and sity. Both of these studies demonstrate the importance species richness to anthropogenic contamination in of habitat forming organisms as mediators of contami- relation to broad habitat classifications. In our meta- nation impacts on fish assemblages. analysis, fish farms were almost always located over bare sediment and displayed the most consistent and largest responses to contamination (see section Species richness responses ‘Results of meta-analysis’). Coral reefs were the only habitat associated with a slight trend towards reduced We observed weak species richness responses for all fish abundance. This may be due to a comparatively contaminant sources and sampling habitats. The lack high contaminant sensitivity of the habitat forming of any consistent species richness response associated organisms (corals) and/or the relatively high propor- with contamination could suggest that contaminants tion of reef fish, which are resident specialists in coral are not having a major impact on marine fish diversity. reef environments. Because corals rely on high water However, further research is needed before such con- clarity to conduct photosynthesis, coral reefs are known clusions can be drawn, especially given that some 184 Mar Ecol Prog Ser 420: 175–191, 2010

studies reported decreases in species richness of up to 2009). While the average species richness response was 60% (Gray et al. 1992, Smith et al. 1999a). While incon- weak, many studies reported substantial changes to the sistent responses were obtained for species richness, it composition of fish assemblages that are not well cap- is possible that contamination may act to alter fish tured or described by species richness measures. Some assemblages in ways that are not detected by simple of these observations are discussed in more detail in diversity measurements. Several studies have shown ‘Linking contaminant effects to fish ecology’ below. significant changes in the abundance and trophic structure of fish assemblages, while little impact was detected by conventional diversity measures (Khalaf & Contrasting invertebrate responses Kochzius 2002, Guidetti et al. 2003, Ribeiro et al. 2008). More complex measures of biodiversity and/or com- Johnston & Roberts (2009) conducted a recent munity structure, such as the Shannon-Wiener diver- review and meta-analysis that examined the impacts of sity index, Pielou’s evenness, or taxonomic relatedness contaminants on the diversity of marine systems. In (Costello et al. 2001) may better characterize commu- that study it was found that many different kinds of nity responses to contamination than simple richness contaminants reduced the diversity of marine assem- or abundance indexes (Washington 1984). However, blages in a variety of habitats. Regardless of the diver- these indices were insufficiently reported in the data sity measure employed, a reduction in diversity of 30 to set for a meta-analysis. 50% was observed in all study systems and contami- Better measurement methods could improve our nant classes. The Johnston & Roberts (2009) results ability to detect and understand changes to fish diver- contrasts strongly with the overall weak species rich- sity. In an excellent study, Khalaf & Kochzius (2002) ness results observed in our study. This is likely due to investigated the impacts of urban and industrial conta- the fact that the vast majority of studies (>90%) consid- mination in a heavily developed port in the Gulf of ered in the Johnston & Roberts (2009) review observed Aqaba, Jordan. They assessed the trophic community diversity changes in sessile invertebrate communities, structure of coral reef fishes near disturbed and undis- while only 1 study in that review observed fish re- turbed sites and analyzed this data using a variety of sponses. The difference in observed species richness univariate and multivariate measures. Univariate mea- response between these 2 studies suggests that sessile sures such as species richness, diversity, and evenness invertebrate assemblages are far more responsive to detected no negative impacts of contamination. How- contamination than fish assemblages. While few stud- ever, a multivariate analysis of species abundance ies have contrasted diversity responses of inverte- (characterized according to trophic categories) clearly brates and fishes within the same research design, it separated disturbed from undisturbed sites. These has been demonstrated in coral reef ecosystems that multivariate measures demonstrated that a major shift the diversity response of sessile coral species is of a in trophic balance had occurred, with significant greater magnitude than associated fish assemblages changes in the relative and absolute abundance of dif- (Fabricius et al. 2005a, Reopanichkul et al. 2009). ferent feeding guilds. Advanced multivariate methods There are a variety of reasons why fish assemblages such as this could be employed to better understand may be less responsive to contaminants than sessile and quantify changes to fish assemblages. invertebrates. Sessile invertebrates are generally con- For fish farms, several studies reported increases in sidered good indicators of contamination as they are abundance and species richness primarily for pelagic immobile and so their contaminant exposure times are fishes with predatory feeding habits, planktivores, and predictable, many are filter feeders or live in the sedi- demersal species that consumed food pellets deposited ment and are thus intimately associated with contami- under cages (Tuya et al. 2005, Valle et al. 2007, nants in the environment, many species readily accu- Fernandez-Jover et al. 2008). Research suggests that mulate contaminants in their tissues and shells, and nutrient enrichment disproportionately increases the their diets are often relatively simple (Linton & Warner dominance of competitively superior species and may 2003). Fish differ in some of these characteristics. hence reduce diversity (Hillebrand et al. 2007). How- While they may accumulate contaminants to a greater ever, where nutrients are a limiting factor an increase degree, due to their high trophic position, they can be in their availability can raise productivity, which results highly mobile so direct exposure times are not certain, in increased resource heterogeneity and hence diver- their diets are comparatively diverse, and they may sity (Hall et al. 2000, Arai 2001). Contaminants may have a higher capacity for physiological resistance and even enhance diversity in some cases by reducing the tolerance (van der Oost et al. 2003, Wirgin & Waldman abundance of competitive dominants, though this has 2004). Several of these characteristics may explain why not yet been demonstrated in fish assemblages (Rohr fish assemblages are less responsive to contaminants & Crumrine 2005, Rohr et al. 2006, Clements & Rohr than sessile invertebrates. McKinley & Johnston: Impacts of contaminant sources on marine fish 185

Linking contaminant effects to fish ecology Mechanisms of contaminant effects

While our analysis of abundance response by broad The majority of research addressing the effect of swimming behavior and feeding guild categories did contamination on marine fish has focused on the not yield significant results, several studies have sug- chemical aspects of contamination, and a great deal gested that fish ecological characteristics are a sig- of literature exists examining the presence, bio- nificant determinant of contamination response. For magnification, toxicology, and biomarker response of example, there is some evidence to suggest that fish contaminants in marine fish populations (Costello & farms disproportionately increase the abundance of Read 1994, Wirgin & Waldman 1998, Austin 1999, pelagic fishes with predatory feeding habits and van der Oost et al. 2003, Hylland 2006b). In this planktivores (Tuya et al. 2005, Valle et al. 2007, Fer- regard, the mechanisms by which contaminants nandez-Jover et al. 2008). In other habitats and conta- affect fish populations have been fairly well investi- minant source groupings, the response of different gated. Fish primarily uptake contaminants through ecological groups appears to be highly variable and ingestion of contaminated food particles and to a at times contradictory For example, in the Khalaf lesser extent from water that passes over the gill & Kochzius (2002) study, herbivores, detrivores, and membranes (Dallinger et al. 1987, Hall et al. 1997). planktivores experienced increased abundance and Some contaminants have also been shown to be diversity in association with industrial contamination, maternally transferred to eggs and larvae (Collier et while fish that feed on invertebrates or other fish al. 1992, Hu et al. 2009). Once ingested, contami- decreased. In contrast, Otway et al. (1996b) found nants move through a wide variety of physiological that sewage contamination disproportionately favored and chemical pathways, many of which have detri- fishes feeding on invertebrates and fish. While it has mental effects for the individual. Some contaminants been demonstrated that many contaminants bioaccu- are readily excreted or breakdown while others are mulate and biomagnify and hence disproportionately considered ‘persistent’ and resist decomposition in accumulate in fishes occupying higher trophic posi- natural systems. Contaminants of this nature have tions, this did not consistently translate into reduced the tendency to accumulate in tissues and may abundance among the higher trophic level species ex- bioaccumulate up the food chain, increasing in con- amined in our review (Burger et al. 2001, van der Oost centration at higher trophic levels (Burger et al. 2001, et al. 2003). Similarly, fish exhibiting different swim- van der Oost et al. 2003). Contaminants may affect ming habits did not appear to be strongly differenti- fish populations and diversity by reducing fish health ated, despite the expectation that demersal species and survivorship (Robinet & Feunteun 2002, Clair- would be more affected than pelagic species, as conta- eaux et al. 2004), by increasing susceptibility to dis- minant concentrations in sediments are normally ease (Arkoosh et al. 1998b), by reducing growth and higher than those in the surrounding water column reproductive success (Waring et al. 1996, Vetemaa et (Daskalakis & O’Connor 1995, Knott et al. 2009). These al. 1997), by reducing the abundance of prey species, findings suggest that insufficient information is avail- and by increasing instances of deformity (Kingsford able to predict the impacts of contamination based on et al. 1996b). Ultimately, any of these mechanisms fish ecological characteristics. It may also be the case could link contaminant exposure to organismal effects that other ecological features, or the type of contami- and population level impacts. nant/habitat, are of greater importance when predict- In many cases, researchers have attempted to use ing the impact of contaminants on fish assemblages. population modeling to extrapolate observed sub- A potentially important ecological variable that was cellular or organismal effects to community level overlooked by the majority of studies in this review is impacts. Numerous studies, and even a book, have how contamination impacts change at different stages been written on this subject (Barnthouse et al. 1987, of a fish’s life cycle. Numerous ecotoxicological studies 1990, Lawrence & Hemingway 2003). While studies of suggest that there is a significant development suscep- this nature are a valuable addition to the literature, tibility in fish species and that eggs and larva are much there is still a need for research that directly investi- more sensitive to contamination than adults (Collier gates and verifies the relationship between organismal et al. 1992, Waring et al. 1996, Kingsford et al. 1996b, effects and population level impacts. Although mecha- Vetemaa et al. 1997, Ganassin et al. 1999, Hu et al. nisms of uptake, toxin persistence, and biochemical 2009). Despite the importance of contamination im- effect are extensively studied within the ecotoxicology pacts at this stage of the life cycle, in our review only literature, we encountered few studies (Roy et al. 2003, three larval fish studies were encountered and all were Claireaux et al. 2004) that linked these measures from a single geographic region (Gray et al. 1992, Gray directly to ecological effects at the population or com- 1996, 1997). munity level. 186 Mar Ecol Prog Ser 420: 175–191, 2010

Claireaux et al. (2004) provide a useful study in this studies have been produced that show the distribution regard. The authors utilized a multi-disciplinary ap- of impacts over both small and large spatial scales proach to evaluate the effects of PAH exposure on the (Dempster et al. 2009) and in terms of depth (Dempster ecology of a common flatfish species Solea solea. They et al. 2005). Describing the scale and distribution of used 3 methods to evaluate the impacts at multiple lev- impacts in this way for other contamination sources els of biological effect. This included several laboratory would allow us to gain a better understanding of the based toxicological experiments (assessing cellular importance of spatial replication in investigations of and organ level impacts), followed by mesocosm this type. experiments investigating fecundity and growth rates In our review relatively few studies were encoun- (organism levels impacts), and finally field surveys to tered that provided both detailed environmental/water provide field validation of toxicological experiments quality data and ecological monitoring data within the and to investigate population structure and abundance same research framework. Coupling fish ecology data (population level impacts). The coupling of field sur- to detailed contaminant monitoring would greatly im- veys with experimental and mesocosm work led the prove our understanding of contaminant impacts and authors to conclude that effects that were readily the role of specific chemicals. Studies of this nature detected within individuals were progressively dimin- would also allow researchers to gain a better under- ished as their research progressed towards higher standing of how contaminants move through the envi- organizational levels. Studies such as this could be ronment and fish assemblages. very useful in linking the ecotoxicology/biomarker and In some instances, aggregation effects (e.g. around fish ecology disciplines and would help researchers warm water outflows from industrial facilities) may better understand the mechanisms behind ecological be attracting and concentrating fish without having impacts. any negative impact on their populations. Carefully Several studies have examined the development of designed experimental studies may be useful in untan- contaminant resistance in marine and freshwater fish gling the relative magnitude of aggregation and non- and some cases of resistance have been demonstrated aggregation population effects (see the discussion of in both laboratory and field conditions (Wirgin & Wald- the role of aggregation effects from fish farm structures man 2004, Xie & Klerks 2004). Resistance may be due and the Tuya et al. 2006 study in ‘Contaminant to either genetic adaptation or physiological acclima- sources’ above). As discussed in ‘Linking contaminant tion in wild fish populations, and current research has effects to fish ecology’ above, future studies could investigated the mechanisms, costs, and persistence of examine the role of fish ecological characteristics such toxicity resistance (Wirgin & Waldman 2004, Xie & as life cycle stage, feeding habitats, and swimming Klerks 2004, Burnett et al. 2007). If the development of characteristics in determining the impacts of contami- contaminant resistance is a regular occurrence in fish nation on fish assemblages. A relatively small number populations chronically exposed to contaminants, then of studies (45) were found for the meta-analysis. This is impacts on abundance or diversity may be difficult to a small sample size compared to other recent meta- observe. However, our current understanding of the analyses addressing similar topics (Micheli 1999, John- mechanisms suggests that the evolution of resistance ston & Roberts 2009). This suggests that studies inves- in fish will be rare (Klerks & Weis 1987, Klerks et al. tigating the effects of contamination on fish abundance 1997, Xie & Klerks 2004). and species richness are under-represented within the scientific literature. Lastly, the implementation of multivariate measurement and analysis methods (as Knowledge gaps opposed to univariate measures) would improve our understanding of changes in fish communities in the One limitation of the meta-analysis is that the scale face of disturbance. of the studies examined varies widely. This is in part due to the nature of the meta-analysis, in that it incor- porates a diverse array of experimental designs, but it CONCLUSIONS is also due to the lack of knowledge of the spatial and temporal scale of contamination impacts in fish popu- Anthropogenic disturbances that provide a source of lations. In truth, the duration and distribution of contamination to marine systems appear to be affect- impacts is not well known for a variety of contaminants ing fish abundances and species richness. Fish farms and so it is difficult to assess the degree to which an are associated with large effects on fish abundance analysis of these impacts should be weighted. The and diversity, and more work needs to be done to scale and distribution of impacts has been best studied understand the ecological ramifications of these alter- in the fish farm literature, where several large scale ations. Clearly, these farms are having a major effect McKinley & Johnston: Impacts of contaminant sources on marine fish 187

on the distribution of wild fish assemblages through Austin B (1999) The effects of pollution on fish health. J Appl the addition of large volumes of biological material into Microbiol 85:234S–242S the sea. Slight negative effects on fish abundance are Baden SP, Loo LO, Pihl L, Rosenberg R (1990) Effects of eutrophication on benthic communities including fish: associated with more toxic sources of contamination Swedish west coast. Ambio 19:113–122 such as run-off and industrial effluent, particularly in ➤ Barnthouse LW, Suter GW II, Rosen AE, Beauchamp JJ (1987) coral reefs. A greater number of more targeted field Estimating responses of fish populations to toxic contami- studies are needed before we can estimate the ecolog- nants. Environ Toxicol Chem 6:811–824 ical impacts of toxic contaminants in these systems; ➤ Barnthouse LW, Suter GW II, Rosen AE (1990) Risks of toxic contaminants to exploited fish populations: influence of however, it is likely that marine invertebrates are con- life history, data uncertainty and exploitation intensity. sistently more sensitive to contamination than fish pop- Environ Toxicol Chem 9:297–311 ulations. Strong trends in species richness were not ➤ Barry KL, Grout JA, Levings CD, Nidle BH, Piercey GE (2000) observed in any categories other than fish farms; this Impacts of acid mine drainage on juvenile salmonids in an could suggest that contaminants are not having a estuary near Britannia Beach in Howe Sound, British Columbia. Can J Fish Aquat Sci 57:2032–2043 major impact on fish diversity. There is also significant ➤ Bishop MJ, Kelaher BP, Smith MP, York PH, Booth DJ (2006a) evidence supporting the idea that features of fish Ratio-dependent response of a temperate Australian estu- assemblages other than species richness and total arine system to sustained nitrogen loading. Oecologia 149: abundance (such as trophic balance and evenness) are 701–708 being altered by contaminants. Our results suggest ➤ Bishop MJ, Powers SP, Porter HJ, Peterson CH (2006b) Ben- thic biological effects of seasonal hypoxia in a eutrophic that some sources of contamination are having size- estuary predate rapid coastal development. Estuar Coast able effects on fish assemblages and may be of com- Shelf Sci 70:415–422 mercial and conservation significance. Further re- ➤ Boyra A, Sanchez-Jerez P, Tuya F, Espino F, Haroun R (2004) search is needed to quantify the impacts of contamina- Attraction of wild coastal fishes to an Atlantic subtropical cage fish farms, Gran Canaria, Canary Islands. Environ tion on marine fish assemblages, to differentiate the Biol Fishes 70:393–401 impacts of various contaminants, and to identify taxa ➤ Breitburg DL, Craig JK, Fulford RS, Rose KA and others that are most sensitive. A better understanding of these (2009) Nutrient enrichment and fisheries exploitation: issues will improve the monitoring and management of interactive effects on estuarine living resources and their marine contamination, fish stocks, and biodiversity. management. Hydrobiologia 629:31–47 ➤ Brewer DT, Milton DA, Fry GC, Dennis DM, Heale SDS, Venables WN (2007) Impacts of gold mine waste disposal Acknowledgements. The authors were supported by the Aus- on deepwater fish in a pristine tropical marine system. tralian Research Council while preparing this review through Mar Pollut Bull 54:309–321 an Australian Research Fellowship and a Linkage Grant ➤ Browman HI (1999) Negative results: the uncertain position, awarded to E.L.J. We thank Dr. Iain Suthers and Dr. Matt status and impact of negative results in marine ecology: Taylor for comments which improved earlier drafts of this philosophical and practical considerations. Mar Ecol Prog manuscript. Ser 191:301–302 ➤ Bundy MH, Breitburg DL, Sellner KG (2003) The responses of Patuxent River upper trophic levels to nutrient and trace LITERATURE CITED element induced changes in the lower food web. Estuaries 26:365–384 ➤ Aguilar C, González-Sansón G, Munkittrick KR, MacLatchy ➤ Burger J, Gaines KF, Boring CS, Stephens WL, Snodgrass J, DL (2004) Fish assemblages on fringe coral reefs of the Gochfeld M (2001) Mercury and selenium in fish from the northern coast of Cuba near Havana Harbor. Ecotoxicol Savannah River: species, trophic level, and locational Environ Saf 58:126–138 differences. Environ Res 87:108–118 ➤ Aguilar C, Gonzalez-Sanson G, Hernandez I, MacLatchy DL, ➤ Burnett KG, Bain LJ, Baldwin WS, Callard GV and others Munkittrick KR (2007) Effects-based assessment in a trop- (2007) Fundulus as the premier teleost model in environ- ical coastal system: status of bicolor damselfish (Stegastes mental biology: opportunities for new insights using partitus) on the north shore of Cuba. Ecotoxicol Environ genomics. Comp Biochem Physiol Part D Genomics Saf 67:459–471 Proteomics 2:257–286 ➤ Arai MN (2001) Pelagic coelenterates and eutrophication: a ➤ Carss DN (1990) Concentrations of wild and escaped fishes review. Hydrobiologia 451:69–87 immediately adjacent to fish farm cages. Aquaculture 90: ➤ Araújo F, Williams W, Bailey R (2000) Fish assemblages as 29–40 indicators of water quality in the middle Thames estuary, Claireaux G, Desaunay Y, Akcha F, Auperin B and others England (1980–1989). Estuaries Coasts 23:305–317 (2004) Influence of oil exposure on the physiology and ➤ Arkoosh MR, Casillas E, Clemons E, Kagley AN, Olson R, ecology of the common sole Solea solea: experimental and Reno P, Stein JE (1998a) Effect of pollution on fish dis- field approaches. Aquat Living Resour 17:335–351 eases: potential impacts on salmonid populations. J Aquat ➤ Clements WH, Rohr JR (2009) Community responses to cont- Anim Health 10:182–190 aminants: using basic ecological principles to predict eco- ➤ Arkoosh MR, Casillas E, Huffman P, Clemons E, Evered J, toxicological effects. Environ Toxicol Chem 28:1789–1800 Stein JE, Varanasi U (1998b) Increased susceptibility of ➤ Collier TK, Stein JE, Sanborn HR, Hom T, Myers MS, juvenile chinook salmon from a contaminated estuary to Varanasi U (1992) Field studies of reproductive success Vibrio anguillarum. Trans Am Fish Soc 127:360–374 and bioindicators of maternal contaminant exposure in 188 Mar Ecol Prog Ser 420: 175–191, 2010

English sole (Parophrys vetulus). Sci Total Environ 116: 308:1621–1623 169–185 ➤ Ganassin RC, Sanders SM, Kennedy CJ, Joyce EM, Bols NC Connolly RM, Jones GK (1996) Determining effects of an oil (1999) Development and characterization of a cell line spill on fish communities in a mangrove-seagrass ecosys- from Pacific herring, Clupea harengus pallasi, sensitive tem in southern Australia. Aust J Ecotoxicol 2:3–15 to both naphthalene cytotoxicity and infection by viral ➤ Costello MJ, Read P (1994) Toxicity of sewage sludge to hemorrhagic septicemia virus. Cell Biol Toxicol 15: marine organisms: a review. Mar Environ Res 37:23–46 299–309 Costello MJ, Pohle G, Martin A (2001) Evaluating biodiversity ➤ Gray CA (1996) Intrusions of surface sewage plumes into con- in marine environmental assessments. Research and tinental shelf waters: interactions with larval and preset- Development Monograph Series. Canadian Environmen- tlement juvenile fishes. Mar Ecol Prog Ser 139:31–45 tal Assessment Agency, Ottawa ➤ Gray CA (1997) Field assessment of numerical impacts of ➤ Dallinger R, Prosi F, Segner H, Back H (1987) Contaminated coastal sewage disposal on fish larvae relative to natural food and uptake of heavy metals by fish: a review and a variability. Environ Biol Fishes 50:415–434 proposal for further research. Oecologia 73:91–98 ➤ Gray CA, Otway NM, Laurenson FA, Miskiewicz AG, Pethe- ➤ Daskalakis KD, O’Connor TP (1995) Normalization and ele- bridge RL (1992) Distribution and abundance of marine mental sediment contamination in the coastal United fish larvae in relation to effluent plumes from sewage out- States. Environ Sci Technol 29:470–477 falls and depth of water. Mar Biol 113:549–559 ➤ Deegan LA, Wright A, Ayvazian SG, Finn JT, Golden H, Mer- ➤ Grigg RW (1994) Effects of sewage discharge, fishing pres- son RR, Harrison J (2002) Nitrogen loading alters seagrass sure and habitat complexity on coral ecosystems and reef ecosystem structure and support of higher trophic levels. fishes in Hawaii. Mar Ecol Prog Ser 103:25–34 Aquat Cons Mar Freshw Ecosyst 12:193–212 Guidetti P, Terlizzi A, Fraschetti S, Boero F (2003) Changes in ➤ Deegan LA, Bowen JL, Drake D, Fleeger JW and others Mediterranean rocky-reef fish assemblages exposed to (2007) Susceptibility of salt marshes to nutrient enrich- sewage pollution. Mar Ecol Prog Ser 253:269–278 ment and predator removal. Ecol Appl 17:S42–S63 Hall BD, Bodaly RA, Fudge RJP, Rudd JWM, Rosenberg DM Dempster T, Sanchez-Jerez P, Bayle-Sempere JT, Giménez- (1997) Food as the dominant pathway of methylmercury Casalduero F, Valle C (2002) Attraction of wild fish to sea- uptake by fish. Water Air Soil Pollut 100:13–24 cage fish farms in the south-western Mediterranean Sea: ➤ Hall SJ, Gray SA, Hammett ZL (2000) Biodiversity-productiv- spatial and short-term temporal variability. Mar Ecol Prog ity relations: an experimental evaluation of mechanisms. Ser 242:237–252 Oecologia 122:545–555 ➤ Dempster T, Sanchez-Jerez P, Sempere JB, Kingsford M ➤ Hedges LV, Gurevitch J, Curtis PS (1999) The meta-analysis (2004) Extensive aggregations of wild fish at coastal sea- of response ratios in experimental ecology. Ecology 80: cage fish farms. Hydrobiologia 525:245–248 1150–1156 ➤ Dempster T, Fernandez-Jover D, Sanchez-Jerez P, Tuya F, ➤ Hillebrand H, Gruner DS, Borer ET, Bracken MES and others Bayle-Sempere J, Boyra A, Haroun RJ (2005) Vertical vari- (2007) Consumer versus resource control of producer ability of wild fish assemblages around sea-cage fish diversity depends on ecosystem type and producer com- farms: implications for management. Mar Ecol Prog Ser munity structure. Proc Natl Acad Sci USA 104:10904–10909 304:15–29 ➤ Hoisington G, Lowe CG (2005) Abundance and distribution of ➤ Dempster T, Uglem I, Sanchez-Jerez P, Fernandez-Jover D, the round stingray, Urobatis halleri, near a heated effluent Bayle-Sempere J, Nilsen R, Bjørn PA (2009) Coastal outfall. Mar Environ Res 60:437–453 salmon farms attract large and persistent aggregations of ➤ Hu JY, Zhang ZB, Wei QW, Zhen HJ and others (2009) Mal- wild fish: an ecosystem effect. Mar Ecol Prog Ser 385:1–14 formations of the endangered Chinese sturgeon, Acipen- Dempster T, Sanchez-Jerez P, Uglem I, Bjorn P (2010) Spe- ser sinensis, and its causal agent. Proc Natl Acad Sci USA cies-specific patterns of aggregation of wild fish around 106:9339–9344 fish farms. Estuar Coast Shelf Sci 86(2):271–275 ➤ Hylland K, Beyer J, Berntssen M, Klungsoyr J, Lang T, Balk L Dethlefsen V, Tiews K (1985) Review on the effects of pollu- (2006a) May organic pollutants affect fish populations in tion on marine fish life and fisheries in the North Sea. the North Sea? J Toxicol Environ Health A 69:125–138 J Appl Ichthyol 1:97–118 ➤ Hylland K (2006b) Biological effects in the management of ➤ Fabricius KE (2005b) Effects of terrestrial runoff on the chemicals in the marine environment. Mar Pollut Bull 53: ecology of corals and coral reefs: review and synthesis. 614–619 Mar Pollut Bull 50:125–146 ➤ Islam MS, Tanaka M (2004) Impacts of pollution on coastal ➤ Fabricius K, De’ath G, McCook L, Turak E, Williams DM and marine ecosystems including coastal and marine fish- (2005a) Changes in algal, coral and fish assemblages eries and approach for management: a review and synthe- along water quality gradients on the inshore Great Barrier sis. Mar Pollut Bull 48:624–649 Reef. Mar Pollut Bull 51:384–398 ➤ Jacobsson A, Neuman E (1991) Fish recruitment around a ➤ Fernandez-Jover D, Jimenez JAL, Sanchez-Jerez P, Bayle- petrochemical centre in the North Sea. Mar Pollut Bull 22: Sempere J, Casalduero FG, Lopez FJM, Dempster T 269–272 (2007) Changes in body condition and fatty acid composi- ➤ Jiao Y (2009) Regime shift in marine ecosystems and implica- tion of wild Mediterranean horse mackerel (Trachurus tions for fisheries management, a review. Rev Fish Biol mediterraneus, Steindachner, 1868) associated to sea cage Fish 19:177–191 fish farms. Mar Environ Res 63:1–18 ➤ Johnston EL, Roberts DA (2009) Contaminants reduce the ➤ Fernandez-Jover D, Sanchez-Jerez P, Bayle-Sempere JT, richness and evenness of marine communities: a review Valle C, Dempster T (2008) Seasonal patterns and diets of and meta-analysis. Environ Pollut 157:1745–1752 wild fish assemblages associated with Mediterranean ➤ Jones JC, Reynolds JD (1997) Effects of pollution on repro- coastal fish farms. ICES J Mar Sci 65:1153–1160 ductive behaviour of fishes. Rev Fish Biol Fish 7:463–491 ➤ Frank KT, Petrie B, Choi JS, Leggett WC (2005) Trophic cas- ➤ Jones GK, Baker JL, Edyvane K, Wright GJ (1996) Nearshore cades in a formerly cod-dominated ecosystem. Science fish community of the Port River-Barker Inlet Estuary, McKinley & Johnston: Impacts of contaminant sources on marine fish 189

South Australia. I. Effect of thermal effluent on the fish tions and the rise and fall of a coastal fishery; a review of community structure, and distribution and growth of data from the Nile Delta, Egypt. Estuar Coast Shelf Sci 77: economically important fish species. Mar Freshw Res 47: 309–319 785–799 ➤ Otway NM, Gray CA, Craig JR, McVea TA, Ling JE (1996a) ➤ Karas P, Neuman E, Sandstrom O (1991) Effects of a pulp mill Assessing the impacts of deepwater sewage outfalls on effluent on the population dynamics of perch, Perca-fluvi- spatially- and temporally-variable marine communities. atilis. Can J Fish Aquat Sci 48:28–34 Mar Environ Res 41:45–71 ➤ Khalaf MA, Kochzius M (2002) Changes in trophic community ➤ Otway NM, Sullings DJ, Lenehan NW (1996b) Trophically- structure of shore fishes at an industrial site in the Gulf of based assessment of the impacts of deepwater sewage dis- Aqaba, Red Sea. Mar Ecol Prog Ser 239:287–299 posal on a demersal fish community. Environ Biol Fishes ➤ Kingsford MJ, Suthers IM, Gray CA (1996a) Exposure to 46:167–183 sewage plumes and the incidence of deformities in larval ➤ Pastorok RA, Bilyard GR (1985) Effects of sewage pollution on fishes. Mar Pollut Bull 33:201–212 coral reef communities. Mar Ecol Prog Ser 21:175–189 ➤ Kingsford MJ, Suthers IM, Gray CA (1996b) Exposure to Pauly D, Christensen V, Guenette S, Pitcher TJ and others sewage plumes and the incidence of deformities in larval (2002) Towards sustainability in world fisheries. Nature fishes. Mar Pollut Bull 33:201–212 418:689–695 ➤ Klerks PL, Weis JS (1987) Genetic adaptation to heavy metals ➤ Perry AL, Low PJ, Ellis JR, Reynolds JD (2005) Climate in aquatic organisms: a review. Environ Pollut 45:173–205 change and distribution shifts in marine fishes. Science ➤ Klerks PL, Leberg PL, Lance RF, McMillin DJ, Means JC 308:1912–1915 (1997) Lack of development of pollutant-resistance or ➤ Reopanichkul P, Schlacher TA, Carter RW, Worachananant S genetic differentiation in darter gobies (Gobionellus boleo- (2009) Sewage impacts coral reefs at multiple levels of soma) inhabiting a produced-water discharge site. Mar ecological organization. Mar Pollut Bull 58:1356–1362 Environ Res 44:377–395 ➤ Ribeiro J, Monteiro CC, Monteiro P, Bentes L and others ➤ Knott NA, Aulbury J, Brown T, Johnston EL (2009) Contem- (2008) Long-term changes in fish communities of the Ria porary ecological threats from historical pollution sources: Formosa coastal lagoon (southern Portugal) based on two impacts of large-scale resuspension of contaminated sedi- studies made 20 years apart. Estuar Coast Shelf Sci 76: ments on sessile invertebrate recruitment. J Appl Ecol 46: 57–68 770–781 Roberts DA, Johnston EL, Knott NA (2010) Impacts of de- ➤ Landner L, Grahn O, Hardig J, Lehtinen KJ, Monfelt C, Tana salination plant discharges on the marine environment: J (1994) A field study of environmental impacts at a a critical review of published studies. Water Res 44: bleached kraft pulp mill site on the Baltic Sea coast. 5117–5128 Ecotoxicol Environ Saf 27:128–157 ➤ Robinet TT, Feunteun EE (2002) Sublethal effects of exposure Lawrence AJ, Hemingway KL (eds) (2003) Effects of pollution to chemical compounds: a cause for the decline in Atlantic on fish: molecular effects and population responses. eels? Ecotoxicology 11:265–277 Blackwell Science, Oxford ➤ Rohr JR, Crumrine PW (2005) Effects of an herbicide and an ➤ Linton DM, Warner GF (2003) Biological indicators in the insecticide on pond community structure and processes. Caribbean coastal zone and their role in integrated coastal Ecol Appl 15:1135–1147 management. Ocean Coast Manag 46:261–276 ➤ Rohr JR, Kerby JL, Sih A (2006) Community ecology as a ➤ Machias A, Karakassis I, Labropoulou M, Somarakis S, framework for predicting contaminant effects. Trends Ecol Papadopoulou KN, Papaconstantinou C (2004) Changes in Evol 21:606–613 wild fish assemblages after the establishment of a fish ➤ Rose KA (2000) Why are quantitative relationships between farming zone in an oligotrophic marine ecosystem. Estuar environmental quality and fish populations so elusive? Coast Shelf Sci 60:771–779 Ecol Appl 10:367–385 ➤ Machias A, Karakassis I, Giannoulaki M, Papadopoulou KN, ➤ Roy LA, Armstrong JL, Sakamoto K, Steinert S and others Smith CJ, Somarakis S (2005) Response of demersal fish (2003) The relationships of biochemical endpoints to communities to the presence of fish farms. Mar Ecol Prog histopathology and population metrics in feral flatfish Ser 288:241–250 species collected near the municipal wastewater outfall of ➤ Machias A, Giannoulaki M, Somarakis S, Maravelias CD and Orange County, California, USA. Environ Toxicol Chem others (2006) Fish farming effects on local fisheries land- 22:1309–1317 ings in oligotrophic seas. Aquaculture 261:809–816 ➤ Russo AR (1982) Temporal changes in fish community struc- ➤ Micheli F (1999) Eutrophication, fisheries, and consumer- ture near a sewage ocean outfall, Mokapu, Oahu, Hawaii. resource dynamics in marine pelagic ecosystems. Science Mar Environ Res 6:83–98 285:1396–1398 Sandström O (1994) Incomplete recovery in a coastal fish ➤ Murphy CA, Rose KA, Alvarez M, Fuiman LA (2008) Model- community exposed to effluent from a modernized ing larval fish behavior: scaling the sublethal effects of Swedish bleached kraft mill. Can J Fish Aquat Sci 51: methylmercury to population-relevant endpoints. Aquat 2195–2202 Toxicol 86:470–484 ➤ Sandström O, Neuman E (2003) Long-term development in a ➤ Nagai T (2003) Recovery of fish stocks in the Seto Inland Sea. Baltic fish community exposed to bleached pulp mill efflu- Mar Pollut Bull 47:126–131 ent. Aquat Ecol 37:267–276 ➤ Nixon S, Buckley B (2002) ‘A strikingly rich zone’—nutrient ➤ Smith AK, Suthers IM (1999b) Effects of sewage effluent dis- enrichment and secondary production in coastal marine charge on the abundance, condition and mortality of hula- ecosystems. Estuaries Coasts 25:782–796 fish, Trachinops taeniatus (Plesiopidae). Environ Pollut ➤ Oakes CT, Pondella DJ (2009) The value of a net-cage as a 106:97–106 fish aggregating device in southern California. J World ➤ Smith AK, Ajani PA, Roberts DE (1999a) Spatial and temporal Aquacult Soc 40:1–21 variation in fish assemblages exposed to sewage and im- ➤ Oczkowski A, Nixon S (2008) Increasing nutrient concentra- plications for management. Mar Environ Res 47:241–260 190 Mar Ecol Prog Ser 420: 175–191, 2010

➤ Tober JD, Griffin MPA, Valiela I (2000) Growth and abun- effluent impacts on reproduction and biochemistry in a dance of Fundulus heteroclitus and Menidia menidia North Sea population of viviparous blenny (Zoarces vivip- in estuaries of Waquoit Bay, Massachusetts exposed to arus). J Aquat Ecosyst Stress Recovery 6:33–41 different rates of nitrogen loading. Aquat Ecol 34: ➤ Waring CP, Stagg RM, Fretwell K, McLay HA, Costello MJ 299–306 (1996) The impact of sewage sludge exposure on the ➤ Tuya F, Boyra A, Sanchez-Jerez P, Haroun RJ (2005) Multi- reproduction of the sand goby, Pomatoschistus minutus. variate analysis of the bentho-demersal ichthyofauna Environ Pollut 93:17–25 along soft bottoms of the eastern Atlantic: comparison ➤ Washington HG (1984) Diversity, biotic and similarity indices: between unvegetated substrates, seagrass meadows and a review with special relevance to aquatic ecosystems. sandy bottoms beneath sea-cage fish farms. Mar Biol 147: Water Res 18:653–694 1229–1237 ➤ Wirgin I, Waldman JR (1998) Altered gene expression and ➤ Tuya F, Sanchez-Jerez P, Dempster T, Boyra A, Haroun RJ genetic damage in North American fish populations. (2006) Changes in demersal wild fish aggregations Mutat Res 399:193–219 beneath a sea-cage fish farm after the cessation of farm- Wirgin I, Waldman JR (2004) Resistance to contaminants in ing. J Fish Biol 69:682–697 North American fish populations. Mutation Research Fun- ➤ Valle C, Bayle-Sempere JT, Dempster T, Sanchez-Jerez P, damental and Molecular Mechanisms of Mutagenesis 552: Giménez-Casalduero F (2007) Temporal variability of wild 73–100 fish assemblages associated with a sea-cage fish farm in ➤ Wu RSS, Lam KS, MacKay DW, Lau TC, Yam V (1994) Impact the south-western Mediterranean Sea. Estuar Coast Shelf of marine fish farming on water quality and bottom sedi- Sci 72:299–307 ment: a case study in the sub-tropical environment. Mar ➤ van der Oost R, Beyer J, Vermeulen NPE (2003) Fish bioaccu- Environ Res 38:115–145 mulation and biomarkers in environmental risk assess- ➤ Xie L, Klerks PL (2004) Changes in cadmium accumulation as ment: a review. Environ Toxicol Pharmacol 13:57–149 a mechanism for cadmium resistance in the least killifish ➤ Vetemaa M, Sandström O, Förlin L (1997) Chemical industry Heterandria formosa. Aquat Toxicol 66:73–81

Appendix 1. Meta-analysis studies

Abundance: Richness: Contaminant Study Research Source direction direction source system approach of impact of impact

No data Increase Sewage Coral reef Survey Aguilar et al. (2004) Decrease No data Sewage Coral reef Survey Aguilar et al. (2007) Decrease Decrease Sewage Vegetated habitats Survey Araújo et al. (2000) Decrease No data Run-off Bare sediment Survey Baden et al. (1990) Decrease No data Industrial effluent Bare sediment Survey Barry et al. (2000) Increase No data Sewage Bare sediment Survey Bishop et al. (2006a) Increase No data Fish farm Vegetated habitats Survey Boyra et al. (2004) Decrease No data Industrial effluent Coral reef Survey Brewer et al. (2007) No change No data Run-off Vegetated habitats Field experiment Bundy et al. (2003) Increase Increase Fish farm Bare sediment Survey Carss (1990) No change No data Industrial effluent Bare sediment Survey Claireaux et al. (2004) No change No data Industrial effluent Vegetated habitats Survey Connolly & Jones (1996) Decrease Decrease Run-off Vegetated habitats Field experiment Deegan et al. (2002) Increase Increase Fish farm Bare sediment Survey Dempster et al. (2002) Increase No data Fish farm Bare sediment Survey Dempster et al. (2004) Increase No data Fish farm Bare sediment Survey Dempster et al. (2004) Decrease Decrease Run-off Coral reef Survey Fabricius et al. (2005a) Increase Increase Sewage Rocky bottom and rocky reef Survey Gray (1996) Decrease Decrease Sewage Rocky bottom and rocky reef Survey Gray et al. (1992) Increase Increase Sewage Coral reef Survey Grigg (1994) Increase No data Sewage Rocky bottom and rocky reef Survey Guidetti et al. (2003) Increase No data Industrial effluent Bare sediment Survey Hoisington & Lowe (2005) Decrease No data Industrial effluent Bare sediment Survey Jacobsson & Neuman (1991) Increase Decrease Industrial effluent Bare sediment Survey Jones et al. (1996) Decrease No data Industrial effluent Poorly defined Survey Karas et al. (1991) Decrease No change Industrial effluent Coral reef Survey Khalaf & Kochzius (2002) Decrease Increase Industrial effluent Bare sediment Survey Landner et al. (1994) Increase No data Fish farm Bare sediment Survey Machias et al. (2004) Increase No data Fish farm Bare sediment Survey Machias et al. (2005) McKinley & Johnston: Impacts of contaminant sources on marine fish 191

Appendix 1 (continued)

Abundance: Richness: Contaminant Study Research Source direction direction source system approach of impact of impact

Decrease No data Run-off Bare sediment Fisheries catch study Nagai (2003) No change No data Fish farm Rocky bottom and rocky reef Survey Oakes & Pondella (2009) Increase No data Run-off Vegetated habitats Fisheries catch study Oczkowski & Nixon (2008) Increase Increase Sewage Rocky bottom and rocky reef Survey Otway et al. (1996a) Increase Increase Sewage Bare sediment Survey Otway et al. (1996b) Decrease No data Sewage Coral reef Survey Reopanichkul et al. (2009) Increase Decrease Sewage Vegetated habitats Survey Ribeiro et al. (2008) Decrease No data Sewage Bare sediment Survey Roy et al. (2003) Increase No data Sewage Bare sediment Survey Russo (1982) Decrease Decrease Industrial effluent Poorly defined Survey Sandström & Neuman (2003) Decrease No data Industrial effluent Poorly defined Survey Sandström (1994) Decrease No data Sewage Rocky bottom and rocky reef Survey Smith & Suthers (1999) Decrease Decrease Sewage Rocky bottom and rocky reef Survey Smith et al. (1999) Increase No data Run-off Vegetated habitats Survey Tober et al. (2000) Increase Increase Fish farm Bare sediment Survey Tuya et al. (2005) Increase Increase Fish farm Bare sediment Survey Valle et al. (2007)

Editorial responsibility: Jana Davis, Submitted: April 1, 2010; Accepted: September 28, 2010 Annapolis, Maryland, USA Proofs received from author(s): December 8, 2010

Chapter 2

HIGH LEVELS OF SEDIMENT CONTAMINATION HAVE LITTLE INFLUENCE ON ESTUARINE BEACH FISH COMMUNITIES

Final Version:

McKinley, A.C. Dafforn K.A., Taylor, M.D. & E.L. Johnston (2011) High levels of sediment contamination have little influence on estuarine beach fish communities. PLoS One, 6(10). High Levels of Sediment Contamination Have Little Influence on Estuarine Beach Fish Communities

Andrew C. McKinley*, Katherine A. Dafforn, Matthew D. Taylor, Emma L. Johnston Evolution and Ecology Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, New South Wales, Australia

Abstract While contaminants are predicted to have measurable impacts on fish assemblages, studies have rarely assessed this potential in the context of natural variability in physico-chemical conditions within and between estuaries. We investigated links between the distribution of sediment contamination (metals and PAHs), physico-chemical variables (pH, salinity, temperature, turbidity) and beach fish assemblages in estuarine environments. Fish communities were sampled using a beach seine within the inner and outer zones of six estuaries that were either heavily modified or relatively unmodified by urbanization and industrial activity. All sampling was replicated over two years with two periods sampled each year. Shannon diversity, biomass and abundance were all significantly higher in the inner zone of estuaries while fish were larger on average in the outer zone. Strong differences in community composition were also detected between the inner and outer zones. Few differences were detected between fish assemblages in heavily modified versus relatively unmodified estuaries despite high concentrations of sediment contaminants in the inner zones of modified estuaries that exceeded recognized sediment quality guidelines. Trends in species distributions, community composition, abundance, Shannon diversity, and average fish weight were strongly correlated to physico-chemical variables and showed a weaker relationship to sediment metal contamination. Sediment PAH concentrations were not significantly related to the fish assemblage. These findings suggest that variation in some physico-chemical factors (salinity, temperature, pH) or variables that co-vary with these factors (e.g., wave activity or grain size) have a much greater influence on this fish assemblage than anthropogenic stressors such as contamination.

Citation: McKinley AC, Dafforn KA, Taylor MD, Johnston EL (2011) High Levels of Sediment Contamination Have Little Influence on Estuarine Beach Fish Communities. PLoS ONE 6(10): e26353. doi:10.1371/journal.pone.0026353 Editor: Myron Peck, University of Hamburg, Germany Received June 17, 2011; Accepted September 25, 2011; Published October 19, 2011 Copyright: ß 2011 McKinley et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This research was primarily supported by the Australian Research Council through an Australian Research Fellowship awarded to ELJ and a Linkage Grant awarded to ELJ. The writing of this manuscript was also supported through the Canadian National Sciences and Engineering Research Council through an award given to ACM. http://www.arc.gov.au/. http://www.nserc-crsng.gc.ca/. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: [email protected]

Introduction settlement estuarine fish communities is critical to managing and conserving native biodiversity in these systems. A variety of anthropogenic activities contribute to widespread Toxicants such as metals and organic Polycyclic Aromatic modification of estuarine environments, and estuaries are among Hydrocarbons (PAHs) are found in fish at various stages of their the most highly impacted of all marine ecosystems [1]. life cycle, often at levels that may potentially reduce growth or Contamination is a major form of anthropogenic impact in survivorship [7]. Toxic substances may have adverse effects on fish estuarine systems, acting as a stressor which influences the by interfering with reproductive processes and by causing composition and health of ecological communities. Estuaries are developmental problems [8]. However, impacts of contaminants generally believed to contain the highest levels of contamination of on post-settlement fish assemblages have been shown to be highly any marine environment due to their proximity to human variable; many studies have reported either localized impacts or no settlements and their position directly downstream of agricultural effect of contaminants on marine fish assemblages, and negative and industrial activities [2]. Many of these complex estuarine impacts at large scales are rarely described [6]. Different types and habitats provide a ‘nursery’ function for ecologically and concentrations of contaminants may have either toxic or enriching economically important species of fish [3]. It has been demon- effects on fish assemblages, however the effects differ between strated that contaminants in these systems can have substantial different guilds of fish [9]. In many cases, contaminants with impacts on larval fish [4], and that they generally reduce the enrichment properties (such as nutrients or sewage) have a largely richness and evenness of marine invertebrate communities [5]. positive effect on the abundance and diversity of post-settlement Despite this, few studies have identified impacts of contamination fish [6]. on post-settlement fish or within the context of natural variability While toxicants are thought to have significant impacts on wild in physico-chemical conditions [6]. As a result, the relative fish communities, natural variation in physico-chemical conditions importance of contaminant impacts on fish assemblages compared such as changes in turbidity, salinity, temperature and pH have to natural hydrographic variability is poorly understood. Identi- consistently been shown to have a large influence on the fying stressors and monitoring ecological impacts in post- composition and species richness of fish assemblages [10]. Spatial

PLoS ONE | www.plosone.org 1 October 2011 | Volume 6 | Issue 10 | e26353 Contamination Has Little Influence on Beach Fish variation in physico-chemical factors can manifest as gradients the biogenic habitat itself may be degraded by these stressors within estuarine systems, influencing the distribution of fish species [18,19]. along the length of an estuary [11]. Similarly, seasonal and We explore the impacts of large-scale anthropogenic distur- temporal variability in physico-chemical factors can influence fish bance on estuarine beach fish communities across heavily modified communities [12]. Estuary geomorphology and physical structure and relatively unmodified estuaries in New South Wales, Australia. will also affect estuarine ecology through variation in entrance Specifically, we investigate whether high levels of modification and conditions (e.g. permanently or intermittently open estuaries), the sediment contamination in the estuarine environment influence relative size of the fluvial and tidal deltas, and the evolutionary the composition, abundance, and Shannon diversity of the post- maturity (stage of sediment filling) of the estuarine system [13]. settlement fish assemblage. We assess these impacts within the The way in which these factors influence the fish assemblage is context of environmental variability both within and between increasingly juxtaposed against the effects of anthropogenic estuaries in order to understand the scale of anthropogenic modification to estuaries. As such, identifying the major drivers impacts relative to variation in environmental conditions [20]. of fish distribution is likely to be more complicated in modified While this would ideally be assessed using a Before After Control habitats. Impact (BACI) sampling design, baseline data was not available Estuarine beaches are a dynamic environment representing a for the study estuaries. As such, we utilize a spatial comparison of juncture between terrestrial and marine systems. These environ- heavily modified vs. relatively unmodified estuaries to test our ments are heavily influenced by both wave action and tidal forces hypotheses [21]. [14]. Due to their shallow nature and position at the shoreline, they are also likely to be relatively heavily impacted by Methods anthropogenic developments situated onshore or within estuarine waters. Beach environments may be directly influenced by run-off Study Sites from urban environments, shoreline alteration, changes to Fish were sampled in six permanently open estuaries along the terrestrial detritus patterns, beach fishing, changes to sediment south coast of New South Wales, Australia. These included three quality, and physical disturbance from recreational activities heavily modified estuaries, Port Jackson (33u44.2589S, [15,16]. Fish living in these habitats represent a diverse 151u16.5429E), Botany Bay (33u59.3529S, 151u11.4339E), and community which is primarily small bodied species or juveniles, Port Kembla (34u28.1219S, 150u54.4109E), and three relatively feeding on a diverse array of food items including terrestrial and unmodified estuaries, Port Hacking (34u04.6809S, 151u09.3119E), marine detritus, plankton and sediment dwelling invertebrates, Jervis Bay (35u04.7629S, 150u44.8589E), and the Clyde River marine vegetation, and other fish [17]. As such, estuarine beaches (35u44.2339S, 150u14.2729E) (Figure 1). The three heavily represent a potentially important environment for the study of modified estuaries are all anthropogenically disturbed environ- anthropogenic impacts on fish. While beach fish may be ments close to large urban and industrial areas and are subject to responsive to anthropogenic modification for the reasons dis- intense commercial and recreational boating traffic, historic and cussed, it should be noted that fish which live in sensitive biogenic ongoing contamination, concentrated recreational fishing activity, habitats such as coral reefs and seagrass beds maybe more sensitive frequent dredging for navigation and construction, and substantial to modification and contamination than those in bare habitats, as urbanization of their shoreline and catchment. In comparison, the

Figure 1. Location of study sites in the six focal estuaries: a) Port Jackson (heavily modified), b) Botany Bay (heavily modified), c) Port Hacking (relatively unmodified), d) Port Kembla (heavily modified), e) Jervis Bay (relatively unmodified), and f) Clyde River (relatively unmodified). Filled diamonds (¤) indicates outer zone sites. Filled circles (N) indicates inner zone sites. doi:10.1371/journal.pone.0026353.g001

PLoS ONE | www.plosone.org 2 October 2011 | Volume 6 | Issue 10 | e26353 Contamination Has Little Influence on Beach Fish relatively unmodified estuaries have less concentrated fishing out on each beach and the net was pulled out straight for 10 m activity, less boating traffic (almost none of which is commercial), prior to encircling the beach. In this way an area of approximately less urbanization of the coastline and catchment, and virtually no 100 m2 was sampled at each time of sampling. Only sites where heavy industry [22,23]. While these estuaries do have some degree the researchers could pull the beach seine out 10 m from shore of agricultural land use in their catchment, the majority of the while keeping their head above water were sampled. As such, all catchment in all of the relatively unmodified estuaries is within sites had a shallow slope (slope #20.2) and the sampling depth conservation areas, forestry zones, or areas where anthropogenic was 1–1.8 m (at its deepest point, 10 m from shore). All fish were 2 utilization is negligible [24]. Both the Clyde River (within euthanized using a 100 mg L 1 benzocaine solution and frozen Bateman’s Bay Marine Park) and Jervis Bay (Jervis Bay Marine for transportation back to the laboratory. Fish were sorted to Park) are within marine parks. Port Hacking is located between the species and standard length and wet weight measurements were suburbs of southern Sydney and the forested slopes of Royal taken. Due to identification difficulties, Sillago sp. (whiting) National Park, which lines the southern border of the estuary. ,10 cm in length were classified to genus only. Individuals larger While not strictly within a marine park, Port Hacking’s catchment than this were identified as either Sillago maculata or Sillago ciliata. is largely intact due to its proximity to the Royal National Park and there is no major industrial activity within the estuary, though Physico-chemical Sampling Methods navigation channels in the outer zone are periodically dredged At each sampling time and location physico-chemical data [23]. Previous monitoring indicates that the heavily modified (temperature, salinity, pH, turbidity) were collected using a YSI- estuaries are also nutrient enriched, whilst nutrient levels in the Sonde 6820-V2 (Yellow Springs, USA) (calibrated weekly). At relatively unmodified estuaries are less elevated [22]. each site benthic sediments were collected once at 5 m depth Each estuary was divided into an inner and outer zone (see site between Feb–Mar 2010 using a sediment grab. Sediments used for coding in Figure 1) based on qualitative observations of physical metal analyses were oven dried at 50uC before being homogenized characteristics. The inner zone is further away from the ocean and to a fine powder in a ball mill. A 0.5 g sub-sample from each site represents the lower reaches of the estuarine tributary. In this zone was digested according to Method 3051A [27]. Specifically, the turbidity and temperatures are generally higher than in the outer sediments were digested in 9 mL HNO3 and 3 mL HCl for zone [13]. The outer zone sites are near the marine entrance to 20 min at 200uC in a 1000 W microwave. Following digestion, the estuaries where salinity, coastal flushing, wave exposure, and samples were diluted up to 30 mL with Milli-Q water and oceanic current systems have greater influence. In this zone analyzed for metal content (As, Co, Cr, Cu, Fe, Mn, Ni, Pb, Zn) sediment grain sizes are also larger and there is greater tidal using inductively coupled plasma-optical emission spectroscopy influence [13]. While all of the estuaries examined in this study are (ICP-OES). The instrument was calibrated with matrix-matched 2 permanently open tidal systems, estuary size varies. As such, standards and had limits of detection (LOD) of 3 mg kg 1 for Cd, 2 estuary size is included in our analysis as a measure of structural Co, Mn and Zn, and 5–25 mg kg 1 for As, Cu, Ni and Pb. variability. Analysis of certified reference materials (sediment CRM – It should also be noted that Jervis Bay and the Clyde River are LGC6137 and oyster CRM – 1566b, Graham B. Jackson Pty within a different bioregion than the other four estuaries, Ltd, Australia) indicated adequate recoveries (within 615%) for according to the Interim Biogeographic Regionalization of most metals. Where recoveries were outside this range, the data Australia (IBRA) system, though the maximum distance between were omitted from analysis. Outer zone sites were primarily sandy estuaries is only 275 km (Batemans Bay to Port Jackson) [25]. and so only inner zone sites of each estuary were analyzed for While this indicates that some differences exist in the biological organic contaminants (all sites were analyzed for metals). Samples and environmental conditions between these areas, most of the fish were analyzed for 16 priority PAHs: naphthalene, acenaphthylene, species examined in this study are known to occur in all the acenaphthene, fluorene, phenanthrene, anthracene, fluoranthene, estuaries examined in this study. Notably, the six estuaries pyrene, benz(a)anthracene, chrysene, benzo(a)pyrene, benzo(b)- examined in this study are at least several hundred kilometers fluoranthene, benzo(k)fluoranthene, indeno(1,2,3-cd)pyrene, di- within the known range of the major species which drive the trends benz(a,h)anthracene and benzo(g,h,i)perylene by the National in this analysis [26]. In addition, differences in physico-chemical Measurement Institute (Sydney, Australia). Values were then variables between zones were greater than between estuaries and normalized to 1% total organic carbon for comparison with the habitat sampled was judged to be reasonably similar in all sediment quality guidelines. estuaries (beaches). For these reasons, we believe that comparisons Sediments were selected to measure contaminants in these between these estuaries are valid for this analysis, despite different systems (rather than a water column measure of contamination) bioregional classification. Other studies have utilized similar for several reasons. First, it is well known that fish accumulate comparisons between these estuaries [4,9]. contaminants through their food to a much greater degree than through their gills or through interaction with contaminated water Fish Sampling Methods [28,29]. The majority of species in this study are benthic or Within each estuary six beach sites consisting of bare sediment benthopelagic foragers and so most would interact with sediments were selected on an ad hoc basis, with three sites in each zone. In regularly during feeding [17]. Second, contaminants accumulate order to capture temporal variability in these physico-chemical in estuarine sediments over the long-term, as such sediment metals conditions, we conducted four sampling rounds over two years. values are less temporally variable and represent a contemporary Thus, there were a total of four rounds of sampling; November– threat from historical pollution sources [30]. December 2009 (Early Summer), February–March 2010 (Sum- For univariate analysis and graphical presentation of total mer), November–December 2010 (Early Summer) and February– metals and PAH contamination a combined sediment metals and March 2011 (Summer). On the south-east coast of Australia, the PAH quotient were calculated following Long et al. (2006) [31]. ‘‘Summer’’ season coincides with the warmest ocean water Each individual metal contaminant load was divided by the low temperatures. Fish were sampled using a beach seine net with a and high trigger values from the ANZECC sediment quality 20 m headline, a 2 m drop, a 1.5 m cod end, constructed from a guidelines [32]. The high and low quotients for each contaminant 12 mm mesh. Before deployment a 10 m segment was measured were then summed for each sample and divided by two to give a

PLoS ONE | www.plosone.org 3 October 2011 | Volume 6 | Issue 10 | e26353 Contamination Has Little Influence on Beach Fish mean sediment quality guideline quotient (mSQGQ) for each Ecological Classification sample. The ANZECC trigger values are threshold values which In order to evaluate the role of ecological characteristics in are meant to provide a baseline for assessing the impacts of marine determining the relative sensitivity of different functional groups, and freshwater contaminants. Where concentrations of contam- species were classified both according to their trophic level and inants exceed the trigger values, it is believed that there is a risk of their usage of estuarine environments during their life cycle adverse environmental effects [32]. (Appendix S1). The numeric trophic level of each species was

Figure 2. Mean (±SE) physico-chemical and sediment contaminant variables by zone and estuary. Including a) Temperature, b) Salinity, c) pH, d) Mean Quotient of Sediment Metals, and e) Mean Quotient of Sediment PAH values. doi:10.1371/journal.pone.0026353.g002

PLoS ONE | www.plosone.org 4 October 2011 | Volume 6 | Issue 10 | e26353 Contamination Has Little Influence on Beach Fish

Table 1. Univariate analysis of physico-chemical variables and estuary size covariates.

a) Temperature b) Salinity c) pH d) Turbidity

Source dF MS F p-value MS F p-value MS F p-value MS F p-value

Dis 1 4.84 1.71 0.250 0.01 0.28 0.961 0.17 0.61 0.738 0.00 0.45 0.944 Zo 1 20.05 6.71 0.004 16.34 3.79 0.038 23.25 10.63 0.004 0.74 0.90 0.557 Ye 1 7.83 0.34 0.835 0.01 0.14 0.974 8.70 2.44 0.194 0.02 0.38 0.890 Ti 1 26.87 1.00 0.517 5.93 0.36 0.844 1.68 0.83 0.557 2.25 1.29 0.376 Es(Dis) 4 1.20 0.73 0.797 5.86 3.62 0.001 9.39 3.11 0.001 3.81 1.45 0.089 DisxZo 1 0.51 0.64 0.713 0.42 0.18 0.988 1.77 0.67 0.703 6.46 2.40 0.061 DisxYe 1 0.11 0.86 0.530 0.16 3.31 0.133 8.06 0.80 0.594 1.67 1.36 0.358 DisxTi 1 2.79 1.11 0.468 0.21 1.72 0.306 3.19 0.41 0.803 0.51 1.34 0.368 ZoxYe 1 0.57 2.43 0.217 0.37 0.32 0.876 0.61 1.65 0.316 2.02 1.14 0.434 ZoxTi 1 0.80 1.48 0.362 0.81 0.37 0.831 0.08 1.82 0.275 0.12 0.44 0.821 YexTi 1 25.35 25.88 0.007 24.98 7.10 0.048 2.45 1.58 0.285 1.74 1.26 0.350 ZoxEs(Dis) 4 0.90 1.67 0.078 3.26 4.46 0.001 1.42 1.93 0.042 1.49 0.85 0.738 Es(Dis)xYe 4 0.41 0.38 0.815 0.52 0.18 0.938 1.75 1.12 0.378 1.91 1.16 0.382 Es(Dis)xTi 4 2.56 2.36 0.091 1.58 0.47 0.747 1.43 0.98 0.443 1.07 0.84 0.563 DisxZoxYe 1 0.15 9.03 0.035 0.30 6.98 0.030 1.77 3.34 0.146 1.69 0.81 0.620 DisxZoxTi 1 0.06 1.47 0.367 1.05 4.56 0.057 0.31 1.84 0.296 0.47 0.52 0.788 DisxYexTi 1 0.85 0.87 0.402 0.59 0.17 0.708 10.25 6.61 0.066 0.34 0.25 0.740 ZoxYexTi 1 0.49 0.54 0.501 6.22 3.74 0.127 0.03 0.21 0.694 0.55 1.17 0.349 Si(Es(Dis)xZo) 24 0.46 1.46 0.107 0.42 2.93 0.006 0.25 1.71 0.034 0.97 1.11 0.248 ZoxEs(Dis)xYe 4 0.12 0.15 1.000 0.14 0.15 1.000 0.42 1.97 0.104 1.63 1.76 0.065 ZoxEs(Dis)xTi 4 0.66 0.69 0.793 0.46 0.33 0.986 0.09 1.22 0.341 0.78 1.28 0.264 Es(Dis)xYexTi 4 0.98 18.34 0.001 3.52 29.69 0.001 1.55 9.47 0.001 1.38 2.18 0.032 DisxZoxYexTi 1 0.00 0.00 0.952 0.14 0.08 0.774 0.15 1.12 0.355 1.03 2.19 0.193 YexSi(Es(Dis)xZo) 24 0.23 4.27 0.001 0.08 0.70 0.796 0.16 1.01 0.517 0.82 1.29 0.149 TixSi(Es(Dis)xZo) 24 0.12 2.33 0.026 0.10 0.86 0.657 0.08 0.48 0.957 0.63 1.00 0.495 ZoxEs(Dis)xYexTi 4 0.91 17.07 0.001 1.66 14.05 0.001 0.13 0.81 0.539 0.47 0.74 0.665 Res 24 0.05 0.12 0.16 0.63 a) Temperature, b) Salinity, c) pH, and d) Turbidity. Factors: Dis = Disturbance Category (Heavily Modified vs. Relatively Unmodified), Zo = Zone (Inner vs. Outer), Ti = Time of Sampling, Ye = Year, Es = Estuary, Si = Site. Bold values correspond to significant values for higher-level factors or interactions between non-random factors. doi:10.1371/journal.pone.0026353.t001

Table 2. Univariate analysis of sediment metals quotient in determined using the fish database website Fishbase [33] and this the full model. data was used to calculate a Marine Trophic Index for each sample [34]. In addition, each species was classified into discrete guilds based on their usage of estuaries during their life cycle Metals Quotient - Full Model following Elliot et al. (2007). Three broad categories were used in this study: Source dF MS F p-value

Dis 1 32.05 5.18 *0.084 N Estuarine Species: Species which spawn within the estuary and which normally complete their entire life cycle within the Zo 1 0.15 0.02 0.926 estuarine environment. Es(Dis) 4 6.18 2.65 0.032 N Estuarine Opportunists: Species which primarily spawn in DisxZo 1 0.02 0.00 0.987 marine coastal waters but enter the estuary either in their Es(Dis)xZo 4 7.52 3.23 0.014 larvae or juvenile stages. Many of these species require Si(Es(Dis)xZo) 24 2.33 Den = 0 sheltered estuarine habitats during their larval and juvenile Res 108 0.00 stages and are hence dependent on the estuarine environments for reproduction. Most species spend part of their adult stage Factors: Dis = Disturbance Category (Heavily Modified vs. Relatively outside of estuaries. Unmodified), Zo = Zone (Inner vs. Outer), Es = Estuary, Si = Site. Bold values correspond to significant values for higher-level factors or interactions between N Marine Stragglers: Species which spawn at sea and normally non-random factors. enter estuaries only in low numbers, occurring most frequently *Indicates Monte Carlo p value. in the lower reaches of the estuary. Many are stenohaline and doi:10.1371/journal.pone.0026353.t002 are primarily associated with coastal marine waters.

PLoS ONE | www.plosone.org 5 October 2011 | Volume 6 | Issue 10 | e26353 Contamination Has Little Influence on Beach Fish

These classifications were made through a review of existing Results literature for these species [10,35,36] and through consultation with regional fish experts (A.G. Miskiewicz, personal communi- Fish Assemblage Characteristics cation, 2011). In total more than 10,350 fish representing 51 species were collected and identified during the study. Thirty of these species Statistical Analysis were relatively rare, represented by only 1–10 individuals over the All multivariate and univariate datasets were analyzed as mixed- course of the study. By abundance the 10 most common species , model PERMANOVA in PRIMER v.6.4 [37]. Prior to analysis, accounted for 92.5% of the fish assemblage. In order of abundance and biomass data were log(x+1) transformed. Bray- abundance these were Ambassis jacksoniensis (39.8%), Myxus elongatus Curtis similarity matrices were calculated for multivariate data (18.7%), Sillago sp. (9.3%), Leptatherina presbyteroides (7.8%), Hyperlo- while Euclidean similarity matrices were used for univariate phus vittatus (3.7%), Gerres subfasciatus (3.2%), Atherinomorus vaigiensis measures. A dummy variable of 1 was added when calculating the (3.1%), Favonigobius lentiginosus (2.8%), Sillago maculata (2.3%), and similarity matrices in order to compensate for zero values. The Torquigener pleurogramma (1.9%). The summarized biological dataset PERMANOVA design employed in the course of this analysis can be found in Appendix S1. consisted of the following factors: Dis - Disturbance category – Heavily Modified or Relatively Physico-chemical variables and Sediment Contamination Unmodified (2 levels, Fixed) In most estuaries, temperature (Figure 2a) was higher in the Zo - Zone – Inner or Outer (2 levels, Fixed) inner zone sites while salinity (Figure 2b) and pH (Figure 2c) Ti – Time of Year – Early Summer or Summer (2 levels, were lower in the inner zone (Table 1 a–c). Physico-chemical Random) variables also showed significant temporal variation (Table 1, Ye – Year – 2009–2010 or 2010–2011 (2 levels, Random) Ye6Ti). While this variation was significant, the trends in Es – Estuary (Disturbance Category) – (6 estuaries, Random) physico-chemical variables between zones remained consistent. Si – Site (Estuary(Disturbance Category)6Zone) – (36 sites, Turbidity did not show a significant trend by zone (Table 1d). Random) There appeared to be higher concentrations of sediment metals Reduced versions of this model were used to analyze the in the inner zones of the heavily modified estuaries as well as the sediment metals and PAH data. In these reduced models the outer zone of the heavily modified Port Kembla (Figure 2d). This Time and Year factors were removed as these covariates were not resulted in a significant interaction between zone and estuary replicated. Monte-Carlo p-values were used in some places where nested within disturbance category (Table 2, 3a). The outer the number of unique permutations was less than 20 (these values zones of all other estuaries (heavily modified or relatively are marked in tables) [37]. Analysis of covariation of physico- unmodified) and the inner zone of the relatively unmodified chemical, metal, and PAH covariates was conducted using the estuaries displayed much lower levels of sediment metal distance-based linear model (DistLM) in PERMANOVA. This contamination (Figure 2d). PAH contamination was only program calculates a distance-based multivariate multiple measured in the inner zones and did not show a clear trend by regression (e.g. dbRDA) for any linear model on the basis of disturbance category, but did differ by estuary (Table 3b). The any distance measure, using permutation procedures [38]. The heavily modified estuaries Port Jackson and Port Kembla ‘Best’ selection procedure was employed using BIC selection displayed relatively high PAH values (Table 1c, Figure 2e). criteria. Covariate factors were then analyzed graphically using Sediment metals values at many of the inner zones within the Principal Coordinated Ordination (PCO). PCO is a PERMA- heavily modified estuaries and in the outer zone of Port Kembla NOVA function that performs a principal coordinate analysis of were above levels predicted to have biological effects on infauna any symmetric distance matrix. This analysis is also called metric according to water quality guidelines [32]. multi-dimensional scaling [39]. All covariate factors were plotted In some analyses several of the random interaction terms were in the PCO charts, however, turbidity did not correlate strongly also significantly different (e.g. Ti, Ye, Si(Es(Dis)xZo) and Es(Dis)). enough to show a discernable vector line. The highest correlating Here and elsewhere, the test of the main effects are still considered, species (those with a multiple correlation factor .0.2) were also as higher level fixed-factor effects remain relevant despite an included in the PCO charts. interaction with a random factor [40].

Table 3. Univariate analysis of a) sediment metals quotient and b) sediment PAH quotient under a reduced model (inner zone only).

a) Metal Quotient - Inner Only b) PAH Quotient - Inner Only

Source dF MS F p-value MS F p-value

Dis 1 39.13 34.95 *0.006 9.89 1.14 *0.349 Es(Dis) 4 1.12 0.49 0.795 8.65 3.91 0.029 Si(Es(Dis)) 12 2.28 Den = 0 2.21 Den = 0 Res 54 0.00 0.00

Factors: Dis = Disturbance Category (Heavily Modified vs. Relatively Unmodified), Es = Estuary, Si = Site. Bold values correspond to significant values for higher-level factors or interactions between non-random factors. *Indicates Monte Carlo p value. doi:10.1371/journal.pone.0026353.t003

PLoS ONE | www.plosone.org 6 October 2011 | Volume 6 | Issue 10 | e26353 Contamination Has Little Influence on Beach Fish

Table 4. Univariate analysis of a) Species Richness, b) Shannon Diversity, c) Biomass, d) Average Fish Weight, and e) Abundance.

a) Species Richness b) Shannon Diversity c) Biomass d) Fish Size e) Abundance

p- p- p- p- p- Source dF MS F value MS F value MS F value MS F value MS F value

Dis 1 23.36 3.08 0.079 1.01 2.12 0.170 6.17 2.85 0.080 497.02 1.10 0.463 0.03 2.25 0.136 Zo 1 318.03 10.47 0.005 11.12 10.75 0.003 66.32 4.31 0.032 660.14 5.75 0.011 53.08 7.51 0.006 Ye 1 17.36 1.16 0.432 0.08 0.24 0.894 4.65 1.14 0.467 738.15 1.09 0.469 21.44 10.82 0.015 Ti 1 0.69 0.18 0.943 0.10 0.28 0.878 0.77 0.63 0.693 39.37 0.56 0.506 0.81 3.23 0.139 Es(Dis) 4 12.66 0.85 0.674 0.35 0.64 0.883 5.22 0.99 0.508 526.49 0.79 0.714 5.08 1.61 0.108 DisxZo 1 0.44 0.41 0.884 0.22 0.54 0.777 4.70 0.68 0.679 0.81 0.40 0.883 2.91 1.20 0.411 DisxYe 1 0.11 0.10 0.982 0.15 0.78 0.587 0.20 0.14 0.959 528.89 0.82 0.562 0.21 0.18 0.944 DisxTi 1 1.00 0.17 0.945 0.15 0.94 0.494 0.18 0.16 0.950 270.78 1.36 0.317 0.03 0.18 0.946 ZoxYe 1 5.44 1.81 0.306 0.00 0.36 0.811 0.00 3.36 0.138 47.80 0.26 0.885 0.00 0.17 0.938 ZoxTi 1 1.00 0.52 0.726 0.00 0.35 0.831 5.47 3.14 0.157 10.37 0.15 0.976 4.35 0.79 0.561 YexTi 1 9.00 4.97 0.098 0.54 4.07 0.135 2.33 1.49 0.286 5.57 0.05 0.812 0.01 0.00 0.962 ZoxEs(Dis) 4 22.30 1.34 0.219 0.90 1.68 0.069 5.72 1.56 0.112 114.94 0.80 0.705 2.90 1.08 0.435 Es(Dis)xYe 4 7.59 2.52 0.075 0.33 2.27 0.093 3.12 1.00 0.417 810.73 2.51 0.064 2.23 0.65 0.621 Es(Dis)xTi 4 4.74 1.37 0.281 0.28 1.54 0.225 1.39 1.03 0.407 287.14 1.28 0.324 1.12 0.53 0.736 DisxZoxYe 1 0.69 0.80 0.568 0.15 1.03 0.483 0.17 7.59 0.040 26.21 0.41 0.809 0.90 2.17 0.248 DisxZoxTi 1 2.25 0.73 0.635 0.05 0.69 0.604 0.01 1.99 0.255 926.89 3.74 0.108 0.61 0.53 0.720 DisxYexTi 1 12.25 6.76 0.060 0.02 0.16 0.701 9.78 6.24 0.092 8.48 0.07 0.777 14.91 5.29 0.098 ZoxYexTi 1 0.25 0.09 0.772 0.42 2.16 0.219 1.28 0.26 0.635 198.57 3.16 0.131 3.98 5.08 0.086 Si(Es(Dis)xZo) 24 9.20 1.90 0.016 0.36 2.02 0.006 6.11 1.39 0.122 286.73 0.81 0.720 2.96 0.97 0.552 ZoxEs(Dis)xYe 4 4.26 1.44 0.219 0.12 1.05 0.469 0.21 0.29 0.987 197.69 1.26 0.336 0.69 0.59 0.855 ZoxEs(Dis)xTi 4 6.85 1.49 0.218 0.14 0.91 0.594 2.06 0.62 0.830 259.05 1.43 0.279 2.54 1.50 0.221 Es(Dis)xYexTi 4 1.81 0.54 0.703 0.13 0.69 0.590 1.57 0.65 0.631 110.07 0.41 0.803 2.82 2.16 0.097 DisxZoxYexTi 1 0.00 0.00 1.000 0.21 1.10 0.342 0.47 0.09 0.782 11.67 0.19 0.703 0.09 0.11 0.740 YexSi(Es(Dis)xZo) 24 2.52 0.76 0.766 0.10 0.52 0.937 3.97 1.64 0.109 344.63 1.27 0.328 2.61 2.00 0.046 TixSi(Es(Dis)xZo) 24 4.09 1.23 0.331 0.17 0.91 0.579 2.15 0.89 0.627 342.85 1.26 0.339 1.78 1.36 0.212 ZoxEs(Dis)xYexTi 4 2.73 0.82 0.533 0.19 1.01 0.431 5.03 2.08 0.125 49.70 0.18 0.944 0.78 0.60 0.650 Res 24 3.33 0.19 2.42 271.34 1.30

Factors: Dis = Disturbance Category (Heavily Modified vs. Relatively Unmodified), Zo = Zone (Inner vs. Outer), Ti = Time of Sampling, Ye = Year, Es = Estuary, Si = Site. Bold values correspond to significant values for higher-level factors or interactions between non-random factors. doi:10.1371/journal.pone.0026353.t004

Fish Assemblages – Within Estuary Variation Fish abundance was significantly higher in the inner zones and in Species richness, Shannon diversity, and fish biomass were all the first year of sampling (Table 4, Figure 3 e,f). While fish significantly greater in the inner zones compared to the outer abundance was significantly greater in the inner zone overall, zones (Table 4 Figure 3a,b,c), while no main effects or interaction occasionally the difference between zones was small, and Botany terms were detected for any of these measures for disturbance Bay appeared to have a greater abundance in the outer zone during category (Table 4). Species richness and Shannon diversity also year 2 (Figure 3 e,f). Multivariate analysis of the community showed significant variation by site (Table 4a,b). Average fish composition found that inner and outer zone fish communities weight was greater in the outer zones of all estuaries except Port differed significantly (Table 5, Figure 4). There was also significant Kembla and Jervis Bay, where the average fish weight was variation in community composition between sites (Table 5). Simper approximately equal between zones (Table 4, Figure 3d). Port analysis revealed that the top six species contributing to differences Jackson displayed a trend towards having higher fish biomass than between zones were M. elongatus, Sillago sp. (,10 cm), A. jacksoniensis, other estuaries across both zones (Figure 3c) while Port Kembla S. ciliata, G. subfasciatus, and F. lentiginosus. All of these species were displayed a trend towards higher average fish weight than other more abundant in the inner zone and collectively they contributed estuaries (Figure 3d). The pattern of increased biomass, species to approximately 59% of the difference between zones. However, richness and Shannon diversity in the inner estuary was weakest several of these species also varied significantly by other factors. for Port Kembla, where we also observed the smallest difference Differences in abundance between zones were significant for the between zones for physico-chemical variables. Interestingly, the Sillago sp., S. ciliata, G. subfasciatus, F. lentiginosus, and nearly significant outer harbor of Port Kembla was the only outer zone to contain for A. jacksoniensis (Table 6 a–f, Figure 5). F. lentiginosus, S. ciliata,and substantial sediment contamination but this did not relate to Sillago sp. also varied significantly by site. A. jacksoniensis showed a reduced fish biomass, species richness, or Shannon diversity significant zone6estuary interaction, which was driven by higher relative to other outer zones. abundances in the inner zone for all estuaries except Botany Bay,

PLoS ONE | www.plosone.org 7 October 2011 | Volume 6 | Issue 10 | e26353 Contamination Has Little Influence on Beach Fish

Figure 3. Mean (±SE) community level indicators by zone/estuary. Including a) Species Richness, b) Shannon Diversity, c) Biomass, d) Average Fish Weight, e) Year 1 Abundance, f) Year 2 Abundance. doi:10.1371/journal.pone.0026353.g003 where it was more abundant in the outer zone. Of these species, relatively consistent across sampling categories and did not differ only G. subfasciatus differed by disturbance category with significantly significantly by zone, estuary, year of sampling, or disturbance more individuals in the inner zone of the heavily modified estuaries category (p.0.05). (Table 6c). This resulted in a near-significant zone6disturbance In contrast, some variation was displayed in the relative category interaction. abundance of estuary usage guilds. Estuarine opportunists accounted for the majority of the dataset, with 23 species Marine Trophic Index and Ecological Characteristics comprising 82% of the fish assemblage. Estuarine Opportunists Within estuary variation in the fish assemblage was not reflected were significantly more abundant in the inner zone of all estuaries by changes in the Marine Trophic Index (MTI). The MTI was (MS1 = 79.92, p = 0.002) except Botany Bay. They also trended

PLoS ONE | www.plosone.org 8 October 2011 | Volume 6 | Issue 10 | e26353 Contamination Has Little Influence on Beach Fish

Table 5. Multivariate analysis of community composition in the full model.

Community Composition - Full Model

Source dF MS F p-value

Dis 1 6513.20 1.28 0.269 Zo 1 32951.00 3.85 0.026 Ye 1 5943.40 1.49 0.238 Ti 1 1815.80 0.75 0.666 Es(Dis) 4 4793.30 1.20 0.088 DisxZo 1 5344.70 0.87 0.627 DisxYe 1 1468.40 0.66 0.768 DisxTi 1 2400.90 0.83 0.614 ZoxYe 1 2451.00 1.29 0.321 Figure 4. Two dimensional MDS plot of multivariate assem- ZoxTi 1 1716.60 0.90 0.576 blage composition by zone. Symbols represent centroids of the YexTi 1 3692.20 1.61 0.229 assemblage composition. Stress value of 0.24 represents a relatively ZoxEs(Dis) 4 3732.20 1.14 0.182 weak ordination of the multivariate data, which is not the result of dispersion (p = 0.625). Es(Dis)xYe 4 1824.20 0.84 0.700 doi:10.1371/journal.pone.0026353.g004 Es(Dis)xTi 4 1790.20 0.82 0.685 DisxZoxYe 1 2140.60 2.82 0.033 inner and outer zones differed significantly. PCO plots indicate DisxZoxTi 1 2036.10 1.89 0.144 that the major cluster of outer zone sites was strongly related to DisxYexTi 1 3891.70 1.69 0.196 both increased pH and salinity. In contrast, the major cluster of inner zone sites corresponded strongly to increased temperature. ZoxYexTi 1 2337.40 1.16 0.333 Sediment metals and estuary size did not correlate with the Si(Es(Dis)xZo) 24 3625.40 1.48 0.001 separation of zones (Figure 6a). ZoxEs(Dis)xYe 4 1127.00 0.71 0.955 Figure 6b plots the species which were highly correlated with ZoxEs(Dis)xTi 4 1798.50 0.88 0.728 the major clusters of inner and outer zone sites (those with a Es(Dis)xYexTi 4 2300.40 1.49 0.056 correlation factor .0.2). In order of the decreasing strength of this , DisxZoxYexTi 1 347.56 0.17 0.938 relationship these were: M. elongatus, Sillago sp. ( 10 cm), A. jacksoniensis, S. ciliata, S. maculata, and T. glaber. These species were YexSi(Es(Dis)xZo) 24 1720.80 1.12 0.243 all more abundant in the inner zone (Figure 5, Figure 6b). The TixSi(Es(Dis)xZo) 24 1763.80 1.15 0.218 distributions of these species (with the exception of M. elongatus) ZoxEs(Dis)xYexTi 4 2021.40 1.31 0.139 approximate the vector lines of the temperature, salinity, and pH Res 24 1540.20 covariates. This suggests that there is a strong association between the distributions of Sillago sp. (,10 cm), A. jacksoniensis, S. ciliata, S. Factors: Dis = Disturbance Category (Heavily Modified vs. Relatively maculata, and T. glaber and temperature, salinity, and pH. Unmodified), Zo = Zone (Inner vs. Outer), Ti = Time of Sampling, Ye = Year, Es = Estuary, Si = Site. Bold values correspond to significant values for higher- A second covariate analysis was undertaken considering inner level factors or interactions between non-random factors. zone data only. This allowed the inclusion of sediment PAH data doi:10.1371/journal.pone.0026353.t005 as an additional covariate (which was available only from inner zone sites). As shown in Table 7b, salinity, pH, temperature, and towards higher abundance in the relatively unmodified estuaries, sediment metal quotient values significantly correlated with the though this was not significant (MS1 = 0.71, p = 0.084). Estuarine fish assemblages when inner zone sites were considered separately opportunists were also significantly more abundant in the first year from the outer zone data (Figure 7a). Within the inner zone sites (MS1 = 25.244, p = 0.016). 12 estuarine species accounted for the distributions of M. elongatus and A. jacksoniensis approximated ,17% of the fish assemblage. No significant differences were the vector lines of the sediment metals quotient. This suggests that found in the abundance of estuarine species (p.0.05). All 17 these species were more abundant in sites with lower sediment species of marine stragglers were comparatively rare, and metals contamination and lower salinity. In contrast, T. glaber, S. accounted for only 1.2% of the dataset. Marine stragglers only maculata, S. ciliata, and Sillago sp. did not show a strong relationship differed significantly by estuary (MS4 = 1.58, p = 0.02). This was to sediment metals or PAH, but approximated the vector lines of due to increased abundance of this guild in Botany Bay and the the temperature and pH covariates. This suggests that these Clyde compared to other estuaries, and was primarily driven by species were more abundant in areas where temperatures were increased abundance of T. bailloni in Botany Bay and L. platycephala higher and pH values lower (Figure 7b). While sediment metals in the Clyde River. correlate significantly in this analysis, salinity, pH, and temperature still contribute a greater proportion of the variance (Table 7b). Covariate Analyses In addition, overall community composition of the inner zone Salinity, pH, and temperature were all found to have a biological data does not differ significantly by disturbance category significant relationship with the fish community composition when (Table 8). There is no correlation between the biological analyzed using a DistLM analysis (Table 7a, Figure 6a). As stated assemblage and sediment PAH levels (Table 7b). Finally, turbidity earlier, multivariate analysis of community composition found that did not correlate strongly in either covariate analysis (Table 7).

PLoS ONE | www.plosone.org 9 October 2011 | Volume 6 | Issue 10 | e26353 Contamination Has Little Influence on Beach Fish

Table 6. Univariate analysis of the abundance of the top six species contributing to differences between zones a) Ambassis jacksoniensis, b) Favonigobius lentiginosus, c) Gerres subfasciatus, d) Myxus elongatus, e) Sillago ciliata, and f) Sillago sp.

a) A. jacksoniensis b) F. lentiginosus c) G. subfasciatus d) M. elongatus e) S. ciliata f) Sillago sp.

p- p- p- p- p- p- Source dF MS F value MS F value MS F value MS F value MS F value MS F value

Dis 1 0.39 0.70 0.687 3.82 1.75 0.216 6.51 4.75 0.028 6.41 2.09 0.151 0.70 0.64 0.728 0.06 0.34 0.926 Zo 1 33.86 3.16 0.058 8.91 9.56 0.002 12.07 4.59 0.028 4.39 1.28 0.387 5.81 3.56 0.049 30.26 5.61 0.015 Ye 1 1.16 0.36 0.833 0.62 1.29 0.419 0.00 0.31 0.871 6.85 3.98 0.104 0.17 0.48 0.753 3.99 3.04 0.159 Ti 1 2.47 0.44 0.798 0.07 1.67 0.328 3.68 3.87 0.115 2.34 1.78 0.322 0.55 0.96 0.503 0.01 1.12 0.459 Es(Dis) 4 6.13 1.83 0.059 0.66 0.84 0.661 1.01 1.01 0.493 0.65 0.93 0.590 1.48 1.50 0.137 3.32 1.09 0.424 DisxZo 1 10.78 1.63 0.265 0.86 1.96 0.170 4.09 3.69 0.051 10.92 1.15 0.435 0.00 0.25 0.964 2.29 0.82 0.562 DisxYe 1 0.11 0.45 0.764 0.56 1.49 0.378 0.19 2.56 0.191 0.01 0.25 0.893 0.23 1.52 0.378 0.96 0.74 0.591 DisxTi 1 1.32 0.54 0.716 0.80 5.08 0.071 0.22 0.47 0.770 7.75 0.64 0.660 0.00 1.28 0.439 2.10 3.00 0.160 ZoxYe 1 3.96 0.74 0.601 0.16 0.29 0.894 0.03 0.21 0.923 1.23 1.75 0.307 0.17 0.50 0.740 2.95 7.33 0.039 ZoxTi 1 2.10 0.45 0.765 0.01 0.63 0.677 1.93 1.74 0.291 0.45 0.77 0.623 0.55 0.98 0.517 0.03 0.49 0.770 YexTi 1 5.85 4.54 0.093 0.34 0.45 0.544 0.29 2.39 0.197 1.34 0.29 0.603 0.66 2.07 0.228 0.07 0.16 0.715 ZoxEs(Dis) 4 6.05 2.04 0.036 0.69 0.56 0.959 0.80 0.96 0.528 2.55 1.08 0.406 0.83 1.36 0.195 2.02 0.88 0.631 Es(Dis)xYe 4 0.92 0.87 0.514 0.73 1.10 0.379 0.10 0.78 0.584 1.56 0.44 0.778 0.36 1.19 0.337 1.39 1.20 0.339 Es(Dis)xTi 4 2.71 0.76 0.562 0.16 0.90 0.496 0.69 1.70 0.172 2.61 0.58 0.685 0.25 0.88 0.522 0.34 1.49 0.250 DisxZoxYe 1 0.33 0.64 0.687 0.14 0.23 0.910 0.26 1.20 0.403 0.32 4.88 0.076 0.01 1.41 0.351 1.40 3.25 0.132 DisxZoxTi 1 1.62 0.85 0.560 0.27 0.84 0.541 0.03 0.71 0.628 6.56 4.74 0.079 0.38 7.43 0.033 1.25 1.14 0.454 DisxYexTi 1 2.17 1.68 0.259 0.15 0.20 0.667 0.02 0.18 0.682 16.83 3.60 0.132 0.00 0.00 0.951 0.51 1.14 0.340 ZoxYexTi 1 6.34 4.06 0.124 0.20 0.92 0.418 0.87 3.69 0.121 1.44 0.72 0.433 0.96 2.03 0.202 0.10 0.13 0.742 Si(Es(Dis)xZo) 24 2.12 0.62 0.964 1.00 2.22 0.012 0.92 1.34 0.142 4.39 1.47 0.120 0.64 1.96 0.017 2.65 1.66 0.047 ZoxEs(Dis)xYe 4 1.08 0.84 0.637 1.13 2.39 0.045 0.41 1.10 0.393 0.40 0.44 0.943 0.31 0.84 0.654 0.41 0.65 0.778 ZoxEs(Dis)xTi 4 1.87 0.59 0.852 0.17 2.05 0.076 0.37 0.99 0.481 1.73 0.75 0.716 0.08 0.46 0.926 1.50 2.02 0.100 Es(Dis)xYexTi 4 1.29 0.79 0.562 0.76 1.06 0.390 0.12 0.34 0.855 4.68 3.66 0.019 0.32 1.22 0.327 0.45 0.37 0.811 DisxZoxYexTi 1 1.89 1.21 0.340 0.41 1.89 0.243 0.00 0.00 0.975 0.07 0.04 0.854 0.04 0.08 0.784 0.25 0.34 0.609 YexSi(Es(Dis)xZo) 24 1.65 1.02 0.502 0.56 0.78 0.763 0.46 1.31 0.258 1.83 1.43 0.209 0.20 0.78 0.721 1.73 1.42 0.189 TixSi(Es(Dis)xZo) 24 4.40 2.71 0.013 0.21 0.30 0.998 0.49 1.41 0.168 2.03 1.59 0.123 0.26 0.99 0.515 0.59 0.49 0.952 ZoxEs(Dis)xYexTi 4 1.56 0.96 0.462 0.22 0.30 0.871 0.23 0.67 0.617 1.99 1.56 0.213 0.47 1.83 0.172 0.76 0.62 0.644 Res 24 1.62 0.72 0.35 1.28 0.26 1.22

Discussion within the estuarine system, irrespective of anthropogenic modification, substantial contamination levels, or individual High levels of anthropogenic modification and sediment variation between estuaries. contamination in the estuarine environment did not strongly influence the composition, abundance, or Shannon diversity of the Physico-chemical and Contamination Variables beach fish assemblages. We assessed these relationships within the The differences in physico-chemical variables documented in context of environmental variability both within and between this study are consistent with the general description and estuaries. Variation in environmental conditions between the inner understanding of environmental conditions in south-east Austra- and outer estuary zones were more strongly related to fish lian estuaries. It is well documented that the interplay between assemblages than high concentrations of contaminants. Inner fluvial and tidal forces in these systems creates consistent zones had greater Shannon diversity, species richness, abundance, differences in physico-chemical conditions within most estuaries and biomass of fish while the average weight of fish was slightly in the region [13]. The physico-chemical parameters that we higher in the outer zones. Community composition also differed measured do not encompass the full range of environmental between zones and some species were strongly associated with conditions that are expected to differ between the two zones. Some inner zones; this association was strongly correlated to pH, salinity, additional variables of interest that may covary with our physico- and temperature (but not metals or PAH contamination, turbidity, chemical measures include: wave exposure, flow rates and grain or estuary size). None of these measures of the beach fish size (expected to be higher in outer zones), phytoplankton assemblage differed between heavily modified and relatively productivity, predator/prey density, sedimentation rates, and unmodified estuaries. This indicates that differences in the fish coverage of submerged aquatic vegetation (expected to be higher assemblage largely follow variation in physico-chemical conditions in the inner zones) [41,42]. Experimental studies would be

PLoS ONE | www.plosone.org 10 October 2011 | Volume 6 | Issue 10 | e26353 Contamination Has Little Influence on Beach Fish

Figure 5. Mean (±SE) abundance by zone/estuary for 100 m2 beach seine samples. Plots of top six species contributing to differences between inner and outer zones. doi:10.1371/journal.pone.0026353.g005 required to determine the extent to which any or all of these estuaries [43]. It is well known that these contaminants are highly variables are the direct cause of the patterns we observed. dispersive and found in significant quantities even in otherwise Higher levels of sediment metals in the heavily modified pristine systems [43]. estuaries is consistent with the idea that urbanization, industrial development, run-off, and other sources of anthropogenic Relationships Between Covariates and the Fish modification increase the flow of contaminants into these estuaries Assemblage [22,30]. In addition, our findings for PAHs are consistent with All biological indicators of the beach fish assemblage (commu- previous studies in the region which have found comparable PAH nity composition, abundance, species richness, Shannon diversity contamination in both relatively unmodified and heavily modified and fish weight) displayed significant differences between inner

PLoS ONE | www.plosone.org 11 October 2011 | Volume 6 | Issue 10 | e26353 Contamination Has Little Influence on Beach Fish

Table 7. Results of DistLM covariate analysis for a) Physico-chemical covariates and sediment metals in the full model and d) Physico-chemical covariates, sediment metals, and sediment PAH under a reduced model (inner zone only).

a) DISTLM Covariate Analysis - Full Model b) DISTLM Covariate Analysis - Inner Zone

Variable SS F p-value Prop. SS F p-value Prop.

Temperature 17511.00 7.26 0.001 0.049 5936.60 2.59 0.008 0.036 Salinity 8929.80 3.61 0.004 0.025 4673.50 2.02 0.045 0.028 pH 20679.00 8.65 0.001 0.057 6975.00 3.07 0.003 0.042 Turbidity 2094.60 0.83 0.642 0.006 1543.80 0.66 0.761 0.009 Metals Quotient 3440.80 1.37 0.176 0.010 6283.10 2.75 0.006 0.038 Estuary Size 2504.60 0.99 0.426 0.007 4368.80 1.89 0.053 0.026 PAH Quotient NA NA NA NA 3498.80 1.50 0.125 0.021

doi:10.1371/journal.pone.0026353.t007 and outer zones. This is consistent with previous studies, which the inner zone data was analyzed separately, sediment metals did have generally found a strong relationship between fish commu- correlate significantly with the biological assemblage, however, this nities and physico-chemical variables such as salinity and turbidity correlation still accounted for a much smaller proportion of the [10]. While correlations between physico-chemical conditions and variance than the physico-chemical variables. In addition, no fish distributions are expected, it is surprising that no biological difference was found in community composition between the inner indicators were found to differ significantly by anthropogenic zones of the heavily modified and relatively unmodified estuaries. modification and that correlations between physico-chemical Thus, even where sediment metal contamination is severe, metals covariates were always stronger than correlations with contami- do not appear to relate to community level impacts in this post- nant covariates. While it is well known that correlative studies are settlement fish assemblage. limited in their ability to identify causal relationships, there The extent to which entire fish communities or populations may appears to be strong evidence to support the idea that variation in be affected when contaminants have negative consequences for the physico-chemical factors within the estuary are more closely individuals is poorly understood [6]. The physiological mecha- related to differences in the beach fish assemblage than nisms by which contaminants affect the health of individual fish contaminant concentrations. This is despite the fact that at many have been previously investigated and a great deal of literature sites both PAH and trace metals concentrations were found to be examines the presence, biomagnification, toxicology, and bio- higher than sediment quality guideline values, above which the marker response of contaminants in marine fishes [44,45]. Fish contaminants are expected to have biological effects [32]. When primarily take up contaminants through ingestion of contaminated

Figure 6. Principal Coordinated Ordination (PCO) of correlations between covariate factors and two dimensional plots of community composition by zone. a) Overlaid with physico-chemical and sediment metal vectors. b) Overlaid with vectors of top six species contributing to differences between zones. (Multiple Correlation .0.2). doi:10.1371/journal.pone.0026353.g006

PLoS ONE | www.plosone.org 12 October 2011 | Volume 6 | Issue 10 | e26353 Contamination Has Little Influence on Beach Fish

Figure 7. Principal Coordinated Ordination (PCO) of correlations between covariate factors and two dimensional plots of community composition by disturbance category. Data from inner zone sites only. a) Physico-chemical, sediment metals, and sediment PAH covariates. b) Plots of six highest correlating species (Multiple Correlation .0.2). Community composition does not differ significantly by disturbance category but is presented for graphical purposes (p = 0.437). doi:10.1371/journal.pone.0026353.g007 food particles and to a lesser extent from water that passes over the contaminants may affect fish populations and diversity by reducing gill membranes [28]. Once ingested, contaminants move through fish health and survivorship [46], by reducing growth and a wide variety of physiological and chemical pathways, many of reproductive success [47], by reducing the abundance of prey which have detrimental effects for the individual. However, the species, and by increasing instances of deformity [8]. Ultimately extent to which these organismal effects translate into community any of these mechanisms could link contaminant exposure to or population level impacts is rarely studied [6]. In theory, organismal effects and ultimately population level impacts. Response mechanisms may be species specific and a lack of knowledge in this area somewhat hinders our ability to detect and Table 8. Multivariate analysis of community composition understand community level impacts of contaminants. under a reduced model (inner zone only). Ecological Characteristics and Life History Stages Community Composition - Inner Zone The findings of this study contrast directly with our findings for larval fish communities in these same estuaries. In previous studies Source dF MS F p-value we have shown that early life-history stages (larval fish) varied Dis 1 8768.40 1.70 *0.154 substantially between heavily modified and relatively unmodified Ye 1 3752.10 2.78 0.093 estuaries, while showing a strong relationship to sediment metals Ti 1 2152.60 1.56 0.236 [4]. We have also demonstrated that in these estuaries anthropogenic stressors appear to primarily affect estuarine taxa and Es(Dis) 4 5168.70 1.46 0.074 benthic egg layers [9]. In the current study estuarine opportunist DisxYe 1 2131.40 1.58 0.237 species accounted for the majority of the dataset and were the DisxTi 1 3409.80 2.46 0.098 major driver of differences in fish assemblages between the inner YexTi 1 4165.40 1.93 0.182 and outer zone of estuaries. However, none of the estuary usage Si(Es(Dis)) 12 3546.90 2.40 0.001 guilds differed significantly by disturbance category, and there was YexEs(Dis) 4 1350.90 0.87 0.637 no strong evidence to suggest differential sensitivity among estuarine taxa. TixEs(Dis) 4 1383.40 0.79 0.683 The disjuncture between the results for the larval and beach fish DisxYexTi 1 1624.80 0.75 0.522 assemblages may be due to a variety of factors. First, it should be YexSi(Es(Dis)) 12 1549.50 1.05 0.397 noted that the larval fish assemblage sampled in previous studies TixSi(Es(Dis)) 12 1760.60 1.19 0.251 were more diverse than this beach fish assemblage, and that many YexTixEs(Dis) 4 2160.10 1.46 0.101 of the larval species are closely associated with biogenic habitats Res 12 1476.40 such as seagrass and mangroves (unlike the beach fish). This may explain the more pronounced impacts observed for larvae [4]. Factors: Dis = Disturbance Category (Heavily Modified vs. Relatively Second, because the majority of beach fish species are estuarine Unmodified), Ti = Time of Sampling, Ye = Year, Es = Estuary, Si = Site. Bold values opportunists, many of them are not captured in large numbers by correspond to significant values for higher-level factors or interactions between non-random factors. larval sampling in estuaries. The larval studies indicated that the *Indicates Monte Carlo p value. majority of anthropogenic impacts are seen among estuarine doi:10.1371/journal.pone.0026353.t008 resident taxa, a group which is underrepresented in this study [9].

PLoS ONE | www.plosone.org 13 October 2011 | Volume 6 | Issue 10 | e26353 Contamination Has Little Influence on Beach Fish

The only major species encountered in large numbers in both estuarine beach fish do not represent a sensitive indicator of the beach seine and larvae studies were S. ciliata, S. maculata, G. contaminant impact. Other forms of environmental disturbance subfasciatus, H. vittatus, F. lentiginosus, and A. jacksoniensis. Of these, A. which have a large scale influence on physico-chemical conditions, jacksoniensis was the only species which differed significantly by such as diversion of freshwater flows, coastal alteration, and sea modification in both studies, being more abundant in the relatively level rise, may have greater potential to affect the beach fish unmodified estuaries. However, the scale of difference was not assemblage than the current input of marine contamination. equal; whereas post-settlement A. jacksoniensis were ,1.5 times Ultimately, conservation and management efforts which include a more abundant in the relatively unmodified estuaries, larvae were consideration of physico-chemical variables will be more effective ,14 times more abundant. This indicates that there is a greater at protecting these fish assemblages. difference in the abundance of this species between heavily modified and relatively modified estuaries at the larval stage. This Supporting Information is consistent with the hypothesis that fish are more sensitive at their larval stage and that impacts are more easily detected on larval Appendix S1 Average beach fish abundance identified to assemblages. However, it also suggests that the relationship lowest taxonomic level by estuary and zone. Heavily between larval and post-settlement abundance is not straightfor- modified estuaries – Port Jackson, Botany Bay and Port Kembla. ward, and hence that impacts at the larval stage may not directly Relatively unmodified estuaries – Port Hacking, Jervis Bay and the translate into impacts at the post-settlement stage. The non-linear Clyde River. Abbreviations - Life Cycle Guild: EO = Estuarine relationship between larval and post-settlement abundance has Opportunist, E = Estuarine, MS = Marine Straggler. Trophic level been well documented in the supply-side ecology literature [48]. values taken from [33]. (DOC) Conclusion The importance of chemical contamination as an ecosystem Acknowledgments stressor will depend on site attributes and variability in environmental conditions [20]. Our study suggests that variation We would like to thank Cian Foster-Thorpe, Shinjiro Ushiama, David Day, Valeriya Komyakova, and Anthousa Harris for their assistance in the in physico-chemical factors has a much greater influence on the field. We would also like to thank the Bluescope Steel Company for their beach fish assemblage than the extent of anthropogenic modifi- generous support of this research and Marine Parks NSW for their help cation or pollution in the estuaries examined in this study. and advice. This study was approved and carried out in strict accordance Significant differences in physico-chemical conditions exist within with the recommendations of the Animal Care and Ethics Committee of these estuaries and these factors are highly associated with the the University of New South Wales (Project No. 09/110A) and the New distributions of some fish species and the composition of the fish South Wales Department of Primary Industries (Permit No. P09/0072- community generally. Despite comparatively high levels of 1.0). anthropogenic modification and contamination, there did not appear to be a large effect on the portion of the fish assemblage Author Contributions examined in this study. However, fish living in more sensitive Conceived and designed the experiments: ACM KAD MDT ELJ. biogenic habitats such as coral reefs and seagrass beds may be Performed the experiments: ACM KAD. Analyzed the data: ACM affected to a much greater degree by modification and KAD. Contributed reagents/materials/analysis tools: MDT ELJ. Wrote contamination, as the biogenic habitat itself may be degraded by the paper: ACM. Revisions to manuscript: KAD MDT ELJ. these stressors. Unlike benthic larval fish assemblages, the

References 1. Lotze HK, Lenihan HS, Bourque BJ, Bradbury RH, Cooke RG, et al. (2006) 11. Rakocinski CF, Baltz DM, Fleeger JW (1992) Correspondence between Depletion, degradation, and recovery potential of estuaries and coastal seas. environmental gradients and the community structure of marsh-edge fishes in Science 312: 1806–1809. a Louisiana estuary. Marine Ecology Progress Series 80: 135–148. 2. Kennish MJ (2002) Environmental threats and environmental future of estuaries. 12. Taylor MD, Laffan SD, Fielder DS, Suthers IM (2006) Key habitat and home Environmental Conservation 29: 78–107. range of mulloway Argyrosomus japonicus in a south-east Australian estuary: 3. Beck MW, Heck KL, Able KW, Childers DL, Eggleston DB, et al. (2001) The finding the estuarine niche to optimise stocking. Marine Ecology Progress Series identification, conservation, and management of estuarine and marine nurseries 328: 237–247. for fish and invertebrates. Bio Science 51: 633–641. 13. Roy P, Williams R (2001) Structure and function of south-east Australian 4. McKinley AC, Miskiewicz A, Taylor MD, Johnston EL (2011) Strong links estuaries. Estuarine, Coastal and Shelf Science 53: 351–384. between metal contamination, habitat modification and estuarine larval fish 14. Masselink G, Short AD (1993) The effect of tide range on beach distributions. Environmental Pollution 159: 1499–1509. morphodynamics and morphology: A conceptual beach model. Journal of 5. Johnston E, Roberts DA (2009) Contaminants reduce the richness and evenness Coastal Research 9: 785–800. of marine communities: A review and meta-analysis. Environmental Pollution 15. Castilla JC (1983) Environmental impact in sandy beaches of copper mine tailings at Chan˜aral, Chile. Marine Pollution Bulletin 14: 459–464. 157: 1745–1752. 16. Rice C (2006) Effects of shoreline modification on a Northern Puget Sound 6. McKinley A, Johnston EL (2010) Impacts of contaminant sources on marine fish beach: Microclimate and embryo mortality in surt smelt (Hypomesus pretiosus). abundance and species richness: A review and meta-analysis of evidence from Estuaries and Coasts 29: 63–71. the field. Marine Ecology Progress Series 420: 175–191. 17. Edgar GJ, Shaw C (1995) The production and tropic ecology of shallow-water 7. Miskiewicz AG, Gibbs PJ (1994) Organochlorine pesticides and hexachlor- fish assemblages in southern Australia. III. General relationships between obenzene in tissues of fish and invertebrates caught near a sewage outfall. sediments, seagrasses, invertebrates and fishes. Journal of Experimental Marine Environmental Pollution 84: 269–277. Biology and Ecology 194: 107–131. 8. Kingsford MJ, Suthers IM, Gray CA (1997) Exposure to sewage plumes and 18. Deegan L (2002) Lessons learned: The effects of nutrient enrichment on the the incidence of deformities in larval fishes. Marine Pollution Bulletin 33: support of nekton by seagrass and salt marsh ecosystems. Estuaries and Coasts 201–212. 25: 727–742. 9. McKinley AC, Foster-Thorpe C, Miskiewicz A, Taylor MD, Johnston EL (in 19. Reopanichkul P, Schlacher TA, Carter RW, Worachananant S (2009) Sewage review) Anthropogenic activities differentially impact fish guilds: The importance impacts coral reefs at multiple levels of ecological organization. Marine Pollution of understanding life history characteristics. Journal of Applied Ecology. Bulletin 58: 1356–1362. 10. Potter IC, Hyndes GA (1999) Characteristics of the ichthyofaunas of 20. Burton GA, Johnston EL (2010) Assessing contaminated sediments in the context southwestern Australian estuaries, including comparisons with holarctic estuaries of multiple stressors. Environmental Toxicology and Chemistry 29: 2625–2643. and estuaries elsewhere in temperate Australia: A review. Australian Journal of 21. Underwood AJ (1994) On beyond BACI: Sampling designs that might reliably Ecology 24: 395–421. detect environmental disturbances. Ecological Applications 4: 3–15.

PLoS ONE | www.plosone.org 14 October 2011 | Volume 6 | Issue 10 | e26353 Contamination Has Little Influence on Beach Fish

22. Scanes P (2010) NSW Estuarine catchment disturbance ranks. Sydney: NSW 34. Pauly D, Watson R (2005) Background and interpretation of the ‘Marine Department of Environment, Climate Change, and Water. Trophic Index’ as a measure of biodiversity. Philosophical Transactions of the 23. NSWDNR (2010) Estuaries in New South Wales. New South Wales Department Royal Society BIology 360. of Natural Resources, Sydney, Available from ,http://www.naturalresources. 35. Elliott M, Whitfield AK, Potter IC, Blaber SJM, Cyrus DP, et al. (2007) The nsw.gov.au/estuaries/inventory/index_ns.shtml. (Accessed November 14, guild approach to categorizing estuarine fish assemblages: a global review. Fish 2010). and Fisheries 8: 241–268. 24. ANRA (2009) Land use - Clyde River - Jervis basin. Australian Natural 36. Neira FJ, Miskiewicz AG, Trnski T (1998) Larvae of temperate Australian fishes: Resources Atlas, Canberra, Available from http://www.anra.gov.au/topics/ A laboratory guide for larval fish identification. Perth: University of Western land/landuse/nsw/basin-clyde-river.html (Accessed May 14, 2011). Australia Press. 474 p. 25. DSEWPC (2011) Interim biogeographic regionalization of Australia. Depart- 37. Anderson MJ (2001) A new method for non-parametric multivariate analysis of ment of Sustainability, Environment, Water, Population and Communities. variance. Austral Ecology 26: 32–46. Canberra, Available from http://www.environment.gov.au/parks/nrs/science/ 38. McArdle BH, Anderson MJ (2001) Fitting multivariate models to community bioregion-framework/ibra/index.html (Accessed July 14, 2011). data: a comment on distance-based redundancy analysis Ecology 82: 290–297. 26. Gomon M, Bray D, Kuiter R (2008) Fishes of Australia’s southern coast. Sydney: 39. Anderson MJ (2003) PCO - Principal Coordinate Analysis: A computer Reed New Holland. program. Aukland: University of Aukland. 27. USEPA (2007) Method 3051A microwave assisted acid digestion of sediments, 40. Quinn G, Keough M (2002) Experimental design and data analysis for sludges and oils. Environmental Protection Agency. Washington: Environmental Protection Agency. biologists. Cambridge: Cambridge University Press. 28. Dallinger R, Prosi F, Segner H, Back H (1987) Contaminated food and uptake of 41. Iverson RL (1990) Control of marine fish production. Limnology and heavy metals by fish: a review and a proposal for further research. Oecologia 73: Oceanography 35: 1593–1604. 91–98. 42. Clark BM (1997) Variation in surf-zone fish community structure across a wave- 29. Hall BD, Bodaly RA, Fudge RJP, Rudd JWM, Rosenberg DM (1997) Food as exposure gradient. Estuarine, Coastal and Shelf Science 44: 659–674. the Dominant Pathway of Methylmercury Uptake by Fish. Water, Air, & Soil 43. Maher WA, Aislabie J (1992) Polycyclic aromatic hydrocarbons in nearshore Pollution 100: 13–24. marine sediments of Australia. Science of the Total Environment 112: 143–164. 30. Knott NA, Aulbury J, Brown T, Johnston EL (2009) Contemporary ecological 44. Costello, Mark J, Read, Paul (1994) Toxicity of sewage sludge to marine threats from historical pollution sources: impacts of large-scale resuspension of organisms: A review. Marine Environmental Research 37: 23–46. contaminated sediments on sessile invertebrate recruitment. Journal of Applied 45. van der Oost R, Beyer J, Vermeulen NPE (2003) Fish bioaccumulation and Ecology 46: 770–781. biomarkers in environmental risk assessment: a review. Environmental 31. Long ER (2006) Calculation and uses of mean sediment quality guideline Toxicology and Pharmacology 13: 57–149. quotients: A critical review. Environmental Science & Technology 40: 46. Robinet TT, Feunteun EE (2002) Sublethal effects of exposure to chemical 1726–1736. compounds: A cause for the decline in Atlantic eels? Ecotoxicology 11: 265–277. 32. ANZECC (2000) Australian and New Zealand guidelines for fresh and marine 47. Waring CP, Stagg RM, Fretwell K, McLay HA, Costello MJ (1996) The impact water quality. Canberra: National Water Quality Management Strategy, of sewage sludge exposure on the reproduction of the sand goby, Pomatoschistus Australian and New Zealand Environment and Conservation Council and minutus. Environmental Pollution 93: 17–25. Agriculture and Resource Management Council of Australia and New Zealand. 48. Roughgarden J, Gaines S, Possingham H (1988) Recruitment dynamics in 33. Froese R, Pauly D (2010) Fishbase. Available from www.fishbase.org (accessed complex life cycles. Science 241: 1460–1466. January 25, 2010).

PLoS ONE | www.plosone.org 15 October 2011 | Volume 6 | Issue 10 | e26353

Chapter 3

PUTTING MARINE SANCTUARIES INTO CONTEXT: A COMPARISON OF ESTUARY FISH ASSEMBLAGES OVER MULTIPLE LEVELS OF PROTECTION AND DISTURBANCE

Final Version:

McKinley, A.C., Ryan, L., Coleman, M.A, Knott, N.A, Clarke, G., Taylor, M.D. & E.L. Johnston. (2011) Putting marine sanctuaries into context: A comparison of estuary fish assemblages over multiple levels of protection and disturbance. Aquatic Conservation, DOI: 10.1002/aqc.1223. AQUATIC CONSERVATION: MARINE AND FRESHWATER ECOSYSTEMS Aquatic Conserv: Mar. Freshw. Ecosyst. (2011) Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/aqc.1223

Putting marine sanctuaries into context: a comparison of estuary ﬁsh assemblages over multiple levels of protection and modiﬁcation

ANDREW C. MCKINLEYa,*, LAURA RYANa, MELINDA A. COLEMANb, NATHAN A. KNOTTc, GRAEME CLARKa, MATTHEW D. TAYLORa and EMMA L. JOHNSTONa aEvolution & Ecology Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, New South Wales 2052, Australia bNSW Marine Park Authority, Batemans Bay Marine Park, Burrawang St., Narooma, NSW 2546, Australia cNSW Marine Park Authority, Jervis Bay Marine Park, 4 Woollamia Rd. Huskisson, NSW 2450, Australia

ABSTRACT 1. In recent decades there has been a significant effort to establish marine sanctuaries for the purpose of protecting marine biodiversity and ecological processes. While many studies have demonstrated that marine sanctuaries increase the abundance, diversity, and trophic level of marine fish communities, few have compared these parameters across multiple levels of protection and human modification. 2. This study utilized baited remote underwater video to compare fish assemblages between marine parks, between different levels of protection within parks (sanctuary and habitat protection zones), and between parks and highly modified systems with similar ecological communities. 3. It was demonstrated that sanctuary zones have higher abundance of targeted fish species compared with other areas within some marine parks. 4. The total abundance of targeted species and abundances of some key fisheries species (e.g. pink snapper) were found to be higher in sanctuary zones. This suggests that increased protection may be effective at improving these aspects of the fish assemblage. 5. However, when marine parks were compared with highly modified environments it was found that targeted species were much more abundant in the highly modified systems. 6. Community composition of entire fish assemblages also differed between these levels of modification and economically important fisheries species contributed most to this difference. 7. These findings suggest that while highly protected sanctuary zones may increase the abundance of targeted fish compared with less protected areas within the same estuary, highly industrialized or urbanized systems, not typically chosen as marine parks, may actually support more targeted species of fish. 8. It was demonstrated that forms of modification in addition to fishing pressure are having large effects on fish assemblages and productivity. Copyright # 2011 John Wiley & Sons, Ltd.

Received 31 March 2011; Revised 28 July 2011; Accepted 14 August 2011

KEY WORDS: estuary; marine protected area; marine park; marine reserve; fish; fishing; urban development; modification; baited remote underwater videos; marine sanctuary

INTRODUCTION precipitated large and signiﬁcant changes to the distribution, abundance, and diversity of marine organisms. Marine Human stressors including pollution (Costello and Read, 1994; sanctuaries aim to reduce human stressors by providing areas Johnston and Roberts, 2009; McKinley and Johnston, 2010; set aside for the protection and maintenance of biological McKinley et al., 2011), overharvesting (Pauly et al., 2002), diversity and ecological processes (Kelleher et al., 1995; habitat modiﬁcation (Dafforn et al., 2009b; Kaiser et al., Claudet et al., 2006; MPA, 2010). A substantial number of 2002) and introduced species (Mack et al., 2000) have studies have established that marine sanctuaries are successful

*Correspondence to: A. C. McKinley, Evolution & Ecology Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, New South Wales 2052, Australia. E-mail: [email protected]

Copyright # 2011 John Wiley & Sons, Ltd. A. C. MCKINLEY ET AL. at achieving a variety of conservation goals (Halpern, 2003; modification and catchment alteration, and reduce susceptibility Pérez-Ruzafa et al., 2006; Harmelin-Vivien et al., 2008; Edgar to colonization by invasive organisms (Stachowicz et al., 1999; et al., 2009). Studies examining the success of sanctuaries Marchetti et al., 2006) compared with unprotected areas. In have assessed impacts primarily over small spatial and effect, these conservation areas can generally be characterized temporal scales and rarely have results been compared with as ‘less anthropogenically modified’ environments compared modified reference locations. Conversely, few studies have with other managed ecosystems. Some of these other examined the impacts of human stressors on fish assemblages modification factors may act synergistically with fishing by systematically comparing heavily modified environments pressure to alter fish assemblages (Micheli, 1999; Islam and with relatively unmodified systems (Agardy et al., 2003). Tanaka, 2004; Breitburg et al., 2009). Because marine sanctuaries are often located in relatively Currently, literature addressing the impacts of marine pristine environments and are protected from many forms of sanctuaries leaves a variety of unanswered questions. Do modification by legislation, monitoring, and enforcement, legislation and management practices limiting processes of they provide excellent reference locations against which the human modification (other than fishing) affect the fish effects of modification can be compared (McKinley and assemblage in marine parks? Do some forms of modification Johnston, 2010). The comparison of marine sanctuaries with change fish assemblages? Should we expect marine sanctuaries highly modified environments is of interest both to evaluate to support higher abundances of targeted species simply the success of sanctuary zones and the impacts of a variety of because they are less harvested, or will other modification stressors in modified areas. The success of marine sanctuaries factors have a strong influence? Hypothetically, what would in increasing the density (Edgar and Stuart-Smith, 2009), happen if we had sanctuary zones in areas typically not biomass (Williamson et al., 2004), and diversity (Lester chosen for protection such as highly modified or urbanized et al., 2009) of marine fish has been well documented (Rowley, environments? 1994). These sanctuary areas have been shown to increase the This is one of the first studies to address some of these average trophic level of the ecosystem and the abundance of questions by simultaneously comparing fish assemblages harvested species (Evans and Russ, 2004). A comprehensive across several levels of protection and modification. meta-analysis of marine parks worldwide found that 90% of Within relatively unmodified marine parks, the impacts of parks increased biomass of target species by an average of sanctuary zones are examined by comparing these 250%, while the average abundances of organisms doubled with habitat protection zones (where recreational and and the mean size of organisms increased by a third (Halpern, commercial fishing is permitted). This gives an idea of the 2003). response of fish assemblages to protection (from both The creation of sanctuary zones, their design, and biological fishing and other disturbances) over small spatial scales. as well as socio-economic effects has been the primary focus of Second, the effects of large scale human modi fication are much scientific debate. Most studies examining the effects of assessed by comparing highly modified environments in marine parks observed conservation benefits with effects heavily developed estuaries with relatively less disturbed attributed primarily to the creation of marine sanctuaries that environments of marine parks. We predict that sanctuary limit or entirely eliminate fishing pressure as a source of zones and marine parks will have higher abundances of modification (Halpern, 2003). Most of these studies evaluated fish than fished areas and modified estuaries. Further, we the effects of sanctuaries by comparing nearby environments predict that there will be differences in fish community (where fishing is permitted) with sanctuary zones. In many composition among these different places. cases these studies are conducted within the same coastal or estuarine system, such that the sanctuary is compared directly with a nearby commercial or recreational fishing zone within the same study area (Halpern, 2003; Lester et al., 2009) and METHODS presumably similar levels of modification. Many of these studies lack reference to external sites and conditions. As a Study location and modification categories result, the way that removal of fishing pressure compares with other forms of modification, and the effects on fish This study was conducted in four estuaries along the south-eastern assemblages resulting from changes to human disturbance coast of New South Wales, Australia; two modified estuaries regimes have been less well studied. (Port Jackson 3344.258′S, 15116.542′E and Port Hacking Several studies have demonstrated that the continuation of 3404.680′S, 15109.311′E) and two relatively unmodified recreational fishing within marine protected areas reduces the estuaries within marine parks (Jervis Bay 3504.762′S, effectiveness of conservation initiatives and that recreational 15044.858′E and Batemans Bay 3544.233′S, 15014.272′E). fishing has a major impact on fish assemblages (Jennings et al., Both Batemans Bay and Jervis Bay are considered to be 1996; Denny and Babcock, 2004; Samoilys et al., 2007). While within relatively pristine ecosystems with less foreshore fishing pressure is a conspicuous pathway for human impacts modification, artificial structures, boating traffic, pollution, on fish assemblages, a variety of other forms of modification urbanization of the catchment and lower nutrient loads have been demonstrated to have impacts on marine fish compared with the modified estuaries (Scanes, 2010). In part, assemblages (Kaiser et al., 2002; Bax et al., 2003; McKinley it is because these estuaries are relatively pristine that they and Johnston, 2010; McKinley et al., 2011). In addition to were designated as marine parks. Sanctuary zones within reducing fishing pressure, the establishment of marine parks and these marine parks have been in place for 8 and 4 years for sanctuaries can limit human modification via legislation and Jervis Bay and Batemans Bay, respectively (MPA, 2010). For placement. Legislation governing these areas is generally these reasons, the marine parks are considered here as designed to limit contamination loading, reduce habitat ‘relatively unmodified estuaries’.

Copyright # 2011 John Wiley & Sons, Ltd. Aquatic Conserv: Mar. Freshw. Ecosyst. (2011) PUTTING MARINE SANCTUARIES INTO CONTEXT

Port Jackson is the heavily urbanized port of the city of fishing occur), and the modified estuaries represent an area of Sydney. It is classified as highly modified due to intense high fishing activity (where some forms of commercial fishing commercial and recreational boat traffic, historic and ongoing occur and where the greatest concentrations of recreational pollution, and widespread urbanization of its shoreline and fishers are found). catchment (Birch and Taylor, 1999; McKinley et al., 2011) Port Hacking is located between the heavily urbanized Sampling design suburbs of southern Sydney and the forested slopes of Royal National Park, which touches the southern border of the In each of the four estuaries, four sites were sampled (Figure 1). estuary. Port Hacking is also subject to heavy recreational In modified estuaries no sites were protected by any special boat traffic and frequent dredging. Monitoring programmes conservation designation and all sites were located in the rank Port Jackson as heavily nutrient enriched while Port outer zone of the estuary in locations that were directly Hacking is less so, but both have higher nutrient loads, comparable with marine parks (Figure 1). In contrast, within catchment modification, and urban run-off than the marine each marine park two sites were placed within sanctuary areas parks (Birch et al., 2010; Scanes, 2010). These monitoring and two within habitat protection zones. In Batemans Bay the programmes use verified land-use and pollutant run-off data sanctuary sites were within the North Head and Tollgate to model catchment pollution levels and to predict total Islands sanctuary zones (Figure 1(d)) and in Jervis Bay they nitrogen, phosphorous, and sediment loading (Scanes, 2010). were located with the Hyams Beach and Huskisson sanctuary For these reasons Port Jackson and Port Hacking are zones (Figure 1(c)). All sites were located at least 500 m away classified as ‘modified estuaries’. from the edge of adjacent zones to avoid edge effects. The sanctuary sites were compared with nearby sites within the Zone definitions and fishing pressure categories habitat protection zones at Judges Beach and Lillipilli Point in Batemans Bay Marine Park and Plantation Point and For the purposes of this paper, the term ‘sanctuary zone’ refers Callala Beach in Jervis Bay Marine Park. to areas where recreational and commercial fishing is All fish sampling was conducted using baited remote completely prohibited. The term ‘sanctuary’ or ‘sanctuary underwater video stations (BRUVS), which were assembled in zone’ has a similar meaning to the terms ‘no-take area’ or a standard configuration (Cappo et al., 2004). Non-destructive ‘reserve’ used in other studies. This contrasts with ‘habitat methods such as BRUVS are preferable in marine sanctuaries protection’ zones within marine parks where recreational and where fish are protected. A single fixed camera was suspended some forms of commercial fishing are allowed. In the marine on a quadrapod approximately 15 cm above the benthos. The parks examined in this paper, most commercial fishing is camera was horizontally oriented in the direction of a baited completely banned with the exception of some commercial bag containing 500 g of crushed pilchards Sardinops sagax and bait fish collection using purse seine nets, and commercial extended 1 m from the base of the camera quadrapod. In each beach seining within both parks. The commercial purse seine estuary four sites were selected close to the mouth of the and beach seine activities are limited to the habitat protection estuary. All sites were over bare sediment 5 m to 10 m from zones and are prohibited within the sanctuary areas (NSW, rocky reef, and in waters between 5 m and 12 m deep. This 1999). Habitat protection zones within marine parks are gave a clear field of view for the cameras and ensured therefore comparable with many other estuaries in terms of consistency in the type of habitat sampled (e.g. bare sediment overall fishing pressure. ‘Marine park’ is used to describe adjacent to rocky reef). Each site was randomly sampled designated conservation areas that are zoned to encompass a twice, from November 2009 to March 2010, with four mixture of multiple sanctuary and habitat protection zones, as replicate BRUVS deployed at each time and site. As sampling is the case with the marine parks examined in this study. was randomized, tidal phase was not taken into account. Both Port Jackson and Port Hacking are subject to intense Temperature, salinity, and pH were sampled at all sites using a recreational fishing pressure and are among the most calibrated YSI 6820 V2 sonde. Visibility was also evaluated recreationally fished estuaries in Australia (Henry and Lyle, during image analysis. These parameters were similar among 2003; DPI, 2010). While exact fishing effort data were not times of sampling across the four estuaries (ANOVA, available at the time of this study, surveys indicated that P > 0.05). Moreover, preliminary analyses revealed that there approximately 10 times as many individuals practise were never any differences in fish abundances or diversity recreational fishing in the Port Jackson/Port Hacking region between these sampling times so ‘time’ was pooled for compared with Batemans Bay and Jervis Bay (DPI, 2010). analyses, giving n = 8 BRUVS drops per site. All BRUVS Most commercial fishing is banned from Port Jackson deployments were spaced at least 200 m apart and each and Port Hacking, although some commercial bait fishing is recording lasted for approximately 35 min. permitted. Although some forms of recreational fishing It should be noted that in NSW, modified estuaries are (netting and trapping) are also banned in some parts of Port usually adjacent to major cities such as Sydney. The location Hacking, the selected study sites fell outside of these areas. of estuaries is therefore necessarily spatially pseudoreplicated ‘Fishing pressure’ in this study is meant to refer to the level along the coastline with modified estuaries around the of harvesting that occurs in these environments. This includes urbanized shores of Sydney and marine parks in the less both recreational and commercial fishing. In this study it is populated south (there are no marine parks in the Sydney argued that sanctuary zones within the marine parks area). Nevertheless, with proper replication of estuaries within constitute an area of low fishing activity (because most forms these modification regimes (see sampling design above) and of fishing are prohibited), habitat protection zones within consideration of species with cosmopolitan distributions along marine parks constitute an area of moderate fishing activity this coastline, this issue can be largely overcome. It should (where recreational fishing and some forms of commercial also be noted that the marine parks and modified estuaries

Port Jackson Port Hacking -34.040 -33.780 a b -34.050 -33.800 -34.060 -33.820 -34.070 -33.840 -34.080 -33.860 -34.090 -33.880 Port Jackson -34.100 -33.900 0 km 5km 0 km 3km Port Hacking -34.110 151.140 151.160 151.180 151.200 151.090 151.110 151.130 151.150 151.170

Jervis Bay Batemans Bay -34.980

c -35.690 d Jervis Bay -35.000 -35.700

-35.020 -35.710

Batemans Bay -35.720 -35.040 -35.730

0 km 100km -35.060 -35.740

-35.750 -35.080 -35.760 -35.100 0 km 5km -35.770 -35.780 -35.120 0 km 3km -35.790

-35.140 150.650 150.700 150.750 150.800 150.850 150.160 150.180 150.200 150.220 150.240 150.260 150.280

Figure 1. Locations of study sites in (a) Port Jackson, (b) Port Hacking, (c) Jervis Bay Marine Park, and (d) Batemans Bay Marine Park. Filled triangles (▲) indicate sampling sites within modified estuaries. For the relatively unmodified estuaries (marine parks), filled diamonds (♦) indicate sampling sites within habitat protection zones, filled circles (●) indicate sampling sites within sanctuary zones. occur in different bioregions according to the Interim maximum number of fish from each species present at any Biogeographic Regionalization of Australia (IBRA) system, one time in the field of view were counted and summarized in though the maximum distance between estuaries is only a relative measure of abundance - Max N. All individuals 275 km (Batemans Bay to Port Jackson) (DSEWPC, 2011). which could be clearly seen and identified to species were While this indicates that some differences exist in the counted. Individuals which were too far from the camera to biological and environmental conditions between these areas, be identified or were otherwise not clearly visible were most of the fish species examined in this study are known to ignored. All individuals were identified to species with the occur in all the estuaries examined in this study. Notably, the exception of the Platycephalus genus (Flatheads). These fish four estuaries examined in this study are at least several were grouped as ‘Platycephalus spp.’ due to the difficult hundred kilometres within the known range of the major nature of species identification without close examination of a species which drive the trends in this analysis (e.g. pink specimen. Each species was also identified according to snapper, silver trevally and yellowfin bream) (Edgar and whether it is ‘targeted’ by commercial or recreational fishing. Shaw, 1995; Gomon et al., 2008). In addition, no differences Any species which was listed as a major game species by the were found in physico-chemical variables between these Department of Primary Industries, NSW was classified as a estuaries at the time of sampling (see above) and the habitat targeted species (DPI, 2010). This source was used to classify sampled was judged to be reasonably similar in all estuaries the majority of species in this study. Species not listed by DPI (bare sediment near rocky reef). For these reasons, it is were classified using the international fisheries database believed that comparisons between these estuaries are valid Fishbase, which classifies species as commercially or for this analysis, despite the existence of some differences in recreationally targeted ‘game fish’ based on international biological and environmental conditions. Other studies have fisheries monitoring data and the scientific literature (Froese utilized similar comparisons between these estuaries (Dafforn and Pauly, 2010). These classifications are summarized in et al., 2009a; McKinley et al., 2011). Appendix 1.

Video image analysis Statistical analysis For all BRUVS footage the ﬁrst minute of tape was disregarded All multivariate and univariate analyses were conducted using to allow time for the BRUVS to settle and disturbed sediment mixed model PERMANOVA in PRIMER v.6 (Anderson, to clear. Following this the next 30 min of tape were analysed. 2001). Before analysis, data were fourth root transformed To avoid repeatedly counting the same individuals, the and Bray–Curtis similarity matrices were calculated for

Copyright # 2011 John Wiley & Sons, Ltd. Aquatic Conserv: Mar. Freshw. Ecosyst. (2011) PUTTING MARINE SANCTUARIES INTO CONTEXT multivariate analyses. SIMPER analysis was used to determine in the marine parks) and 19 species occurred only in one the contribution of fish species to the average dissimilarity modified estuary. The majority of these species were rare between significant factors in multivariate analyses (Clarke, within the study and encountered in less than one or two 1993). The four highly abundant species yellowtail scad samples. (Trachurus novaezelandiae), mado (Atypichthys strigatus), silver sweep (Scorpis lineolata) and ocean leatherjacket Effects of protection (Nelusetta ayraud) were excluded from the multivariate The abundance of targeted fish was greater in sanctuary zones analysis of community composition. This was done because relative to habitat protection zones but there were no these predominately schooling species of fish occurred in differences in either the average Max N of all fish or the extremely high abundances in each estuary and across almost number of fish species (Table 1, Figure 2). When individual all samples and so obscured differences in assemblage species were analysed separately, only pink snapper (Pagrus composition within the remainder of the fish assemblage. auratus) showed significant differences in abundance, with Univariate tests of these species indicated that only silver more in sanctuary zones than in habitat protection zones sweep differed significantly by modification, with increased (Table 2, Figure 3). There was, however, a trend for more abundance in the heavily modified estuaries (P = 0.028), silver trevally (Pseudocaranx georgianus) in sanctuary relative though it was still present in large numbers in the relatively to habitat protection zones (Figure 3) but variation at the site unmodified estuaries. Yellowtail scad, mado and ocean level probably obscured this pattern (Table 2). There was no leatherjacket did not differ significantly by modification and difference in assemblage composition between zones (Table 3, were abundant in all estuaries (P = 0.059, 0.295, 0.907 Figure 4). However, there was variation in community respectively). None of these species differed significantly composition between pairs of sanctuary zones within each between sanctuary and habitat protection zones within marine estuary, but not between pairs of habitat protection zones. parks (P > 0.05). These species were only excluded from the This indicates more variation in fish communities among multivariate community composition analysis and are sanctuary zones. Similarly, relative abundances of silver included elsewhere. trevally varied between sanctuary zones within Batemans Bay Univariate analyses were performed using the same but nowhere else. PERMANOVA design as the assemblage data but with Euclidean distance as the measure of dissimilarity. These models were used to analyse total population data (average Max N and total number of species) as well as abundance of fi speci c population sub-groups (targeted species). In cases 50 Modified where the site factor was insignificant (P >0.25) it was Unmodified SZ pooled. Monte Carlo P-values were used where there were HPZ low numbers (< 50) of possible unique permutations in 40 analyses. All analyses were subdivided into two separate fi fi parts. The rst compared sh assemblages and variables 30 between highly modified and marine park estuaries, and the second compared sanctuary and habitat protection zones within the marine park estuaries. 20 *

10 *

RESULTS Mean number per BRUV drop (SE) In total, 5508 fish from 59 species were recorded in this study. 0 Targeted Species Total Max N Number Species Fourteen species occurred in all estuaries and 26 species were found in at least one modified estuary and one marine park. Figure 2. (a) Average Max N of targeted species and (b) species richness Thirteen species of fish occurred only in the marine parks, per BRUV drop in modified versus relatively unmodified estuaries while 20 species occurred only in the modified estuaries. Of (marine parks) (n = 64 drops per estuary) and in sanctuary and habitat fi fi protection zones within unmodi ed estuaries (n = 32 drops per zone). these, ve species occurred in both marine parks (but not the Error bars are Standard Error. Lines above bars indicate categories modified estuaries) and eight species occurred only in one which do not significantly differ from one another. * Indicates park. One species occurred in both modified estuaries (but not categories which differ significantly from all other categories.

Table 1. Univariate analysis of the impacts of protection on overall fish abundance (Max N), number of species, and abundance of targeted fish. Factors: Zo = Zone (habitat protection and sanctuary), Es = Estuary, Si = Site. Significant values are indicated in bold

Average Max N Number of Species Targeted Fish (Max N)

Source dF MS F p-value MS F p-value MS F p-value

Zone 1 0.065 0.049 0.873 39.063 3.307 0.161 100 7.98 0.028 Estuary 1 0.048 0.036 0.902 14.063 1.191 0.339 45.563 3.636 0.144 Zo x Es 1 3.219 2.425 0.17 0.563 0.048 0.776 68.063 5.431 0.068 Site (EsxZo) 4 1.328 3.891 0.008 11.813 1.755 0.152 12.531 1.198 0.324 Residual 56 0.341 6.732 10.464

Table 2. Univariate analysis of the mean Max N of (a) silver trevally, (b) yellowﬁn bream, (c) pink snapper by zone. Factors: Zo = Zone (habitat protection and sanctuary), Es = Estuary, Si = Site. Signiﬁcant values are indicated in bold. * Indicates Monte Carlo P-value. Presented when less than 20 unique permutations

Silvery trevally Yellowﬁn bream Pink snapper

Source dF MS F p-value MS F p-value MS F p-value

Zone 1 2.954 3.127 *0.154 0.149 2.238 *0.215 1.241 14.087 *0.020 Estuary 1 0.351 0.382 *0.565 0.24 3.598 *0.127 1.071 12.159 *0.031 Zo x Es 1 0.351 0.382 *0.544 0.019 0.279 *0.648 1.241 14.087 *0.023 Site (EsxZo) 4 0.918 6.85 0.001 0.067 0.631 0.668 0.088 0.466 0.796 Residual 56 0.134 0.106

5 2D Stress: 0.26

Snapper Trevally 4 Bream

Mean Max N per BRUV drop (±SE) p = 0.001 (estuaries) 0 p = 0.018 (modification) Modified Unmodified SZ HPZ Port Hacking Port Jackson Batemans Bay Jervis Bay Figure 3. Average abundance per BRUV drop of three recreationally targeted fish in modified versus relatively unmodified estuaries Figure 4. nMDS plots of multivariate assemblage composition by (marine parks) (n = 64 drops per estuary) and in sanctuary and habitat estuary. Symbols represent centroids of the assemblage composition. protection zones within unmodified estuaries (n = 32 drops per zone). Batemans Bay and Jervis Bay are marine parks, Port Hacking and Error bars are Standard Error. Port Jackson are modified estuaries.

Table 3. Multivariate analysis of the impacts of zone on assemblage relatively unmodified marine parks (Figure 3). The four composition. Factors: Zo = Zone (habitat protection and sanctuary), estuaries examined in this study are at least several hundred Es = Estuary, Si = Site. Significant values are indicated in bold kilometres within the known range of these species and so Source dF MS F p-value observed differences are not likely to be due to range effects (Edgar and Shaw, 1995; Gomon et al., 2008) Zone 1 3907.6 1.499 0.26 Overall assemblage composition was different between Estuary 1 12152 2.798 0.049 fi fi Zo x Es 1 2163 0.498 0.83 modi ed estuaries and the relatively unmodi ed marine parks Site (EsxZo) 4 4342.4 2.521 0.001 (Table 6, Figure 4). Each estuary occupied a distinct cluster in Residual 56 1722.8 multivariate space (Figure 4). SIMPER analysis revealed that the three species that contributed most to differences between the modified estuaries and marine parks were silver trevally, yellowfin bream, and pink snapper. These species collectively Effects of modification contributed approximately 35% of the difference (SIMPER) and there was a trend for them to be more abundant in the The relative abundance and number of fish species did not modified estuaries despite insignificant univariate analyses differ between modified estuaries and marine parks (Table 4, (Table 5, Figure 3). Figure 2). Modified estuaries, however, had more than four times higher abundance of targeted fish species (Table 4, Figure 2). Variation among sites within estuaries in both DISCUSSION average Max N and Max N of targeted fish was largely confined to sites within Jervis Bay Marine Park. There were This is one of the first studies to examine the impacts of no differences in abundances of most individual species of fish sanctuary zones and modification in marine systems by between levels of modification, in many cases this was due to simultaneously comparing fish assemblages across several variability between sites within each estuary (Table 5). levels of protection and modification. The findings support However, there was a clear trend towards more pink snapper, the idea that sanctuary zones increase the abundance of yellowfin bream and silver trevally in modified compared with targeted species as a whole and abundances of some

Copyright # 2011 John Wiley & Sons, Ltd. Aquatic Conserv: Mar. Freshw. Ecosyst. (2011) PUTTING MARINE SANCTUARIES INTO CONTEXT

Table 4. Univariate analyses of the impacts of modification on overall fish abundance (Max N), number of species, and abundance of targeted fish. Factors: Mo = Modification (Modified vs. Marine Parks Estuaries), Es = Estuary, Si = Site. Significant values are indicated in bold. * Indicates Monte Carlo P-value. Presented when less than 20 unique permutations

Average Max N Number of Species Targeted Fish (Max N)

Source dF MS F p-value MS F p-value MS F p-value

Mo 1 0.11 0.259 *0.665 11.281 1.249 *0.358 19.061 23.46 *0.039 Es (Mo) 2 0.425 0.432 0.679 9.031 0.813 0.459 0.812 1.33 0.298 Si (Es(Mo)) 12 0.983 2.9 0.002 11.115 1.75 0.07 0.608 3.12 0.002 Residual 112 0.339 6.353 0.195

Table 5. Univariate analysis of the mean Max N of (a) silver trevally, (b) yellowfin bream, (c) pink snapper by modification. Factors: Mo = Modification (Modified vs. Marine Parks Estuaries), Es = Estuary, Si = Site. Significant values are indicated in bold. * Indicates Monte Carlo P-value. Presented when less than 20 unique permutations

Silver trevally Yellowﬁn bream Pink snapper

Source dF MS F p-value MS F p-value MS F p-value

Mo 1 14.361 8.714 *0.103 19.607 3.757 *0.912 6.672 8.583 *0.101 Es (Mo) 2 1.628 2.199 0.148 5.219 12.611 0.004 0.777 0.998 0.396 Si (Es(Mo)) 12 0.749 2.138 0.024 0.414 1.831 0.046 0.779 2.139 0.02 Residual 112 0.35 0.226 0.364

Table 6. Multivariate analysis of the impacts of modification on fishing (in these estuaries, elsewhere they are all commercially assemblage composition. Factors: Mo = Modification (Modified vs. targeted). Given that the density of recreational anglers is Marine Parks Estuaries), Es = Estuary, Si = Site. * Indicates Monte highest in the estuaries where the greatest abundance of Carlo P-value. Presented when less than 20 unique permutations targeted fish was observed, these findings suggest that Source dF MS F p-value differences in recreational fishing pressure alone are insufficient to explain the trends observed in this study and Mo 1 44581 3.622 *0.018 that other conditions or stressors are having a substantial Es (Mo) 2 12309 3.39 0.001 Si(Es(Mo)) 12 3631.5 2.285 0.001 impact. Residual 112 1589.1 These findings do not imply that sanctuary zones or marine parks are ineffective at achieving conservation goals. Sanctuary zones contained a greater abundance of targeted species and individual species. However, contrary to our predictions, we more pink snapper than habitat protection zones. Moreover, also found that modified areas had a substantially higher there was a trend for more silver trevally in sanctuary zones abundance of targeted species and supported different fish but site specific differences in abundance probably obscured communities than relatively unmodified marine parks. these patterns. The greater abundances of pink snapper found Differences in the abundance of targeted species are highly in sanctuary zones is encouraging and appears to be a general relevant for assessing the relative importance of recreational pattern for this species with the same result seen over a fishing pressure as an ecological stressor. It is well known that number of years of sampling in deeper offshore waters in commercial and recreational fishing preferentially targets Batemans Bay Marine Park (M. A. Coleman, unpbl. data). In species that are predatory, larger bodied, and higher up the past studies snapper have been shown to respond well to the food chain (Pauly et al., 1998; Essington et al., 2006). A establishment of marine protected areas and older marine greater abundance of high trophic level species is often parks in New Zealand have shown increases in snapper density interpreted as an indication of increased productivity (Ryther, of up to 14 times compared with fished areas (Willis et al., 1969; Pauly and Christensen, 1995) and ecosystem health 2003). It should be noted that the current study was not (Munawar et al., 1989). In this study, large bodied predatory temporally replicated over the long term and could reflect species which are most sensitive to over-fishing were more differences in fish abundances that were present before marine abundant in the highly modified systems, despite greater park establishment. Long-term monitoring data contrasting recreational fishing activity in these areas (Henry and Lyle, sanctuary zones within parks with comparable outside areas 2003; DPI, 2010). Yellowfin bream, pink snapper and silver are required to ascertain definitively whether increased trevally are the 3rd, 13th and 16th species most harvested by abundance of pink snapper is due to increased protection recreational fishing activity in the state (respectively) (Henry (Edgar et al., 2009). While the present study agrees with the and Lyle, 2003) and there was a trend (significant for snapper) general assertion in the literature that marine parks and for each of these to be more abundant in modified estuaries. specifically sanctuary zones are effective at achieving a variety The types of commercial fishing undertaken in the study of conservation targets (one of which is the protection of estuaries (beach seining and purse seine bait collection) do not populations of targeted species), findings also suggest that typically target these species and so the majority of fishing protected areas need to be properly contextualized and pressure for these fish would probably come from recreational compared with modified environments (Halpern, 2003; Lester

Copyright # 2011 John Wiley & Sons, Ltd. Aquatic Conserv: Mar. Freshw. Ecosyst. (2011) A. C. MCKINLEY ET AL. et al., 2009). The role of modification factors other than fishing, A fourth possible explanation for these findings could be and the impacts that conservation measures have on general differences in the abundance and activity of apex predators. modification regimes, are clearly important issues which may In several cases increased abundance of predatory species in have a large effect on conservation outcomes (Greene and marine parks have been shown to have a ‘top down’ affect on Shenker, 1993; McKinley and Johnston, 2010). It is likely that aspects of the marine community (Shears and Babcock, 2002; modification factors other than fishing will also have an effect Micheli et al., 2004). These methods did not produce a on conservation outcomes in other kinds of marine sufficient sample size to understand the distributions of these environments. Coral reefs, offshore environments, and coastal large predators, but it is likely that they are more abundant in systems are also affected by many of the same human stressors the marine parks than in the modified estuaries. Both Jervis as estuaries. It is therefore likely that modification will also Bay and Batemans Bay have well documented resident affect conservation outcomes for marine parks in these populations of grey nurse sharks (Carcharias taurus) and systems. The degree to which sanctuary zones will translate common bottlenose dolphins (Tursiops truncatus). Both of into increased abundance of targeted species in these other these species are significant predators of a variety of targeted habitats cannot be addressed by this study, though a variety of fish species and both are believed to be largely absent from other studies have examined sanctuary zones in these systems Port Jackson and Port Hacking (Gomon et al., 2008). (Halpern, 2003). However, Port Jackson also has a well documented breeding Nutrient enrichment is a possible explanation for the population of dusky whaler sharks (Carcharhinus obscurus) increased abundance of targeted species in the modified so it is difficult to speculate about differences in the overall estuaries (Nixon and Buckley, 2002; Breitburg et al., 2009). abundance and activity of apex predators (McGrouther, The modified estuaries are nutrient enriched relative to the 2010). Differences in apex predator activity could not explain estuaries sampled in marine parks (Birch et al., 2010; Scanes, the observed differences between sanctuary and non-sanctuary 2010). It is likely that urbanization, land-use alteration and zones, as apex predators such as dolphins and sharks are run-off are largely responsible for the elevated nutrient levels likely to be active in both zones within a park (Shane et al., in these estuaries (Nixon, 1995; Scanes, 2010). Increased 1986; Last and Stevens, 2009). nutrient levels may be enhancing the productivity of the Impacts of human modifications on the environment and system and hence the abundance of fish. Several studies have associated ecological assemblages are complex, and so it is demonstrated that nutrient enrichment (at pre-eutrophication likely that a combination of these factors have influenced the levels) can enhance the abundance of fish and can results of this study. Regardless of which combination of substantially increase fisheries yields (Micheli, 1999; factors is responsible for the differences observed in this Oczkowski and Nixon, 2008; McKinley and Johnston, 2010). analysis, it is clear that there is significant ecological value in However, it is also possible that the modified estuaries are the modified estuaries and that they are highly diverse and naturally more productive than the marine parks systems. productive systems. Despite the abundance and diversity of While historic data on productivity do not go back far the fish assemblages in Port Jackson and Port Hacking, there enough to assess this quantitatively, it is likely that the are no significant sanctuary zones in either of these estuaries, placement of the major cities are not random and that they although many forms of commercial fishing have been have been somewhat influenced by natural productivity. It is limited. In fact, there are no major marine protected areas in well documented that cities are preferentially built in areas any of the heavily modified estuaries in south-eastern which are naturally highly productive as the availability of Australia (MPA, 2010) other than small Aquatic Reserves, natural resources (such as large fish populations) are a major many of which allow line fishing. A similar trend can be incentive for early economic and urban growth (Folke et al., observed worldwide, as very few heavily modified systems 1997; Haberl et al., 2004; Lotze et al., 2006). It is also have been protected by international marine parks systems possible that marine parks and sanctuary zones are selectively (Kelleher et al., 1995; IUCN, 2010). While modified areas are established in areas that are not heavily used by local not highly valued in traditional conservation thinking, these recreational fishing (e.g. poor fishing locations) as creating systems may harbour significant biodiversity and may be sanctuary zones in such places is more politically feasible heavily influenced by patterns of human activity. This study (Agardy et al., 2003; Ray, 2004; Edgar et al., 2008). supports the idea that the ecological characteristics of these Another possible explanation for these trends is increased highly modified systems warrant further investigation and habitat complexity in modified estuaries. Owing to the greater conservation efforts. degree of development and boat traffic, a large amount of artificial habitat (maritime structures) exists in the modified estuaries (Connell and Glasby, 1999). It has been CONCLUSION demonstrated that these structures can harbour diverse communities of invertebrates and plants (Connell and Glasby, These findings support the idea that sanctuary zones increase the 1999; Glasby and Connell, 1999; Glasby et al., 2007) and may abundance of targeted species overall as well as the abundance of aggregate or enhance fish abundances (Tuya et al., 2006). predatory species such as pink snapper (P. auratus) relative to Several of the species that were more abundant in the fished areas. However, it was also found that modified areas modified estuaries are known to feed on both sessile and had a substantially higher abundance of targeted species than mobile invertebrates (Coleman and Mobley, 1984; Froese and the marine parks. Stressors other than fishing pressure may be Pauly, 2010). It is therefore possible that artificial structures causing increased abundance of targeted fish in the modified support a higher abundance of these invertebrate food items estuaries and it is suggested that nutrient enrichment, increased and that this in turn has led to an increased abundance of the habitat complexity, and/or differences in the abundance of apex recreational fish species. predators could be responsible. However, further investigation

Copyright # 2011 John Wiley & Sons, Ltd. Aquatic Conserv: Mar. Freshw. Ecosyst. (2011) PUTTING MARINE SANCTUARIES INTO CONTEXT is needed to clarify the role of these variables. This study clearly Hagy JD, et al. 2009. Nutrient enrichment and fisheries supports the idea that modification factors other than fishing exploitation: interactive effects on estuarine living pressure are major determinants of fish assemblage structure. resources and their management. Hydrobiologia 629: – The minimization of these other stressors through the 31 47. Cappo M, Speare P, De’ath G. 2004. Comparison of baited establishment of marine parks is possible via legislation, remote underwater video stations (BRUVS) and prawn although it is unlikely that they can ever be totally removed. (shrimp) trawls for assessments of fish biodiversity in Reducing the impact of non-extractive marine stressors within inter-reefal areas of the Great Barrier Reef Marine Park. marine parks may have a substantial yet poorly understood Journal of Experimental Marine Biology and Ecology impact on fish assemblages. 302: 123–152. There is substantial evidence that sanctuary zones are an Clarke KR. 1993. Non-parametric multivariate analyses of efficient and successful conservation tool. The findings will changes in community structure. Austral Ecology 18: help to contextualize the performance of marine sanctuaries 117–143. and the metrics by which those parks are evaluated. The Claudet J, Pelletier D, Jouvenel JY, Bachet F, Galzin R. 2006. findings suggest that emphasizing the importance of marine Assessing the effects of marine protected area (MPA) on a reef fish assemblage in a northwestern Mediterranean parks as a method of bolstering populations of economically marine reserve: identifying community-based indicators. valuable species, an argument which is articulated in many Biological Conservation 130: 349–369. studies and management plans, is perhaps not the best Coleman N, Mobley M. 1984. Diets of commercially measure of a marine park’s success (McClanahan and exploited fish from Bass Strait and adjacent waters, Kaunda-Arara, 1996). In the estuaries studied, the findings Southeastern Australia. Marine and Freshwater Research suggest that the pervasive belief that the most natural and 35: 549–560. pristine conditions will produce the most fish is not necessarily Connell SD, Glasby TM. 1999. Do urban structures influence true, as some forms of modification may have large but local abundance and diversity of subtidal epibiota? A case poorly understood indirect effects on fish population study from Sydney Harbour, Australia. Marine – productivity. A higher abundance of commercially and Environmental Research 47: 373 387. Costello MJ, Read P. 1994. Toxicity of sewage sludge to marine recreationally important species does not necessarily reflect fi organisms: a review. Marine Environmental Research 37: natural conditions and reduced modi cation should not 23–46. fi necessarily be expected to increase sheries yields. Dafforn KA, Glasby TM, Johnston EL. 2009a. Links between estuarine condition and spatial distributions of marine invaders. Diversity and Distributions 15: 807–821. ACKNOWLEDGEMENTS Dafforn KA, Johnston EL, Glasby TM. 2009b. Shallow moving structures promote marine invader dominance. – This research was supported by the Australian Research Biofouling 25: 277 287. Denny CM, Babcock RC. 2004. Do partial marine reserves Council through an Australian Research Fellowship awarded protect reef fish assemblages? Biological Conservation 116: to ELJ and a Linkage Grant awarded to ELJ and MC. The 119–129. writing of this manuscript was also supported through the DPI. 2010. Survey of recreational fishing in New South Wales. Canadian National Sciences and Engineering Research Department of Primary Industries, Sydney. Council through a postgraduate award given to AM. We DSEWPC. 2011. Interim biogeographic regionalization of would like to thank Cian Foster-Thorpe and Shinjiro Australia. Department of Sustainability, Environment, Water, Ushiama for their help with the project. Comments from Population and Communities. Canberra, Available from http:// three anonymous reviewers greatly improved the manuscript. www.environment.gov.au/parks/nrs/science/bioregion-framework/ ibra/index.html (Accessed July 14, 2011). Edgar GJ, Shaw C. 1995. The production and tropic ecology of shallow-water fish assemblages in southern Australia. III. REFERENCES General relationships between sediments, seagrasses, invertebrates and fishes. Journal of Experimental Marine Agardy T, Bridgewater P, Crosby MP, Day J, Dayton PK, Biology and Ecology 194: 107–131. Kenchington R, Laffoley D, McConney P, Murray PA, Edgar GJ, Stuart-Smith RD. 2009. Ecological effects of Parks JE, Peau L. 2003. Dangerous targets? Unresolved marine protected areas on rocky reef communities: a issues and ideological clashes around marine protected continental-scale analysis. Marine Ecology Progress areas. Aquatic Conservation: Marine and Freshwater Series 388:51–62. Ecosystems 13: 353–367. Edgar GJ, Langhammer PF, Allen G, Brooks TM, Brodie J, Anderson MJ. 2001. A new method for non-parametric Crosse W, De Silva N, Fishpool LDC, Foster MN, Knox multivariate analysis of variance. Austral Ecology 26:32–46. DH, et al. 2008. Key biodiversity areas as globally Bax N, Williamson A, Aguero M, Gonzalez E, Geeves W. significant target sites for the conservation of marine 2003. Marine invasive alien species: a threat to global biological diversity. Aquatic Conservation: Marine and biodiversity. Marine Policy 27: 313–323. Freshwater Ecosystems 18: 969–983. Birch G, Taylor S. 1999. Source of heavy metals in sediments of Edgar GJ, Barrett NS, Stuart-Smith RD. 2009. Exploited reefs the Port Jackson estuary, Australia. The Science of the Total protected from fishing transform over decades into Environment 227: 123–138. conservation features otherwise absent from seascapes. Birch G, Cruickshank B, Davis B. 2010. Modelling nutrient Ecological Applications 19: 1967–1974. loads to Sydney estuary (Australia). Environmental Essington TE, Beaudreau AH, Wiedenmann J. 2006. Fishing Monitoring and Assessment 167: 333–348. through marine food webs. Proceedings of the National Breitburg DL, Craig JK, Fulford RS, Rose KA, Boynton Academy of Sciences of the United States of America 103: WR, Brady DC, Ciotti BJ, Diaz RJ, Friedland KD, 3171–3175.

Evans R, Russ G. 2004. Larger biomass of targeted reef fish in potential of estuaries and coastal seas. Science 312: no-take marine reserves on the Great Barrier Reef, Australia. 1806–1809. Aquatic Conservation: Marine and Freshwater Ecosystems 14: Mack RN, Simberloff D, Mark Lonsdale W, Evans H, Clout 505–519. M, Bazzaz FA. 2000. Biotic invasions: causes, Folke C, Jansson A, Larsson J, Costanza R. 1997. Ecosystem epidemiology, global consequences, and control. Ecological appropriation by cities. Ambio 26: 167–172. Applications 10: 689–710. Froese R, Pauly D. 2010. Fishbase. Available from www. Marchetti MP, Lockwood JL, Light T. 2006. Effects of fishbase.org (accessed January 25, 2010). urbanization on California’s fish diversity: differentiation, Glasby T, Connell S. 1999. Urban structures as marine homogenization and the influence of spatial scale. Biological habitats. Ambio 28: 595–598. Conservation 127: 310–318. Glasby T, Connell S, Holloway M, Hewitt C. 2007. Nonindigenous McClanahan T, Kaunda-Arara B. 1996. Fishery recovery in a biota on artificial structures: could habitat creation facilitate coral-reef marine park and its effect on the adjacent fishery. biological invasions? Marine Biology 151: 887–895. Conservation Biology 10: 1187–1199. Gomon M, Bray D, Kuiter R. 2008. Fishes of Australia’s McGrouther M. 2010. Dusky whaler juveniles in Sydney. Southern Coast. Reed New Holland: Sydney. Australia Museum, Melbourne, Available from http:// Greene LE, Shenker JM. 1993. The effects of human australianmuseum.net.au/Dusky-Shark-juveniles-in-Sydney/ activity on the temporal variability of coral reef fish (Accessed September 10, 2010). assemblages in the Key Largo National Marine Sanctuary. McKinley A, Johnston EL. 2010. Impacts of contaminant Aquatic Conservation: Marine and Freshwater Ecosystems 3: sources on marine fish abundance and species richness: a 189–205. review and meta-analysis of evidence from the field. Marine Haberl H, Wackernagel M, Krausmann F, Erb KHK-H, Ecology Progress Series 420: 175–191. Monfreda C. 2004. Ecological footprints and human McKinley AC, Miskiewicz A, Taylor MD, Johnston EL. 2011. appropriation of net primary production: a comparison. Strong links between metal contamination, habitat Land Use Policy 21: 279–288. modification and estuarine larval fish distributions. Halpern BS. 2003. The impact of marine reserves: do reserves Environmental Pollution 159: 1499–1509. work and does reserve size matter? Ecological Applications Micheli F. 1999. Eutrophication, fisheries, and consumer- 13: 117–137. resource dynamics in marine pelagic ecosystems. Science 21: Harmelin-Vivien M, Le Diréach L, Bayle-Sempere J, 1396–1398. Charbonnel E, García-Charton JA, Ody D, Pérez-Ruzafa Micheli F, Halpern BS, Botsford LW, Warner RR. 2004. A, Reñones O, Sánchez-Jerez P, Valle C. 2008. Gradients of Trajectories and correlates of community change in no-take abundance and biomass across reserve boundaries in six marine reserves. Ecological Applications 14: 1709–1723. Mediterranean marine protected areas: evidence of fish MPA. 2010. Marine Parks Authority of New South Wales, spillover? Biological Conservation 141: 1829–1839. Sydney, Available from, http://www.mpa.nsw.gov.au/ Henry GW, Lyle JM. 2003. The national recreational and (Accessed September 15, 2010). indigenous fishing survey. Commonwealth of Australia, Munawar M, Munawar IF, Mayfield CI, McCarthy LH. Department of Agriculture, Fisheries, Forestry, Canberra. 1989. Probing ecosystem health: a multi-disciplinary and Islam M, Tanaka M. 2004. Impacts of pollution on coastal and multi-trophic assay strategy. Hydrobiologia 188–189:93–115. marine ecosystems including coastal and marine fisheries and Nixon SW. 1995. Coastal marine eutrophication: a definition, approach for management: a review and synthesis. Marine social causes, and future concerns. Ophelia 41: 199–219. Pollution Bulletin 48: 624–649. Nixon S, Buckley B. 2002. “A strikingly rich zone”–nutrient IUCN. 2010. Marine protected areas. International Union for the enrichment and secondary production in coastal marine Conservation of Nature, Gland, Available from http://iucn. ecosystems. Estuaries and Coasts 25: 782–796. org/about/work/programmes/marine/marine_our_work/ NSW. 1999. Marine parks (zoning plans) regulation of 1999 marine_mpas/ (Accessed October 31, 2010). division marine park zones. New South Wales Government, Jennings S, Marshall SS, Polunin NVC. 1996. Seychelles’ marine Sydney, Available from www.austlii.edu.au/au/legis/nsw/ protected areas: comparative structure and status of reef fish consol_reg/mppr1999362/ (Accessed November 12, 2010). communities. Biological Conservation 75:201–209. Oczkowski A, Nixon S. 2008. Increasing nutrient Johnston E, Roberts DA. 2009. Contaminants reduce the concentrations and the rise and fall of a coastal fishery; a richness and evenness of marine communities: a review and review of data from the Nile Delta, Egypt. Estuarine meta-analysis. Environmental Pollution 157:1745–1752. Coastal and Shelf Science 77: 309–319. Kaiser MJ, Collie JS, Hall SJ, Jennings S, Poiner IR. 2002. Pauly D, Christensen V. 1995. Primary production required to Modification of marine habitats by trawling activities: sustain global fisheries. Nature 374: 255–257. prognosis and solutions. Fish and Fisheries 3: 114–136. Pauly D, Christensen V, Dalsgaard J, Froese R, Torres F, Kelleher G, Bleakley C, Wells S. 1995. A global representative Jr. 1998. Fishing down marine food webs. Science 279: system of marine protected areas: Antarctic, Arctic, 860–863. Mediterranean, Northwest Atlantic, Northeast Atlantic and Pauly D, Christensen V, Guenette S, Pitcher TJ, Sumaila UR, Baltic. Great Barrier Reef Marine Park Authority, World Walters CJ, Watson R, Zeller D. 2002. Towards Conservation Union (IUCN), and World Bank, sustainability in world fisheries. Nature 418:689–695. Washington, DC. Pérez-Ruzafa Á, González-Wangüemert M, Lenfant P, Marcos Last PR, Stevens PD. 2009. Sharks and Rays of Australia. C, García-Charton JA. 2006. Effects of fishing protection on CSIRO Publishing: Collingwood, Victoria. the genetic structure of fish populations. Biological Lester SE, Halpern BS, Grorud-Colvert K, Lubchenco J, Conservation 129: 244–255. Ruttenberg BI, Gaines SD, Airame S, Warner RR. 2009. Ray GC. 2004. Reconsidering ‘dangerous targets’ for marine Biological effects within no-take marine reserves: a global protected areas. Aquatic Conservation: Marine and synthesis. Marine Ecology Progress Series 384:33–46. Freshwater Ecosystems 14: 211–215. Lotze HK, Lenihan HS, Bourque BJ, Bradbury RH, Cooke Rowley RJ. 1994. Marine reserves in fisheries management. RG, Kay MC, Kidwell SM, Kirby MX, Peterson CH, Aquatic Conservation: Marine and Freshwater Ecosystems 4: Jackson JBC. 2006. Depletion, degradation, and recovery 233–254.

Copyright # 2011 John Wiley & Sons, Ltd. Aquatic Conserv: Mar. Freshw. Ecosyst. (2011) PUTTING MARINE SANCTUARIES INTO CONTEXT

Ryther JH. 1969. Photosynthesis and fish production in the sea. Stachowicz JJ, Whitlatch RB, Osman RW. 1999. Species Science 166:72–76. diversity and invasion resistance in a marine ecosystem. Samoilys MA, Martin-Smith KM, Giles BG, Cabrera B, Anticamara Science 286: 1577–1579. JA, Brunio EO, Vincent ACJ. 2007. Effectiveness of five small Tuya F, Sanchez-Jerez P, Dempster T, Boyra A, Haroun RJ. Philippines’ coral reef reserves for fish populations 2006. Changes in demersal wild fish aggregations beneath a depends on site-specific factors, particularly enforcement sea-cage fish farm after the cessation of farming. Journal of history. Biological Conservation 136: 584–601. Fish Biology 69: 682–697. Scanes P. 2010. NSW Estuarine catchment disturbance ranks. Williamson DH, Russ GR, Ayling AM. 2004. No-take marine NSW Department of Environment, Climate Change, and reserves increase abundance and biomass of reef fish on Water, Sydney. inshore fringing reefs of the Great Barrier Reef. Shane SH, Wells RS, Würsig B. 1986. Ecology, behavior and Environmental Conservation 31: 149–159. social organization of the bottlenose dolphin: a review. Willis TJ, Millar RB, Babcock RC. 2003. Protection of Marine Mammal Science 2:34–63. exploited fish in temperate regions: high density and Shears N, Babcock R. 2002. Marine reserves demonstrate biomass of snapper Pagrus auratus (Sparidae) in northern top-down control of community structure on temperate New Zealand marine reserves. Journal of Applied Ecology reefs. Oecologia 132: 131–142. 40: 214–227.

APPENDIX 1. AVERAGE SPECIES ABUNDANCE DATA BY PROTECTION ZONE AND ESTUARY. SPECIES ARE DISPLAYED IN ORDER OF TOTAL ABUNDANCE ACROSS ALL ESTUARIES. VALUES REPRESENT THE AVERAGE OF BRUV TAPES FOR EACH CATEGORY. HABITAT PROTECTION AND SANCTUARY ZONES ARE SUBSETS OF THE MARINE PARKS. TARGETED SPECIES FROM DPI (2010) AND FROESE & PAULY (2010)

Batemans HPZ Batemans SZ Jervis HPZ Jervis SZ Port Hacking Port Jackson Targeted Species Yellowtail Scad 31.19 11.56 21.13 23.06 13.66 9.94 N Trachurus novaezelandiae Mado 12.13 7.81 8.63 10.50 15.03 1.42 N Atypichthys strigatus Ocean Leatherjacket 1.56 0.06 4.19 10.63 5.00 3.80 N Nelusetta ayraud Silver Sweep 4.38 1.31 1.19 1.69 3.69 2.95 N Scorpis lineolata Silver Trevally 0.00 1.63 0.00 0.44 4.00 3.45 Y Pseudocaranx georgianus Pink Snapper 0.06 1.56 0.19 0.19 2.84 2.60 Y Pagrus auratus Yellowﬁn Bream 0.25 0.38 0.00 0.06 1.94 4.81 Y Acanthopagrus australis Weaping Toadﬁsh 0.00 0.00 0.19 3.06 0.84 0.99 N Torquigener pleurogramma Maori Wrasse 1.25 2.13 0.19 0.50 0.19 0.06 N Ophthalmolepis lineolatus Flathead sp. 0.25 0.06 1.00 0.69 0.72 0.04 Y Platycephalus sp. Fiddler Ray 0.31 0.38 0.81 0.75 0.03 0.06 N Trygonorrhina fasciata Green Moray Eel 0.38 0.69 0.06 0.00 0.31 0.15 N Gymnothorax funebris Common Eagle Ray 0.00 0.31 0.31 0.44 0.16 0.00 N Myliobatus aquila Long Finned Pike 0.19 0.13 0.00 0.00 0.50 0.21 Y Dinolestes lewini Luderick 0.00 0.63 0.00 0.00 0.22 0.03 Y Girella tricuspidata Senator Wrasse 0.06 0.13 0.25 0.25 0.06 0.07 N Pictilabrus laticlavius Tarwhine 0.19 0.25 0.06 0.25 0.00 0.00 Y Rhabdosargus sarba Red Rock Cod 0.19 0.44 0.06 0.00 0.00 0.00 Y Scorpaena cardinalis White Ear 0.25 0.13 0.06 0.06 0.13 0.03 N Parma microlepis Common Stingaree 0.06 0.06 0.19 0.19 0.47 0.16 N Trygonoptera imitata Wirrah 0.13 0.25 0.00 0.13 0.09 0.00 Y Acanthistius ocellatus

(Continues)

Table 1. (Continued)

Batemans HPZ Batemans SZ Jervis HPZ Jervis SZ Port Hacking Port Jackson Targeted Species Eastern Shovelnose Stingray 0.00 0.00 0.25 0.31 0.00 0.00 Y Aptychotrema rostrata Yellowtail Kingfish 0.00 0.00 0.00 0.00 0.00 0.54 Y Seriola lalandi Striped Trumpeter 0.00 0.00 0.00 0.00 0.47 0.00 N Pelates sexlineatus Eastern Kelpfish 0.13 0.25 0.00 0.06 0.00 0.00 N Chironemus marmoratus Tailor 0.00 0.00 0.00 0.00 0.03 0.38 Y Pomatomus saltatrix Blue Spotted Goatfish 0.00 0.00 0.00 0.00 0.38 0.00 N Upeneichthys vlamingii Blue Stripped Goatfish 0.00 0.13 0.00 0.06 0.19 0.00 N Upeneichthys lineatus Old Wife 0.00 0.00 0.06 0.25 0.00 0.00 N Enoplosus armatus Crimson Banded Wrasse 0.13 0.13 0.00 0.00 0.03 0.03 N Notolabrus gymnogenis Yellowfin Leatherjacket 0.00 0.13 0.00 0.00 0.03 0.11 Y Meuschenia trachylepis Port Jackson Shark 0.00 0.13 0.06 0.00 0.00 0.00 N Heterodontus portusjacksoni Red Morwong 0.06 0.00 0.06 0.06 0.00 0.00 Y Cheilodactlus fuscus Blind Shark 0.00 0.19 0.00 0.00 0.00 0.00 Y Brachaelurus waddi Eastern Blue Groper 0.00 0.06 0.06 0.00 0.03 0.00 N Achoerodus viridis Eastern Smooth Boxfish 0.00 0.00 0.00 0.00 0.13 0.00 N Anoplocapros inermis Amber Jack 0.00 0.13 0.00 0.00 0.00 0.00 Y Seriola dumerili Comb Wrasse 0.00 0.06 0.00 0.00 0.03 0.00 N Coris picta Girdled Parma 0.06 0.00 0.00 0.00 0.00 0.00 N Parma unifasciata Sergeant Baker 0.00 0.06 0.00 0.00 0.00 0.00 N Aulopus purpurissatus Eastern Garfish 0.00 0.06 0.00 0.00 0.00 0.00 Y Hyporhamphus australis Bronze Whaler Shark 0.00 0.06 0.00 0.00 0.00 0.00 Y Carcharhinus brachyurus Banded Wobbegong Shark 0.00 0.00 0.00 0.00 0.00 0.06 Y Orectolobus ornatus Blue Morwong 0.00 0.06 0.00 0.00 0.00 0.00 Y Nemadactylus douglasii Sergeant Major 0.00 0.00 0.00 0.00 0.00 0.04 N Abudefduf vaigiensis Scaly Tail Toadfish 0.00 0.00 0.00 0.00 0.00 0.04 N Torquigener squamicauda Fan Belly Leatherjacket 0.00 0.00 0.00 0.00 0.00 0.04 Y Monacanthus chinensis Sand Whiting 0.00 0.00 0.00 0.00 0.00 0.04 Y Sillago ciliate Herring Cale 0.00 0.00 0.00 0.00 0.03 0.00 N Odax cyanomelas Common Beardie 0.00 0.00 0.00 0.00 0.03 0.00 N Lotella rhacina Eastern Hulafish 0.00 0.00 0.00 0.00 0.00 0.03 N Trachinops taeniatus Smooth Flutemouth 0.00 0.00 0.00 0.00 0.03 0.00 N Fistularia commersonii Smooth Ray 0.00 0.00 0.00 0.00 0.03 0.00 N Dasyatis brevicaudata Half Banded Seaperch 0.00 0.00 0.00 0.00 0.03 0.00 N Hypoplectrodes maccullochi Globe Fish 0.00 0.00 0.00 0.00 0.03 0.00 N Diodon nicthemerus Roach 0.00 0.00 0.00 0.00 0.00 0.03 Y Gerres subfasciatus

(Continues)

Copyright # 2011 John Wiley & Sons, Ltd. Aquatic Conserv: Mar. Freshw. Ecosyst. (2011) PUTTING MARINE SANCTUARIES INTO CONTEXT

Table 1. (Continued)

Batemans HPZ Batemans SZ Jervis HPZ Jervis SZ Port Hacking Port Jackson Targeted Species Small Toothed Flounder 0.00 0.00 0.00 0.00 0.00 0.03 Y Pseudorhombus jenynsii Silver Drummer 0.00 0.00 0.00 0.00 0.03 0.00 Y Kyphosus sydneyanus Average Max N – Targeted 1.25 5.81 1.63 2.06 10.44 12.39 Species Average Max N - All 53.19 31.31 39.00 53.63 51.41 32.16 Species Richness 5.06 6.81 4.38 5.75 5.13 4.63

Chapter 4

STRONG LINKS BETWEEN METAL CONTAMINATION, HABITAT MODIFICATION AND ESTUARINE LARVAL FISH DISTRIBUTIONS

Final Version:

McKinley, A.C., Miskiewicz, A., Taylor, M.D. & E.L. Johnston. (2011) Strong links between metal contamination, habitat modification and estuarine larval fish distributions. Environmental Pollution, 159: 1499-1509. This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright Author's personal copy

Environmental Pollution 159 (2011) 1499e1509

Contents lists available at ScienceDirect

Environmental Pollution

journal homepage: www.elsevier.com/locate/envpol

Strong links between metal contamination, habitat modiﬁcation and estuarine larval ﬁsh distributions

Andrew C. McKinley a,*, Anthony Miskiewicz b, Matthew D. Taylor a, Emma L. Johnston a a Evolution & Ecology Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, New South Wales 2052, Australia b Environment and Recreation, Wollongong City Council, 41 Burelli Street, Wollongong, New South Wales 2500, Australia article info abstract

Article history: Changes to larval fish assemblages may have far reaching ecological impacts. Correlations between Received 4 February 2011 habitat modification, contamination and marine larval fish communities have rarely been assessed in Accepted 14 March 2011 situ. We investigated links between the large-scale distribution of stressors and larval fish assemblages in estuarine environments. Larval fish communities were sampled using a benthic sled within the inner and Keywords: outer zones of three heavily modified and three relatively unmodified estuaries. Larval abundances were Contaminants significantly greater in modified estuaries, and there were trends towards greater diversity in these Pollution systems. Differences in larval community composition were strongly related to sediment metal levels Fish larvae Habitat modification and reduced seagrass cover. The differences observed were driven by two abundant species, Paedogobius Sediment metals kimurai and Ambassis jacksoniensis, which occurred in large numbers almost exclusively in highly contaminated and pristine locations respectively. These findings suggest that contamination and habitat alteration manifest in substantial differences in the composition of estuarine larval fish assemblages. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction It is well documented that toxic contaminants such as metals are found in fish at various stages of their life cycle, often at levels that A variety of anthropogenic activities contribute to widespread may potentially reduce growth or survivorship (Alquezar et al., pollution and contamination in the marine environment, which 2006; Guo et al., 2008; Isosaari et al., 2006; Kojadinovic et al., influences the composition and health of ecological communities 2007; Miskiewicz and Gibbs, 1994). Evidence also points to the (Johnston and Roberts, 2009). Estuaries are generally believed to be potential adverse effects of toxic substances on reproduction and exposed to the highest levels of contamination of any marine envi- development of fishes (Arkoosh et al., 1998; Hose et al., 1989; Jones ronment due to their proximity to human settlements and their and Reynolds, 1997; Kingsford et al., 1997; Robinet and Feunteun, position directly downstream of agricultural and industrial activities 2002). The less toxic enriching contaminants (such as nutrients or (Kennish, 2002; Lotze et al., 2006). Similarly, habitat modification in sewage) may have either a weakly negative or largely positive effect estuarine systems is widespread, and many estuaries around the on abundance and diversity of adult fish (McKinley and Johnston, world have experienced losses of seagrass, mangrove, saltmarsh and 2010) however, the effects on the egg and larval stages have rarely other vegetated habitats (Duke et al., 2007; Lotze et al., 2006; been studied in situ. There are a small number of quantitative field Waycott et al., 2009). Many of these complex estuarine habitats studies examining these effects at the population and community provide a ‘nursery’ function for ecologically and economically level (Bervoets et al., 2005). This includes several studies demon- important species of fish (Beck et al., 2001; Boesch and Turner,1984; strating changes to the composition and distribution of larval fish Dorenbosch et al., 2004; Robertson and Duke, 1987; Taylor et al., communities around sewage plumes (Gray, 1996, 1997; Gray et al., 2005). Thus, it is imperative to understand how the modification 1992; Kingsford et al., 1997) and one study which found negative of estuaries through contamination and loss of habitat could be effects on embryo and larval development in relation to pulp mill impacting the early life stages of estuarine fish. Identifying stressors effluent (Karas et al., 1991). and monitoring ecological impacts in these communities is critical to Modification to marine habitats represents another potential managing and conserving native biodiversity in these systems. stressor of larval fish communities. Habitat degradation is the largest source of ecological modification globally and the greatest threat to * Corresponding author. biodiversity (Tilman et al., 1994). Australian estuaries have experi- E-mail address: [email protected] (A.C. McKinley). enced widespread changes to vegetative habitats over the last century,

1500 A.C. McKinley et al. / Environmental Pollution 159 (2011) 1499e1509 with well documented losses of mangrove (Valiela et al., 2001), salt- (Jervis Bay Marine Park) are within marine parks (NSW, 1999). Port Hacking is marsh (Saintilan and Williams, 1999), and seagrass habitats (Walker located between the suburbs of southern Sydney and the forested slopes of Royal National Park, which lines the southern border of the estuary. While not strictly and McComb, 1992) in south-eastern Australia. The reduced extent within a marine park, Port Hacking’s catchment is largely intact due to its proximity of these habitats within heavily modified estuaries arises due to to the Royal National Park and there is no major industrial activity within the estuary a variety of anthropogenic activities including dredging, increased (NSWDNR, 2010). Previous monitoring indicates that the heavily modified estuaries siltation, nutrient enrichment, contamination, clearing for coastal are nutrient enriched while nutrient levels in the relatively unmodified estuaries are development, and alterations to natural tidal or fluvial patterns less elevated (Scanes, 2010). Each estuary was divided into an inner and outer zone which reflected predicted (Saintilan and Williams, 1999; Valiela et al., 2001; Walker and physio-chemical and contamination gradients. These zones were defined based on McComb, 1992). Degradation of estuarine macrophytes is likely to their physical and biological characteristics. The inner zone is further up the estuary lead to changes to larval fish communities, as fishes require estuarine and represents the lower reaches of the estuarine tributary where brackish waters habitats to survive early stages of their life cycle (Beck et al., 2001; occur. In this zone turbidity, temperatures, and nutrient levels are higher than in the outer zone (Dafforn et al., in press). The outer zone sites are near the marine Boesch and Turner, 1984; Dorenbosch et al., 2004; Robertson and entrance to the estuaries where salinity, coastal flushing, wave exposure and oceanic Duke, 1987). current systems have greater influence. In this zone sediment grain sizes are also We explore the impacts of large-scale anthropogenic effects on larger and there is greater tidal influence (Dafforn et al., in press). Within each estuarine larval fish communities across heavily modified and rela- estuary six sites with bottom characteristics that allowed uninterrupted trawling e tively unmodified estuaries in New South Wales, Australia. Specifi- were selected, three in each zone. All sampling was replicated over two seasons the first in the Spring of 2009 and again in the late summer of 2010. cally, we examine how high levels of modification and contamination in the estuarine environment affect the composition, abundance, 2.2. Sampling methods and diversity of larvae. In addition, we utilize water quality, sediment metals, and habitat coverage data to estimate the relative importance Larvae were sampled using a benthic sled trawl towed along bare sediment behind a powerboat. Trawls were conducted along relatively flat profiles at a depth of these stressors within the broader modification regime. of 3e12 m. GPS was used to ensure that all trawls were 250 m in length, towed at a speed of 1.5 knots for approximately 5 min. The trawl was rigged to a four point 2. Methods bridle using an approximate 3:1 warp to depth ratio. The trawl frame consists of a stainless steel sled measuring approximately 1.5 m across and 2 m long. Within the 2.1. Study sites and sampling design sled frame two plankton nets were mounted in 50 cm diameter stainless steel rings 15 cm off the bottom. Each of these nets consisted of a 50 cm 300 cm (long) conical Larval fish were sampled in six estuaries along the south coast of New South plankton net with 250 mm mesh. A 1 L plastic sample jar was affixed to each cod end. Wales, Australia. These included three heavily modified estuaries e Port Jackson This yielded two sample jars for each trawl, one of which was processed for data (3344.2580S, 15116.5420E), Botany Bay (3359.3520S, 15111.4330E), and Port Kem- while the other was retained as a backup. bla (3428.1210S,15054.4100E), as well as three relatively unmodified estuaries e Port Because it is well known that the vertical and spatial distribution of larvae can be Hacking (3404.6800S, 15109.3110E), Jervis Bay (3504.7620S, 15044.8580E), and the influenced by light conditions and diel period, several precautions were taken to Clyde River (3544.2330S, 15014.2720E) (Fig. 1). The three heavily modified estuaries ensure consistency in these variables (Bridger, 1956; Pittman and McAlpine, 2003). are all highly anthropogenically disturbed environments near large urban and All sampling was conducted at night following the incoming tide up the estuary industrial areas and are subject to intense commercial and recreational boating (ie. starting in the outer zone and moving inwards). In order to standardize light traffic, historic and ongoing contamination, greater recreational fishing activity, and conditions sampling was conducted each month within a two week window around widespread urbanization of their shoreline and catchment (Birch and Taylor, 1999; the new moon (one week before and one week after the new moon). All samples DPI, 2010; Henry and Lyle, 2003; Scanes, 2010). Compared to the modified estu- were immediately preserved in a buffered 5% formalin/seawater mixture for aries, the relatively unmodified estuaries have fewer recreational fishermen, less transportation back to the laboratory. Where possible larvae were sorted and boating traffic, less urbanization of the coastline and catchment, and virtually no identified to species using the current taxonomic standard (Neira et al., 1998). For heavy industry (Birch and Taylor, 1999; DPI, 2010; Henry and Lyle, 2003; Scanes, some taxa larvae were only sorted to genus or family where current taxonomic 2010). Both the Clyde River (within Bateman’s Bay Marine Park) and Jervis Bay knowledge is insufficient for species level identification. A variety of Gobiidae sp.

Fig. 1. Location of study sites in a) Port Jackson, b) Botany Bay, c) Port Hacking, d) Port Kembla, e) Jervis Bay, and f) Clyde River estuaries. A Indicates outer zone sites C Indicates inner zone sites. Author's personal copy

A.C. McKinley et al. / Environmental Pollution 159 (2011) 1499e1509 1501

Table 1 Mean SE. Water quality and benthic sediment metals values. ‘All Metals’ represents the normalized total of values for Co, Cr, Cu, Fe, Mn, Ni, Pb and Zn from Dafforn et al. (in press). Individual values are presented for Cu, Ni, and Pb.

Temp (C) Sal & pH

Outer Inner Outer Inner Outer Inner Heavily Port Jackson 22 0.2 23.5 0.6 36.3 0.3 34.7 0.6 8.1 0.0 7.9 8.5 Modiﬁed Botany Bay 23.5 0.5 24.8 0.3 34.7 1.2 29.5 2.5 8.2 0.0 8.0 0.0 Estuaries Port Kembla 22.8 0.2 23.3 0.2 36.3 0.2 36.1 0.2 8.2 0.0 8.2 0.0

Relatively Port Hacking 21.8 0.6 23.6 0.5 35.4 0.2 34.9 0.7 8.1 0.0 8.0 0.0 Unmodiﬁed Jervis Bay 22.3 0.4 22.7 0.3 35.4 0.4 35.1 0.6 8.1 0.0 8.1 0.0 Estuaries Clyde River 20.1 0.6 23.0 0.8 35.5 0.7 30.8 2.5 8.1 0.0 7.9 0.0 All Metals (Normalized Total) Cu (mg kg 1) Ni (mg kg 1) Pb (mg kg 1)

Outer Inner Outer Inner Outer Inner Outer Inner Heavily Port Jackson 3.9 0.1 6.3 3.7 9.5 3.2 151.7 37.7 1.0 0.3 16.2 4.1 11.0 0.4 243.1 68.7 Modiﬁed Botany Bay 2.7 0.3 5.3 1.0 15.2 5.4 58.6 1.9 1.6 0.1 25.0 1.6 13.2 4.5 86.0 5.3 Estuaries Port Kembla 15.9 1.8 7.5 3.4 118.1 17.4 163.4 64.3 10.5 1.9 16.0 2.0 81.9 8.6 110.1 37.4

Relatively Port Hacking 1.6 0.9 7.4 0.8 31.6 20.0 0.0 7.0 1.8 1.0 0.9 2.2 17.3 10.6 5.2 5.4 Unmodified Jervis Bay 5.2 0.0 7.9 0.0 7.6 2.8 9.8 3.2 0.8 0.3 0.4 0.1 0.9 0.6 1.3 0.6 Estuaries Clyde River 2.5 0.2 3.7 1.3 1.2 0.8 6.9 4.4 2.4 0.4 9.8 3.0 2.2 0.3 10.0 3.0 were distinguished as morphologically distinct varieties and were included as Principal Coordinated Ordination (PCO). PCO is a computer program that performs separate ‘species’ within the analysis. Due to unresolved taxonomic issues, these a principal coordinate analysis of any symmetric distance matrix. This analysis is also unidentified goby species are presented as ‘unidentified Gobiidae sp.’. See Appendix called metric multi-dimensional scaling (Anderson, 2003). 1 for detailed species information. A calibrated flow meter was affixed to the trawl and readings were used to standardize all larval abundance results to 100 m3 of sea water. At each sampling 3. Results location four replicate water samples of 1 L each were taken at a depth of 1 m using a water sampling tube. These samples were combined in a plastic container on the 3.1. Estuary characteristics vessel and basic water quality data was measured from this water at the time of sampling using a calibrated YSI 6820 V2 sonde. In a parallel study surficial sediments were collected from sites in close proximity to the trawls. Samples were oven-dried Estuaries displayed similar average water quality conditions, before being digested and analyzed with ICP-OES following the methods outlined by though differences were found in most parameters between fi Hill et al. (2009). Recoveries were calculated against certi ed reference materials zones (Table 1). Higher sediment metals values were recorded in the and all metals used in this study were within accepted recovery limits. Arsenic and fi mercury recoveries were insufficient for further analysis (Dafforn et al., in press). modi ed estuaries, particularly in the inner zone sites where Vegetative habitat size and cover were calculated using estuarine fact sheet moni- anthropogenic contamination is greater. See Dafforn et al. (in press) toring values (NSWDNR, 2010). for detailed description and analysis of the sediment metals data. In many of the modified sites sediment metals values were above 2.3. Statistical analysis levels predicted to have biological effects according to water quality guidelines (ANZECC, 2010; Dafforn et al., in press). On average All multivariate and univariate analysis was conducted using mixed model fi PERMANOVA in PRIMER v.6 (Anderson, 2001). Prior to analysis abundance data was relatively unmodi ed estuaries had greater coverage of mangroves log(x þ 1) transformed. BrayeCurtis similarity matrices were calculated for multi- (22.4%) and seagrass (10.5%) relative to the modified estuaries (1.8%, variate data while Euclidean similarity matrices were used for univariate measures. 2.6% respectively). These vegetated habitats are virtually absent The PERMANOVA design employed in the course of this analysis consisted of the from Port Kembla and Port Jackson (NSWDNR, 2010). The coverage following factors: of saltmarsh was similar between relatively unmodified (4%) and modified (4.8%) estuaries though this was due to large saltmarsh Mo e Modification e Heavily Modified or Relatively Unmodified (2 levels, Fixed). Ti e Time e November or February (2 levels, Fixed). patches in Botany Bay (Table 2). Zo e Zone e Inner or Outer (2 levels, Fixed). Es e Estuary(Modification) e (6 estuaries, Random). Si e Site(Estuary(Modification) Zone) e (36 sites, Random). 3.2. Larval fish assemblages

Monte Carlo p-values were used in some places where the number of unique In total more than 10,200 fish larvae were collected and iden- permutations was less than 20. Analysis of water quality, metals, and habitat cover tified during the study. The summarized larval dataset can be found covariates was conducted using the DistLM function of PERMANOVA. This program in Appendix 1. The abundance of larval fish was significantly greater calculates a distance-based multivariate multiple regression (e.g. dbRDA) for any fi ¼ linear model on the basis of any distance measure, using permutation procedures in the heavily modi ed estuaries (p 0.003). There was a non- (McArdle and Anderson, 2001). Covariate factors were analyzed graphically using significant trend towards increased species richness (p ¼ 0.19) and

Table 2 Size of vegetative habitats and % of estuary with vegetative habitats (NSWDNR, 2010).

Estuary Size Habitat Size (km2) % Habitat Cover

Seagrass Mangrove Saltmarsh Seagrass Mangrove Saltmarsh Heavily Port Jackson 49.7 0 0 0 0.00 0.00 0.00 Modiﬁed Botany Bay 80 6.238 4.227 11.573 7.80 5.28 14.47 Estuaries Port Kembla 1.6 0 0 0 0.00 0.00 0.00

Relatively Port Hacking 11 0.807 0.307 0.082 7.34 2.79 0.75 Unmodiﬁed Jervis Bay 5.3 0.972 3.314 0.521 18.34 62.53 9.83 Estuaries Clyde River 103.2 6.05 1.999 1.486 5.86 1.94 1.44 Author's personal copy

1502 A.C. McKinley et al. / Environmental Pollution 159 (2011) 1499e1509

Table 3 Univariate analysis of the impacts of modiﬁcation on a) Larval abundance, b) Species richness, c) Shannon diversity. Factors: Mo ¼ Modiﬁcation, Zo ¼ Zone (Inner vs. Outer), Ti ¼ Time of Sampling, Es ¼ Estuary, Si ¼ Site. Bold values correspond to plots in Fig. 2. , Indicates Monte Carlo p value.

Source dF a) Abundance b) Species Richness c) Shannon Diversity

MS F p-value MS F p-value MS F p-value Mo 1 19.36 36.598 ,0.003 46.722 2.273 ,0.19 0.36275 5.7006 ,0.074 Zo 1 11.955 2.1936 0.189 4.5 0.13776 0.767 0.28464 0.54786 0.471 Ti 1 8.1882 1.9609 0.254 200 5.4381 0.091 1.5799 9.8475 0.027 Es(Mo) 4 0.529 0.18135 0.958 20.556 0.88729 0.473 6.36E-02 0.18753 0.949 MoxZo 1 10.619 1.9486 0.243 12.5 0.38265 0.605 1.2568 2.419 0.212 MoxTi 1 5.6766 1.3594 0.29 80.222 2.1813 0.204 0.58704 3.659 0.117 ZoxTi 1 1.238 3.0933 0.143 10.889 1.4 0.317 4.29E-03 6.59E-02 0.816 Es(Mo)xZo 4 5.4498 1.8683 0.131 32.667 1.4101 0.247 0.51954 1.5311 0.218 Es(Mo)xTi 4 4.1757 3.3578 0.033 36.778 5.1719 0.007 0.16044 0.96128 0.454 MoxZoxTi 1 3.7201 9.295 0.042 18 2.3143 0.18 0.27969 4.2995 0.116 Si(Es(Mo)xZo) 24 2.917 2.3457 0.024 23.167 3.2578 0.003 0.33933 2.0331 0.044 Es(Mo)xZoxTi 4 0.40023 0.32184 0.855 7.7778 1.0937 0.364 6.51E-02 0.38977 0.812 Res 24 1.2436 7.1111 0.1669

Shannon diversity (0.074) in the heavily modified estuaries (Fig. 2, (p ¼ 0.028). PCO plots indicate that the major cluster of modified Table 3). sites correspond strongly to both increased sediment metals levels Multivariate analysis of the community composition found that and decreased coverage of seagrass. Salinity also correlates strongly modified and unmodified estuaries differed significantly but did not show a clear trend by modification (Fig. 5a). All sedi- (p ¼ 0.028) (Fig. 3). Several of the random interaction terms were ment metals trended in approximately the same direction and were also significantly different (e.g. Mo Zo Ti). Here and elsewhere, inversely related to vegetative cover, such that the most contami- the test of the main effects can still be considered, as the higher nated sites occurred primarily in the estuaries with the lowest level fixed factor effect remains relevant regardless of the outcome seagrass cover. This suggests that there is a strong relationship of the interaction with a random factor (Quinn and Keough, 2002). between sediment metals levels, reduced vegetative cover, and Simper analysis revealed that the top six species contributing to community composition in the heavily modified estuaries. In this difference were Paedogobius kimurai (wide gape paedomorphic contrast, the major clustering of relatively unmodified sites showed goby), Gobiopterus semivestita (transparent goby), Arenigobius spp. little relationship to the metals and vegetative cover covariates (bridled goby spp.), Hyperlophus vittatus (sandy sprat), Hyperlophus (Fig. 5a, Table 4). transucidus (translucent sprat) and Ambassis jacksoniensis (Port Fig. 5b plots the species which were highly correlated with the Jackson glassfish) (Fig. 4, Table 5). Collectively these species major clusters of modified and relatively unmodified sites (those accounted for 37% of the difference between the modified and with a correlation factor >0.6). In order of the power of this rela- relatively unmodified estuaries. tionship, P. kimurai, G. semivestita, A. bifrenatus, and H. transucidus are more abundant in the sites that are highly metal contaminated and have lower seagrass cover (Fig. 5b). In order of the power of this 3.3. Covariates analysis e modified vs. relatively relationship, A. jacksoniensis, and H. vittatus are more abundant in unmodified estuaries the relatively unmodified sties (Fig. 5b). Notably, P. kimurai and A. jacksoniensis were the 2nd and 3rd most abundant species in this Salinity, pH, temperature, all sediment metals, % seagrass, % study, and each was encountered almost exclusively in modified/ mangrove, and % saltmarsh cover were all found to have a signifi- relatively unmodified sites (respectively). cant relationship with the larval community composition according to the DistLM analysis (Table 4b). However, DistLM is considered a poor predictor of the relative strength of these effects and so it 4. Discussion was used only to identify appropriate covariates to test in the PCO (Anderson, 2003; McArdle and Anderson, 2001). As stated earlier, We documented large differences between larval fish assem- multivariate analysis of community composition found that blages living in heavily modified and relatively unmodified estu- modified and relatively unmodified sites differed significantly aries. Total abundance of fish larvae was significantly greater in the

Fig. 2. Mean SE a) Larval abundance, b) Species richness, and c) Shannon diversity in estuaries of differing levels of modiﬁcation. Dotted line is average across all samples. Author's personal copy

A.C. McKinley et al. / Environmental Pollution 159 (2011) 1499e1509 1503

4.1. Positive effects of estuary modiﬁcation

Increased abundances of fishes in modified environments are 3D Stress: 0.17 unlikely to result directly from either increased levels of metal contamination or reduced cover of seagrass habitat. A far more likely cause of this pattern would be nutrient enrichment. Monitoring indicates that nutrient levels in the three heavily modified estuaries are elevated compared to the relatively unmodified estuaries (Scanes, 2010) although data was not available at sufficient resolution to formally analyze this relationship. Trends in this study suggest that larval fish communities may also be more diverse in the modified estuaries. Several studies have demonstrated that forms of contamination which have an enriching effect (e.g. nutrient run-off, fish farms, sewage, hydrocarbons, etc.) increase both the abundance and diversity of adult fish assemblages (McKinley and Johnston, 2010; McKinley et al., in review). This is the first study to observe fi p = 0.028 positive relationships between anthropogenic modi cation of estuaries and the abundance of larval fish. Heavily Modified Relatively Unmodified 4.2. Impacts of metals contamination

Fig. 3. Three dimensional MDS plot of multivariate assemblage composition by modification. Symbols represent centroids of the assemblage composition. Heavily A variety of studies indicate that adult fish are fairly resilient in modified includes sites in Port Jackson, Botany Bay, and Port Kembla. Relatively the face of anthropogenic contaminants and adult fish assemblages unmodified includes sites in Port Hacking, the Clyde River, and Jervis Bay. do not appear to be as sensitive to contaminants as invertebrates or fish larvae (Johnston and Roberts, 2009; McKinley and Johnston, 2010). In most field studies contaminants have been shown to have either weakly negative or a largely positive effect (where heavily modified estuaries and trends suggest that diversity could enriching contaminants are present) on adult fish abundance and also be higher in these modified environments. However, certain diversity (McKinley and Johnston, 2010). However, most of these species were strongly negatively associated with estuary modifica- studies have focused primarily on adults of large bodied predatory tion. Differences in community composition were strongly related to fishes which are highly mobile (McKinley and Johnston, 2010). both sediment metal levels and reduced seagrass cover in heavily This contrasts to the larvae examined in this study, which are modified estuaries. Contamination levels and seagrass cover are primarily small bodied species that are comparatively less mobile inversely related and experimental work is needed to establish at maturity (e.g. Gobiidae, Clupeidae and Apogonidae spp.). Whilst causation and to parse out the relative contribution of each factor to large species may accumulate contaminants to a greater degree the observed patterns. than larvae or invertebrates due to their high trophic position, they

Fig. 4. Mean SE larval abundance by estuary for top six species contributing to differences between heavily modiﬁed and relatively unmodiﬁed estuaries. Author's personal copy

1504 A.C. McKinley et al. / Environmental Pollution 159 (2011) 1499e1509

Table 4 a) Multivariate analysis of the impacts of modiﬁcation on larval community ¼ ﬁ ¼ ¼ composition. Factors: Mo Modi cation, Zo Zone (Inner vs. Outer), Ti Time of 0.875 -value , 0.002 Sampling, Es ¼ Estuary, Si ¼ Site. b) Results of DistLM covariate analysis. Bold values p correspond to Fig. 3 plot. , Indicates Monte Carlo p value. Zone (Inner ¼

Source dF a) Community Composition

MS F p-value cation, Zo

Mo 1 18,770 2.8208 ,0.028 ﬁ Zo 1 7057.6 1.6089 0.219 Hyperlophus vittatus Ti 1 8514.7 2.3778 0.087 Modi 14.19 7.4158 f) Sandy Sprat Es(Mo) 4 6654.1 3.0881 0.001 ¼ MoxZo 1 5412.3 1.2338 0.349 MoxTi 1 4482.7 1.2519 0.319 0.145 0.4408 3.11E-02

ZoxTi 1 1968.9 1.1223 0.371 -value MS F , 0.001 p Es(Mo)xZo 4 4386.7 2.0358 0.001 Es(Mo)xTi 4 3580.9 2.6272 0.001 MoxZoxTi 1 2231.7 1.2721 0.316 Si(Es(Mo)xZo) 24 2154.8 1.5809 0.001 Es(Mo)xZoxTi 4 1754.3 1.2871 0.14 ed estuaries. Factors: Mo

Res 24 1363 ﬁ

Variable b) DistLM Covariate Results Hyperlophus transucidus 10.016 22.901 e) Translucent Sprat SS F p-value Temp (C) 8.51Eþ03 3.1395 0.003

Sal 7.00Eþ03 2.5608 0.004 0.107 29.698 2.9652 -value MS F , 0.001 pH 5.49Eþ03 1.992 0.029 p Co 228.616 1.82Eþ04 7.0587 0.001 þ

Cr 267.716 1.96E 04 7.6637 0.001 ed and relatively unmodi Cu 327.393 1.06Eþ04 3.9677 0.001 ﬁ Fe 238.204 1.68Eþ04 6.4673 0.001 Mn 257.610 1.42Eþ04 5.4071 0.001 Ni 231.604 1.95Eþ04 7.6265 0.001 Gobiopterus semivestita Pb 220.353 1.36Eþ04 5.1363 0.001 value. p 11.876 9.1728 d) Transparent Goby Zn 206.200 1.55Eþ04 5.9537 0.001 %Seagrass 1.10Eþ04 4.1003 0.001 þ

%Mangrove 6.20E 03 2.2589 0.011 0.077 47.307 3.9835 -value MS F , 0.007 %Saltmarsh 9409.6 3.4858 0.001 p Estuary Area 7.35Eþ03 2.6935 0.003 spp. Indicates Monte Carlo , can be highly mobile so direct exposure times are not certain, their . Arenigobius

diets are comparatively diverse, and they have a higher capacity for Fig. 4 7.3611 4.3679 physiological resistance and tolerance (van der Oost et al., 2003; c) Bridled Goby spp. Wirgin and Waldman, 1998).

ﬁ 0.092 45.827 6.2256

In contrast, signi cant evidence points to developmental and -value MS F , 0.001 p reproductive susceptibility to contaminants in ﬁsh populations Wide (Arkoosh et al., 1998; Hose et al., 1989; Jones and Reynolds, 1997; Kingsford et al., 1997; Robinet and Feunteun, 2002). Studies have found reduced egg and larval abundance due to exposure to sewage sludge (Waring et al., 1996), behavioral and metabolic changes with heavy metal exposure (Kienle et al., 2008; Sreedevi et al., 1992), increased incidents of larval deformity due to sewage Paedogobius kimurai 106.8 5.0101 plumes, chemical efﬂuent and tributylin exposure (Hu et al., b) Gape Paedomorphic Goby 2009; Kingsford et al., 1997; Vetemaa et al., 1997), and reduced Site. Bold values correspond to plots in ¼

fish condition, survivorship and growth with exposure to metals 0.005 -value MS F p (Bervoets and Blust, 2003; Canli and Atli, 2003; Hutchinson et al., , sh cation on the abundance of the top six species contributing to differences between heavily modi fi 1994). While few of these effects have been verified in at the fi Estuary, Si

community level in wild populations, it is not unreasonable to ¼ expect that many of these impacts could be detected in wild larval fish assemblages. As such, it is probable that contamination impacts have directly contributed to differences in the larval fish assemblage between modified and relatively unmodified Ambassis jacksoniensis MS F Port Jackson Glass estuaries. In this study the abundance of several species were positively Time of Sampling, Es

correlated with highly contaminated areas, which could imply that ¼ these species favor these sites or are comparatively resistant to pollution effects. Notably, P. kimurai were very strongly associated with sediment metals levels. This species was extremely abundant ZoTiEs(Mo) 1 4 1 5.3218 1.3239 4.9414 1.0103 0.58894 0.79944 0.657 0.38 0.393 21.317 11.16 3.3772 17.73 1.4373 2.5078 0.301 0.178 11.025 1.1148 41.378 0.60947 0.459 0.003 41.333 0.81232 1.1328 5.1813 0.356 0.048 1.116 20.957 3.4281 1.6502 0.079 0.275 2.25E-02 2.5402 7.88E-03 0.941 0.22698 0.689 MoMoxZoMoxTi 1 27.921ZoxTiEs(Mo)xZo 21.09 1Es(Mo)xTiMoxZoxTi 1 2.756 4Si(Es(Mo)xZo) 2.2305 1 4 5.2677 0.5232 24Es(Mo)xZoxTi 0.36086 1.4214 1 6.181Res 2.3433 2.248 4 0.517 0.583 10.022 6.0243 0.083 0.14182 3.6434 1.3251 42.478 0.047 8.36E-02 12.333 0.02 9.1559 0.99 0.255 0.003 4.4502 24 2.7714 1.3765 3.8965 1.6965 3.7012 0.172 1.2024 2.3498 0.132 1.4678 0.67833 0.93777 0.02 0.409 2.257 0.46214 2.4041 5.13E-02 4.6984 2.9349 0.515 0.021 0.19262 0.61466 0.008 1.8291 0.042 0.663 1.2339 6.9228 1.6853 0.2738 0.26644 2.2449 1.0854 0.298 0.1117 0.634 3.0837 0.16116 0.365 1.0194 1.3579 15.455 0.15577 0.952 0.167 0.74144 0.494 0.693 7.9772 0.29 1.9374 0.8802 0.71707 8.64E-02 0.50012 1.2947 0.58669 6.1616 0.314 0.1026 1.083 0.409 0.84235 0.004 1.9554 0.771 0.86754 1.2723 23.701 0.47051 0.375 0.362 0.052 6.1134 0.289 0.95335 0.63674 0.67627 3.8769 0.43735 3.7085 0.492 13.978 0.73894 1.2902 0.048 8.6253 5.578 3.13E-01 0.33138 0.001 9.4246 2.9029 0.001 9.446 0.21389 0.601 0.001 0.002 0.66492 7.51E-02 2.8487 1.6532 11.191 2.1637 0.817 0.457 4.3658 1.9135 1.4887 6.6171 0.204 2.5813 1.1314 0.24 0.002 0.064 0.401 0.66209 7.84E-02 1.6913 Source dF a) vs. Outer), Ti in highly contaminated sites; in such areas they accounted for Table 5 Univariate analysis of the impacts of modi Author's personal copy

A.C. McKinley et al. / Environmental Pollution 159 (2011) 1499e1509 1505

a 40 b 40

20 20 Hyperlophus vittatus Ambassis jacksoniensis ) on ti a ri

0 va l 0 a %Seagrass t Arenigobius spp. o ft Mn Ni Paedogobius kimurai Cu Salinity

Zn 9% o . -20 Co -20 14

Pb ( Gobiopterus semivestita Fe Cr Hyperlophus transucidus PCO2 (14.9% of total variation) PCO2

-40 -40

Heavily Modified Heavily Modified Relatively Unmodified Relatively Unmodified -60 -60 -40 -20 0 20 40 60 -40 -20 0 20 40 60 PCO1 (17.3% of total variation) PCO1 (17.3% of total variation)

Fig. 5. Principal Coordinated Ordination (PCO) of correlations between covariate factors and two dimensional plots of community composition by modification. a) Metals contamination, habitat modification, and water quality covariates. (Pearson Correlation > 0.2). b) Plots of top six species contributing to differences between heavily modified and relatively unmodified estuaries (Pearson Correlation > 0.6).

60e95% of the larval assemblage and they were hyper-abundant nutrient availability has been shown to increase the relative particularly in the most highly contaminated inner zone sites of dominance of epiphytic plants in seagrass beds, often to the Port Jackson and Port Kembla. In contrast, this species was rarely detriment of the seagrass community (Harlin and Thorne-Miller, encountered in the relatively unmodified estuaries. P. kimurai 1981). For these reasons losses of some estuarine vegetative occurred in the highly metals contaminated sites both in estuaries habitats may be strongly correlated with levels of contaminant where there were virtually no vegetative habitats (Port Jackson and exposure. Contaminants such as metals are also acutely toxic to Port Kembla) and directly adjacent to mangrove and seagrass some seagrass and other plant species, and trace metal run-off is patches (Botany Bay). This implies that this species is particularly a well documented cause of seagrass habitat loss (Macinnis-Ng and successful in highly contaminated sites and may be unusually Ralph, 2002; Prange and Dennison, 2000; Warnau et al., 1995). It is resistant to contaminants. Contamination resistance in fish has therefore difficult to distinguish the relative effects of contamina- been demonstrated in some cases (Wirgin and Waldman, 2004; Xie tion vs. habitat loss as these stressors may be intimately related in and Klerks, 2004). In addition, the distribution of this species is the estuarine system. unusually patchy and there is some speculation that it is an invasive Many of the species which contributed strongly to the trends in originating from South East Asia (eg. Thailand). However, this has this study feed on vegetative matter, lay their eggs on plants, or use not been confirmed (Iwata et al., 2001; Neira et al., 1998). P. kimurai estuarine vegetation for shelter during the larval and post settle- is also unusual among fish as it is sexually mature at a very small ment stages of their life cycle (Miskiewicz, 1987). It is therefore size e adults average approximately 1.5 cm in length and females possible that changes to the larval assemblage have directly pregnant with eggs are approximately the same size as other resulted from loss of vegetative habitats in the modified estuaries. gobies’ larvae (Iwata et al., 2001). It has been demonstrated that For example, A. jacksoniensis and F. lentiginosus primarily settle in a variety of highly invasive invertebrate and fish species are rapidly seagrass beds during their juvenile and adult stages while juvenile maturing and unusually contamination resistant (Alcaraz et al., and adult Gerres subfasciatus (roach) utilize mangroves, so changes 2005; Dafforn et al., 2009; Piola and Johnston, 2008). It is to, or the absence of these habitats could explain decreased possible that this species is also a marine invader displaying these abundance of these species in the modified estuaries (Gray et al., characteristics. 1996; Jelbart et al., 2007; Jenkins et al., 1997; Neira et al., 1998). While most of the larvae sampled in this study were taken over bare sediment and were at the pre-settlement (planktonic) stage of 4.3. Impacts of vegetative habitat alteration their life cycle, changes to these vegetative habitats could impact these larvae when they reach their juvenile and adult stages. It is Correlative studies are limited in their ability to predict the therefore possible that losses of vegetative habitats have reduced magnitude and relative importance of covarying factors (Shipley, the available habitat for these species, which would reduce the 2002). In this study, sites with high levels of metals in the sedi- population size and hence larval abundance in the long term. ment also tended to have reduced coverage of vegetative habitats; this is likely the case because the core mechanisms of estuary contamination exposure (e.g. run-off, urbanization of shoreline/ 4.4. Estuarine opportunists vs. truly estuarine species catchment, outflows, etc.) also tend to precipitate habitat alteration in estuarine systems (Drinkwater and Frank, 1994; Rogers et al., It is well known that a variety of fish species utilize estuaries 2002). Run-off which carries metals and other contaminants is during their life cycle. Many ‘estuarine opportunist’ species spawn also known to increase turbidity in many cases, hence lowering at sea and find their way into the estuarine environment using light levels and impacting plant growth (Longstaff and Dennison, oceanic currents for transport, entering the estuary at the preflex- 1999). In many cases increased nutrient levels accompany other ion and postflexion stages (after hatching) based on a variety of forms of contamination (McKinley and Johnston, 2010). Increased environmental cues (Neira and Potter, 1992; Norcross and Shaw, Author's personal copy

1506 A.C. McKinley et al. / Environmental Pollution 159 (2011) 1499e1509

1984; Potter et al., 1988). In contrast, truly estuarine species live reasons we do not believe our results to be strongly spatially or within the estuary for their entire life cycle, including when temporally restricted. spawning. Of the species which contributed most to the differences between heavily modified and relatively unmodified estuaries four are truly estuarine species (P. kimurai, G. semivestita, Arenigobius 5. Conclusion spp., and H. transucidus) while the other two (A. jacksoniensis and H. vittatus) live primarily within the estuary but spawn at sea It is clear that there are large-scale differences between the (Miskiewicz, 1987; Neira et al., 1998). All of these species spend the larval fish assemblages living in heavily modified and relatively majority of their life cycle in the estuarine environment and so unmodified estuaries. Differences in larval fish community changes to that environment are likely to have an impact, regard- composition were strongly related to both sediment metal levels less of whether or not they spawn at sea or within the estuary. and reduced seagrass cover in heavily modified sites. We believe However, it is possible that high levels of contamination and that it is likely that habitat alteration and estuarine sediment habitat modification disproportionately impact species which contamination are interrelated stressors which have contributed to spawn in the estuary environment itself. Estuarine opportunist the observed differences between modified and relatively species which spawn at sea and enter estuaries later in their life unmodified estuaries. Ultimately changes to larval fish assemblages cycle may be less affected by these stressors as they may be less may have far reaching ecological impacts both for the adult fish exposed to contaminants during their early (pre-hatching) growth community and other organisms. Notably, the impacts of stressors stages. It is therefore possible that impacts experienced during at the larval stage of economically and culturally important fish larval stages of estuarine fishes’ could reduce their relative species are poorly studied and little understood. The absence of competitiveness and hence increase the dominance of taxa which studies examining anthropogenic impacts on estuarine fish larvae spawn at sea. It is also possible that some of the estuarine oppor- represents a major gap in the environmental impact literature and tunist species are descended from parents who lived in other further investigation and monitoring is warranted. estuaries and may not have been exposed to contamination levels reflective of the conditions of the estuary in which they were found during the larval phase of their life cycle. Acknowledgements It should be noted that the temporal and spatial variability of larval fishes has been found to be a significant issue in previous This research was primarily supported by the Australian studies (Gray, 1996, 1997; Gray and Miskiewicz, 2000). In this study Research Council through an Australian Research Fellowship we sampled over a large spatial scale with a relatively high level of awarded to ELJ and a Linkage Grant awarded to ELJ. The writing of spatial replication both within and between estuaries. Despite this manuscript was also supported through the Canadian National significant differences between the random factor of sites for many Sciences and Engineering Research Council through an award given analyses, our sampling design and level of replication was suffi- to AM. We would like to thank Dr. Katherine Dafforn, Cian Foster- ciently robust to show clear differences by estuary and modifica- Thorpe and Shinjiro Ushiama for their help and contributions of tion. We have also sampled across two seasons, during which data for the project. We would also like to thank the Bluescope Steel a large proportion of estuarine species would have been breeding Company and Marine Parks NSW for their generous support. (Neira et al., 1998). Despite a high degree of temporal variability between these rounds of sampling, time was not sufficiently strong to produce a significant result in most of our analyses. For these Appendix

Appendix 1 Average abundance data identiﬁed to lowest taxonomic level by zone and estuary. Gobiidae sp. represent the total of all observed morphologically distinct taxa which could not be identiﬁed to species.

Family Taxon Botany Bay Port Jackson Port Kembla Clyde River Jervis Bay Port Hacking

Outer Inner Outer Inner Outer Inner Outer Inner Outer Inner Outer Inner Ambassidae Ambassis jacksoniensis 26.32 1.12 5.02 0.33 6.62 3.35 20.78 10.86 570.10 33.34 46.69 6.66 Glassfish Port Jackson Glassfish Ambassis marianus 0.00 0.00 0.34 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Estuary Glassfish Apogonidae Foa sp. 0.00 0.00 0.00 3.08 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Cardinalfish Cardinalfish sp. Siphamia cephalotes 0.36 0.00 3.47 0.00 0.00 0.00 0.00 0.00 18.96 2.68 0.00 0.00 Wood’s Siphonfish Apogonidae sp. A 2.51 0.00 1.98 0.00 0.00 0.00 0.00 0.00 0.37 0.00 0.00 0.64 Cardinalfish sp. A Apogonidae sp. B 0.00 0.00 2.36 0.33 0.00 0.00 0.00 0.00 0.67 0.27 0.36 0.00 Cardinalfish sp. B Apogonidae sp. C 0.00 0.00 0.99 0.99 0.00 0.42 0.00 0.00 0.00 0.00 0.00 0.00 Cardinalfish sp. C Atherinidae sp. Atherinidae sp. 0.73 1.25 0.00 0.00 0.00 0.00 0.00 0.00 0.00 2.15 0.39 0.00 Old World Silversides Hardyhead sp. Belonidae Hempheridae sp. 0.73 1.25 0.00 0.00 0.00 0.00 0.76 0.49 0.00 0.00 0.39 0.00 Needlefish Garfish sp. Author's personal copy

A.C. McKinley et al. / Environmental Pollution 159 (2011) 1499e1509 1507

Appendix 1 (continued)

Family Taxon Botany Bay Port Jackson Port Kembla Clyde River Jervis Bay Port Hacking Outer Inner Outer Inner Outer Inner Outer Inner Outer Inner Outer Inner Blenniidae Omobranchus anolius 5.03 3.73 0.50 0.00 0.35 0.00 0.00 0.60 0.00 0.00 0.00 0.00 Blennies Oyster Blenny Omobranchus rotundiceps 1.09 0.00 0.00 5.30 0.00 0.00 0.38 0.77 0.00 0.00 0.00 0.00 Combtooth Blenny Petroscirtes lupus 0.00 0.00 0.00 0.34 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Brown Sabretooth Blenny Callionymidae Callionymidae sp. 0.00 0.00 0.34 0.00 0.34 0.00 0.38 0.24 0.00 0.00 0.00 0.00 Dragonets Dragonet sp. Carangidae Trachurus novaezelandiae 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.40 0.00 Jacks/Jack Mackerels Yellowtail Scad Clinidae Cristiceps sp. 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.79 0.00 Clinids Weedfish sp. Clupeidae Etrumeus teres 0.00 0.00 0.00 0.00 0.71 1.24 0.00 0.00 0.00 0.00 0.00 0.00 Herrings/Sprats Maray Herklotsichthys castelnaui 0.00 0.00 0.00 4.88 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Southern Herring Hyperlophus transucidus 1.82 109.62 0.00 34.29 0.00 0.00 0.00 0.00 0.00 0.00 0.40 0.00 Translucent Sprat Hyperlophus vittatus 8.33 5.26 7.06 0.34 16.15 42.96 67.92 58.51 116.68 40.50 7.34 0.79 Sandy Sprat Sardinops sagax 0.00 0.00 0.00 0.00 5.89 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Pilchard Spratelloides robustus 0.00 0.00 0.00 0.00 0.00 3.34 0.00 0.00 0.00 0.00 1.57 0.00 Blue Sprat Creediidae Creedia haswelli 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.37 0.00 0.00 0.00 Sandburrower Slender Sandburrower Engraulidae Engraulis australis 0.00 0.50 0.00 2.35 3.20 7.66 0.00 0.00 0.00 0.70 0.00 0.00 Anchovies Australian Anchovy Gerreidae Gerres subfasciatus 30.54 0.00 6.74 7.02 12.78 0.82 27.76 10.77 8.33 5.77 12.54 1.52 Silver Biddies Roach Gobiesocidae Gobiesocidae sp. 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.34 0.00 1.59 0.00 Clingfish Clingfishes sp. Gobiidae Afurcagobius tamarensis 0.00 0.00 0.00 0.00 0.00 0.00 0.39 0.36 0.00 0.00 0.00 0.00 Gobies Tamar Goby Arenigobius spp. 9.67 48.63 10.20 18.65 43.89 74.49 3.06 8.25 16.50 1.12 8.09 2.50 Half Bridled and Bridled Goby Favonigobius lentiginosus 2.96 0.00 4.84 0.33 1.13 0.41 0.39 5.06 3.95 7.19 36.41 0.96 Long Finned Goby Gobiidae sp. 13.62 13.62 13.62 13.62 13.62 13.62 13.62 13.62 13.62 13.62 13.62 13.62 Unidentified Goby sp. Gobiopterus semivestita 5.29 214.77 0.68 244.13 1.39 1.10 0.00 1.50 0.00 7.39 0.00 0.32 Transparent Goby Paedogobius kimurai 0.78 1.55 26.29 477.99 37.67 164.78 0.35 0.00 0.34 0.00 0.79 4.18 Wide Gape Paedomorphic Goby Gobiidae Pseudogobius sp. 0.00 1.25 12.73 1.16 0.00 2.06 2.23 1.30 0.34 0.00 0.00 0.00 Gobies Eastern Bluespot Goby Redigobius macrostoma 3.60 4.31 0.00 1.74 1.06 4.56 5.42 9.45 13.18 22.01 0.00 0.53 Large Mouth Goby Schindleriidae sp. 2.51 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Schindler’s Goby Tridentiger trigonocephalus 0.00 0.00 0.00 0.00 1.68 3.24 0.00 0.00 0.00 0.00 0.00 0.00 Trident Goby Gonostomatidae Gonostomatidae sp. 0.00 0.00 0.00 0.00 0.00 0.00 0.38 0.00 0.00 0.00 0.00 0.00 Bristlemouths Bristlemouth sp. Kyphosidae Girella tricuspidata 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.63 0.00 0.00 Sea Chubs Luderick Leptoscopidae Lesueurina platycephala 0.00 0.00 0.34 0.00 0.34 0.00 0.00 0.00 0.00 0.00 5.09 0.00 Sandfish Common Sandfish Lutjanidae Lutjanidae sp. 0.00 0.00 0.34 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Snappers Snapper sp. Monacanthidae Monacanthus chinensis 0.00 1.68 0.00 1.09 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Leatherjackets Fan Belly Leatherjacket Monacanthidae sp. 0.00 0.00 0.34 0.00 0.00 0.00 0.38 0.00 0.00 0.00 0.40 0.00 Leatherjacket sp. Monodactylidae Monodactylus argenteus 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.36 0.00 0.27 1.98 0.00 Moonfish Diamondfish Schuettea scalaripinnis 0.00 0.00 0.00 0.00 0.00 0.00 0.76 0.00 0.00 0.00 0.00 0.00 Eastern Pomfret Mugilidae Liza argentea 0.00 0.00 0.00 0.00 0.35 0.42 0.00 0.00 0.00 0.00 0.00 0.00 Mullets Flat Tail Mullet Odacidae Odacidae sp. 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.38 0.00 0.00 0.00 Weed Whitings/Cales Weed Whiting sp. (continued on next page) Author's personal copy

1508 A.C. McKinley et al. / Environmental Pollution 159 (2011) 1499e1509

Appendix 1 (continued)

Family Taxon Botany Bay Port Jackson Port Kembla Clyde River Jervis Bay Port Hacking Outer Inner Outer Inner Outer Inner Outer Inner Outer Inner Outer Inner Paralichthyidae Pseudorhombus jenynsii 0.00 0.00 0.00 0.00 0.00 0.00 1.83 0.94 0.00 0.27 0.00 0.47 Large Tooth Flounders Small Tooth Flounder Pempheridae Pempheridae sp. 0.42 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.37 0.00 0.00 0.00 Sweepers/Bullseyes Bullseye sp. Platycephalidae Platycephalus fuscus 0.00 0.00 0.00 0.00 0.00 0.35 1.14 2.20 1.60 0.00 2.29 0.00 Flatheads Dusky Flathead Platycephalus sp. 0.00 0.00 0.87 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Flathead sp. Sciaenidae Argyrosomus japonicus 0.00 0.00 0.00 0.00 0.35 0.00 0.76 0.00 0.00 0.00 0.00 0.00 Drums and Croakers Mulloway Atractoscion aequidens 0.00 0.00 0.00 0.00 0.35 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Teraglin Silliginidae Sillago ciliata 5.45 13.44 1.01 0.43 2.22 0.00 4.93 0.47 35.50 10.80 3.61 1.46 Whitings Sand Whiting Sillago flindersai 17.99 1.00 2.82 0.00 0.00 1.20 1.14 0.00 2.27 0.00 3.90 0.00 Eastern School Whiting Sillago maculata 1.09 34.15 0.00 8.46 1.41 0.41 4.12 0.24 1.68 12.61 1.57 0.32 Trumpeter Whiting Soleidae Soleidae sp. 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.36 0.37 0.00 1.98 0.53 True Soles Sole sp. Sparidae Acanthopagrus australis 0.00 0.00 0.00 0.00 0.00 0.00 0.39 0.00 3.20 4.59 0.00 0.00 Sea Breams Yellowfin Bream Sphyraenidae Sphyraena sp. 0.00 0.00 0.00 0.00 0.00 0.00 0.34 0.00 0.00 0.00 0.00 0.00 Barracudas Barracuda sp. Syngnathidae Stigmatopora nigra 0.00 0.00 0.00 0.00 0.00 0.51 0.00 0.83 0.40 0.27 3.49 0.00 Pipefish/Seahorses Wide Bodied Pipefish Urocampus carinirostris 0.79 0.00 0.00 0.00 0.00 0.00 0.38 0.47 0.00 0.96 1.96 0.00 Hairy Pipefish Vanacampus margaritifer 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.40 0.00 Mother of Pearl Pipefish Synodontidae Synodontidae sp. 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.45 0.00 0.00 0.00 0.00 Lizardfishes Lizardfish sp. Terapontidae Pelates sexlineatus 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.36 1.57 0.47 Grunter Perch Six Lined Trumpeter Tetraodontidae Tetraodontidae sp. 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.60 0.00 0.00 0.00 Pufferfish Toadfish sp. Tetrarogidae Centropogon australis 0.00 3.18 1.01 0.00 0.34 0.00 0.39 0.45 1.60 1.26 1.59 0.00 Waspfish Fortesque Tripterygiidae Tripterygiidae sp. 28.05 0.00 1.04 4.58 1.33 0.00 0.38 0.00 2.97 1.84 1.93 0.00 Triplefin Blennies Triplefin Blenny sp. Total Abundance 169.70 460.31 104.91 831.40 153.17 326.95 160.77 128.57 816.66 170.60 163.16 34.96

References Boesch, D., Turner, R., 1984. Dependence of fishery species on salt marshes: the role of food and refuge. Estuaries and Coasts 7, 460e468. Bridger, J.P., 1956. On day and night variation in catches of fish larvae. Journal du Alcaraz, C., Vila-Gispert, A., García-Berthou, E., 2005. Profiling invasive fish species: Conseil 22, 42e57. the importance of phylogeny and human use. Diversity and Distributions 11, Canli, M., Atli, G., 2003. The relationships between heavy metal (Cd, Cr, Cu, Fe, Pb, 289e298. Zn) levels and the size of six Mediterranean fish species. Environmental Alquezar, R., Markich, S.J., Booth, D.J., 2006. Metal accumulation in the smooth Pollution 121, 129e136. toadfish, Tetractenos glaber, in estuaries around Sydney, Australia. Environ- Dafforn, K.A., Glasby, T.M., Johnston, E.L., 2009. Links between estuarine condition and mental Pollution 142, 123e131. spatial distributions of marine invaders. Diversity and Distributions 15, 807e821. Anderson, M.J., 2001. A new method for non-parametric multivariate analysis of Dafforn, K.A., Simpson, S.L., Wong, C., Komyakova, V., Johnston, E.L. Comparing variance. Austral Ecology 26, 32e46. contaminant loads in urbanised estuaries: assessing the efficacy of a common Anderson, M.J., 2003. PCO e Principal Coordinate Analysis: A Computer Program. biomonitor, in press. University of Aukland, Aukland. Dorenbosch, M., van Riel, M.C., Nagelkerken, I., van der Velde, G., 2004. The rela- ANZECC, 2010. Chapter 3: Aquatic Ecosystems, Australian and New Zealand tionship of reef fish densities to the proximity of mangrove and seagrass Guidelines for Fresh and Marine Water Quality. Available from:. Australian and nurseries. Estuarine, Coastal and Shelf Science 60, 37e48. New Zealand Environmental Conservation Council, Canberra, Canberra http:// DPI, 2010. Survey of Recreational Fishing in New South Wales. Department of www.environment.nsw.gov.au/water/usinganzeccandwqos.htm (accessed Primary Industries, Sydney. 14.11.10). Drinkwater, K.F., Frank, K.T., 1994. Effects of river regulation and diversion on Arkoosh, M.R., Casillas, E., Clemons, E., Kagley, A.N., Olson, R., Reno, P., Stein, J.E., marine fish and invertebrates. Aquatic Conservation: Marine and Freshwater 1998. Effect of pollution on fish diseases: potential impacts on salmonid pop- Ecosystems 4, 135e151. ulations. Journal of Aquatic Animal Health 10, 182e190. Duke, N.C., Meynecke, J.-O., Dittmann, S., Ellison, A.M., Anger, K., Berger, U., Cannicci, S., Beck, M.W., Heck, K.L., Able, K.W., Childers, D.L., Eggleston, D.B., Gillanders, B.M., Diele, K., Ewel, K.C., Field, C.D., Koedam, N., Lee, S.Y., Marchand, C., Nordhaus, I., Halpern, B., Hays, C.G., Hoshino, K., Minello, T.J., Orth, R.J., Sheridan, P.F., Dahdouh-Guebas, F., 2007. A world without mangroves? Science 317, 41e42. Weinstein, M.P., 2001. The identification, conservation, and management of estu- Gray, C., McElligott, D., Chick, R., 1996. Intra- and inter-estuary differences in arine and marine nurseries for fish and invertebrates. BioScience 51, 633e641. assemblages of fishes associated with shallow seagrass and bare sand. Marine Bervoets, L., Blust, R., 2003. Metal concentrations in water, sediment and gudgeon and Freshwater Research 47, 723e735. (Gobio gobio) from a pollution gradient: relationship with fish condition factor. Gray, C.A., 1996. Intrusions of surface sewage plumes into continental shelf waters: Environmental Pollution 126, 9e19. interactions with larval and presettlement juvenile fishes. Marine Ecology Bervoets, L., Knaepkens, G., Eens, M., Blust, R., 2005. Fish community responses to Progress Series 139, 31e45. metal pollution. Environmental Pollution 138, 338e349. Gray, C.A., 1997. Field assessment of numerical impacts of coastal sewage disposal Birch, G., Taylor, S., 1999. Source of heavy metals in sediments of the Port Jackson on fish larvae relative to natural variability. Environmental Biology of Fishes 50, estuary, Australia. Science of the Total Environment 227, 123e138. 415e434. Author's personal copy

A.C. McKinley et al. / Environmental Pollution 159 (2011) 1499e1509 1509

Gray, C.A., Miskiewicz, A.G., 2000. Larval fish assemblages in South-east Australian Miskiewicz, A.G., Gibbs, P.J., 1994. Organochlorine pesticides and hexa- coastal waters: seasonal and spatial structure. Estuarine, Coastal and Shelf chlorobenzene in tissues of fish and invertebrates caught near a sewage outfall. Science 50, 549e570. Environmental Pollution 84, 269e277. Gray, C.A., Otway, N.M., Laurenson, F.A., Miskiewicz, A.G., Pethebridge, R.L., 1992. Neira, F.J., Miskiewicz, A.G., Trnski, T., 1998. Larvae of Temperate Australian Fishes: A Distribution and abundance of marine fish larvae in relation to effluent plumes Laboratory Guide for Larval Fish Identification. University of Western Australia from sewage outfalls and depth of water. Marine Biology 113, 549e559. Press, Perth. Guo, Y., Meng, X.-Z., Tang, H.-L., Zeng, E.Y., 2008. Tissue distribution of organo- Neira, F.J., Potter, I.C., 1992. Movement of larval fishes through the entrance channel chlorine pesticides in fish collected from the Pearl River Delta, China: impli- of a seasonally open estuary in Western Australia. Estuarine, Coastal and Shelf cations for fishery input source and bioaccumulation. Environmental Pollution Science 35, 213e224. 155, 150e156. Norcross, B.L., Shaw, R.F., 1984. Oceanic and estuarine transport of fish eggs and Harlin, M.M., Thorne-Miller, B., 1981. Nutrient enrichment of seagrass beds in larvae: a review. Transactions of the American Fisheries Society 113, 153e165. a Rhode Island coastal lagoon. Marine Biology 65, 221e229. NSW, 1999. Marine Parks (Zoning Plans) Regulation of 1999 Division Marine Park Henry, G.W., Lyle, J.M., 2003. The National Recreational and Indigenous Fishing Zones. New South Wales Government, Sydney. Available from: www.austlii. Survey. Commonwealth of Australia, Department of Agriculture, Fisheries, edu.au/au/legis/nsw/consol_reg/mppr1999362/ (accessed 12.11.10). Forestry, Canberra. NSWDNR, 2010. Estuaries in New South Wales. New South Wales Department of Hill, N.A., King, C.K., Perrett, L.A., Johnston, E.L., 2009. Contaminated suspended Natural Resources, Sydney. Available from: (accessed 14.11.10). effects. Environmental Toxicology and Chemistry 28, 409e417. Piola, R.F., Johnston, E.L., 2008. Pollution reduces native diversity and increases Hose, J.E., Cross, J.N., Smith, S.G., Diehl, D., 1989. Reproductive impairment in a fish invader dominance in marine hard-substrate communities. Diversity and inhabiting a contaminated coastal environment off Southern California. Envi- Distributions 14, 329e342. ronmental Pollution 57, 139e148. Pittman, S.J., McAlpine, C.A., 2003. Movements of marine fish and decapod crusta- Hu, J.Y., Zhang, Z.B., Wei, Q.W., Zhen, H.J., Zhao, Y.B., Peng, H., Wan, Y., Giesy, J.P., ceans: process, theory and application. Advances in Marine Biology 44, 205e294. Li, L.X., Zhang, B., 2009. Malformations of the endangered Chinese sturgeon, Potter, I.C., Cheal, A.J., Loneragan, N.P., 1988. Protracted estuarine phase in the life Acipenser sinensis, and its causal agent. Proceedings of the National Academy of cycle of the marine pufferfish Torquigener pleurogramma. Marine Biology 98, Sciences of the United States of America 106, 9339e9344. 317e329. Hutchinson, T.H., Williams, T.D., Eales, G.J., 1994. Toxicity of cadmium, hexavalent Prange, J.A., Dennison, W.C., 2000. Physiological responses of five seagrass species chromium and copper to marine fish larvae (Cyprinodon variegatus) and to trace metals. Marine Pollution Bulletin 41, 327e336. copepods (Tisbe battagliai). Marine Environmental Research 38, 275e290. Quinn, G., Keough, M., 2002. Experimental Design and Data Analysis for Biologists. Isosaari, P., Hallikainen, A., Kiviranta, H., Vuorinen, P.J., Parmanne, R., Koistinen, J., Cambridge University Press, Cambridge. Vartiainen, T., 2006. Polychlorinated dibenzo-p-dioxins, dibenzofurans, biphe- Robertson, A., Duke, N., 1987. Mangroves as nursery sites: comparisons of the nyls, naphthalenes and polybrominated diphenyl ethers in the edible fish caught abundance and species composition of fish and crustaceans in mangroves and from the Baltic Sea and lakes in Finland. Environmental Pollution 141, 213e225. other nearshore habitats in tropical Australia. Marine Biology 96, 193e205. Iwata, A., Hosoya, S., Larson, H.K., 2001. Paedogobius kimurai, a new genus and Robinet, T.T., Feunteun, E.E., 2002. Sublethal effects of exposure to chemical species of goby (Teleostei: Gobioidei: Gobiidae) from the West Pacific. Records compounds: a cause for the decline in Atlantic eels? Ecotoxicology 11, 265e277. of the Australian Museum 53, 103e122. Rogers, C.E., Brabander, D.J., Barbour, M.T., Hemond, H.F., 2002. Use of physical, Jelbart, J., Ross, P., Connolly, R., 2007. Fish assemblages in seagrass beds are influ- chemical, and biological indices to assess impacts of contaminants and physical enced by the proximity of mangrove forests. Marine Biology 150, 993e1002. habitat alteration in urban streams. Environmental Toxicology and Chemistry Jenkins, G.P., May, H.M.A., Wheatley, M.J., Holloway, M.G., 1997. Comparison of fish 21, 1156e1167. assemblages associated with seagrass and adjacent unvegetated habitats of Port Saintilan, N., Williams, R.J., 1999. Mangrove transgression into saltmarsh environ- Phillip Bay and Corner Inlet, Victoria, Australia, with emphasis on commercial ments in south-east Australia. Global Ecology and Biogeography 8, 117e124. species. Estuarine, Coastal and Shelf Science 44, 569e588. Scanes, P., 2010. NSW Estuarine Catchment Disturbance Ranks. NSW Department of Johnston, E., Roberts, D.A., 2009. Contaminants reduce the richness and evenness of Environment, Climate Change, and Water, Sydney. marine communities: a review and meta-analysis. Environmental Pollution 157, Shipley, B., 2002. Cause and Correlation in Biology: A User’s Guide to Path Analysis, 1745e1752. Structural Equations and Causal Inference. Cambridge University Press, London. Jones, J.C., Reynolds, J.D., 1997. Effects of pollution on reproductive behaviour of Sreedevi, P., Sivaramakrishna, B., Suresh, A., Radhakrishnaiah, K., 1992. Effect of fishes. Reviews in Fish Biology and Fisheries 7, 463e491. nickel on some aspects of protein metabolism in the gill and kidney of the Karas, P., Neuman, E., Sandstrom, O., 1991. Effects of a pulp mill effluent on the freshwater fish, Cyprinus carpio L. Environmental Pollution 77, 59e63. population dynamics of perch Perca fluviatilis. Canadian Journal of Fisheries & Taylor, M.D., Palmer, P.J., Fielder, D.S., Suthers, I.M., 2005. Responsible estuarine finfish Aquatic Sciences 48, 28e34. stock enhancement: an Australian perspective. Journal of Fish Biology 67, 299e331. Kennish, M.J., 2002. Environmental threats and environmental future of estuaries. Tilman, D., May, R.M., Lehman, C.L., Nowak, M.A., 1994. Habitat destruction and the Environmental Conservation 29, 78e107. extinction debt. Nature 371, 65e66. Kienle, C., Köhler, H.R., Filser, J., Gerhardt, A., 2008. Effects of nickel chloride and Valiela, I., Bowen, J.L., York, J.K., 2001. Mangrove forests: one of the world’s oxygen depletion on behaviour and vitality of zebrafish (Danio rerio, Hamilton, threatened major tropical environments. BioScience 51, 807 e815. 1822) (Pisces, Cypriniformes) embryos and larvae. Environmental Pollution 152, van der Oost, R., Beyer, J., Vermeulen, N.P.E., 2003. Fish bioaccumulation and 612e620. biomarkers in environmental risk assessment: a review. Environmental Toxi- Kingsford, Michael J., Suthers, Iain M., Gray, Charles A., 1997. Exposure to sewage cology and Pharmacology 13, 57e149. plumes and the incidence of deformities in larval fishes. Marine Pollution Vetemaa, Markus, Forlin, Lars, Sandstrom, Olof, 1997. Chemical industry effluent Bulletin 33, 201e212. impacts on reproduction and biochemistry in a North Sea population of Kojadinovic, J., Potier, M., Le Corre, M., Cosson, R.P., Bustamante, P., 2007. viviparous blenny (Zoarces viviparus). Journal of Aquatic Ecosystem Stress & Bioaccumulation of trace elements in pelagic fish from the Western Indian Recovery 6, 33e41. Ocean. Environmental Pollution 146, 548e566. Walker, D.I., McComb, A.J., 1992. Seagrass degradation in Australian coastal waters. Longstaff, B.J., Dennison, W.C., 1999. Seagrass survival during pulsed turbidity Marine Pollution Bulletin 25, 191e195. events: the effects of light deprivation on the seagrasses Halodule pinifolia and Waring, C.P., Stagg, R.M., Fretwell, K., McLay, H.A., Costello, M.J., 1996. The impact of Halophila ovalis. Aquatic Botany 65, 105e121. sewage sludge exposure on the reproduction of the sand goby, Pomatoschistus Lotze, H.K., Lenihan, H.S., Bourque, B.J., Bradbury, R.H., Cooke, R.G., Kay, M.C., minutus. Environmental Pollution 93, 17e25. Kidwell, S.M., Kirby, M.X., Peterson, C.H., Jackson, J.B.C., 2006. Depletion, degrada- Warnau, M., Ledent, G., Temara, A., Bouquegneau, J.-M., Jangoux, M., Dubois, P., tion, and recovery potential of estuaries and coastal seas. Science 312, 1806e1809. 1995. Heavy metals in Posidonia oceanica and Paracentrotus lividus from sea- Macinnis-Ng, C.M.O., Ralph, P.J., 2002. Towards a more ecologically relevant grass beds of the north-western Mediterranean. Science of the Total Environ- assessment of the impact of heavy metals on the photosynthesis of the sea- ment 171, 95e99. grass, Zostera capricorni. Marine Pollution Bulletin 45, 100e106. Waycott, M., Duarte, C.M., Carruthers, T.J.B., Orth, R.J., Dennison, W.C., Olyarnik, S., McArdle, B.H., Anderson, M.J., 2001. Fitting multivariate models to community data: Calladine, A., Fourqurean, J.W., Heck, K.L., Hughes, A.R., Kendrick, G.A., a comment on distance-based redundancy analysis. Ecology 82, 290e297. Kenworthy, W.J., Short, F.T., Williams, S.L., 2009. Accelerating loss of seagrasses McKinley, A., Johnston, E.L., 2010. Impacts of contaminant sources on marine fish across the globe threatens coastal ecosystems. Proceedings of the National abundance and species richness: a review and meta-analysis of evidence from Academy of Sciences of the United States of America 106, 12377e12381. the field. Marine Ecology Progress Series 420, 175e191. Wirgin, I., Waldman, J.R., 1998. Altered gene expression and genetic damage in McKinley, A., Ryan, L., Coleman, M., Knott, N., Clarke, G., Taylor, M., Johnston, E.L. North American fish populations. Mutation Research 399, 193e219. Putting marine sanctuaries into context: a comparison of estuary fish assem- Wirgin, I., Waldman, J.R., 2004. Resistance to contaminants in North American fish blages over multiple levels of protection and disturbance. Estuarine Coastal & populations. Mutation Research/Fundamental and Molecular Mechanisms of Shelf Science, in review. Mutagenesis 552, 73e100. Miskiewicz, A.G., 1987. Taxonomy and ecology of fish larvae in Lake Macquarie and Xie, L., Klerks, P.L., 2004. Changes in cadmium accumulation as a mechanism for New South Wales coastal waters, PhD thesis, University of New South Wales, cadmium resistance in the least killifish Heterandria formosa. Aquatic Toxi- Sydney, p. 191. cology 66, 73e81.

Chapter 5

ANTHROPOGENIC ACTIVITIES DIFFERENTIALLY IMPACT FISH GUILDS: THE IMPORTANCE OF UNDERSTANDING LIFE HISTORY CHARACTERISTICS

Unpublished Version:

Anthropogenic activities differentially impact fish guilds: The importance of understanding life history characteristics

Andrew C. McKinley**, Cian Foster-Thorpe*, Anthony Miskiewicz1, Matthew D. Taylor* & Emma L. Johnston*

** [email protected] +61 04 2586 7081 *Evolution & Ecology Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, New South Wales 2052, Australia 1Environment and Recreation, Wollongong City Council, 41 Burelli Street, Wollongong, New South Wales 2500, Australia

Keywords: contaminants, pollution, fish larvae, habitat modification, fish guilds, early life history

Abstract

Significant literature exists which classifies fish into various guilds according to reproductive and life cycle characteristics. However, guild approaches focusing on these characteristics have never been used to investigate impact of anthropogenic activities on sensitive fish larvae. Our study represents the first use of fish larvae reproductive and estuary usage guilds to evaluate the impacts of anthropogenic modification. Larval fish communities were collected at two depths using a surface tow and a bottom trawl. We sampled multiple sites within the inner and outer zones of three heavily modified and three relatively unmodified estuaries. The relative abundance of major guilds differed significantly between the surface and bottom samples, and between heavily modified vs. relatively unmodified estuaries. Larvae of estuarine species and benthic spawners were significantly more abundant in the bottom waters of heavily modified estuaries. The abundance of estuarine species and benthic spawners was strongly related to sediment metal contamination and seagrass cover. In contrast, estuarine opportunist species trended towards higher abundance in the surface waters of relatively unmodified systems, while other guilds did not respond significantly to sampling depth or modification. The abundance of the estuarine opportunist guild was strongly related to decreased temperatures, salinity, and sediment metals; and increased coverage of seagrass. This guild also showed a strong relationship to the width of the estuary mouth, in contrast to estuarine and benthic guilds which showed no relationship to the width of the estuary mouth. Overall, greater impacts were observed in the epibenthic compared to surface samples, suggesting that epibenthic sampling alone is sufficient for future monitoring. This study demonstrates that habitat modification and sediment contamination have substantially different relationships to larvae of different ecological guilds. The life history characteristics of early life stages represent a potentially valuable tool for predicting the sensitivity of fish to anthropogenic impacts.

Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Introduction

Robust methods are needed to monitor the impacts of anthropogenic activities in marine ecosystems. Central to these methods is the ability to effectively identify and evaluate impacts on sensitive ecosystem components. Guild approaches which classify species into discrete groups based on shared ecological characteristics have been used extensively in terrestrial systems to quantify and understand anthropogenic impacts on bats (Klingbeil and Willig, 2009), birds (Poulin et al., 2010), insects (Woodcock et al., 2009), and plants (Holmes and Cowling, 1997). In marine environments, feeding, trophic, and foraging guilds have been used extensively to evaluate anthropogenic impacts on adult fish and other organisms (Pinnegar et al., 2002). Significant literature also exists in marine systems which classify species according to reproductive and life cycle characteristics (Elliott et al., 2007). Guild approaches focusing on these characteristics may be a valuable tool for monitoring impacts on sensitive early life stages and for identifying susceptible functional groups. Our study is the first to utilize guild classifications to examine the response of fish larvae to anthropogenic modification.

The relative abundance of larval guilds may be impacted by a variety of natural conditions including the depth sampled, estuarine geomorphology, vegetative habitat coverage, and physico- chemical conditions. Previous studies have shown that some taxa are strongly associated with a portion of the water column, and so sampling depth can strongly influence larval abundance and distribution (Gray et al., 1992). Since many larvae are dispersed by ocean currents and tidal forces, the size and geomorphology of an estuary and the width of the estuary mouth may also play a role in determining assemblage composition (Neira and Potter, 1992; Neira et al., 1992), while spatial and temporal variability in physico-chemical parameters (e.g. pH, salinity, and temperature) have been shown to influence short-term variability in fish distributions and assemblages (Jaureguizar et al., 2004; Kupschus and Tremain, 2001; McKinley et al., 2011b; Prista et al., 2003). In addition, the

2 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012. extent of seagrass, mangrove, and saltmarsh coverage in an estuarine system may also impact larval distributions. Many fish utilize these vegetative habitats for egg laying and larvae may rely on them for shelter and food (Miskiewicz, 1987; Potter et al., 1990).

The way in which these factors influence fish assemblages can be impacted by anthropogenic modification, including the addition of non-natural stressors such as contaminants. It is well known that many estuarine systems are exposed to high levels of human disturbance and modification, due to their proximity to human settlement and upstream agricultural and industrial activity (Kennish, 2002; Lotze et al., 2006). Consequently, some estuaries are highly contaminated, and the presence of these contaminants may impact the composition and diversity of ecological communities (Johnston and Roberts, 2009). Moreover, toxicological evidence points to developmental and larval susceptibility of fish to contaminants (Jones and Reynolds, 1997;

Kingsford et al., 1997).

Despite the potential vulnerability of fish larvae to environmental disturbance and the high degree of management, conservation, and scientific interest in fish communities, relatively little research has been conducted examining anthropogenic impacts on larval assemblages (McKinley and Johnston, 2010). Our study represents the first comparison of reproductive and estuary usage guilds across two sampling depths (surface and bottom), in relation to anthropogenic modification in heavily modified vs. relatively unmodified estuaries. For the first time, we examine guild abundance and incorporate a variety of natural factors (estuary geomorphology, vegetative habitat coverage, physico-chemical conditions) and measures of anthropogenic stress (estuarine modification and sediment trace metals contamination). These covariates are utilized in a multivariate analysis in order to determine their influence on guild abundance, community composition, and the distribution of taxa. In addition, our study takes place in both heavily modified

3 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012. and relatively unmodified estuaries in New South Wales, Australia. We hypothesize several outcomes:

1. We predict that the relative abundance of guilds will be significantly different between

surface and bottom samples, and also between heavily modified and relatively unmodified

estuaries.

2. We predict that composition of the community will vary significantly by sampling depth,

and also between heavily modified and relatively unmodified estuaries.

3. We predict that the abundance of pelagic spawners and estuarine opportunists guilds will be

strongly correlated to estuary geomorphology. In contrast, we predict that the benthic

spawners and estuarine guilds will show a stronger relationship to vegetative cover and

sediment metal contamination.

4 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Methods

Estuarine Characteristics, Location, and Sampling Methods

Sampling was undertaken in six estuaries along the south coast of New South Wales,

Australia (Figure 1). These estuaries are all permanently open tidal systems, though they vary in their size, geomorphic structure, the width of the estuary mouth, and in the proportion of the estuary that contains vegetative habitats. In order to estimate the influence of these factors on observed larvae distribution and abundance patterns, measures of estuarine structure and habitat characteristics have been included in the multivariate analysis (Table 2). Estuary size was included in this analysis as a measure of structural variability, while the width of the estuary mouth is included in order to approximate the exposure of the estuary to coastal currents and tidal flushing.

The extent of vegetative habitats was calculated using estuarine fact sheet monitoring values

(NSWDNR, 2010).

The six estuaries are also exposed to varying levels of anthropogenic modification. Three of these estuaries (Port Jackson - 33°44.258’S, 151°16.542’E, Botany Bay - 33°59.352’S,

151°11.433’E, and Port Kembla - 34°28.121’S, 150°54.410’E) were classified as heavily modified estuaries. In contrast, three estuaries were classified as relatively unmodified (Port Hacking -

34°04.680’S, 151°09.311’E, Jervis Bay - 35°04.762’S, 150°44.858’E, and the Clyde River -

35°44.233’S, 150°14.272’E) (Figure 1). The three heavily modified estuaries are subject to high levels of historic and contemporary pollution, greater recreational fishing activity, urban and industrial development in the majority of their shorelines and catchments, and greater commercial and recreational boating activity (Birch and Taylor, 1999; DPI, 2010; Henry and Lyle, 2003;

Scanes, 2010). In contrast, all of these stressors are substantially reduced in the relatively unmodified estuaries.

5 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Within each estuary, six sites were identified with bottom characteristics which allowed uninterrupted trawling. Sampling was conducted in both the inner and outer areas of the estuary, which were defined as distinct sampling zones. Sampling was divided evenly between these inner and outer ‘zones’, with three sampling sites in each. Zones were defined according to their physical and biological characteristics. The outer zone sites are close to the marine entrance of the estuaries where salinity, coastal flushing, wave exposure and oceanic current systems are expected to have greater influence. In the outer zone, there is also greater tidal influence and sediment grain sizes are larger (Dafforn et al., 2012; McKinley et al., 2011b). In contrast, the inner zone sites are further from the estuary mouth, within the lower reaches of the estuarine tributary. In this area anthropogenic disturbance factors are expected to be more significant. Turbidity, temperatures, and nutrient levels are expected to be higher than in the outer zone (Dafforn et al., 2012; McKinley et al., 2011b). Sampling was temporally replicated, with the first sampling event occurring in

November 2009, and the second in February 2010.

Larvae were sampled in both the surface and bottom waters using a surface plankton tow and a bottom trawl. For each method the net frame housed two nets (i.e. bongo nets), and each consisted of 50 cm x 300 cm (long) conical plankton net with 250 μm mesh. The surface sample was conducted with the net approximately 45 degrees off the starboard side, 20 m behind the boat throughout a clockwise arc, at a depth of 1-1.5 m. At the same sites and times a bottom sled trawl was towed along bare sediment along flat profiles at 5-12 m depth. GPS was used for both tows and trawls to ensure that all sampling occurred at a speed of 1.5-2.5 knots for 250 m. The sled was rigged to a four-point bridle using an approximate 3:1 warp to depth ratio. The sled consisted of a stainless steel frame measuring approximately 1.5 m across and 2 m long. The net frame was mounted within the sled such that the net was 15 cm off the bottom during trawls, and a series of 15 cm vertical “tickler-chains” dangled from the base of the net and made contact with the sediment. A

6 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

1 L plastic sample jar was affixed to the net cod end of all nets, as well as a calibrated flow meter

(Model P/N 2030R Flow Meter, Underwater Video Systems PTY LTD, Sydney, Australia) to standardize larval abundance results to 100 m3 of sea water. All sampling was conducted at night following the incoming tide up the estuary. In order to standardize light conditions sampling was conducted each month within a two-week window around the new moon (Pittman and McAlpine,

2003). All samples were immediately preserved in a buffered 5% formalin/seawater mixture for transportation back to the laboratory.

7 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

-33.920 -34.040 -33.780 abc -33.940 -34.050 -33.800 -33.960 -34.060 -33.820 -33.980 -34.070 -33.840 -34.000 -34.080 -33.860 -34.020 -34.090 Port Jackson -33.880 -34.040 Botany Bay -34.100 -33.900 0 km 3km Port Hacking 0 km 5km 0 km 5km -34.060 -34.110 151.140 151.160 151.180 151.200 151.100 151.150 151.200 151.250 Port Kembla 151.090 151.110 151.130 151.150 151.170

-34.440 -34.980 Jervis Bay de-35.690 f -35.000 -34.445 -35.700

-35.020 Clyde River -35.710 -34.450 -35.720 -35.040 0 km 100km -35.730 -34.455 -35.060 -35.740

-35.750 -35.080 -34.460 -35.760

-35.100 0 km 5km -35.770 -34.465 -35.780 -35.120 0 km 1km -35.790 0 km 3km -34.470 -35.140 150.880 150.890 150.900 150.910 1 150.650 150.700 150.750 150.800 150.850 150.160 150.180 150.200 150.220 150.240 150.260 150.280

Figure 1: Location of study sites in a) Port Jackson, b) Botany Bay, c) Port Hacking, d) Port Kembla, e) Jervis Bay, and f) Clyde River estuaries. ♦ Indicates outer zone sites ● Indicates inner zone sites.

8 Physico-Chemical and Sediment Metals Covariates

In addition to the estuary structure and habitat characteristics discussed above, basic water quality variables were also measured and included in multivariate analysis. This was done in order to estimate the influence of phyisco-chemical factors on observed results. At each sampling location temperature, salinity, and pH were measured using a YSI 6820 V2 sonde (calibrated weekly).

Surficial sediments were collected by Dafforn et al. (2012) from sites close to the larval samples, and analyzed for concentrations of trace metals. These samples were oven-dried, digested, and analyzed with Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) following the methods outlined by Hill et al. (2009). Recoveries were calculated against certified reference materials and were within accepted recovery limits. Full details of analyses and contaminant datasets are presented in Dafforn et al. (2012). Prior to analysis a total metals contamination quotient was calculated. This included values for Co, Cr , Cu, Fe, Mn, Ni, Pb and

Zn. Each individual metal contaminant load was divided by the low and high trigger values from the

ANZECC sediment quality guidelines (ANZECC, 2010). The high and low quotients for each metal were then summed for each sample and divided by two to give a mean trace metal sediment quality guideline quotient (mSQGQ) for each sample (Long, 2006).

Sediments were selected to measure contaminants in these systems (rather than a water column measure of contamination) for several reasons. First, it is well known that fish accumulate contaminants through their food to a much greater degree than through their gills or through interaction with contaminated water (Dallinger et al., 1987; Hall et al., 1997). The majority of species in this study are benthic or benthopelagic foragers and so most would interact with sediments regularly during feeding (Edgar and Shaw, 1995). Second, contaminants accumulate in estuarine sediments over the long-term, as such sediment metals values are less temporally variable and represent a contemporary threat from historical pollution sources (Knott et al., 2009).

Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Larval Sorting and Classification

Larvae samples were sorted and identified to species following Neira et al. (1998). Current taxonomic knowledge was insufficient for all taxa to be identified to the species level, such taxa were instead identified to genus or family (few taxa could not be identified to species level;

Appendix 1). A variety of Fa. Gobiidae were distinguished as morphologically distinct taxa and were included as separate ‘species’ within the analysis, even though a particular species could not be assigned to these groups.

Each species was classified into discrete larval guilds based on their early life history characteristics. Two classifications were made for each species; one based on their pattern of estuary usage during their life cycle and one based on their spawning strategy. These classifications were made through a review of existing literature for relevant species (Elliott et al., 2007;

Loneragan and Potter, 1990; Loneragan et al., 1989; Neira et al., 1998; Neira et al., 1992; Potter et al., 1990; Potter and Hyndes, 1999) and through consultation with regional larval fish experts

(Miskiewicz, personal communication, 2011).

Species were classified by estuarine usage during their life cycle following the standard classification system outlined by Elliot et al. (2007). Three broad categories were used in this study: x Estuarine guild: Species which spawn within the estuary and which normally complete their

entire life cycle within the estuarine environment. x Estuarine opportunist guild: Species which primarily spawn in marine coastal waters but enter

the estuary either in their larvae or juvenile stages. Many of these species require sheltered

estuarine habitats during their larval and juvenile stages and are hence dependent on the

estuarine environments for reproduction. Most species spend part of their adult stage outside of

estuaries. Opportunist in this case refers to the ability of the guild to colonize estuarine

environments, and does not refer to r/k strategists.

10 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012. x Marine straggler guild: Species which spawn at sea and normally enter estuaries only in low

numbers, occurring most frequently in the lower reaches of the estuary. Many are stenohaline

and are primarily associated with coastal marine waters.

Species were also classified according to their spawning strategy (Elliott et al., 2007). Four categories were identified in this study. The majority of species were benthic spawners (species which deposit their eggs on the sediment, vegetation or hard surfaces) or pelagic spawners (species which broadcast floating eggs in the water column). A small number of species were also identified as brooders (species which construct and guard benthic nests) or viviparous (species which bear live young). Detailed classifications and species information are presented in Appendix 1.

Statistical Analysis

Mixed-model PERMANOVA was utilized for all multivariate and univariate analysis

(Anderson, 2001). All analysis was conducted with the following statistical design:

Dis – Disturbance – Heavily Modified or Relatively Unmodified (2 levels, Fixed) De – Depth – Surface Tow or Bottom Trawl (2 levels, Fixed) Ti – Time – November or February (2 levels, Fixed) Zo – Zone – Inner or Outer (2 levels, Fixed) Es – Estuary (Modification) – (6 estuaries, Random) Si – Site (Estuary(Modification) x Zone) – (36 sites, Random)

In order to create a more normal distribution and to reduce the influence of highly dominant taxa, a log(x+1) transformation was applied to all abundance data prior to analysis. All univariate tests were conducted using Euclidean distance matrices, while multivariate tests were conducted with Bray-Curtis similarity matrices. Monte Carlo P-values are presented for some tests where the number of unique permutations was less than 20. These are marked in the text as p*.

PERMANOVA results are summarized and presented in text for higher level fixed factors. Only significant interactions between fixed factors are presented in text, while non-significant

11 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012. interactions or interactions between random factors are not discussed. Interactions with random factors are ignored as the test of higher level fixed factors remains relevant regardless of interactions with random factors (Quinn and Keough, 2002).

Covariate factors were analyzed in a multivariate analysis known as metric multi- dimensional scaling. This is also known as Principal Coordinated Ordination (PCO) (Anderson,

2003). PCO is a PERMANOVA function that performs a principal coordinate analysis of any symmetric distance matrix. All variables in Table 1 & 2 were included in the PCO analysis, however, only those which correlated highly are displayed in the final PCO charts. This includes physico-chemical and estuary habitat characteristics with a multiple correlation factor >0.2 (Figure

5a) and species with a Pearson correlation factor >0.35 (Figure 5c,d). Variables which correlated below these thresholds did not show clearly discernible trends in the PCO charts, and are thus omitted.

12 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Results

Larval Fish Assemblages

Over 101 taxa were represented by 24,600 larvae collected in epibenthic samples(~73%) and

9,000 larvae collected in the surface samples (~27%). Larval species richness was greater in the bottom trawl than the surface trawl (MS1 = 124.69, p = 0.022), but did not differ significantly by disturbance category, time, or zone (p > 0.05). Shannon diversity did not differ significantly by zone, disturbance category, or sampling depth (p > 0.05), but was significantly higher in the second round of sampling (MS1 = 3.10, p = 0.015). Multivariate analysis of the community composition found that the larval assemblage differed significantly by disturbance category (MS1 = 14110, p* =

0.016), time of sampling (MS1 = 10653, p = 0.021), and by sampling depth (MS1 = 28270, p =

0.017), but not by zone (p > 0.05) (Figure 2). Differences in community composition between sampling times were driven primarily by a significant increase in the abundance of two common species in the second round of sampling. These were Arenigobius bifrenatus (MS1 = 21.50, p =

0.002) and Gerres subfasciatus (MS1 = 59.08, p = 0.003). Other highly abundant species and all guilds did not differ significantly between sampling times (discussed below).

13 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Figure 2: Three dimensional MDS plot of multivariate assemblage composition by disturbance category and sampling depth. Symbols represent centroids of the assemblage composition. Heavily modified includes sites in Port Jackson, Botany Bay, and Port Kembla. Relatively unmodified includes sites in Port Hacking, the Clyde River, and Jervis Bay.

Abundance of Larval Guilds

Significant differences were found in the relative abundance of larval guilds both between heavily modified and relatively unmodified estuaries and by sampling depth. When classified by estuarine usage, the relative abundance of larval guilds differed significantly between heavily modified and relatively unmodified estuaries (MS1 = 2536.4, p* = 0.012) but not by sampling depth,

14 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012. time, or zone (p > 0.05). Differences between heavily modified and relatively unmodified estuaries were due primarily to a significant increase in the abundance of estuarine guild larvae in the heavily modified estuaries (MS1 = 23.6, p* = 0.037). This guild was particularly abundant in the bottom samples of the heavily modified estuaries, resulting in a significant disturbance x depth interaction

(MS1 = 17.5, p = 0.019) (Figure 3a). There was also a trend towards increased abundance of the estuarine-opportunist guild in the surface samples of the relatively unmodified estuaries, however, this was not significant (disturbance x depth interaction, MS1 = 9.87, p* = 0.108; Figure 3b). Marine stragglers did not change with modification (Figure 3a,b).

When classified by spawning strategy, the relative abundance of spawning guilds did not differ significantly by any factor (p > 0.05). When analyzed individually, however, the abundance of the benthic spawners guild was significantly higher in the bottom samples of the heavily modified estuaries. This produced a significant disturbance x depth interaction (MS1 = 17.4, p =

0.010) (Figure 3c,d). Other spawning guilds did not differ significantly by depth or modification (p

> 0.05).

15 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

a – otto – ur ace

ea ly o e stuar es elat e l y n o e stuar es ea ly o e stuar es elat e l y n o e stuar es

c – otto – ur ace

ea ly o e stuar es elat e l y n o e stuar es ea ly o e stuar es elat e l y n o e stuar es

Figure 3: Mean ± SE larval abundance by estuary for each larval guild. Classified as a) Bottom by estuary usage, b) Surface by estuary usage, c) Bottom by spawning guild, and d) Surface by spawning guild. Significant difference between heavily modified and relatively unmodified estuaries for relative abundance of estuary usage guilds (MS1 = 2536.4, p* = 0.012), estuarine taxa in bottom samples (MS1 = 17.5, p = 0.019), and benthic spawners in bottom samples (MS1 = 17.4, p = 0.010).

16 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Individual Species Distributions

Univariate tests were conducted on the six most abundant species in the study, which collectively accounted for approximately 60% of the assemblage. Ambassis jacksoniensis was significantly more abundant in the relatively unmodified estuaries (MS1 = 48.77, p* = 0.003), but did not differ significantly by time, zone, or depth (p >0.05) (Figure 4a). Arenigobius bifrenatus was more abundant in the heavily modified estuaries (MS1 = 27.83, p* = 0.049), while it also showed a significant increase in the second round of sampling (MS1 = 21.50, p = 0.002). This species did not differ significantly by zone or depth (p > 0.05) (Figure 4c). Gobiopterus semivestita was approximately equally abundant between depths, but was significantly more abundant in bottom trawls conducted in heavily modified estuaries. This resulted in a significant disturbance x depth interaction (MS1 = 19.44, p = 0.048), while this species did not differ significantly by zone or time

(p > 0.05) (Figure 4e). Similarly, Paedogobius kimurai also showed a significant disturbance x depth interaction (MS1 = 62.17, p = 0.020), while it also did not differ significantly by other factors

(p > 0.05). Paedogobius kimurai was most abundant in the bottom trawl and only differed significantly by disturbance category at this depth, being more abundant in the heavily modified estuaries (Figure 4f). Pseudogobius sp. was significantly more abundant in the surface waters (MS1

= 73.52, p = 0.001) and did not show any significant interactions. Pseudogobius did not differ significantly by other factors (p > 0.05) (Figure 4d). In contrast, Hypherlophus vittatus did not differ significantly by any high-level factors (p > 0.05) and was the most uniformly distributed species in this study (Figure 4b).

17 a c ort ac son lass sh an y prat al r l e o y Ambassis jacksoniensis Hypherlophus vittatus Arenigobius bifrenatus

ea ly o e stuar es elat e l y n o e stuar es ea ly o e stuar es elat e l y n o e stuar es ea ly o e stuar es elat e l y n o e stuar es

e astern luespot o y ransparent o y e ape ae o orph c o y Pseudogobius sp. Gobiopterus semivestita Paedogobius kimurai

ea ly o e stuar es elat e l y n o e stuar es ea ly o e stuar es elat e l y n o e stuar es ea ly o e stuar es elat e l y n o e stuar es

Figure 4: Occurrence of six most abundant species by depth and modification category. 18

Estuary Characteristics

Overall, average physico-chemical conditions were similar between heavily modified and relatively unmodified disturbance categories (Table 1). However, temperature (MS4 = 2.23, p =

0.001), salinity (MS4 = 3.44, p = 0.001), and pH (MS4 = 3.75, p = 0.001) did differ significantly by estuary. Temperature (MS1 = 14.76, p = 0.026) and pH (MS1 = 20.11, p = 0.051) also differed significantly by zone, with higher temperatures and lower pH values in inner zone sites (Table 1).

Significantly higher sediment metal values were recorded in the heavily modified estuaries

(MS1 = 2.23, p* = 0.001) (Table 1). Metals values showed a trend towards being particularly high in the inner zone of the heavily modified estuaries, though this did not produce a significant disturbance x zone interaction (p > 0.05). At sites in many of the heavily modified estuaries, sediment metal values were above levels predicted to have biological effects according to national guidelines (ANZECC, 2010; Dafforn et al., 2012).

Relatively unmodified estuaries had greater coverage of mangroves (7.9%) and seagrass

(6.3%) compared to the heavily modified estuaries (1.8%, 2.6% respectively) (NSWDNR, 2010).

The relatively unmodified estuaries had lower coverage of saltmarsh (1.7%) compared to the heavily modified estuaries (4.8%). The comparatively high coverage of saltmarsh in the group of heavily modified estuaries was due exclusively to the very large saltmarsh patches in Botany Bay

(Table 2).

Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Table 1: Mean ± SE. Physico-chemical and benthic sediment metals values. Metals quotient represents combined values of Co, Cr , Cu, Fe, Mn, Ni, Pb and Zn. * denotes areas where some metals values exceeded ANZECC guidelines at some sites (ANZECC, 2010). See Dafforn et al. (2012) for more detail.

Temp (°C) Sal ‰ pH Metals Quotient Outer Inner Outer Inner Outer Inner Outer Inner Port Jackson 22.0 ± 0.2 23.5 ± 0.6 36.3 ± 0.3 34.7 ± 0.6 8.1 ± 0.0 7.9 ± 8.5 0.35 ± 0.03 *7.42 ± 1.95 Heavily Botany Bay 23.5 ± 0.5 24.8 ± 0.3 34.7 ± 1.2 29.5 ± 2.5 8.2 ± 0.0 8.0 ± 0.0 0.71 ± 0.25 *4.18 ± 0.27 Modified Estuaries Port Kembla 22.8 ± 0.2 23.3 ± 0.2 36.3 ± 0.2 36.1 ± 0.2 8.2 ± 0.0 8.2 ± 0.0 3.75 ± 0.51 *5.61 ± 1.65 Port Hacking 21.8 ± 0.6 23.6 ± 0.5 35.4 ± 0.2 34.9 ± 0.7 8.1 ± 0.0 8.0 ± 0.0 0.11 ± 0.01 0.16 ± 0.03 Relatively Jervis Bay 22.3 ± 0.4 22.7 ± 0.3 35.4 ± 0.4 35.1 ± 0.6 8.1 ± 0.0 8.1 ± 0.0 0.14 ± 0.03 0.14 ± 0.02 Unmodified Estuaries Clyde River 20.1 ± 0.6 23.0 ± 0.8 35.5 ± 0.7 30.8 ± 2.5 8.1 ± 0.0 7.9 ± 0.0 0.23 ± 0.03 0.83 ± 0.29

Table 2: Estuarine size and habitat characteristics (NSWDNR, 2010).

Habitat Size (km^2) % Habitat Cover Estuary Mouth Estuary Size Seagrass Mangrove Saltmarsh Port Jackson 1.38 49.7 0.0 0.0 0.0 Heavily Botany Bay 1.08 80.0 7.8 5.3 14.5 Modified Estuaries Port Kembla 0.35 1.6 0.0 0.0 0.0 Port Hacking 0.83 11.0 7.3 2.8 0.7 Relatively Jervis Bay 3.45 103.2 5.9 1.9 1.4 Unmodified Estuaries Clyde River 4.51 17.5 5.6 18.9 3.0

20 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Relationship Amongst Covariates and Larval Guilds

As stated earlier, multivariate analysis of community composition found that heavily modified and relatively unmodified sites differed significantly, as did the larval community in surface waters compared to the bottom (Figure 2). PCO plots indicated that the biological data had the strongest correlation with sediment metals, seagrass cover, temperature, salinity, and the width of the estuary mouth. Mangrove cover, saltmarsh cover, pH, and the estuary size did not correlate strongly with the larval data (multiple correlation <0.2) (Figure 5a). Heavily modified sites displayed increased sediment metal levels and temperature and decreased coverage of seagrass. Salinity and the width of the estuary mouth also correlated strongly with the larval data, but did not show a clear trend by modification (Figure 5a). All sediment metals trended in approximately the same direction and were inversely related to vegetative cover, such that the most contaminated sites occurred primarily in the estuaries with the lowest seagrass cover. This suggests that there was a strong relationship between sediment metals levels, reduced vegetative cover, and community composition in the heavily modified estuaries.

Figure 5b plots the overall abundance vectors of the largest larval guilds; this includes estuarine species and estuarine opportunists (for estuary usage analysis) and benthic/pelagic spawners. These vector lines suggest similar trends between the abundance of estuarine- opportunists and pelagic spawners; and similar trends between estuarine species and benthic spawners. Vector lines suggest that estuarine-opportunists/pelagic spawners were more abundant in sites where temperatures, sediment metals, and salinity were lower and seagrass cover was higher. Vector lines also suggested a strong relationship between the abundance of estuarine- opportunists/pelagic spawners and the width of the estuary mouth. Estuarine species/benthic spawners appeared more abundant where sediment contamination was higher and also where

21 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012. salinity and seagrass cover were lower. The vector lines of the estuarine species/benthic spawners and the width of the estuary mouth were nearly perpendicular, indicating a weak relationship between these variables (Figure 5a,b).

Plots of the highest correlating species for each group confirm that almost all of the species of the estuarine guild which drove the observed trends were also benthic spawners (the exception being H. transucidus) (Figure 5c). Most of these species trended towards the same sector of the graph as the group average, the exception being Favonigobius lentiginosus and

Redigobius macrostoma. Similarly, plots of the highest correlating species of estuarine- opportunists confirm that all of the larvae which drove this trend were pelagic spawners and that all these species trended in approximately the same direction (Figure 5d).

22 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

a - Covariates b – Life History Groups

Estuarine-Opportunist

Pelagic

Estuary Mouth Size %Seagrass Sal e – MDS Plot of Community Composition 40

Temp (°C) Sediment Metals Estuarine 20 Benthic

Estuary Use Category n)n) oo ii Spawning Strategy atat ii 0 alal var var otot tt

-20 c – Estuarine Larvae d – Estuarine Opportunists O2O2 (12.7% (12.7% of of CC PP

-40 Heavily Modified - Surface Ambassis jacksoniensis Heavily Modified - Bottom Hyperlophus vittatus Relatively Unmodified - Surface Favonigobius lateralis Relatively Unmodified - Bottom Platycephalus fuscus Sillago ciliata -60 Gerres subfasciatus -60 -40 -20 0 20 40 Redigobius macrostoma PCO1 (16.1% of total variation)

Arenigobius bifrenatus Paedogobius kimurai

Omobranchus anolius Pseudogobius sp. Hyperlophus transucidus Gobiopterus semivestita

Benthic Spawner Benthic Spawner Pelagic Spawner Pelagic Spawner

23 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Figure 5: Principal Coordinated Ordination (PCO) of correlations between covariate factors and two dimensional plots of community composition by modification. a) Covariate factors. b) Vectors of largest larval guilds. (Plots a/b - Multiple Correlation >0.2). For Plot A % mangrove, % saltmarsh, estuary size, and pH were included in analysis but correlated <0.2 and so are not shown. c) Plots of top eight estuarine species contributing to differences between heavily modified and relatively unmodified estuaries, subdivided by spawning strategy. d) Plots of top five estuary opportunist species contributing to differences between heavily modified and relatively unmodified estuaries, subdivided by spawning strategy (Plots c/d – Pearson Correlation > 0.35). e) Two dimensional MDS plot of community composition by disturbance category and sampling depth.

24 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Discussion Abundance of Estuary Usage Guilds

In this study, truly estuarine species were the dominant guild and significantly more abundant in the bottom waters of heavily modified estuaries compared to relatively unmodified estuaries. Increased abundance and dominance of estuarine species in heavily modified areas may indicate that high levels of contamination and habitat modification disproportionately affect species which complete their entire life cycle within the estuary. Surprisingly, this affect appeared to be largely positive as a decrease in seagrass cover and increased sediment contamination correlated with increased dominance of estuarine taxa. As expected, these impacts were concentrated in the bottom samples and few significant differences were observed in the surface waters of heavily modified vs. relatively unmodified estuaries. Larger differences in the bottom samples are likely due to the fact that changes to vegetative habitats have their greatest impact on the benthos, and contaminant levels are far greater in the sediment than in the water column (Knott et al., 2009).

While it is surprising that these potential stressors are correlated with increased abundance of estuarine species, other studies have identified positive effects on fish assemblages from estuarine modification (McKinley et al., 2011a; McKinley et al., 2011c). Increased abundances of fishes in heavily modified environments may have resulted from nutrient enrichment, since nutrients are often released with other contaminants (e.g. in sewage and urban runoff). Several studies have demonstrated that forms of contamination which have an enriching effect (e.g. nutrient run-off, fish farms, sewage, hydrocarbons, etc.) increase both the abundance and diversity of adult fish assemblages (McKinley and Johnston, 2010). Monitoring indicates that nutrient levels in the three heavily modified estuaries were elevated compared to the relatively unmodified estuaries (Scanes, 2010) although data was not available at sufficient

25 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012. resolution to formally analyze this relationship. Positive effects due to nutrient enrichment would be more likely to influence estuarine taxa than estuarine opportunists or marine stragglers, as estuarine taxa would benefit from increased food availability throughout their life cycle.

Increased food supply could increase larval abundance both by enhancing the fitness and fecundity of parents or by reducing postflexion larval mortality (Breitburg et al., 2009; Nixon and Buckley, 2002). It is unclear whether nutrient enrichment effects vary by depth, and further study would be needed to evaluate the role of nutrient enrichment in driving differences between the surface and bottom in these systems.

Another possible explanation for increased abundance of the estuarine species in the heavily modified estuaries is that this guild is more tolerant of environmental disturbance than other guilds. Estuarine species exhibit a greater ability to tolerate rapid changes in physico- chemical and habitat conditions, as this is required for survival in estuary habitats that have exceptionally high levels of natural variability. Some authors have argued that the highly adaptable characteristics of estuarine taxa make them more resilient to anthropogenic stressors

(Elliott and Quintino, 2007). Differences in availability of vegetative habitats may also act to concentrate some species, though larvae may also be highly responsive to oceanographic factors

(discussed below). The abundance of estuarine species showed a very weak relationship to the width of the estuary mouth; this suggests that the abundance of this guild was not strongly influenced by oceanographic factors, current patterns or coastal recruitment sources. This was an expected result, given that estuarine species are largely self-recruiting within an estuary (Elliott et al., 2007).

In contrast, estuarine opportunists trended towards increased abundance in the surface waters of relatively unmodified estuaries. Notably, this guild was approximately as abundant as

26 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012. estuarine species in relatively unmodified estuaries, where neither guild was clearly dominant.

Estuarine opportunist species spend less of their life cycle in the estuary than truly estuarine species, but are able to colonize estuarine environments for breeding (Elliott et al., 2007).

Because estuarine opportunists spawn at sea and enter estuaries later in their life cycle, potentially sensitive early life stages (e.g. embryogenesis, egg, and preflexion stages) may be less exposed to the effects of estuarine modification. It is interesting to note that estuarine opportunists displayed a reasonably strong relationship to the width of the estuary mouth (in addition to other covariates). Trends in the abundance of this guild may be more strongly influenced by oceanographic factors including current and tidal patterns. This is likely the case because this guild is dependent on recruitment from coastal areas much more than estuarine species, which are largely self-recruiting within estuaries (Elliott et al., 2007). This may explain the trend towards increased abundance of the estuarine opportunist guild in the relatively unmodified estuaries, two of which have exceptionally wide estuary entrances (Clyde River and

Jervis Bay). Further investigations of circulation and current patterns in these systems would be needed to fully explore this relationship.

Marine stragglers were equally abundant across modification and depth categories. This should be expected both because their occurrence in estuaries is largely incidental, and because their predominantly marine life cycle should insulate them from the effects of estuarine modification.

Abundance of Spawning Guilds

Benthic spawners were more abundant and dominant in the bottom samples of the heavily modified estuaries while there was no effect of modification on pelagic spawners.

27 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Brooders and viviparous species did not show strong trends by modification and represented a very small portion of the larval assemblage. Covariate analysis indicates that benthic spawners correlated very strongly with estuarine species and that pelagic spawners correlated very strongly with estuarine opportunists. This is because most estuarine species were in fact benthic spawners

(89.7% by abundance) while most estuarine opportunists were pelagic spawners (94.5% by abundance) (see Appendix 1). This is logical, as the majority of estuarine species will employ benthic eggs in order to increase the chances that their larvae are retained within the estuarine environment (Elliott et al., 2007; Miskiewicz, 1987). In contrast, the majority of estuarine opportunists will employ pelagic eggs in order to facilitate dispersal from coastal spawning grounds into the estuaries (Norcross and Shaw, 1984). The only species which significantly differed from this trend was H. translucidus which was the only highly abundant pelagic spawner in the estuarine group. This species was unusual among pelagic spawners, as it showed a strong increase in areas of highly contaminated sediment and reduced seagrass cover.

Impacts of Metals Contamination

Significant evidence points to developmental and reproductive susceptibility to contaminants in fish populations (Jones and Reynolds, 1997; Kingsford et al., 1997; McKinley et al., 2011c). However, we did not find strong evidence of negative impacts at the community level in this study. We found that larval groups which are likely to be subject to higher levels of contaminant exposure, namely estuarine species and benthic spawners who lay their eggs on the sediment, were in fact more abundant in heavily modified estuaries. This was the case despite comparatively high sediment contamination in these areas that exceed sediment quality guidelines (ANZECC, 2010). This suggests either that negative toxicological effects are

28 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012. negligible, or that they are being masked by related positive effects from disturbance in heavily modified estuaries (e.g. nutrient enrichment). It should be noted however that not all species conformed to the group average, and some individual species appeared to be less abundant in the heavily modified estuaries. Notably, one of the most abundant species encountered in this study,

Ambassis jacksoniensis, was significantly less abundant in heavily modified areas. This species may be relatively sensitive to anthropogenic disturbance in these systems, and may be a useful indicator species for measuring the effects of habitat modification.

Impacts of Vegetative Habitat Alteration

The study sites which displayed the highest levels of metals in the sediment also had the lowest coverage of vegetative habitats, particularly seagrass. This outcome likely results from a close relationship between the drivers of estuarine contaminant exposure and vegetative loss.

The primary causes of estuary contamination (such as run-off, urbanization of shoreline/catchment, sedimentation, outflows, etc.) also tend to negatively impact estuarine vegetative communities (Rogers et al., 2002). While natural morphological differences between estuaries and the availability of appropriate areas for vegetative growth are likely to play a role in the extent of submerged aquatic vegetation within these habitats, it is well documented that losses of seagrass, saltmarsh, and mangroves have occurred in all three heavily modified estuaries (Saintilan and Williams, 1999; Walker and McComb, 1992). Thus, differences in the abundance of vegetative habitats cannot be attributed to differences in natural conditions alone.

It is well known that estuarine vegetative habitats can function as ‘nurseries’ for a variety of fish taxa (Robertson and Duke, 1987). These habitats can provide food, shelter, and spawning areas for a variety of larvae, including many of the species which contributed strongly to the

29 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012. trends in this study (Miskiewicz, 1987). It is therefore possible that alteration or destruction of these habitats would affect the composition of the larval community. Notably, estuarine opportunists trended towards higher abundance in areas with greater seagrass cover. Many of these species occur in vegetative habitats during their juvenile and adult stages. As a result, the degradation or removal of these habitats could explain decreased abundance of these estuarine opportunist species in the heavily modified estuaries (Miskiewicz, 1987).

While nutrient enrichment could explain the increased abundance of estuarine species, changes to vegetative habitats may also play a major role. It is possible that estuarine species were more abundant in heavily modified estuaries because they are relatively adaptable to the loss of aquatic vegetation, because they prefer bare sediment, or because their estuarine opportunist competitors/predators are excluded by vegetative loss. In this study, the large-scale loss of seagrass and mangrove habitat in the heavily modified estuaries was highly correlated with increased abundance of several estuarine species including Paedogobius kimurai,

Gobiopterus semivestita, Psuedogobius sp., Hypherlophus vittatus, and Omobranchus anolius. In contrast, one highly abundant estuarine species, Favonigobius lentiginosus, decreased with seagrass loss.

It should be noted that one limitation of this study is that potentially synergistic or antagonistic effects of covariates cannot be differentiated with the methodology used. This is particularly true in cases where factors were found to strongly covary. For example, it was shown in this study that sites which have high levels of sediment contamination also tend to have lower seagrass cover. This is likely the case because the core mechanisms of estuary contamination exposure (e.g. run-off, urbanization of shoreline/ catchment, outflows, etc.) also tend to precipitate habitat alteration in estuarine systems (Drinkwater and Frank, 1994; Rogers et

30 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012. al., 2002). Because these factors covary strongly, it is unclear whether the relationship between loss of seagrass cover, increased sediment contamination, and increased abundance of the estuarine guild is due to one or both of these factors. It is possible that only one of these factors is responsible for the observed trend but that they both correlate in Figure 5 due to covariation. It is also possible that they act synergistically and hence increase the strength of the trend observed.

The converse may also be true of antagonistic factors. Unfortunately, this is a limitation of correlative field studies and the relative contributions of covariates are difficult to parse out. It is also possible that the trends in this study are influenced by differential tolerance between taxa or guilds. Differential tolerance and contamination resistance in fish has been demonstrated in some cases (Wirgin and Waldman, 2004; Xie and Klerks, 2004).

This study confirms the species specific patterns observed for bottom dwelling fish larvae in McKinley et al. (2011c). In that study the impacts of anthropogenic modification on benthic larvae were examined, and differences in the abundance of species such as Paedogobius kimurai and Ambassis jacksoniensis were associated with changes in seagrass cover and sediment metals contamination. The current study provides evidence that bottom dwelling fish are more strongly influenced by estuary modification than surface dwelling communities, and highlights for the first time the role that guild characteristics play in determining relative sensitivity. The examination of physical characteristics such as the size of the estuary and width of the estuary mouth, have also illustrated the potential for complex interactions between anthropogenic modifications and estuary structure in determining larval fish distributions. If the substantial changes we observed in larval fish communities were to translate into changes in adult fish assemblages, then anthropogenic modification would have the potential to substantially affect

31 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012. not only the biodiversity of estuarine systems but also ecosystem function. These novel findings will help to better target future monitoring in estuarine systems.

32 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Conclusion

Estuarine modification appears to have the greatest effect on the larvae of truly estuarine fish and benthic spawners compared to other estuary usage and spawning guilds. These effects were largely concentrated at the seafloor while few impacts were detected in surface waters. This suggests that impacts on larvae in these systems can be effectively monitored through sampling of the bottom waters alone. It is clear that there are large-scale differences between the larval fish assemblages living in the surface and bottom waters of heavily modified vs. relatively unmodified estuaries. Increased abundance of estuarine species and benthic spawners was strongly related to both sediment metal levels and reduced seagrass cover in heavily modified sites. However, we believe that these relationships are indicative of the impact of a broad suite of anthropogenic contamination, including increased nutrient enrichment and sedimentation. We believe that a combination of anthropogenic stressors have contributed to a substantial difference in larval fish communities between heavily modified and relatively unmodified systems.

Utilizing a guild approach focused on early life history characteristics has enabled us to identify sensitive traits and to quantify which groups are most clearly associated with environmental modifications. Increased utilization of an early life history guild approach would contribute significantly to the environmental impact literature and future monitoring of fish assemblages.

Ultimately changes to larval fish assemblages may have far reaching ecological impacts both for the adult fish community and other organisms.

33 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Acknowledgements

This research was primarily supported by the Australian Research Council through an

Australian Research Fellowship awarded to ELJ and a Linkage Grant awarded to Johnston,

Kelaher and Coleman. We would like to thank Dr. Katherine Dafforn and Shinjiro Ushiama for their help with the project. We would also like to thank the Bluescope Steel Company and

Marine Parks NSW for their generous support. This study was approved and carried out in strict accordance with the recommendations of the Animal Care and Ethics Committee of the

University of New South Wales (Project No. 09/110A) and the New South Wales Department of

Primary Industries (Permit No. P09/0072-1.0).

34 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

References

Anderson, M.J. 2001. A new method for non-parametric multivariate analysis of variance. Austral Ecology 26: 32-46. Anderson, M.J. 2003. PCO - Principal Coordinate Analysis: A computer program. University of Aukland. ANZECC. 2010. Chapter 3: Aquatic Ecosystems, Australian and New Zealand guidelines for fresh and marine water quality. Canberra: Australian and New Zealand Environmental Conservation Council. http://www.environment.nsw.gov.au/water/usinganzeccandwqos.htm Accessed November 14, 2010. Birch, G. and S. Taylor. 1999. Source of heavy metals in sediments of the Port Jackson estuary, Australia. The Science of The Total Environment 227: 123-138. Breitburg, D.L., J.K. Craig, R.S. Fulford, K.A. Rose, W.R. Boynton, D.C. Brady, B.J. Ciotti, R.J. Diaz, K.D. Friedland, J.D. Hagy, D.R. Hart, A.H. Hines, E.D. Houde, S.E. Kolesar, S.W. Nixon, J.A. Rice, D.H. Secor and T.E. Targett. 2009. Nutrient enrichment and fisheries exploitation: interactive effects on estuarine living resources and their management. Hydrobiologia 629: 31- 47. Dafforn, K.A., S.L. Simpson, B.P. Kelaher, G. Clarke, V. Komyakova, C.K.C. Wong and E.L. Johnston. 2012. The challenge of choosing environmental indicators of anthropogenic impacts in estuaries. Environmental Pollution 163: 207-217. Dallinger, R., F. Prosi, H. Segner and H. Back. 1987. Contaminated food and uptake of heavy metals by fish: a review and a proposal for further research. Oecologia 73: 91-98. DPI. 2010. Survey of recreational fishing in New South Wales. Sydney: Department of Primary Industries. Drinkwater, K.F. and K.T. Frank. 1994. Effects of river regulation and diversion on marine fish and invertebrates. Aquatic Conservation: Marine and Freshwater Ecosystems 4: 135-151. Edgar, G.J. and C. Shaw. 1995. The production and tropic ecology of shallow-water fish assemblages in southern Australia. III. General relationships between sediments, seagrasses, invertebrates and fishes. Journal of Experimental Marine Biology and Ecology 194: 107-131. Elliott, M. and V. Quintino. 2007. The estuarine quality paradox, environmental homeostasis and the difficulty of detecting anthropogenic stress in naturally stressed areas. Marine Pollution Bulletin 54: 640-645. Elliott, M., A.K. Whitfield, I.C. Potter, S.J.M. Blaber, D.P. Cyrus, F.G. Nordlie and T.D. Harrison. 2007. The guild approach to categorizing estuarine fish assemblages: a global review. Fish and Fisheries 8: 241-268. Gray, C.A., N.M. Otway, F.A. Laurenson, A.G. Miskiewicz and R.L. Pethebridge. 1992. Distribution and abundance of marine fish larvae in relation to effluent plumes from sewage outfalls and depth of water. Marine Biology 113: 549-559. Hall, B.D., R.A. Bodaly, R.J.P. Fudge, J.W.M. Rudd and D.M. Rosenberg. 1997. Food as the dominant pathway of methylmercury uptake by Fish. Water, Air, & Soil Pollution 100: 13-24. Henry, G.W. and J.M. Lyle. 2003. The national recreational and indigenous fishing survey. Commonwealth of Australia, Department of Agriculture, Fisheries, Forestry, Canberra. Hill, N.A., C.K. King, L.A. Perrett and E.L. Johnston. 2009. Contaminated suspended sediments toxic to an Antarctic filter feeder: Aqueous- and particulate-phase effects. Environmental Toxicology and Chemistry 28: 409-417.

35 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Holmes, P.M. and R.M. Cowling. 1997. The effects of invasion by Acacia saligna on the guild structure and regeneration capabilities of South African fynbos shrublands. Journal of Applied Ecology 34: 317-332. Jaureguizar, A.J., R. Menni, R. Guerrero and C. Lasta. 2004. Environmental factors structuring fish communities of the Río de la Plata estuary. Fisheries Research 66: 195-211. Johnston, E. and D.A. Roberts. 2009. Contaminants reduce the richness and evenness of marine communities: A review and meta-analysis. Environmental Pollution 157: 1745-1752. Jones, J.C. and J.D. Reynolds. 1997. Effects of pollution on reproductive behaviour of fishes. Reviews in Fish Biology and Fisheries 7: 463-491. Kennish, M.J. 2002. Environmental threats and environmental future of estuaries. Environmental Conservation 29: 78-107. Kingsford, M.J., I.M. Suthers and C.A. Gray. 1997. Exposure to sewage plumes and the incidence of deformities in larval fishes. Marine Pollution Bulletin 33: 201-212. Klingbeil, B.T. and M.R. Willig. 2009. Guild-specific responses of bats to landscape composition and configuration in fragmented Amazonian rainforest. Journal of Applied Ecology 46: 203-213. Knott, N.A., J. Aulbury, T. Brown and E.L. Johnston. 2009. Contemporary ecological threats from historical pollution sources: impacts of large-scale resuspension of contaminated sediments on sessile invertebrate recruitment. Journal of Applied Ecology 46: 770-781. Kupschus, S. and D. Tremain. 2001. Associations between fish assemblages and environmental factors in nearshore habitats of a subtropical estuary. Journal of Fish Biology 58: 1383-1403. Loneragan, N.R. and I.C. Potter. 1990. Factors influencing community structure and distribution of different life-cycle categories of fishes in shallow waters of a large Australian estuary. Marine biology 106: 25-37. Loneragan, N.R., I.C. Potter and R.C.J. Lenanton. 1989. Influence of site, season and year on contributions made by marine, estuarine, and diadromous and freshwater-water species to the fish fauna of a temperate Australian estuary. Marine biology 103: 461-479. Long, E.R. 2006. Calculation and uses of mean sediment quality guideline quotients: A critical review. Environmental Science & Technology 40: 1726-1736. Lotze, H.K., H.S. Lenihan, B.J. Bourque, R.H. Bradbury, R.G. Cooke, M.C. Kay, S.M. Kidwell, M.X. Kirby, C.H. Peterson and J.B.C. Jackson. 2006. Depletion, degradation, and recovery potential of estuaries and coastal seas. Science 312: 1806-1809. McKinley, A. and E.L. Johnston. 2010. Impacts of contaminant sources on marine fish abundance and species richness: A review and meta-analysis of evidence from the field. Marine Ecology Progress Series 420: 175-191. McKinley, A., L. Ryan, M. Coleman, N. Knott, G. Clarke, M. Taylor and E.L. Johnston. 2011a. Putting marine sanctuaries into context: A comparison of estuary fish assemblages over multiple levels of protection and disturbance. Aquatic Conservation. McKinley, A.C., K.A. Dafforn, M.D. Taylor and E.L. Johnston. 2011b. High levels of sediment contamination have little influence on estuarine beach fish community indices compared to physico-chemical variation. PLoS ONE 6(10). McKinley, A.C., A. Miskiewicz, M.D. Taylor and E.L. Johnston. 2011c. Strong links between metal contamination, habitat modification and estuarine larval fish distributions. Environmental Pollution 159: 1499-1509. Miskiewicz, A.G. 1987. Taxonomy and ecology of fish larvae in Lake Macquarie and New South Wales coastal waters, PhD Thesis. Sydney: University of New South Wales.

36 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Neira, F.J., A.G. Miskiewicz and T. Trnski. 1998. Larvae of temperate Australian fishes: A laboratory guide for larval fish identification. Perth: University of Western Australia Press. Neira, F.J. and I.C. Potter. 1992. Movement of larval fishes through the entrance channel of a seasonally open estuary in Western Australia. Estuarine, Coastal and Shelf Science 35: 213-224. Neira, F.J., I.C. Potter and J.S. Bradley. 1992. Seasonal and spatial changes in the larval fish fauna within a large temperate Australian estuary. Marine biology 112: 1-16. Nixon, S. and B. Buckley. 2002. “A strikingly rich zone”—Nutrient enrichment and secondary production in coastal marine ecosystems. Estuaries and Coasts 25: 782-796. Norcross, B.L. and R.F. Shaw. 1984. Oceanic and estuarine transport of fish eggs and larvae: A review. Transactions of the American Fisheries Society 113: 153-165. NSWDNR. 2010. Estuaries in New South Wales. Sydney: New South Wales Department of Natural Resources. http://www.naturalresources.nsw.gov.au/estuaries/inventory/index_ns.shtml Accessed November 14, 2010. Pinnegar, J.K., S. Jennings, C.M. O’Brien and N.V.C. Polunin. 2002. Long-term changes in the trophic level of the Celtic Sea fish community and fish market price distribution. Journal of Applied Ecology 39: 377-390. Pittman, S.J. and C.A. McAlpine. 2003. Movements of marine fish and decapod crustaceans: Process, theory and application. Advances in Marine Biology 44: 205-294. Potter, I.C., L.E. Beckley, A.K. Whitfield and R.C.J. Lenanton. 1990. Comparisons between the roles played by estuaries in the life-cycles of fishes in temperate Western Australia and Southern Africa. Environmental Biology of Fishes 28: 143-178. Potter, I.C. and G.A. Hyndes. 1999. Characteristics of the ichthyofaunas of southwestern Australian estuaries, including comparisons with holarctic estuaries and estuaries elsewhere in temperate Australia: A review. Australian Journal of Ecology 24: 395-421. Poulin, B., G. Lefebvre and L. Paz. 2010. Red flag for green spray: adverse trophic effects of Bti on breeding birds. Journal of Applied Ecology 47: 884-889. Prista, N., R.P. Vasconcelos, M.J. Costa and H. Cabral. 2003. The demersal fish assemblage of the coastal area adjacent to the Tagus estuary (Portugal): Relationships with environmental conditions. Oceanologica Acta 26: 525-536. Quinn, G. and M. Keough. 2002. Experimental design and data analysis for biologists. Cambridge: Cambridge University Press. Robertson, A. and N. Duke. 1987. Mangroves as nursery sites: comparisons of the abundance and species composition of fish and crustaceans in mangroves and other nearshore habitats in tropical Australia. Marine biology 96: 193-205. Rogers, C.E., D.J. Brabander, M.T. Barbour and H.F. Hemond. 2002. Use of physical, chemical, and biological indices to assess impacts of contaminants and physical habitat alteration in urban streams. Environmental Toxicology and Chemistry 21: 1156-1167. Saintilan, N. and R.J. Williams. 1999. Mangrove transgression into saltmarsh environments in south-east Australia. Global Ecology and Biogeography 8: 117-124. Scanes, P. 2010. NSW Estuarine catchment disturbance ranks. Sydney: NSW Department of Environment, Climate Change, and Water. Walker, D.I. and A.J. McComb. 1992. Seagrass degradation in Australian coastal waters. Marine Pollution Bulletin 25: 191-195. Wirgin, I. and J.R. Waldman. 2004. Resistance to contaminants in North American fish populations. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis 552: 73-100.

37 Anthropogenic activities differentially impact fish guilds. McKinley et al. 2012.

Woodcock, B.A., S.G. Potts, T. Tscheulin, E. Pilgrim, A.J. Ramsey, J. Harrison-Cripps, V.K. Brown and J.R. Tallowin. 2009. Responses of invertebrate trophic level, feeding guild and body size to the management of improved grassland field margins. Journal of Applied Ecology 46: 920-929. Xie, L. and P.L. Klerks. 2004. Changes in cadmium accumulation as a mechanism for cadmium resistance in the least killifish Heterandria formosa. Aquatic Toxicology 66: 73-81.

38 Appendix 1 - Average abundance data identified to lowest taxonomic level by estuary and depth. Gobiidae sp. represent the total of all observed morphologically distinct taxa which could not be identified to species. Abbreviations - Life Cycle Guild: EO = Estuarine Opportunist, E = Estuarine, MS = Marine Straggler. Spawning Guild: P = Pelagic, BE = Benthic, BR = Brooders, V = Viviparous.

Port Jackson Botany Bay Port Kembla Port Hacking Jervis Bay Clyde River Family Taxon Surface Bottom Surface Bottom Surface Bottom Surface Bottom Surface Bottom Surface Bottom Life Cycle Spawning Ambassidae Ambassis jacksoniensis Glassfish Port Jackson Glassfish 5.07 2.67 3.12 13.72 1.71 4.98 3.72 26.68 19.02 301.72 61.08 15.82 EO P Ambassis marianus Estuary Glassfish 0.00 0.17 0.00 0.00 0.00 0.00 0.71 0.00 0.35 0.00 1.02 0.00 E P Apogonidae Siphamia cephalotes Cardinalfish Wood's Siphonfish 0.00 1.73 0.00 0.18 0.51 0.00 0.00 0.00 0.00 10.82 0.00 0.00 E BR Apogonid A 1.85 0.99 0.00 1.26 6.34 0.00 3.36 0.32 0.00 0.18 0.79 0.00 E BR Apogonid B 0.00 1.34 0.00 0.00 1.04 0.00 0.00 0.18 0.00 0.47 0.00 0.00 E BR Apogonid C 0.76 0.99 0.00 0.00 0.00 0.21 0.00 0.00 0.00 0.00 0.00 0.00 E BR Apogonid D 0.00 1.54 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 E BR Atherinidae Atherinid A Old World Silversides . 0.70 0.00 0.82 0.00 0.00 0.00 0.39 0.00 0.47 0.00 1.05 0.00 E BE Atherinid B 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.41 0.00 0.00 0.00 E BE Atherinid C 0.00 0.00 0.00 0.99 0.00 0.00 0.00 0.19 0.00 1.07 0.00 0.00 E BE Hemiramphidae Hyporhamphus sp. Garfishes 0.00 0.00 0.00 0.99 0.00 0.00 0.00 0.19 0.00 0.00 0.00 0.62 E BE

Blenniidae Omobranchus anolius Blennies Oyster Blenny 2.15 0.25 12.84 4.38 6.37 0.18 2.06 0.00 1.05 0.00 1.60 0.30 E BE

Omobranchus rotundiceps Combtooth Blenny 0.00 2.65 0.69 0.54 11.13 0.00 0.00 0.00 0.00 0.00 0.27 0.57 E BE

Parablennius tasmanianus Tasmanian Blenny 0.49 0.00 1.67 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 E BE

Petroscirtes lupus Brown Sabretooth Blenny 0.00 0.17 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EO BE Callionymidae Callionymids Dragonets 0.00 0.17 0.00 0.00 0.00 0.17 0.00 0.00 0.00 0.00 0.00 0.31 E P

Port Jackson Botany Bay Port Kembla Port Hacking Jervis Bay Clyde River Family Taxon Surface Bottom Surface Bottom Surface Bottom Surface Bottom Surface Bottom Surface Bottom Life Cycle Spawning Carangidae Trachurus novaezelandiae Scads Yellowtail Scad 0.73 0.00 0.59 0.00 0.26 0.00 0.00 0.20 0.24 0.00 0.00 0.00 MS P Clinidae Cristiceps sp. Clinids 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.40 0.00 0.00 0.00 0.00 EO BE

Clupeidae Etrumeus teres Herrings and Sprats Maray 0.00 0.00 0.00 0.00 0.00 0.98 0.00 0.00 0.00 0.00 0.00 0.00 MS P

Herklotsichthys castelnaui Southern Herring 0.00 2.44 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 E P

Hyperlophus transucidus Translucent Sprat 6.99 17.14 8.84 55.72 0.55 0.00 3.82 0.20 0.78 0.00 0.00 0.00 E P

Hyperlophus vittatus Sandy Sprat 2.95 3.70 0.66 6.79 1.78 29.55 0.89 4.06 10.28 78.59 12.86 63.22 EO P

Sardinops sagax Pilchard 0.24 0.00 1.25 0.00 0.29 2.94 0.00 0.00 0.35 0.00 0.00 0.00 MS P

Spratelloides robustus Blue Sprat 1.95 0.00 0.00 0.00 0.00 1.67 0.00 0.79 1.22 0.00 0.00 0.00 EO P Creediidae Creedia haswelli Sandburrower Slender Sandburrower 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.18 0.00 0.00 MS BE Eleotridae Philypnodon sp. Gudgeons Flathead Gudgeon sp. 0.60 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 E BE Engraulidae Engraulis australis Anchovies Australian Anchovy 0.24 1.18 0.24 0.25 0.27 5.43 0.00 0.00 0.88 0.35 0.00 0.00 EO P

Gerreidae Gerres subfasciatus Silver Biddies Roach 1.93 6.88 6.08 15.27 13.34 6.80 1.06 7.03 10.12 7.05 27.88 19.27 EO P Gobiesocidae Gobiesocid sp. Clingfish 0.00 0.00 0.00 0.00 0.26 0.00 0.00 0.79 0.70 0.17 0.27 0.00 MS BE Gobiidae Afurcagobius tamarensis Gobies Tamar Goby 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.38 E BE Arenigobius spp. Half Bridled and Bridled Goby 11.29 14.43 3.34 29.15 2.95 59.19 2.96 5.29 1.50 8.81 13.06 5.65 E BE Callogobius producta Elongate goby 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.18 0.00 0.00 E BE Favonigobius lentiginosus Long Finned Goby 7.40 2.58 1.18 1.48 0.82 0.77 6.74 18.68 1.08 5.57 10.38 2.73 E BE Gobiopterus semivestita Transparent Goby 16.72 122.40 26.50 110.03 4.63 1.25 8.20 0.16 9.84 3.69 16.11 0.75 E BE

Port Jackson Botany Bay Port Kembla Port Hacking Jervis Bay Clyde River Family Taxon Surface Bottom Surface Bottom Surface Bottom Surface Bottom Surface Bottom Surface Bottom Life Cycle Spawning Gobiidae Paedogobius kimurai Gobies WG Paedomorphic Goby 13.24 252.14 0.00 1.17 0.00 101.22 6.52 2.49 0.00 0.17 0.86 0.17 E BE Pseudogobius sp. Eastern Bluespot Goby 11.18 6.95 14.17 0.63 9.84 1.03 8.37 0.00 12.10 0.17 34.07 1.77 E BE Redigobius macrostoma Large Mouth Goby 5.85 0.87 1.75 3.96 0.30 2.81 3.71 0.26 2.21 17.60 21.00 7.44 E BE Tridentiger trigonocephalus Trident Goby 0.00 0.00 0.00 0.00 0.00 2.46 0.00 0.00 0.00 0.00 0.00 0.00 E BE Gobiids 55.68 49.18 10.84 27.99 9.98 23.29 14.98 12.82 9.44 180.30 19.43 75.43 E BE Gonostomatidae Gonostomatids Bristlemouths 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.19 MS P Kyphosidae Atypichthys strigatus Sea Chubs Mado 0.00 0.00 0.30 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MS P Girella tricuspidata Luderick 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.31 0.00 0.00 EO P Kyphosis sp. Drummer sp. 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.19 0.00 MS P

Scorpididae lineolata Silver Sweep 0.00 0.00 0.00 0.00 0.52 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MS P

Labridae Labrids Wrasse 0.00 0.00 0.00 0.00 0.81 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MS P

Leptoscopidae Lesueurina platycephala Sandfish Common Sandfish 0.00 0.17 0.00 0.00 0.00 0.17 0.00 2.55 0.00 0.00 0.00 0.00 MS BE

Lutjanidae Lutjanids Snappers 0.00 0.17 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MS P Monacanthidae Monacanthus chinensis Leatherjackets Fan Belly Leatherjacket 0.00 0.55 0.41 0.84 0.00 0.00 0.00 0.00 0.24 0.00 0.00 0.00 E BE Monacanthids 0.24 0.17 0.00 0.00 0.00 0.00 0.45 0.20 0.00 0.00 0.00 0.19 EO BE Monodactylidae Monodactylus argenteus Moonfish Diamondfish 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.99 0.25 0.14 0.19 0.18 EO P Schuettea scalaripinnis Eastern Pomfret 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.38 MS P Mugilidae Liza argentea Mullets Flat Tail Mullet 0.00 0.00 0.27 0.00 0.00 0.39 0.00 0.00 0.00 0.00 0.00 0.00 EO P Mullidae Mullids Goatfish 0.00 0.00 0.00 0.00 0.58 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MS P

Port Jackson Botany Bay Port Kembla Port Hacking Jervis Bay Clyde River Family Taxon Surface Bottom Surface Bottom Surface Bottom Surface Bottom Surface Bottom Surface Bottom Life Cycle Spawning Myctophidae Myctophids Lanternfish 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.24 0.00 0.00 0.00 MS P Odacidae Odacids Weed Whitings 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.69 0.00 0.00 MS P Paralichthyidae Pseudorhombus A Large Tooth Flounders 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.23 0.00 0.14 0.34 1.38 EO P Pseudorhombus B 0.37 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MS P Pempheridae Pempherids Sweepers/Bullseyes 0.00 0.00 0.00 0.21 0.00 0.00 0.00 0.00 0.00 0.18 0.00 0.00 MS P Percophidae Enigmapercis reducta Duckbills Broad Sandfish 0.00 0.00 0.00 0.00 0.00 0.00 0.24 0.00 0.00 0.00 0.00 0.00 MS BE Platycephalidae Platycephalus fuscus Flatheads Dusky Flathead 0.25 0.00 0.00 0.00 0.00 0.17 0.00 1.15 0.96 0.80 0.56 1.67 EO P

Platycephalus sp. 0.00 0.43 0.69 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MS P

Schindleriidae Schindleriidae sp. Infrantfishes Schindler's Goby 0.00 0.00 0.00 1.26 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MS BE Sciaenidae Argyrosomus japonicus Drums/Croakers Mulloway 0.00 0.00 0.00 0.00 0.00 0.18 0.00 0.00 0.00 0.00 0.00 0.38 EO P Atractoscion aequidens Teraglin 0.00 0.00 0.00 0.00 0.00 0.18 0.00 0.00 0.00 0.00 0.00 0.00 EO P Scorpaenidae Scorpaenids Scorpionfish 0.00 0.00 0.00 0.00 0.51 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MS P Nesosebastidae Neosebastes scorpaenoides Gurnard Perches Common gurnard perch 0.00 0.00 0.60 0.00 0.30 0.00 0.00 0.00 0.49 0.00 0.45 0.00 MS P Serranidae Anthiids Basslets 0.00 0.00 0.53 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MS P

Silliginidae Sillago ciliata Whitings Sand Whiting 0.25 0.72 1.09 9.44 0.89 1.11 1.13 2.53 4.53 23.15 27.16 2.70 EO P

Sillago flindersai Eastern School Whiting 0.00 1.41 0.00 9.50 0.30 0.60 0.00 1.95 0.00 1.14 0.00 0.57 MS P Sillago maculata Trumpeter Whiting 1.27 4.23 0.85 17.62 0.00 0.91 0.74 0.95 0.72 7.15 1.06 2.18 E P

Port Jackson Botany Bay Port Kembla Port Hacking Jervis Bay Clyde River Family Taxon Surface Bottom Surface Bottom Surface Bottom Surface Bottom Surface Bottom Surface Bottom Life Cycle Spawning

Soleidae Soleids True Soles 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.25 0.00 0.18 0.00 0.18 MS P

Sparidae Acanthopagrus australis Sea Breams Yellowfin Bream 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 3.89 0.00 0.20 EO P

Sphyraenidae Sphyraena sp. Barracudas Barracuda sp. 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.17 MS P

Syngnathidae Hippocampus abdominalis Pipefish/Seahorses Big Belly Seahorse 0.00 0.00 0.31 0.00 0.00 0.00 0.37 0.00 0.00 0.00 0.00 0.00 E V

Stigmatopora nigra Wide Bodied Pipefish 0.00 0.00 0.00 0.00 0.00 0.26 0.00 1.74 0.00 0.33 0.00 0.41 E V

Urocampus carinirostris Hairy Pipefish 0.82 0.00 1.25 0.39 0.53 0.00 0.24 0.98 0.00 0.48 0.00 0.43 E V

Vanacampus margaritifer Mother of Pearl Pipefish 0.24 0.00 0.00 0.00 0.00 0.00 0.00 0.20 0.00 0.00 0.00 0.00 E V

Synodontidae Synodontids Lizardfishes 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.23 MS P

Terapontidae Pelates sexlineatus Grunters Six Lined Trumpeter 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.02 0.00 0.18 0.22 0.00 EO P

Tetraodontidae Tetraodontids Pufferfish 0.00 0.00 0.00 0.00 0.26 0.00 0.00 0.00 0.00 0.80 0.22 0.00 E BR

Tetrarogidae Centropogon australis Waspfish Fortesque 0.00 0.50 0.00 1.59 0.00 0.17 0.00 0.80 0.00 1.43 0.00 0.42 EO P

Tripterygiidae Tripterygid A Triplefin Blennies 0.24 2.81 4.57 14.03 2.15 0.66 3.23 0.97 2.10 2.41 1.44 0.19 MS BE

Tripterygid B 0.00 0.00 1.11 0.00 0.24 0.00 0.00 0.00 0.00 0.00 0.19 0.00 MS BE

Total Abundance 151.72 503.71 106.53 329.38 79.45 249.73 73.91 98.26 91.58 660.50 253.75 206.47

Chapter 6

RELATIONSHIPS BETWEEN BODY BURDENS OF TRACE METALS (As, Cu, Fe, Hg, Mn, Se, Zn) AND THE RELATIVE BODY SIZE OF SMALL TOOTH FLOUNDER (PSEUDORHOMBUS JENYNSII)

Final Version:

Science of the Total Environment 423 (2012) 84–94

Contents lists available at SciVerse ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Relationships between body burdens of trace metals (As, Cu, Fe, Hg, Mn, Se, and Zn) and the relative body size of small tooth ﬂounder (Pseudorhombus jenynsii)

Andrew C. McKinley ⁎, Matthew D. Taylor, Emma L. Johnston

Evolution & Ecology Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, New South Wales 2052, Australia article info abstract

Article history: Several studies have described strong relationships between body size and the accumulation of trace metals Received 3 October 2011 in animal tissues. However, few of these studies have utilized aging techniques to control for age related ef- Received in revised form 3 February 2012 fects. We utilized relative body size (g y−1) of a model flounder species, Pseudorhombus jenynsii, in order to Accepted 6 February 2012 control for age related effects on growth and size measurements. We investigated links between relative Available online 3 March 2012 body size, concentrations of trace metals in flounder muscle tissue, physico-chemical variables (temperature, salinity, pH, and turbidity), and levels of trace metals in the sediment. Flounder were sampled using an otter Keywords: fi fi Trace metals trawl net in the inner areas of eight estuaries that were either heavily modi ed or relatively unmodi ed by Body size urbanization and industrial activity. Our results indicate that this commonly eaten fish is accumulating signif- Flounder icant levels of some trace metals in their muscle tissue, both in relatively unmodified and heavily modified Sediment contamination estuaries. Concentrations of Cu, Zn and Fe in muscle tissue, as well as temperature, showed a negative rela- Pollution tionship to the relative body size of flounder. In contrast, Se and Hg in muscle showed a positive relationship Fish aging to relative body size. Observed growth patterns indicate that these effects are not driven by age related differences in metabolic activity. Instead, our results suggest that differences in food supply or toxicological effects may be responsible for the observed relationships between relative body size and concentrations of Cu, Zn, and Se in muscle tissues. The use of otolith aging and growth measurement techniques represents a novel method for assessing the relationships between trace metal accumulation and the relative body size of fish in a field environment. © 2012 Elsevier B.V. All rights reserved.

1. Introduction controlling for differential growth rates or age related effects (Canli and Atli, 2003; Liang et al., 1999). By controlling for the age of fish, The accumulation of pollutants in fish species is of significant in- it is possible to determine if the relationship between trace metal ac- terest to ecotoxicologists and fisheries managers. This is due both to cumulation and body size is due to other factors unrelated to the age the potential ecological impacts of those pollutants (Johnston and of the animal. Prey-mediated effects (Deb and Fukushima, 1999)or Roberts, 2009) and because of human health risks associated with physiological impacts due to the acute toxicity of trace metals the consumption of contaminated fish (Campbell et al., 2008). Previ- (Buckley et al., 1982; Kearns and Atchison, 1979; Waiwood and ous field studies have shown a strong relationship between the size of Beamish, 1978) may also play a significant role in driving the rela- fish and the accumulation of various trace metals (Al-Yousuf et al., tionship between trace metal accumulation and body size. 2000; Canli and Atli, 2003; Liang et al., 1999). These relationships Stressors such as trace metals can draw metabolic energy away have often been attributed to changes in metabolic activity as a fish from growth, and changes to an organism's growth rate can be indic- ages, where it is assumed that larger (and hence older) fish are ative of a loss of fitness or health caused by physiological or environ- experiencing a reduction in their metabolic activity (Canli and Atli, mental stress (Brown and Ahsanullah, 1971). Several studies have 2003). Changes in metabolic rate as a fish ages are thought to influ- documented reduced growth rates in fish as a result of exposure to ence the uptake or efflux of trace metals, creating a strong relation- trace metals (Buckley et al., 1982; Kearns and Atchison, 1979; ship between body size, age, and tissue concentrations of trace Waiwood and Beamish, 1978). However, physiological impacts can metals (Widianarko et al., 2000). However, most studies have not be difficult to detect or accurately measure in a field environment assessed these relationships by explicitly determining the age of fish (Mondon et al., 2001). In many fish, age can be determined through and have instead linked body size and tissue concentrations without examination of the otoliths, and enumeration of the accreted calcified layers within these bones. This data can be used to control for age related effects in body size measurements, by dividing the specimen's fi ⁎ Corresponding author. Tel.: +1 613 410 2200. weight by its age. This calculates the relative body size of the sh, E-mail address: [email protected] (A.C. McKinley). which also provides an indirect measure of the relative growth rate

0048-9697/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2012.02.007 A.C. McKinley et al. / Science of the Total Environment 423 (2012) 84–94 85 over the lifespan of an individual. Although these methods are com- 151°57.953′E), Broken Bay (33°32.203′S, 151°12.839′E), Port Hacking monly used in ecological studies, they are seldom used in toxicologi- (34°04.680′S, 151°09.311′E), and the Clyde River (35°44.233′S, cal studies (Francis et al., 1992). As such, the utilization of otolith 150°14.272′E) (Fig. 1). The four heavily modified estuaries are all an- aging and growth measurement techniques represents a novel meth- thropogenically disturbed environments near large urban and industri- od for assessing the relationships between trace metal accumulation al areas and are subject to intense commercial and recreational boating and the relative growth of fish in a field environment. traffic, historic and ongoing contamination, greater recreational fishing In most estuarine environments contamination accumulates in activity, and substantial urbanization of their shoreline and catchment sediments, particularly in the sheltered inner areas of estuaries (Birch and Taylor, 1999; DPI, 2010; Scanes, 2010). In comparison, the where flushing from coastal currents is minimal and fine sediments relatively unmodified estuaries have fewer recreational fishers, less are available to bind contaminants (Hesslein et al., 1980; Knott boating traffic (almost none of which is commercial), less urbanization et al., 2009). Because of this, concentrations of contaminants in sedi- of the coastline and catchment, and virtually no heavy industry (Birch ments are normally far greater than in the surrounding water column and Taylor, 1999; DPI, 2010; Scanes, 2010). While these estuaries do (Dafforn et al., 2012; Knott et al., 2009). Consequently, organisms have some degree of agricultural land use in their catchment, the ma- which are highly associated with the benthos are thought to be ex- jority of the catchment in all of the relatively unmodified estuaries is posed to greater contamination levels than their pelagic or benthope- within conservation areas, forestry zones, or areas where anthropogen- lagic counterparts (Dallinger et al., 1987). Flounders spend their ic utilization is negligible. Most significantly for this study, previous entire life cycle on the benthos and frequently feed or burrow in the studies have indicated high levels of trace metal contamination in the sediment (Gomon et al., 2008). In addition, flatfish of various species four heavily modified estuaries, while the relatively unmodified estuar- have been the subject of a variety of field eco-toxicological studies ies have comparatively low levels of trace metal contamination (Cossa et al., 1992; de Boer et al., 2001; Mondon et al., 2001; (Dafforn et al., 2012). Plaskett and Potter, 1979), and consequently represent an ideal taxa Four sampling sites were initially selected in each estuary with for the evaluation of contamination impacts. Pseudorhombus jenynsii three replicate trawls conducted per site to sample flounder (twelve (small tooth flounder) is a common Paralichthyidae (sand flounder) trawls per estuary). Flounder were captured in 7–12 trawls per estu- in the estuaries of southeastern Australia. P. jenynsii is a fast growing ary, representing 3–4 sites. All sampling was done at a similar dis- species and has high fecundity (minimum population doubling time tance from the mouth of the estuary, within the sheltered inner 15 months). The species is an estuarine resident, and is a secondary areas of the estuary (Fig. 1). consumer feeding at a trophic level of 3.5±0.37 (Froese and Pauly, 2010; Gomon et al., 2008). P. jenynsii is also a predatory generalist 2.2. Fish sampling methods and most of the animals that it preys upon (e.g. sediment infauna, benthic invertebrates, and small benthic fish) also spend the majority Sampling at each site was conducted with an otter trawl (6 m of their life cycle in or close to the sediment (Gomon et al., 2008). mouth, 12 m length, and 3.8 cm diamond mesh, with a fine 0.6 mm Lastly, P. jenynsii is a highly sought after species for both commercial cod-end mesh) pulled over bare sediment at 4–12 m depth. Otter and recreational fishing and is consumed by humans (Froese and boards were rigged into a four point bridle using an approximate Pauly, 2010). For all of these reasons, P. jenynsii represents an excel- 3:1 warp to depth ratio. Trawls were 15 min, at a speed ~1 knot, cov- lent species for the evaluation of contamination impacts in fish ering a distance of ~450 m. Following trawling all fish were sorted on (Plaskett and Potter, 1979). the boat and flounder were euthanized in a 100 mg L− 1 benzocaine We explore the relationship between trace metal contamination solution and frozen for transport back to the laboratory. in sediments, the occurrence of trace metals in the muscle tissues of P. jenynsii, and the relative body size of fish at different ages (as an in- 2.3. Tissue trace metal analysis direct measure of relative growth rate over the lifespan). We assess these relationships within the context of large scale anthropogenic Fish were taken back to the laboratory and rinsed with Milli-Q modification and physico-chemical variability. We hypothesize sever- water before being dissected for muscle tissue samples. All flounder al outcomes: were dry weighed and length measurements were taken in the lab. Acid washed tools were used for dissections and muscle tissue sam- 1. We hypothesize that levels of trace metals in muscle tissues will be ples were placed in polypropylene vials and freeze-dried for 48 h correlated with levels of trace metals in the sediment. Further- prior to trace metal analysis. Freeze-dried samples were digested fl fi more, we predict that ounder living in heavily modi ed estuaries using microwave high pressure digestion (C-225) at 200 °C. All will show higher levels of trace metals in their muscle tissues. trace metals except Fe and Mn were measured using inductively 2. We predict that the occurrence of trace metals in muscle tissues, coupled plasma mass spectrometry (ICP-MS) (C-209). Fe and Mn above concentrations which have been demonstrated to have were measured using inductively couple plasma atomic emissions fi fi physiological signi cance for sh, will show a negative relation- spectrometry (ICP-AES) (C-229). All samples were analyzed in ship to the relative body size of P. jenynsii. batches with blanks and analytical accuracy was determined using 3. Lastly, we predict that the negative relationship between trace certified reference material with every digestion batch. DORM-3 fi metals in muscle tissue and the relative body size of sh will be ob- Fish Muscle and DOLT-4 Dogfish Liver reference material was served for all age groups. obtained from the National Research Council of Canada. All recoveries were within 10% of the certified values. Limits of detection (LOD 3σ) 2. Methods were determined per digestion batch and are listed in the results section; where samples were below the limit of detection this has been 2.1. Study sites noted (Table 1). Throughout the manuscript metal concentrations in both the flounder muscle tissue and in sediment samples are pre- Fish were sampled in eight permanently open estuaries along the sented as μgg− 1. south coast of New South Wales, Australia. These included four heavily modified estuaries — Hunter River (32°55.352′S, 151°46.191′E), Port 2.4. Otolith analysis Jackson (33°44.258′S, 151°16.542′E), Botany Bay (33°59.352′S, 151°11.433′E), and Port Kembla (34°28.121′S, 150°54.410′E), as well Sagittal otoliths were extracted, cleaned, dried, and stored in plas- as four relatively unmodified estuaries — Karuah River (32°38.782′S, tic vials. The dorsal otolith was utilized for analysis while the ventral 86

a b ..MKne ta./Sineo h oa niomn 2 21)84 (2012) 423 Environment Total the of Science / al. et McKinley A.C. c de

fg h – 94

Fig. 1. Locations of study sites in eight focal estuaries: a) Karuah River (relatively unmodified), b) Hunter River (heavily modified), c) Broken Bay (relatively unmodified), d) Port Jackson (heavily modified), e) Botany Bay (heavily modified), f) Port Hacking (relatively unmodified), g) Port Kembla (heavily modified) and g) Clyde River (relatively unmodified). A.C. McKinley et al. / Science of the Total Environment 423 (2012) 84–94 87

Table 1 − 1 ⁎ † Mean (±SE) concentrations (μgg ) of trace metals in ﬂounder muscle tissue. All samples below detection limits (Ag, Cd, and Cr). 50–70% of samples below detection limit (Ni and Pb). Marked trace metals (Ag, Cd, Cr, Ni, and Pb) omitted from further analysis. European Union, United States Federal Drug Administration, and Australian Federal limits for human consumption are listed for regulated metals (Rayment, 1991; USFDA, 2001; EC, 2005).

Ag* As Cd* Cr* Cu Fe Hg Mn Ni† Pb† Se Zn

Heavily modiﬁed Botany Bay b0.030 4.16 b0.05 b0.18 0.81 8.55 0.15 1.03 0.07 0.02 1.56 21.47 estuaries (±0.25) (±0.02) (±0.70) (±0.01) (±0.22) (±0.02) (±0.01) (±0.06) (±0.98) Port Jackson b0.030 1.97 b0.05 b0.18 0.68 6.09 0.37 2.41 0.04 0.04 2.06 17.00 (±0.20) (±0.03) (±0.75) (±0.08) (±1.04) (±0.02) (±0.02) (±0.12) (±1.17) Port Kembla b0.030 4.17 b0.05 b0.18 0.68 8.11 0.11 1.33 0.05 0.03 3.52 16.61 (±0.71) (±0.03) (±0.70) (±0.01) (±0.57) (±0.03) (±0.01) (±0.16) (±0.52) Newcastle b0.030 1.74 b0.05 b0.18 0.68 7.16 0.14 1.48 0.08 0.02 1.83 22.74 (±0.12) (±0.02) (±0.43) (±0.01) (±0.46) (±0.03) (±0.01) (±0.10) (±1.90) Relatively unmodiﬁed Broken Bay b0.030 2.42 b0.05 b0.18 0.66 4.64 0.17 0.65 0.04 0.02 1.90 16.60 estuaries (±0.40) (±0.03) (±0.62) (±0.02) (±0.08) (±0.02) (±0.01) (±0.07) (±0.69) Clyde River b0.030 2.26 b0.05 b0.18 0.63 6.77 0.21 1.02 0.05 0.01 2.13 21.31 (±0.25) (±0.05) (±0.93) (±0.02) (±0.30) (±0.04) (±0.01) (±0.12) (±1.92) Karuah River b0.030 1.82 b0.05 b0.18 0.65 6.69 0.27 1.25 0.05 0.02 1.84 17.01 (±0.16) (±0.03) (±0.69) (±0.02) (±0.46) (±0.02) (±0.01) (±0.07) (±0.43) Port Hacking b0.030 4.49 b0.05 b0.18 0.65 6.91 0.11 0.72 0.03 0.01 1.79 16.98 (±0.57) (±0.02) (±0.45) (±0.01) (±0.19) (±0.01) (±0.01) (±0.05) (±0.45) EU Regulated Limit –– 0.05 –– – 0.50 ––0.20 –– USFDA Regulated Limit – 86.00 4.00 13.00 ––1.00 – 80.00 1.70 –– Australian Federal Limit –– 2.00 – 10.00 – 0.50 ––1.50 – 150.00 Limit of Detection (3σ) 0.03 0.04 0.05 0.18 0.14 4 0.03 0.4 0.09 0.05 0.07 0.30 Analysis Method C-209 C-209 C-209 C-209 C-209 C-229 C-209 C-229 C-209 C-209 C-209 C-209 Digestion Method C-225 C-225 C-225 C-225 C-225 C-225 C-225 C-225 C-225 C-225 C-225 C-225

otolith was retained as a back-up. Otoliths were embedded in a rect- methods (Hill et al., 2009; USEPA, 2007). Recoveries were calculated angular mould using prepared Epofix hardener/resin mixture and against certified reference materials and all trace metals used in this dried for 3 h in a drying oven at 45 °C. The core of the embedded oto- study were within accepted recovery limits. Hg and Se values were lith blocks was then marked and the blocks were sectioned trans- not obtained from this method and so are not utilized in analysis versely through the core using a diamond blade low speed saw. of the sediment. Full details of analyses and contaminant datasets Sections were then mounted on glass slides using a thermoplastic ce- are presented in Dafforn et al. (2012). ment (Crystal Bond, SPI Supplies) and polished using aluminum oxide abrasive paper on a rotating wet sander, followed by lapping film. Sections were photographed using a microscope mounted digital 2.6. Statistical analysis camera with conventional lighting. Otolith pictures were measured along the ventral side of the sulcus and analyzed using ImageJ (v Most multivariate and univariate datasets were analyzed as 1.44p, National Institute of Health, USA). mixed-model PERMANOVA in PRIMER v.6.4 (Anderson, 2001). The There are no published age validation studies for P. jenynsii, so the exception was the Linear Mixed Model (LMM) analysis and associat- number of annuli rings was considered a relative measurement of fish ed linear regressions, which were analyzed in SPSS using the MIXED age amongst individuals (Francis et al., 1992). The relative body size function (Table 5 and Fig. 4 only). For PERMANOVA analysis, Bray– of fish was calculated as the weight of the fish divided by its age. Curtis similarity matrices were calculated for multivariate data This is an estimate of the growth rate of the fish over the entire life- while Euclidean distance matrices were used for univariate measures. span. The relative body size thus approximates the accumulation of A dummy variable of 1 was added when calculating the similarity ma- body mass per year of life. Relative body size increased in a linear trices in order to compensate for zero values. The PERMANOVA and fashion from the first to the fourth year (Fig. 2d). This indicates that LMM designs employed in the course of this analysis were identical the majority of fish captured in this study were at a stage of their and consisted of the following factors: life cycle where mass was linearly increasing with age. Dis Disturbance category — Heavily Modified or Relatively 2.5. Physico-chemical and sediment sampling methods Unmodified (2 levels, Fixed) Es Estuary (Disturbance Category) — (8 estuaries, Random) Physico-chemical variables and sediment grain size data were in- Si Site (Estuary(Disturbance Category) — (38 sites, Random) cluded in the analysis in order to evaluate if these variables had a Tr Trawl (Site(Estuary(Disturbance Category))) — (90 trawls, significant correlation with the relative body size of flounder or ac- Random) cumulation of trace metals. At each sampling time and location physico-chemical data were collected using a YSI-Sonde 6820-V2 Analyses of covariation of physico-chemical, tissue metal, flounder (Yellow Springs, USA) (calibrated weekly). Physico-chemical data relative body size, and sediment size variables were conducted using collected using the YSI-Sonde included temperature, salinity, pH, a Draftsman's Plot in combination with Principal Component Analysis dissolved oxygen, and turbidity values. At each site benthic sedi- (PCA) plots. The Draftsman's Plot computes the correlation co- ments were collected once at 5 m depth between Feb–Mar 2011 efficient between pairs of variables in order to provide a rough esti- using a sediment grab. Grab sediments were sorted through a series mate of the degree to which these variables are interrelated (Clarke of filters to measure the distribution of grain sizes (Dafforn et al., and Ainsworth, 1993). All tissue trace metal values which were en- 2012). For the purposes of this analysis grains were divided into countered above detection limits in the majority of samples were in- two broad categories, coarse grains (>2 mm) and fine grains cluded in this analysis (described in Table 1). Only values for trace (b63 μm). Surficial sediment samples were oven-dried before metals which were highly correlated between muscle tissue and rel- being digested and analyzed with ICP-OES following standard ative body size of flounder are included from the sediment samples 88 A.C. McKinley et al. / Science of the Total Environment 423 (2012) 84–94

a b

Heavily Modified Estuaries Relatively Unmodified Estuaries Heavily Modified Estuaries Relatively Unmodified Estuaries

c d y = 47.098 * x - 60.608 R2 = 0.596

Heavily Modified Estuaries Relatively Unmodified Estuaries

Fig. 2. Mean (±SE) physical characteristics of flounder by estuary and disturbance category, including a) age, b) flounder weight, and c) relative body size. d) Linear regression of flounder weight by age, indicating an approximately linear growth pattern up to fourth year.

(Cu, Fe, and Zn). Hg and Se values were not available from sediment two outlier data points. These outliers were identified through the samples. Among the physico-chemical variables, salinity, tempera- use of exploratory scatter plots and consisted of one sample each ture, and % fine grains were included in the Draftsman's Plot and from Port Hacking and Port Jackson (Fig. 4). PCA analyses. Dissolved oxygen, turbidity, and pH were excluded as they are not expected to have biological relevance for this analysis, 3. Results and initial analysis indicated no correlation between these variables and the relative body size of flounder (Lemly, 1993). The proportion 3.1. Flounder physical characteristics of sediment that was coarse grains were omitted as this value is by definition inversely related to % fine grains. Variables included in In total 184 flounder were captured in the course of this study; the Draftsman's Plot were also analyzed graphically using the multi- 154 fish were retained for analysis. Of these 153 were successfully variate PCA function of PERMANOVA. PCA is an ordination technique processed for tissue trace metal concentrations while otoliths were which analyzes any symmetric distance matrix. This analysis is also obtained from 150 fish (149 fish yielded data for both measures). called multi-dimensional scaling (Anderson, 2003). PERMANOVA sta- Fish were obtained in numbers which were sufficient for statistical tistical results are presented in text as (MSdf =0.00, p=0.000), analysis in every estuary (>8 individuals) and were obtained at 3–4 where MS=mean squares, df=degrees of freedom, and p=proba- sites in each estuary. bility value. The abundance of fish differed by estuary (MS6 =0.70, p=0.008) Physico-chemical variables (temperature) and tissue trace metal and site (MS30 =0.42, p=0.060) with significantly more fish in the variables (Cu, Zn, Hg, and Se) which showed a medium or high corre- Hunter River, Botany Bay, and Karuah River compared to the other es- lation to the relative body size of flounder in the Draftsman's Plot tuaries. The relative body size (MS20 =246.72, p=0.040), age (Table 4) were analyzed in more detail through the use of a Linear (MS20 =0.48, p=0.049), and weight (MS20 =2908.9, p=0.042) of Mixed Model (LMM) and regression plots. Fe was also analyzed in flounder samples all differed significantly by site but not by estuary, the LMM as it showed a comparatively strong correlation to the rela- though there were trends towards differences in these measures tive body size of flounder. The LMM procedure allows a linear regres- amongst estuaries (Fig. 2 a–c). Notably, none of these measures dif- sion between two variables to be calculated within the context of the fered significantly by disturbance category (p>0.05). All flounder larger sampling design. In this analysis fixed effects are evaluated in- displayed 1–4 annuli rings (approximately 1–4 years of age), and dependently of random factors. All data used in the LMM procedure the average age of fish was similar across all estuaries with 92% of was approximately normally distributed following the removal of fish displaying 2 or 3 annuli rings (Fig. 2a). Despite being similar in A.C. McKinley et al. / Science of the Total Environment 423 (2012) 84–94 89 age, the relative body size of flounder ranged widely and was approx- Correlations between the relative body size of flounder, tissue imately 5–60 g y− 1 (Fig. 2c). Relative body size increased in a linear metal concentrations, physico-chemical variables, sediment size, fashion from the first to the fourth year (Fig. 2d). This indicates that and sediment trace metal concentrations (Table 3) were evaluated the majority of fish captured in this study were at a stage of their using a series of bivariate correlations (Draftsman's Plot) (Table 4) life cycle where mass was linearly increasing with age. Most fish in conjunction with a multivariate PCA analysis (Fig. 3). Cu, Zn, and were of similar size, with approximately 80% of samples 10–20 cm temperature displayed strong negative correlations to the relative long, weighing 10–100 g. However, total variation in size was large, body size of flounder. In contrast, Hg and Se displayed strong positive and specimens ranged from 8 to 27 cm total length (TL) and 7– correlations with the relative body size of flounder (Table 4). Fe, Zn, 240 g. As such, the majority of specimens in this study were below and Cu are all highly correlated with one another and trend in the maturity, which is reached at approximately 25–30 cm (Gomon same direction in the PCA (Fig. 3a). Hg and Se, while highly positively et al., 2008). correlated to the relative body size of flounder, are not strongly correlated to one another or most other trace metals. Among the physico-chemical variables, only temperature showed 3.2. Concentrations of trace metals in muscle tissue a strong correlation to the relative body size of flounder. Notably, no other physico-chemical, sediment size, or sediment trace metal vari- Values for As, Cu, Fe, Mn, Hg, Se, and Zn were sufficient for further ables were strongly correlated to the relative body size of flounder analysis. None of the trace metals differed significantly by disturbance (Table 4). In addition, the levels of Cu, Fe, and Zn in the sediment category (p>0.05). Cd, Cr, and Ag concentrations in muscle tissue were not strongly correlated to the concentrations of these trace were below detection limits in all samples, and so were omitted metals in muscle tissue or the relative body size of flounder. However, from further analysis (Table 1). Pb and Ni concentrations were Cu, Fe, and Zn in the sediment were strongly positively correlated to below detection limits in 50–70% of the samples and so were omitted one another, while temperature and salinity were strongly negatively from further analysis due to low replication (Table 1). Hg correlated. (MS =0.08, p=0.001), Se (MS =3.96, p=0.001), and As 6 6 While correlations of this nature suggest relationships between (MS =31.89, p=0.002) differed significantly by estuary, while As 6 the variables discussed, they are insufficient to fully evaluate these re- (MS =5.61, p=0.032) also differed significantly by site. Concentra- 20 lationships within the context of the larger sampling design. Key relations of As in muscle tissue were significantly greater in Botany Bay, tionships between relative body size of flounder and highly Port Kembla, and Port Hacking compared to other estuaries, Hg con- correlated variables (Table 4) were further investigated using a Linear centrations were greater in Port Jackson and Karuah River, and Se Mixed Model (LMM). This approach calculates a linear regression be- concentrations were higher in Port Kembla than other estuaries tween two variables within the context of the larger statistical design (Table 1). Zn, Cu, Fe, and Mn did not differ by site or estuary (Table 5). The relative strength of these relationships were evaluated (p>0.05) (Table 1). through regression plots (Fig. 4). European Union, US Federal Drug Administration and Australian LMM analysis indicated a significant regression between the rela- federal regulated limits for human consumption of some trace metal tive body size of flounder and tissue concentrations of Cu, Zn, Hg, Se, contaminants in fish tissues are listed in Table 1. Regulated limits and Fe, but not temperature (Table 5). The relative body size of floun- have not been set for all trace metals. In no case did the estuary aver- der showed a strong negative relationship to Zn and a moderately age exceed regulated limits for human consumption; although four strong negative relationship to Cu (Fig. 4a,b). Temperature also samples did exceed regulated limits individually. These included showed a moderately strong negative relationship to the relative two samples which exceeded European Union limits for Pb (0.22, − body size of flounder, though this was not significant in the LMM 0.33 μgg 1) both of which came from the Kogarah Bay area of Botany analysis. Lastly, the relationship between Fe and the relative body Bay, and two samples which exceeded European Union and Austra- − size was comparatively weak (Fig. 4e,f). In contrast, Se showed a lian limits for Hg (0.50, 0.81 μgg 1) both originating from the Iron strong positive relationship to the relative body size of flounder, Cove area of Port Jackson. In all of these cases the samples did not ex- while the relationship between Hg and relative body size was weaker ceed the higher USFDA limits (EC, 2005; Rayment, 1991; USFDA, but still positive (Fig. 4c,d). 2001). In summary, only Cu, Zn, and Se were both statistically significant in the LMM analysis, while also displaying a moderate or strong rela- 3.3. Relationships between variables tionship to the relative body size of flounder in the regression plots. The effect sizes of these relationships are comparatively large, as in- Physico-chemical and sediment size values are displayed in creasing concentrations of trace metals in the tissue were correlated

Table 2. Temperature (MS6 =31.89, p=0.002), pH (MS6 =0.08, with substantial changes in the relative body size of ﬂounder p=0.001), and turbidity (MS6 =3.96, p=0.001) were all found to (Fig. 4). Cu showed a large negative effect size with an increase in differ signiﬁcantly by estuary while temperature also differed by Cu concentrations of 250% (0.4–1.0 μgg− 1) corresponding to a de- − 1 site (MS20 =5.61, p=0.032). Salinity and dissolved oxygen did not cline in relative body size of −70% (from ~35 to 10 g y ). Zn differ by estuary or site (p>0.05). showed a large negative effect size with an increase in Zn

Table 2 Mean (±SE) values for physico-chemical and sediment size covariates. Sediment size covariates represent average of inner estuary sites from Dafforn et al. (2012).

Temperature Salinity (‰) DO (%) pH Turbidity % Coarse grains % Fine grains (°C) (NTU+) (> 2 mm) (b 63 μm)

Heavily modiﬁed estuaries Botany Bay 23.83 (±0.05) 27.11 (±0.24) 83.77 (±0.69) 8.12 (±0.01) 3.23 (±0.07) 0.11 53.06 Port Jackson 20.10 (±0.12) 33.73 (±0.21) 138.74 (±8.72) 7.96 (±0.01) 2.48 (±0.26) 0.60 30.73 Port Kembla 18.72 (±0.10) 34.60 (±0.07) 96.30 (±0.99) 8.22 (±0.01) 2.06 (±0.03) 7.43 22.37 Newcastle 20.20 (±0.12) 26.46 (±0.33) 101.00 (±0.52) 8.09 (±0.03) 1.93 (±0.13) 2.39 62.45 Relatively unmodiﬁed Broken Bay 22.46 (±0.07) 25.78 (±0.33) 105.93 (±0.89) 7.92 (±0.01) 12.97 (±1.88) 1.56 71.71 estuaries Clyde River 23.32 (±0.58) 29.11 (±0.89) 125.21 (±3.89) 7.98 (±0.03) 1.63 (±0.18) 1.19 5.75 Karuah River 21.90 (±0.09) 25.88 (±1.14) 109.32 (±4.76) 7.93 (±0.03) 4.54 (±0.32) 6.40 38.92 Port Hacking 19.18 (±0.09) 31.86 (±0.54) 129.54 (±3.33) 7.99 (±0.01) 0.41 (±0.10) 18.23 3.84 90 A.C. McKinley et al. / Science of the Total Environment 423 (2012) 84–94

Table 3 Mean (±SE) concentrations (μgg− 1) of trace metals in benthic sediments. *Sn samples b5.00 indicate all samples below detection limits. Hg and Se values were not available (Daf- forn et al., 2012).

Al As Cd Co Cr Cu Fe Mn Ni Pb Sn* V Zn

Heavily modiﬁed Botany 11476 18.5 0.56 8.1 31.8 34.2 (±0.8) 26983 136 11.5 70.9 1.52 44.6 217 estuaries Bay (±632) (±0.4) (±0.03) (±0.1) (±1.2) (±644) (±4.4) (±0.4) (±1.7) (±0.34) (±1.2) (±4.4) Port 9727 21.4 0.60 7.4 75.8 133 (±10.2) 25956 120 11.5 218 10.8 46.3 498 Jackson (±714) (±1.0) (±0.07) (±0.5) (±7.7) (±1644) (±8.6) (±0.9) (±17.3) (±0.49) (±2.9) (±36) Port 7684 12.1 1.31 10.0 67.7 141 (±13.9) (±3524) 608 18.2 128 201 83.9 638 Kembla (±450) (±0.8) (±0.14) (±0.7) (±7.5) 40596 (±52) (±1.4) (±14.4) (±23.7) (±8.2) (±90) Newcastle 22814 10.9 1.08 16.1 51.1 82.3 (±4.0) 37137 308 40.5 80.5 b5.00 63.1 356 (±279) (±0.2) (±0.02) (±0.3) (±0.9) (±252) (±1.4) (±0.1) (±4.4) (±0.1) (±16) Relatively unmodiﬁed Broken 7752 15.0 0.53 9.6 16.1 12.8 (±0.2) 24791 305 11.4 23.8 b5.00 39.0 63.4 estuaries Bay (±97.2) (±0.1) (±0.01) (±0.1) (±0.2) (±130) (±4.6) (±0.1) (±0.2) (±0.4) (±0.8) Clyde 3896 5.8 0.31 4.0 6.5 2.4 (±0.4) 9506 50.3 6.0 6.6 b5.00 9.7 24.6 River (±256) (±0.4) (±0.01) (±0.2) (±0.5) (±525) (±2.2) (±0.3) (±0.4) (±0.8) (±1.4) Karuah 10310 7.8 0.28 4.8 8.1 4.1 (±0.5) 13330 98.9 4.2 12.3 b5.00 21.4 38.1 River (±895) (±0.6) (±0.02) (±0.1) (±0.6) (±79.4) (±14.4) (±0.4) (±0.9) (±0.3) (±1.6) Port 2852 8.2 0.29 2.0 7.5 11.4 (±0.7) 8228 31.7 2.6 27.3 3.16 18.5 53.5 Hacking (±201) (±0.1) (±0.01) (±0.1) (±0.3) (±196) (±1.5) (±0.1) (±1.4) (±1.72) (±0.7) (±1.9)

concentrations of +67% (from ~15 to 25 μgg−1) corresponding to a concentrations or flounder relative body size differ significantly by decline in relative body size of −80% (~50–10 g y− 1). Beyond disturbance category. This is despite the fact that concentrations of 25 μgg− 1 increased Zn concentrations did not show any significant trace metals in the sediment are known to be more than 10× greater additional impact on the relative body size of flounder. In contrast, in the heavily modified estuaries compared to the relatively unmo- Se showed a large positive effect size with an increase in Se concen- dified estuaries (Dafforn et al., 2012). trations of 300% (1.0–3.0 μgg− 1) correlating to an increase in the relative body size of 600% (~5–30 g y− 1). Beyond 3.0 μgg− 1 an increase 4.1. Negative relationship between Cu, Fe, Zn, temperature, and relative in Se concentrations did not show any significant addition impact on body size the relative body size of flounder. Strong correlations between the relative body size of flounder (as 4. Discussion an indirect measure of the growth rate over time) and the concentrations of trace metals in the tissue of flounder are described in this We have demonstrated that the edible flounder P. jenynsii has study. These relationships may be mediated or influenced by a vari- detectable concentrations of trace metals in their muscle tissue, ety of factors including the bioavailability of the contaminant in the both in relatively unmodified and heavily modified estuaries. sediment, the interrelationships between contaminants, the way While these trace metals were only above regulated safety limits that the contaminant bioaccumulates, and relationships between for human consumption in four fish, they do represent comparative- the accumulation of the contaminant and mediating factors, such ly high body burdens of trace metals contaminants. Some trace as the food supply (Deb and Fukushima, 1999; Klaverkamp et al., metals concentrations (notably Cu and Zn) were several times 1983; van der Oost et al., 2003). greater than have been found in this species in other parts of Austra- Cu, Zn, and Fe are essential for fish metabolism and are taken up lia (Plaskett and Potter, 1979). Concentrations of Cu and Zn were naturally by fish as part of their normal metabolic processes (Bury also comparable to levels which have been shown or predicted to et al., 2003; Wood, 1992). However, uptake of trace metals can lead have biological effects in other flounder species (Cossa et al., 1992; to bioaccumulation in tissues when intake from food or environmen- de Boer et al., 2001). It is significant to note that in no case did tissue tal sources exceeds the rate of efflux (Deb and Fukushima, 1999). It

Table 4 Draftsman's Plot of correlations between relative body size of ﬂounder (g y− 1), concentrations of trace metals in muscle tissue, physico-chemical covariates, and concentrations of trace metals in benthic sediments. Only values for trace metals which highly correlate between muscle tissue and relative body size of ﬂounder are included from the sediment samples (Cu, Fe, and Zn). Hg and Se values were not available from sediment samples.

Relative body As Cu Fe Hg Mn Se Zn Temperature Salinity % Fine Cu Fe size (g y− 1) (Tissue) (Tissue) (Tissue) (Tissue) (Tissue) (Tissue) (Tissue) Grain (Sediment) (Sediment)

Arsenic (tissue) −0.048 Copper (tissue) −0.440 0.218 Iron (tissue) −0.258 0.111 0.330 Mercury (tissue) 0.427 −0.261 −0.147 −0.226 Manganese (tissue) −0.083 −0.088 0.133 −0.015 0.022 Selenium (tissue) 0.376 −0.034 −0.268 −0.146 0.090 0.017 Zinc (tissue) −0.486 0.015 0.462 0.214 −0.299 0.180 −0.384 Temperature −0.378 0.053 0.331 0.079 0.133 −0.077 −0.447 0.155 Salinity 0.209 0.227 −0.116 −0.001 −0.085 0.074 0.475 −0.167 −0.521 % Fine grain −0.116 −0.197 0.220 −0.007 0.000 0.015 −0.317 0.210 0.449 −0.602 Copper (sediment) 0.107 0.018 −0.017 0.068 −0.068 0.182 0.531 0.003 −0.490 0.422 0.009 Iron (sediment) −0.034 −0.020 0.118 0.097 −0.212 0.103 0.311 0.145 −0.138 0.070 0.482 0.798 Zinc (sediment) 0.024 0.055 0.030 0.115 −0.119 0.144 0.478 0.037 −0.382 0.375 0.048 0.962 0.858 High correlation ±0.5–1.0 Medium correlation ±0.3–0.5 A.C. McKinley et al. / Science of the Total Environment 423 (2012) 84–94 91

4 ab4

2 2

Fe (Sediment) Zn Zn (Sediment) Fe Cu % Fine Grains 0 0 Cu (Sediment) Temperatue PC2 PC2 Se Salinity -2 -2 Hg Relative Body Size of Flounder (g y -1)

-4 -4

-6 -6 -8 -6 -4 -2 0 2 -8 -6 -4 -2 0 PC1 PC1

Fig. 3. Principal Components Analysis (PCA) of multivariate relationships between a) concentrations of highly correlating trace metals in flounder muscle tissue and relative body size of flounder (g y− 1) and b) physico-chemical covariates, grain size, and concentrations of trace metals in the sediment. Only values for trace metals which highly correlate (>0.3, Table 4) between muscle tissue and relative body size of flounder are included from the sediment samples (Cu, Fe, and Zn). Hg and Se values were not available from sediment samples. has been well established that metabolic activity is one of the most effects due to differences between year classes, given that we know important factors governing trace metal accumulation in fish and in the growth pattern is approximately linear (Fig. 2d). The approxi- many species metabolic activity declines with age and maturity mately linear growth pattern suggests that fish are not experiencing (Douben, 1989; Nussey et al., 2000; Widianarko et al., 2000). In a decline in their rate of growth (and hence metabolism) up to their some cases concentrations of these trace metals have been shown fourth year (Machiels and Henken, 1985). For these reasons it is un- to reach a steady state in mature fish, where uptake and efflux rates likely that differences in metabolic activity between fish of different are balanced and no further net accumulation occurs (Widianarko ages are the primary driver of the negative relationships observed et al., 2000). This has been attributed to declining metabolic activity in this study. in older fish, which reduces the net accumulation rate of some trace Our results indicate that Cu, Zn, and Fe concentrations in the fish tis- metals including Cu, Zn, and Fe (Douben, 1989; Nussey et al., 2000; sue were negatively correlated to the relative body size of flounder. Cu Widianarko et al., 2000). Therefore, one possible explanation for the and Zn are known to have negative physiological and developmental observed negative relationships between relative body size and tissue impacts on fish at concentrations comparable to those observed in concentrations of Zn, Cu, and Fe may be differences in metabolic ac- this study (Buckley et al., 1982; Clearwater et al., 2002; Kearns and tivity between younger and older fish. However, other studies have Atchison, 1979; Waiwood and Beamish, 1978). Therefore, one possible generally not controlled for differences in the age of the fish and explanation is that the strong negative relationships between Cu, Zn, have instead calculated the relationship between body size and and the relative body size of flounder are at least partially due to the trace metal concentrations directly (Canli and Atli, 2003; Liang toxicity of these trace metals. Reduced relative body size in association et al., 1999). Our measure of relative body size (g y− 1) approximates with Zn and Cu accumulation may imply that the adverse toxicological the mass of the fish per year of life, which should largely control for effects of these trace metals are impacting the growth rate of this

Table 5 Multivariate tests of the relationship between relative body size of flounder (g y− 1) and covariate factors within the context of a linear mixed model (LMM) for a) Cu, b) Zn, c) Hg, d) Se, e) Fe, f) temperature. Factors: Dis=Disturbance Category (Heavily Modified vs. Relatively Unmodified), Es=Estuary, Si=Site, Tr=Trawl. Bold values are significant.

Tests of ﬁxed effects for relative body size (dependant variable)

a) Cu b) Zn c) Hg d) Se e) Fe f) Temperature

Source Denom. Fp-Denom. Fp-Denom. Fp-Denom. Fp-Denom. Fp-Denom. Fp- df value df value df value df value df value df value

Dis 6.10 0.05 0.824 6.77 0.03 0.865 5.54 0.00 1.000 4.99 0.65 0.458 6.99 0.002 0.962 4.53 0.89 0.392 independent 132.91 21.87 0.000 142.60 38.26 0.000 142.63 65.50 0.000 111.41 34.49 0.000 135.26 14.471 0.000 5.64 5.43 0.061 variable Intercept 100.25 80.45 0.000 48.16 140.93 0.000 9.32 2.73 0.132 24.84 0.00 0.949 33.89 109.04 0.000 0.00 12.91 0.012 Tests of random factors covariance parameters

a) Cu b) Zn c) Hg d) Se e) Fe f) Temperature

Estimate Std. error Estimate Std. error Estimate Std. error Estimate Std. error Estimate Std. error Estimate Std. error

Es(Dis) 10.46 13.335 16.74 13.050 70.39 49.040 41.98 31.957 11.46 13.786 5.07 11.167 Si(Es(Dis)) 18.38 12.385 5.56 7.027 18.61 12.355 5.74 7.349 22.74 13.081 20.23 12.151 Tr(Si(Es(Dis))) 0.00 0.000 0.00 0.000 0.00 0.000 0.00 0.000 0.00 0.000 0.00 0.000 Res 73.61 9.484 71.18 9.119 54.96 7.191 70.92 9.035 76.31 9.747 83.63 10.674 92 A.C. McKinley et al. / Science of the Total Environment 423 (2012) 84–94 a b

LMM p-value = 0.000 LMM p-value = 0.000

c d

LMM p-value = 0.000 LMM p-value = 0.000

e f

LMM p-value = 0.000 LMM p-value = 0.061

Fig. 4. Regression of relationship between flounder relative body size and concentrations of trace metals in muscle tissue. Including a) Cu, b) Zn, c) Hg, d) Se, e) Fe, and f) Temper- ature. Linear Mixed Model (LMM) p-value corresponds to multivariate tests of the relationship between flounder relative body size and covariate factors within the context of a linear mixed model, as shown in Table 5. species. The effect sizes associated with these relationships were com- Zn, and Fe were found to be strongly covarying both in the tissue of paratively large, and so it is possible that Zn and Cu accumulation in flounder and in the sediments. It is possible that the weak negative the muscle tissue of P. jenynsii are having a significant impact on the relationship between Fe and the relative body size of flounder is health and ecology of this species. due to covariation with Cu and Zn, rather than any direct effects of In contrast, Fe has generally only been shown to have negative Fe at the concentrations observed. physiological and body size impacts at concentrations substantially While temperature also displayed a moderately strong negative higher than those observed in this study (Canli and Atli, 2003). Cu, relationship to the relative body size of flounder in the Draftsman's A.C. McKinley et al. / Science of the Total Environment 423 (2012) 84–94 93

Plot, this relationship was not significant within the context of the 2000). Uptake mechanisms and pathways are complex and may be LMM analysis (p=0.061). This is possibly due to significant variation mediated by prey species (Deb and Fukushima, 1999), patchy distri- in temperature between estuaries and sites. The relationship between butions of contaminants in the sediment (Dafforn et al., 2012), com- temperature and the relative body size of the fish is not likely to be plex contaminant mixtures (e.g. sewage) which include both food indicative of a latitudinal gradient in growth rates, as the relative subsidies and toxicological contaminants (Connolly et al., 2009; body size of fish did not differ significantly by estuary. It may instead Gaston and Suthers, 2004; McKinley and Johnston, 2010), and be indicative of P. jenysii's preference for the inner areas of an estuary, other environmental factors (Lemly, 1993). Differences in tissue where conditions are usually warmer compared to sites closer to the trace metal concentrations may also be due to differential resistance estuary mouth (McKinley et al., 2011). Both temperature and the rel- and tolerance, both between individuals and populations. Physio- ative body size of the fish were found to differ significantly by site, logical resistance to trace metals uptake and relative tolerance to which could be consistent with this hypothesis. The relative position tracemetalsexposurehasbeendemonstratedinsomefish of sites within the estuary was not explicitly evaluated in this (Wirgin and Waldman, 2004). A variety of mechanisms could ac- analysis. count for differential tolerance to contaminants between individuals and populations, for example, differential capacity to produce 4.2. Positive relationship between Hg, Se, and relative body size metal binding proteins such as metallothioneins (Xie and Klerks, 2004). Se showed a strong positive relationship to the relative body size of flounder (as an indirect measure of the growth rate over time). 5. Conclusion Hg also showed a positive relationship to the relative body size of flounder, though this relationship was relatively weak. We have demonstrated that the use of otolith aging and growth Se concentrations in this study were low compared to values ob- measurement techniques represents an effective and novel method served in similar studies, and are not likely to have large physiological for assessing the relationships between trace metal accumulation and developmental impacts on fish at the levels observed (Hodson and the relative body size of fish in a field environment. Surprising- and Hilton, 1983). Similarly, observed Hg concentrations were ly, concentrations of Cu, Zn, and Fe in muscle tissue showed almost below levels normally associated with adverse physiological impacts no relationship to concentrations of these trace metals in the sedi- in fish (Friedmann et al., 1996; Latif et al., 2001). In spite of any po- ment, and P. jenynsii appears to be accumulating significant levels tential toxicological effects, both trace metals showed a positive rela- of these trace metals in both relatively unmodified and heavily tionship with the flounder relative body size. This implies that any modified estuaries. It is also interesting to note that in no case did toxicity effects associated with the accumulation of these trace metals sediment trace metals concentrations show a strong relationship are being outweighed by other positive growth factors. to the relative body size of flounder, and in no case did relative While it has been argued that Cu, Zn, and Fe concentrations can body size of flounder, tissue trace metals, or physico-chemical vari- reach an equilibrium state where uptake and efflux rates are equiva- ables differ signi ficantly between heavily modified and relatively lent (Dang et al., 2009; Widianarko et al., 2000; Zhang and Wang, unmodified estuaries. 2005), Se and Hg concentrations have generally been shown to con- In our study Se and Hg showed a positive relationship to the rela- tinually increase with continued exposure (Barak and Mason, 1990; tive body size of P. jenynsii, while Cu, Zn, and Fe were negatively relat- Klaverkamp et al., 1983). This is because little efflux of Se and Hg oc- ed to the relative body size. The use of fish aging techniques has curs (Campbell et al., 2008; Deb and Fukushima, 1999; Hodson and allowed us to eliminate the possibility that these effects are driven Hilton, 1983). The main pathway for exposure to Hg and Se in fish by age related differences in metabolic activity. Instead, we suggest is by ingestion of food particles (Creighton and Twining, 2010; that differences in food supply likely explain the relationship be- Dallinger et al., 1987; Mathews and Fisher, 2009). Because there is lit- tween Se concentrations and relative body size. In contrast, we have tle efflux of Se and Hg, fish which have eaten more food are likely to suggested that toxicological effects may be responsible for the ob- have greater concentrations of these trace metals in their tissues (if served relationships between relative body size and concentrations the concentrations in food sources are equal). Because our measure of Cu and Zn in muscle tissues. The effect sizes associated with of relative body size (g y−1) controls for age differences, fish of larger these relationships were comparatively large, and so it is possible relative body size must have grown faster than smaller fish. Fish that Zn and Cu accumulation in the muscle tissue of P. jenynsii is hav- which grow faster are likely to be consuming food quicker than ing a significant impact on the health and ecology of this species. slower growing fish, which increases their ingestion of Hg and Se in contaminated environments. Thus, it is possible that the observed positive relationship between Se, Hg, and the relative body size of Acknowledgements flounder is due to increased food intake by faster growing fish. How- ever, this relationship cannot be confirmed without comparing values This research was primarily supported by the Australian Research of Hg and Se in the sediment, which were not available in this study. Council through an Australian Research Fellowship awarded to ELJ For all studies of trace metals accumulation, dilution of the con- and a Linkage Grant awarded to ELJ. The writing of this manuscript taminant in the tissues of faster growing animals can be a confound- was also supported through the Canadian National Sciences and Engi- ing factor. Dilution may be a substantial issue in situations where neering Research Council through an award given to ACM. We would growth outpaces contaminant uptake and this can be explicitly stud- like to thank Dr. Katherine Dafforn, Shinjiro Ushiama, David Day, and ied in controlled laboratory settings (Sprague, 1973). In our extensive Valeriya Komyakova for their help with this project. We would also field survey we explicitly examine relationships between relative like to thank Stuart Simpson and Chad Jarolimek at the CSIRO Centre body size and tissue concentrations, but cannot distinguish tissue di- for Environmental Contaminants Research for their help with the tis- lution from changes to exposure or uptake in the field. sue trace metals analysis, and Jerome Stocks at the Cronulla Fisheries Research Centre of Excellence for his help with otolith aging tech- 4.3. Relationship between sediment and tissue trace metals niques. Lastly, we would like to thank the Bluescope Steel Company for their generous support of this research and Marine Parks NSW While it is surprising that concentrations of trace metals in the for their help and advice. This study was approved and carried out sediment were not correlated to levels in the tissues of P. jenynsii, in strict accordance with the recommendations of the Animal Care a variety of factors may explain this disjuncture (Widianarko et al., and Ethics Committee of the University of New South Wales (Project 94 A.C. McKinley et al. / Science of the Total Environment 423 (2012) 84–94

No. 09/110A) and the New South Wales Department of Primary Hill NA, King CK, Perrett LA, Johnston EL. Contaminated suspended sediments toxic to an Antarctic filter feeder: aqueous- and particulate-phase effects. Environ Toxicol Industries (Permit No. P09/0072-1.0). Chem 2009;28:409–17. Hodson PV, Hilton JW. The nutritional requirements and toxicity to fish of dietary and References waterborne selenium. Environ Biochem 1983;35:335–40. Johnston E, Roberts DA. Contaminants reduce the richness and evenness of marine – Al-Yousuf MH, El-Shahawi MS, Al-Ghais SM. Trace metals in liver, skin and muscle of communities: a review and meta-analysis. Environ Pollut 2009;157:1745 52. fl Lethrinus lentjan fish species in relation to body length and sex. Sci Total Environ Kearns P, Atchison G. Effects of trace metals on growth of yellow perch (Perca aves- – – 2000;256:87–94. cens) as measured by RNA DNA ratios. Environ Biol Fishes 1979;4:383 7. Anderson MJ. A new method for non-parametric multivariate analysis of variance. Aus- Klaverkamp JF, Turner MA, Harrison SE, Hesslein RH. Fates of metal radiotracers added tral Ecol 2001;26:32–46. to a whole lake: accumulation in slimy sculpin (Cottus cognatus) and white sucker – Anderson MJ. PCO — Principal Coordinate Analysis: A Computer Program. University of (Catostomus commersoni). Sci Total Environ 1983;28:119 28. Aukland; 2003. Knott NA, Aulbury J, Brown T, Johnston EL. Contemporary ecological threats from his- Barak NAE, Mason CF. Mercury, cadmium and lead in eels and roach: the effects of size, torical pollution sources: impacts of large-scale resuspension of contaminated sed- – season and locality on metal concentrations in flesh and liver. Sci Total Environ iments on sessile invertebrate recruitment. J Appl Ecol 2009;46:770 81. 1990;92:249–56. Latif MA, Bodaly RA, Johnston TA, Fudge RJP. Effects of environmental and maternally Birch G, Taylor S. Source of heavy metals in sediments of the Port Jackson estuary, Aus- derived methylmercury on the embryonic and larval stages of walleye (Stizoste- – tralia. Sci Total Environ 1999;227:123–38. dion vitreum). Environ Pollut 2001;111:139 48. fi Brown B, Ahsanullah M. Effect of heavy metals on mortality and growth. Mar Pollut Lemly DA. Metabolic stress during winter increases the toxicity of selenium to sh. – Bull 1971;2:182–7. Aquat Toxicol 1993;27:133 58. Buckley JT, Roch M, McCarter JA, Rendell CA, Matheson AT. Chronic exposure of coho Liang Y, Cheung RYH, Wong MH. Reclamation of wastewater for polyculture of fresh- fi fi – salmon to sublethal concentrations of copper—I. Effect on growth, on accumulation water sh: bioaccumulation of trace metals in sh. Water Res 1999;33:2690 700. and distribution of copper, and on copper tolerance. Comp Biochem Physiol C Machiels MAM, Henken AM. Growth rate, feed utilization and energy metabolism of fi Comp Pharmacol 1982;72:15–9. the African cat sh, Clarias gariepinus (Burchell, 1822), as affected by dietary pro- – Bury NR, Walker PA, Glover CN. Nutritive metal uptake in teleost fish. J Exp Biol tein and energy content. Aquaculture 1985;44:271 84. 2003;206:11–23. Mathews T, Fisher NS. Dominance of dietary intake of metals in marine elasmobranch fi – Campbell L, Verburg P, Dixon DG, Hecky RE. Mercury biomagnification in the food web and teleost sh. Sci Total Environ 2009;407:5156 61. of Lake Tanganyika (Tanzania, East Africa). Sci Total Environ 2008;402:184–91. McKinley AC, Dafforn KA, Taylor MD, Johnston EL. High levels of sediment contamina- fl fi Canli M, Atli G. The relationships between heavy metal (Cd, Cr, Cu, Fe, Pb, Zn) levels tion have little in uence on estuarine beach sh community indices compared to and the size of six Mediterranean fish species. Environ Pollut 2003;121:129–36. physico-chemical variation. PLoS One 2011;6(10). fi Clarke KR, Ainsworth M. A method of linking multivariate community structure to en- McKinley AC, Johnston EL. Impacts of contaminant sources on marine sh abundance fi vironmental variables. Mar Ecol Prog Ser 1993;92:205–19. and species richness: a review and meta-analysis of evidence from the eld. Mar – Clearwater SJ, Farag AM, Meyer JS. Bioavailability and toxicity of dietborne copper and Ecol Prog Ser 2010;420:175 91. fi zinc to fish. Comp Biochem Physiol C Toxicol Pharmacol 2002;132:269–313. Mondon JA, Duda S, Nowak BF. Histological, growth and 7-ethoxyresoru nO- fl Connolly RM, Schlacher TA, Gaston TF. Stable isotope evidence for trophic subsidy of deethylase (EROD) activity responses of greenback ounder Rhombosolea tapirina – coastal benthic fisheries by river discharge plumes off small estuaries. Mar Biol to contaminated marine sediment and diet. Aquat Toxicol 2001;54:231 47. Res 2009;5:164–71. Nussey G, Van Vuren JHJ, Preez HHD. Bioaccumulation of chromium, manganese, nickel Cossa D, Auger D, Averty B, Lucon M, Masselin P, Noël J. Flounder (Platichthys flesus)mus- and lead in the tissues of the moggel, Labeo umbratus (Cyprinidae), from Witbank cle as an indicator of metal and organochlorine contamination of French Atlantic Dam, Mpumalanga, Vol. 26. Pretoria, South Africa: Water Research Commission; coastal waters. Ambio 1992;21:176–82. 2000. Creighton N, Twining J. Bioaccumulation from food and water of cadmium, selenium Plaskett D, Potter IC. Heavy metal concentrations in the muscle tissue of 12 species of and zinc in an estuarine fish, Ambassis jacksoniensis. Mar Pollut Bull 2010;60: teleost from Cockburn Sound, Western Australia. Aust J Mar Freshw Res 1979;30: – 1815–21. 607 16. Dafforn KA, Simpson SL, Kelaher BP, Clarke G, Komyakova V, Wong CKC, et al. The chal- Rayment GE. Australian and some international food standards for heavy metals. lenge of choosing environmental indicators of anthropogenic impacts in estuaries. Queensland Department of Primary Industries; 1991. Available fromhttp://www. fi Environ Pollut 2012;163:207–17. gbrmpa.gov.au/__data/assets/pdf_ le/0012/4143/ws016_paper_09.pdf (Accessed Dallinger R, Prosi F, Segner H, Back H. Contaminated food and uptake of heavy metals June 19, 2011). by fish: a review and a proposal for further research. Oecologia 1987;73:91–8. Scanes P. NSW Estuarine catchment disturbance ranks. NSW Department of Environ- Dang F, Zhong H, Wang WX. Copper uptake kinetics and regulation in a marine fish ment. Sydney: Climate Change, and Water; 2010. fi after waterborne copper acclimation. Aquat Toxicol 2009;94:238–44. Sprague JB. The ABC's of pollutant bioassay using sh. Biological methods for the as- de Boer J, van der Zande TE, Pieters H, Ariese F, Schipper CA, van Brummelen T, et al. sessment of water quality. ASTM STPAmerican Society for Testing and Materials; Organic contaminants and trace metals in flounder liver and sediment from the 1973. p. 6-30. Amsterdam and Rotterdam harbours and off the Dutch coast. J Environ Monit USEPA. Method 3051A microwave assisted acid digestion of sediments, sludges and 2001;3:386–93. oils. Washington: Environmental Protection Agency; 2007. fi Deb SC, Fukushima T. Metals in aquatic ecosystems: mechanisms of uptake, accumula- USFDA. Fish and sheries products hazards and controls guidance. Appendix 5 - tion and release: ecotoxicological perspectives. Int J Environ Stud 1999;56: FDA & EPA safety levels in regulations and guidance. Washington: United 385–417. States Federal Drug Administration; 2001. Available fromhttp://www.fda. Douben PE. Lead and cadmium in stone loach (Noemacheilus barbatulus) from three gov/Food/GuidanceComplianceRegulatoryInformation/GuidanceDocuments/ rivers in Derbyshire. Ecotoxicol Environ Saf 1989;18:35–58. Seafood/FishandFisheriesProductsHazardsandControlsGuide/ucm120108.htm. DPI. Survey of recreational fishing in New South Wales. Sydney: Department of Prima- (Accessed June 12, 2011). ry Industries; 2010. van der Oost R, Beyer J, Vermeulen NPE. Fish bioaccumulation and biomarkers in envi- EC. Commission regulation (EC) No 78/2005 amending regulation (EC) No 466/2001 as ronmental risk assessment: a review. Environ Toxicol Pharmacol 2003;13:57-149. regards heavy metals. European Commission, Brussels. Off J Eur Union 2005;16:43–5. Waiwood KG, Beamish FWH. The effect of copper, hardness and pH on the growth of – Francis RICC, Paul LJ, Mulligan KP. Ageing of adult snapper (Pagrus auratus) from oto- rainbow trout, Salmo gairdneri. J Fish Biol 1978;13:591 8. lith annual ring counts: validation by tagging and oxytetracycline injection. Aust Widianarko B, Van Gestel CA, Verweii RA, Van Straalen NM. Associations between trace J Mar Freshw Res 1992;43:1069–89. metals in sediment, water, and guppy, Poecilia reticulata (Peters), from urban – Friedmann AS, Watzin MC, Brinck-Johnsen T, Leiter JC. Low levels of dietary methyl- streams of Semarang, Indonesia. Ecotoxicol Environ Saf 2000;46:101 7. fi mercury inhibit growth and gonadal development in juvenile walleye (Stizostedion Wirgin I, Waldman JR. Resistance to contaminants in North American sh populations. vitreum). Aquat Toxicol 1996;35:265–78. Mutat Res-Fundam Mol Mech Mutagen 2004;552:73-100. fi Froese R, Pauly D. Fishbase. Available fromwww.fishbase.org2010. (accessed January Wood CM. Flux measurements as indices of H+ and metal effects on freshwater sh. – 25, 2010). Aquat Toxicol 1992;22:239 63. Gaston TF, Suthers IM. Spatial variation in δ13C and δ15N of liver, muscle and bone in a Xie L, Klerks PL. Changes in cadmium accumulation as a mechanism for cadmium resis- fi – rocky reef planktivorous fish: the relative contribution of sewage. J Exp Mar Biol tance in the least killi sh Heterandria formosa. Aquat Toxicol 2004;66:73 81. Ecol 2004;304:17–33. Zhang L, Wang WX. Effects of Zn pre-exposure on Cd and Zn bioaccumulation and fi GomonM,BrayD,KuiterR.FishesofAustralia'ssoutherncoast.Sydney:ReedNew metallothionein levels in two species of marine sh. Aquat Toxicol 2005;73: – Holland; 2008. 353 69. Hesslein RH, Broecker WS, Schindler DW. Fates of metal radiotracers added to a whole lake: sediment–water interactions. Can J Fish Aquat Sci 1980;37:378–86. Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Discussion and Conclusions

Summary The aim of this thesis was to examine the potential impacts of anthropogenic disturbance on the ecology of estuarine fish assemblages. Specifically, I investigated the potential impacts of habitat modification and contamination within the context of broader disturbance regimes. Using a variety of methods I documented large scale patterns of fish communities in heavily modified vs. relatively unmodified estuaries. First, a comprehensive meta-analysis identified trends and knowledge gaps. Studies on adult fish have generally shown weakly negative responses to contamination or largely positive responses where enriching contaminants are present. I conducted two field studies of adult fish, the results of which were broadly consistent with the literature. Surveys of large bentho-pelagic fish using underwater video showed increased abundance of recreationally targeted species in heavily modified/nutrient enriched estuaries. However, beach seine surveys of small bodied species indicated little influence on the beach fish community, even where high levels of modification and contamination were detected. Instead, beach fish were highly correlated to physico-chemical gradients. This suggests that potential impacts are highly variable among adult fish, and that both ecological characteristics and habitat preferences play an important mediating role. Prior to this thesis very little research had been published assessing the potential impacts of stressors on marine larvae. Substantial differences in larval communities were detected in heavily modified vs. relatively unmodified areas, including increased abundance and diversity, large shifts in the occurrence of species, and changes to the overall composition of the community. These trends were highly correlated to contamination of trace metals in the sediment and loss of seagrass cover. The strongest patterns were observed for fully estuarine taxa and those with benthic eggs. Lastly, a final project indicated the potential for the accumulation of some trace metals (Cu, Zn, and Se) in muscle tissue of Psuedorhombus jenynsii to affect the relative body size of the species. Some consistent trends in the results were found across multiple chapters, irrespective of the sampling methods, life cycle stage, and portion of the fish assemblage examined. These ‘common themes’ of the thesis are discussed below. Relatively few studies exist which investigate the potential impacts of contaminants and other anthropogenic stressors on wild estuarine fish assemblages. This thesis has investigated

xv Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

these potential impacts in detail and has contributed significantly to the literature in this field. I utilized both novel approaches to evaluate the possible influence of contamination (e.g. coupling otolith and tissue accumulation methods, utilizing an early life history guild approach, etc.) and established methods in new ways (e.g. Larval impacts studies, BRUV impact studies). The results of this thesis clearly indicate that estuarine fish assemblages are potentially being impacted by anthropogenic stressors. It has been shown that substantial differences exist in assemblages of fish in heavily modified vs. relatively unmodified estuaries. These potential impacts are ecologically significant and may have far reaching consequences for economically important species, the wider ecological system, and for other groups of organisms. Potential impacts on species which are economically and culturally significant have also been demonstrated. For these reasons, the results presented in this thesis should be of significant management interest because they provide a basis for understanding the large scale (community level) impacts of contamination and habitat modification on fish ecology in south east Australian estuaries.

xvi Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Major Findings

Chapter 1: McKinley, A.C. & Johnston, E.L. (2010) Impacts of contaminants on fish abundance and species richness: A review and meta-analysis of evidence from the field. Marine Ecology Progress Series, 420: 175-191.

x The meta-analysis showed that among 45 papers the average change in abundance at contaminated sites was +103% (fish farms), +40% (sewage studies), –52% (industrial effluent), and –65% (run-off). x Weak trends were observed for species richness, which may suggest that contamination is not having an impact on fish diversity. x Some sources of contamination that are also sources of food are having sizeable effects on abundance and richness and are likely to be of commercial and environmental significance. x Overall, marine contaminants appear to have either a weakly negative or strongly positive effect on the abundance and diversity of marine fish communities.

Chapter 2: McKinley, A.C. Dafforn K.A., Taylor, M.D. & E.L. Johnston (2011) High levels of sediment contamination have little influence on estuarine beach fish communities. PLoS One, 6(10).

x Shannon diversity, biomass and abundance were all significantly higher in the inner zone of estuaries while fish were larger on average in the outer zone. x Strong differences in community composition were also detected between the inner and outer zones. x Few differences were detected between fish assemblages in heavily modified versus relatively unmodified estuaries despite high concentrations of sediment contaminants in the inner zones of modified estuaries that exceeded recognized sediment quality guidelines.

xvii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

x Trends in species distributions, community composition, abundance, Shannon diversity, and average fish weight were strongly correlated to physico-chemical variables and showed a weaker relationship to sediment metal contamination. x Sediment PAH concentrations were not significantly related to the fish assemblage. x These findings suggest that variation in some physico-chemical factors (salinity, temperature, pH) or variables that co-vary with these factors (e.g., wave activity or grain size) have a much greater influence on this fish assemblage than anthropogenic stressors such as contamination.

Chapter 3: McKinley, A.C., Ryan, L., Coleman, M.A, Knott, N.A, Clarke, G., Taylor, M.D. & E.L. Johnston. (2011) Putting marine sanctuaries into context: A comparison of estuary fish assemblages over multiple levels of protection and disturbance. Aquatic Conservation, DOI: 10.1002/aqc.1223.

x It was demonstrated that sanctuary zones have higher abundance of targeted fish species compared with other areas within some marine parks. x The total abundance of targeted species and abundances of some key fisheries species (e.g. pink snapper) were found to be higher in sanctuary zones. This suggests that increased protection may be effective at improving these aspects of the fish assemblage. x However, when marine parks were compared with highly modified environments it was found that targeted species were much more abundant in the highly modified systems. x Community composition of entire fish assemblages also differed between these levels of modification and economically important fisheries species contributed most to this difference. x These findings suggest that while highly protected sanctuary zones may increase the abundance of targeted fish compared with less protected areas within the same estuary, highly industrialized or urbanized systems, not typically chosen as marine parks, may actually support more targeted species of fish. x It was demonstrated that forms of modification in addition to fishing pressure are having large effects on fish assemblages and productivity.

xviii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Chapter 4: McKinley, A.C., Miskiewicz, A., Taylor, M.D. & E.L. Johnston. (2011) Strong links between metal contamination, habitat modification and estuarine larval fish distributions. Environmental Pollution, 159: 1499-1509.

x Larval abundances were significantly greater in modified estuaries, and there were trends towards greater diversity in these systems. x Differences in larval community composition were strongly related to sediment metal levels and reduced seagrass cover. x The differences observed were driven by two abundant species, Paedogobius kimurai and Ambassis jacksoniensis, which occurred in large numbers almost exclusively in highly contaminated and pristine locations respectively. x These findings suggest that contamination and habitat alteration manifest in substantial differences in the composition of estuarine larval fish assemblages.

Chapter 5: McKinley, A.C., Foster-Thorpe, C., Miskiewicz, A., Taylor, M.D. & E.L. Johnston. (2012) Anthropogenic activities differentially impact fish guilds: The importance of understanding life history characteristics.

x The relative abundance of major larval guilds differed significantly between the surface and bottom samples, and between heavily modified vs. relatively unmodified estuaries. x Larvae of estuarine species and benthic spawners were significantly more abundant in the bottom waters of heavily modified estuaries. x The abundance of estuarine species and benthic spawners was strongly related to sediment metal contamination and seagrass cover. x In contrast, estuarine opportunist species trended towards higher abundance in the surface waters of relatively unmodified systems, while other guilds did not respond significantly to sampling depth or modification. x The abundance of the estuarine opportunist guild was strongly related to decreased temperatures, salinity, and sediment metals; and increased coverage of seagrass. This

xix Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

guild also showed a strong relationship to the width of the estuary mouth, in contrast to estuarine and benthic guilds which showed no relationship to the width of the estuary mouth. x Overall, greater impacts were observed in the epibenthic compared to surface samples, suggesting that epibenthic sampling alone is sufficient for future monitoring. x This study demonstrates that habitat modification and sediment contamination have substantially different relationships to larvae of different ecological guilds.

Chapter 6: McKinley, A.C., Taylor, M.D. & E.L. Johnston (2012) Relationships between body burdens of trace metals (As, Cu, Fe, Hg, Mn, Se, Zn) and the relative body size of small tooth flounder (Pseudorhombus jenynsii). Science of the Total Environment, 423: 84-94.

x Our results indicate that this commonly eaten fish is accumulating significant levels of some trace metals in their muscle tissue, both in relatively unmodified and heavily modified estuaries. x Concentrations of Cu, Zn and Fe in muscle tissue, as well as temperature, showed a negative relationship to the relative body size of flounder. x In contrast, Se and Hg in muscle showed a positive relationship to relative body size. x Observed growth patterns indicate that these effects are not driven by age related differences in metabolic activity. x Instead, our results suggest that differences in food supply or toxicological effects may be responsible for the observed relationships between relative body size and concentrations of Cu, Zn, and Se in muscle tissues.

Overall Trends: x Past studies on adult fish have generally shown weakly negative responses to contamination or largely positive responses where enriching contaminants are present. x Field surveys were broadly consistent with the literature and showed either weakly negative impacts or largely positive responses. Heavily modified estuaries showed increased abundance, biomass, and diversity except in the beach fish assemblage.

xx Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

x Potential impacts are highly variable among different guilds and functional groups, and both ecological characteristics and habitat preferences play an important mediating role. x The larval fish assemblage appears much more sensitive to anthropogenic modification than the adult fish assemblage. Estuarine larvae account for most of the difference between heavily modified and relatively unmodified estuaries. x Ambassis jacksoniensis may be a useful indicator species. x Strong evidence was presented suggesting physiological impacts from metals accumulation in muscle tissue of Pseudorhombus jenynsii.

xxi Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Community Level Impacts and Positive Effects In Chapter 1 I utilized a meta-analysis to summarize the trends in the literature surrounding contamination impacts on the abundance and species richness of fish. It was found that on average pollution studies describe either weakly negative responses from contamination or largely positive responses where enriching contaminants are present. In this thesis, strong differences were observed in fish communities living in relatively unmodified estuaries vs. heavily modified estuaries in three out of four chapters which measured potential impacts at the community level. In the BRUV study (Chapter 3) the community composition of bentho-pelagic and pelagic fish living in heavily modified estuaries differed significantly from those in marine parks. A strong positive effect of estuarine modification was also described, where recreationally targeted species were approximately four times as abundant in heavily modified estuaries compared to the marine parks. A trend towards increased abundance of recreationally targeted species was also documented in the otter trawl study, though this was not formally presented as the sampling was not temporally replicated (Chapter 6 – Community results not presented). In both larval studies strong differences were observed in community composition of benthic larvae, surface larvae, and all larvae (surface and benthic together) between heavily modified vs. relatively unmodified estuaries (Chapter 4 & 5). A variety of positive effects were described in these studies. Notably, larvae were found to be both more diverse and more abundant in the bottom waters of the heavily modified estuaries. In addition, large scale differences were observed in the distributions of individual species with some species being several orders of magnitude more abundant in the heavily modified estuaries and highly contaminated sites (e.g. Paedogobius kimuraii). As a group, estuarine taxa and benthic spawners displayed increased abundance in the bottom waters of heavily modified estuaries compared to other guilds (Chapter 5). However, some species (e.g. Ambassis jacksoniensis) were consistently less abundant in highly contaminated sites and strong negative impacts were identified for some taxa. In Chapter 2 little difference was observed among beach fish communities by modification, though significant differences were observed between the inner and outer zone of the estuary. However, trends towards increased biomass in Port Jackson and increased average fish size in Port Kembla suggested that some positive effects may be occurring in these heavily

xxii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

modified estuaries. In Chapter 2 some possible explanations of why beach fish may be relatively insensitive to disturbance are discussed. As a whole, the findings of this thesis are broadly consistent with the trends in the literature identified in Chapter 1. That is, pollution and disturbance effects were found to have only weakly negative effects in some cases (e.g. beach seines in Chapter 2) and largely positive effects in others (Chapters 3,4&5). In each chapter it was argued that nutrient enrichment is the likely cause of these positive effects, though other factors may also be important (e.g. increased habitat complexity, changes in trophic dynamics or abundance of predators, reduced competition, etc.). Nutrient enrichment is a possible explanation for the positive effects observed in the modified estuaries (Nixon & Buckley 2002, Breitburg et al. 2009). The modified estuaries are nutrient enriched compared to the relatively unmodified estuaries (Birch et al. 2010, Scanes 2010). It is likely that urbanization, land-use alteration and run-off are largely responsible for the elevated nutrient levels in these estuaries (Nixon 1995, Scanes 2010). Increased nutrient levels may be enhancing the productivity of the system and hence the abundance of fish. Several studies have demonstrated that nutrient enrichment (at pre-eutrophication levels) can enhance the abundance of fish and can substantially increase fisheries yields (Micheli 1999, Oczkowski & Nixon 2008). However, it is also possible that the modified estuaries are naturally more productive than the relatively unmodified systems. While historic data on productivity does not go back far enough to assess this quantitatively, it is likely that the placement of the major cities are not random and that they have been somewhat influenced by natural productivity. It is well documented that cities are preferentially built in areas which are naturally highly productive as the availability of natural resources (such as large fish populations) are a major incentive for early economic and urban growth (Folke et al. 1997, Haberl et al. 2004, Lotze et al. 2006). It is also possible that conservation measures such as marine parks and sanctuary zones are selectively established in areas that are not heavily used by local recreational fishing (e.g. poor fishing locations) as creating sanctuary zones in such places is more politically feasible (Agardy et al. 2003, Ray 2004, Edgar et al. 2008). Another possible explanation for these trends is increased habitat complexity in modified estuaries. Owing to the greater degree of development and boat traffic, a large amount of artificial habitat (maritime structures) exists in the modified estuaries (Connell & Glasby 1999).

xxiii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

It has been demonstrated that these structures can harbour diverse communities of invertebrates and plants (Connell & Glasby 1999, Glasby & Connell 1999, Glasby et al. 2007) and may aggregate or enhance fish abundances (Tuya et al. 2006). Several of the species that were more abundant in the modified estuaries (for example in Chapter 3) are known to feed on both sessile and mobile invertebrates (Coleman & Mobley 1984, Froese & Pauly 2010). It is therefore possible that artificial structures support a higher abundance of these invertebrate food items and that this in turn has led to an increased abundance of the recreational fish species. A fourth possible explanation for these findings could be differences in the abundance and activity of apex predators. In several cases increased abundance of predatory species in marine parks and other unmodified areas have been shown to have a ‘top down’ effect on aspects of the marine community (Shears & Babcock 2002, Micheli et al. 2004). These methods did not produce a sufficient sample size to understand the distributions of these large predators, but it is likely that they are more abundant in the relatively unmodified estuaries (particularly the marine parks) compared to the modified estuaries. For example, both Jervis Bay and Batemans Bay have well documented resident populations of grey nurse sharks (Carcharias taurus) and common bottlenose dolphins (Tursiops truncatus). Both of these species are significant predators of a variety of targeted fish species and both are believed to be largely absent from the heavily modified estuaries examined (e.g. Port Jackson, Botany Bay, Port Kembla, and Hunter River) (Gomon et al. 2008). However, Port Jackson also has a well documented breeding population of dusky whaler sharks (Carcharhinus obscurus) so it is difficult to speculate about differences in the overall abundance and activity of apex predators (McGrouther 2010). Differences in apex predator activity could not explain the observed differences between sanctuary and non- sanctuary zones described in Chapter 3, as apex predators such as dolphins and sharks are likely to be active in both zones within the same estuarine marine park (Shane et al. 1986, Last & Stevens 2009). Impacts of human modifications on the environment and associated ecological assemblages are complex, and so it is likely that a combination of these factors have influenced the results of this study. While a variety of potential positive impacts have been described in this thesis, these results should be interpreted cautiously. It has been argued in each chapter that ‘positive’ effects represent a form of anthropogenic impact and that these effects are not necessarily beneficial. In particular, these results should not be seen as being indicative that

xxiv Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

‘pollution does not harm fish’ as this is completely untrue. In truth, increased abundance and diversity is just as unnatural as decreases in these measures. I believe that the potential positive (e.g. increasing) impacts observed in this study should not be interpreted to signify that contamination and estuarine modification are not harmful to the fish assemblage. In fact, in every chapter some species were shown to be highly negatively correlated to these stressors. I believe that these effects should be interpreted to signify large scale ecological change in fish communities living in heavily modified estuaries. The consequences of these changes for rare species or on other marine organisms have not been described and may be of significant management concern. Overall, in combination the weight of evidence of these results suggests that anthropogenic impacts in these estuaries generally trend towards weakly negative effects or largely positive effects on the portions of the fish assemblage observed. This is of significant management interest as it suggests that despite anthropogenic disturbance, there is in fact significant ecological value in the modified estuaries and that they are highly diverse and productive systems, even compared to marine parks. Despite the abundance and diversity of the fish assemblages in places like Port Jackson, Botany Bay, and Port Kembla, there are no significant sanctuary zones in these estuaries, although many forms of commercial fishing have been limited. In fact, there are no major marine protected areas in any of the heavily modified estuaries in south-eastern Australia (MPA 2010) other than small Aquatic Reserves, many of which allow line fishing. A similar trend can be observed worldwide, as very few heavily modified systems have been protected by international marine parks systems (Kelleher et al. 1995, IUCN 2010).

Larval Susceptibility and Indicator Species Significant laboratory evidence points to larval and reproductive susceptibility to contaminants in fish populations (Jones & Reynolds 1997, Kingsford et al. 1997, Arkoosh et al. 1998, Robinet & Feunteun 2002) (see discussion Chapter 4). Prior to this thesis few of these effects had been verified in a field environment as only a handful of studies existed assessing the potential impacts of anthropogenic stressors on wild larval fish assemblages (Karas et al. 1991, Gray et al. 1992, Gray 1996, 1997, Kingsford et al. 1997).

xxv Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

In this thesis much larger and more consistent potential impacts were described for larval fish communities (Chapters 4&5) compared to adult fishes (Chapter 2&3). Notably, larvae differed significantly by modification regardless of the depth sampled, temporal variability, or differences in physico-chemical conditions between zones and estuaries. These findings suggest that larval fish communities are more sensitive and susceptible to anthropogenic disturbance than adult fish. Unfortunately, the degree to which potential impacts at the larval stage translate to the adult fish assemblage cannot be adequately answered by this thesis. This is because the majority of species which showed strong differences by modification in the beach seine (Chapter 2) and BRUV (Chapter 3) studies were not encountered in large numbers in the larval sampling (Chapter 4&5). In the BRUV study (Chapter 3) the major species driving differences between heavily modified estuaries and marine parks were Pagrus auratus (pink snapper), Acanthopagrus australis (yellowfin bream) and Pseudocaranx georgianus (silver trevally). No larvae were encountered for P. georgianus and P. auratus, and larvae of A. australis were captured only in very low numbers (<10). This is because these species are all estuarine opportunists and enter the estuarine system only as juveniles or adults (Elliott et al. 2007). Sillago ciliata (sand whiting), Sillago maculata (trumpeter whiting), Gerres subfasciatus (Roach), Hypherlophus vittatus (sandy sprat) and Favonigobius lentiginosus (long finned goby) were encountered in large numbers both in the beach seine study (Chapter 2) and as larvae (Chapter 4&5), but none of these differed by modification as both larvae and adults (though some differed by modification at one stage of the life cycle). The disjuncture between the results for the larval and beach fish assemblages may be due to a variety of factors. First, it should be noted that the larval fish assemblage sampled in Chapter 4&5 were more diverse than the beach fish assemblage in Chapter 2, and that many of the larval species are closely associated with biogenic habitats such as seagrass and mangroves (unlike the beach fish). This may explain the more pronounced impacts observed for larvae in Chapter 4&5. Second, because the majority of beach fish species are estuarine opportunists, many of them are not captured in large numbers by larval sampling in estuaries. The larval studies indicated that the majority of anthropogenic impacts are seen among estuarine resident taxa, a group which is underrepresented among the beach fish in Chapter 2.

xxvi Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Ambassis jacksoniensis (Port Jackson Glassfish) was also captured in large numbers as both an adult and larvae. This species was the only common species which differed significantly by modification in all studies. Specifically, A. jacksoniensis was found to be significantly more abundant in the relatively unmodified estuaries both as a larvae and an adult. A. jacksoniensis is a widespread species which is common throughout most estuaries in south-eastern Australia and easy to monitor using beach seines or other methods. For these reasons, A. jacksoniensis may represent a useful indicator species for continued ecological monitoring. While A. jacksoniensis differed between heavily modified and relatively unmodified estuaries both as a larvae and as an adult, the scale of difference was not equal; whereas adult A. jacksoniensis were approximately 1.5x more abundant in the relatively unmodified estuaries, larval A. jacksoniensis were approximately 13.7x more abundant. This indicates that there is a greater difference in the abundance of this species between heavily modified and relatively modified estuaries at the larval stage. This is consistent with the hypothesis that fish are more sensitive at their larval stage. However, it also suggests that the relationship between larval and adult abundance is not linear, and hence that potential impacts at the larval stage may not directly translate into potential impacts at the adult stage. The non-linear relationship between larval and adult abundance has been well documented in the supply side ecology literature (Woodin 1979, Roughgarden et al. 1988). The idea that larvae are much more sensitive to anthropogenic modification than adult fish should be of significant management interest. Our results indicate that future monitoring and research is more likely to detect impacts in the fish assemblage if they focus on sampling the larval community, and that a greater diversity of fish is encountered than with most adult fish sampling methods. Sampling for larvae, while skill intensive, is generally lower cost and less destructive than adult fish sampling. Ultimately, changes to larval fish assemblages may have far reaching ecological impacts both for the adult fish community and other organisms. Notably, the impacts of stressors at the larval stage of economically and culturally important fish species are poorly studied and little understood. The absence of studies examining anthropogenic impacts on estuarine fish larvae represents a major gap in the environmental impact literature and further investigation and monitoring is warranted.

xxvii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Ecological Characteristics and Complex Pathways In several chapters it was argued that the ecological characteristics of fish play an important role in determining the relative sensitivity of taxa to disturbance stressors. This concept is discussed more fully in Chapter 1. Among the adult fish it was argued that ecological characteristics could explain why large bentho-pelagic fish in the BRUV study appear very responsive to estuarine disturbance (Chapter 3), while small bodied beach fishes appear relatively insensitive to disturbance (Chapter 2). In the initial larval study some patterns suggested that potential impacts may be influenced by life cycle and spawning characteristics (Chapter 4). This was addressed explicitly in Chapter 5 through the novel application of a guild approach. In Chapter 5 it was clearly demonstrated that estuarine species/benthic spawners account for nearly all of the difference in the larval assemblage between heavily modified and relatively unmodified estuaries, while estuarine opportunists, marine stragglers, and pelagic spawners do not differ significantly. This suggests that the ecological characteristics of different fish guilds may play a large role in determining their relative sensitivity to stressors. This is of significant management and research interest, as it is clear that future sampling could focus on sampling sensitive functional groups (e.g. estuarine larvae), rather than insensitive functional groups (e.g. adult beach fish), in order to capture the majority of difference in the fish community between heavily modified and relatively unmodified estuaries. As such, the findings that some ecological groups and life stages are more sensitive to anthropogenic modification in south-eastern Australian estuaries may help to better target future research and monitoring. Just as the ecological characteristics of the fish may play a significant role, it was also argued that habitat and prey mediated impacts are likely to be important. In both larval studies it was shown that changes to larval assemblages were highly correlated to both sediment metals contamination and losses of vegetative habitats (particularly seagrass). It was argued that it is highly likely that both direct impacts (e.g. physiological impacts from metals) and indirect impacts (e.g. habitat mediated impacts from the loss of seagrass) are causing the observed differences in larval communities between heavily modified and relatively unmodified estuaries (Chapters 4&5). It should be noted that it is possible that trends in Chapters 4&5 and elsewhere in this thesis are influenced by differential tolerance between taxa or guilds. Several studies have examined the development of contaminant resistance and tolerance in marine and freshwater

xxviii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

fish, and some cases of resistance have been demonstrated in both laboratory and field conditions (Wirgin & Waldman 2004, Xie & Klerks 2004). Resistance may be due to either genetic adaptation or physiological acclimation in wild fish populations, and current research has investigated the mechanisms, costs, and persistence of toxicity resistance (Wirgin & Waldman 2004, Xie & Klerks 2004). If the development of contaminant resistance is a regular occurrence in fish populations chronically exposed to contaminants, then impacts on abundance or diversity may be difficult to observe. However, our current understanding of the mechanisms suggests that the evolution of resistance in fish will be rare (Wirgin & Waldman 2004, Xie & Klerks 2004).

Physico-Chemical Variables and Interrelationships Between Stressors My research has demonstrated clearly that the potential impacts of estuarine modification and contamination must be considered within the context of physico-chemical variability. This idea was addressed explicitly in Chapter 2, where it was argued that variability in physico- chemical conditions between the inner and outer estuary have a much greater potential impact on the beach fish assemblage than sediment contamination or large scale habitat disturbance. In Chapter 2 few differences were detected between fish assemblages in heavily modified versus relatively unmodified estuaries despite high concentrations of sediment contaminants (PAHs and metals) in the inner zones of modified estuaries that exceeded recognized sediment quality guidelines. Trends in species distributions, community composition, abundance, diversity, and fish size were strongly correlated to physico-chemical variables and showed a weaker relationship to sediment metal contamination. Sediment PAH concentrations were not significantly related to the fish assemblage. While this was the only chapter where physico- chemical variability was estimated to have a greater influence than contamination or modification, in every chapter where physico-chemical variables were included they showed some form of significant relationship to the biological data. This included significant relationships between salinity and the benthic larvae (Chapter 4), as well as salinity, temperature, estuary mouth size and various larval guilds (Chapter 5). It is clear that physico-chemical variables play a significant role in the distributions and ecology of fish and that these factors need to be considered in any impact assessment (Rose 2000, Whitfield & Elliott 2002). The differences in physico-chemical variables documented in this study are consistent with the general description and understanding of environmental conditions in south-east

xxix Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Australian estuaries. It is well documented that the interplay between fluvial and tidal forces in these systems creates consistent differences in physico-chemical conditions within most estuaries in the region (Roy & Williams 2001). The physico-chemical parameters that we measured do not encompass the full range of environmental conditions that are expected to differ between the inner and outer zones of these estuaries. Some additional variables of interest that may covary with our physico-chemical measures include: wave exposure, flow rates and grain size (expected to be higher in outer zones), phytoplankton productivity, predator/prey density, sedimentation rates, and coverage of submerged aquatic vegetation (expected to be higher in the inner zones) (Iverson 1990, Clark 1997). Experimental studies would be required to determine the extent to which any or all of these variables are the direct cause of the patterns we observed. The findings of this thesis also clearly indicated that stressors are frequently co-varying and that interrelationships exist between some stressors. For example, in Chapter 4 & 5 it was found that sites which had higher concentrations of trace metals in the sediment also tended to have reduced vegetative cover. It was argued that these stressors are interrelated because the core mechanisms of estuary contamination exposure (e.g. run-off, urbanization of shoreline/catchment, sedimentation, outflows, etc.) also tend to precipitate habitat alteration in estuarine systems (Drinkwater & Frank 1994, Rogers et al. 2002). Because both vegetative loss and sediment contamination are covarying, discerning the relative contribution of these forms of anthropogenic disturbance is difficult. Similarly, it was argued in Chapter 6 that the uptake mechanisms of some metals (e.g. Hg/Se or Cu/Fe/Zn) are strongly interrelated in marine fish (Ganther et al. 1972), which explains the strong covariation of some trace metals observed in this study (e.g. Cu/Fe/Zn). As a result, parsing out the relative contributions of each metal to the observed trends is challenging. These findings suggest that interrelationships exist between many stressors and that no single stressor or form of disturbance is likely to explain the patterns observed. Unfortunately, one limitation of correlative studies is that it is difficult to parse out the relative contributions of multiple covariates.

Mechanisms of Contaminant Impacts The majority of research addressing the effect of contamination on marine fish has focused on the chemical aspects of contamination, and a great deal of literature exists examining the presence, biomagnification, toxicology, and biomarker response of contaminants in marine

xxx Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

fish populations (Costello et al. 1994, Wirgin et al. 1998, Austin 1999, van der Oost et al. 2003, Hylland 2006). In this regard, the mechanisms by which contaminants affect fish populations have been fairly well investigated. Fish primarily uptake contaminants through ingestion of contaminated food particles and to a lesser extent from water that passes over the gill membranes (Dallinger et al. 1987, Hall et al. 1997). Some contaminants have also been shown to be maternally transferred to eggs and larvae (Collier et al. 1992, Hu et al. 2009). Once ingested, contaminants move through a wide variety of physiological and chemical pathways, many of which have detrimental effects for the individual. Some contaminants are readily excreted or breakdown while others are considered ‘persistent’ and resist decomposition in natural systems. Contaminants of this nature have the tendency to accumulate in tissues and may bioaccumulate up the food chain, increasing in concentration at higher trophic levels (Burger et al. 2001, van der Oost et al. 2003). Contaminants may affect fish populations and diversity by reducing fish health and survivorship (Robinet & Feunteun 2002, Claireaux et al. 2004), by increasing susceptibility to disease (Arkoosh et al. 1998), by reducing growth and reproductive success (Waring et al. 1996, Vetemaa et al. 1997), by reducing the abundance of prey species, and by increasing instances of deformity (Kingsford et al. 1997). Ultimately, any of these mechanisms could link contaminant exposure to organismal effects and population level impacts. Unfortunately, while these mechanisms are fairly well understood, very little research exists which links these mechanisms to impacts in wild fish or to impacts at the community level (discussed in detail in Chapter 1). In Chapter 6 it was argued that it is likely that accumulation of Cu, Zn, and Fe in muscle tissue of Pseudorhombus jenynsii had a sufficient physiological impact that it reduced the relative body size of the species. It should be noted that the approach used throughout this thesis, where contaminant levels in sediment are correlated to biological variables, is not able to evaluate the degree to which contaminants are biologically available to the fish community. Sediments were selected to measure contaminants in these systems (rather than a water column measure of contamination) for several reasons. First, it is well known that fish accumulate contaminants through their food to a much greater degree than through their gills or through interaction with contaminated water (Dallinger et al. 1987, Hall et al. 1997). The majority of species examined in this thesis are benthic or benthopelagic foragers and so most would interact with sediments regularly during feeding (Edgar & Shaw 1995). Second, contaminants accumulate in estuarine sediments over the

xxxi Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

long-term, as such, sediment contamination values are less temporally variable and represent a contemporary threat from historical pollution sources (Knott et al. 2009). While sediments are an appropriate way to measure contaminants in these systems, it remains true that the biologically available portion of the contamination, and actual exposure, cannot be determined with the methods used.

xxxii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Future Considerations

Some specific recommendations for future directions of research and management include the following: x Increased research focusing on the larval assemblage. The potential impacts described in Chapter 4&5 are previously undocumented. x Increased monitoring of the larval assemblage. The potential impacts described in Chapter 4&5 are entirely missed under current monitoring programs. x Investigation into the origin of P. kimuraii: Paedogobius kimuraii (Wide gape paedomorphic goby) was found to be super abundant (50-90% of the larval assemblage) in areas of high sediment metals contamination within the heavily modified estuaries. The unusual preference that this goby displayed for areas of very high anthropogenic disturbance and contamination is of significant interest. This species was only formally described in 2001 and there is some speculation that it may be an invasive, though this has not been formally confirmed (Iwata et al. 2001). The goby’s unusual affinity for highly disturbed sites (which may imply some kind of tolerance), its patchy distribution, very high fecundity and growth rates, and maturity at an unusually small size (~2 cm), match the profile of many invertebrate marine invaders (Alcaraz et al. 2005, Piola & Johnston 2008, Dafforn et al. 2009). x Full life cycle studies: Nearly no research exists which explicitly investigates potential anthropogenic impacts on a fish across all stages of the life cycle. This would give new insights into how larval effects translate to the adult stage. Research of this nature would be challenging, as many species do not complete their entire life cycle within the estuarine environment. x Field experimental approaches: While field experiments with fish are logistically very difficult, they could confirm causal relationship and the relative importance of different stressors. x Evaluate nutrient enrichment hypothesis: A sampling regime specifically designed to evaluate nutrient enrichment could provide a better estimate of the potential impact of this phenomenon in the estuaries studied.

xxxiii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

x Increased focus on ecological and field studies: The comprehensive review presented in Chapter 1 identified a bias in the literature towards laboratory and biomarker methods for the evaluation of potential contamination impacts on fish. These approaches far outnumber ecologically focused field studies, especially for difficult taxa such as larvae. While biomarker and laboratory methods are useful, their ability to predict ecological effects in wild fish assemblages is limited, particularly with regards to potential impacts at the population and community level (discussed in Chapter 1).

xxxiv Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Works Cited (Introduction and Discussion)

Agardy T, Bridgewater P, Crosby MP, Day J, Dayton PK, Kenchington R, Laffoley D, McConney P, Murray PA, Parks JE, Peau L (2003) Dangerous targets? Unresolved issues and ideological clashes around marine protected areas. Aquatic Conservation: Marine and Freshwater Ecosystems 13:353-367 Alcaraz C, Vila-Gispert A, García-Berthou E (2005) Profiling invasive fish species: the importance of phylogeny and human use. Diversity and Distributions 11:289-298 Alquezar R, Markich SJ, Booth DJ (2006) Metal accumulation in the smooth toadfish, Tetractenos glaber, in estuaries around Sydney, Australia. Environmental Pollution 142:123-131 Arkoosh, Mary R, Casillas, Edmundo, Clemons, Ethan, Kagley, Anna N, Olson, Robert, Reno, Paul, Stein, John E (1998) Effect of pollution on fish diseases: Potential impacts on salmonid populations. Journal of Aquatic Animal Health 10:182-190 Austin B (1999) The effects of pollution on fish health. Journal of Applied Microbiology 85:234S-242S Beck MW, Heck KL, Able KW, Childers DL, Eggleston DB, Gillanders BM, Halpern B, Hays CG, Hoshino K, Minello TJ, Orth RJ, Sheridan PF, Weinstein MP (2001) The identification, conservation, and management of estuarine and marine nurseries for fish and invertebrates. BioScience 51:633-641 Birch G, Cruickshank B, Davis B (2010) Modelling nutrient loads to Sydney estuary (Australia). Environmental Monitoring and Assessment 167:333-348 Birch G, Taylor S (1999) Source of heavy metals in sediments of the Port Jackson estuary, Australia. The Science of The Total Environment 227:123-138 Boesch D, Turner R (1984) Dependence of fishery species on salt marshes: The role of food and refuge. Estuaries and Coasts 7:460-468 Breitburg DL, Craig JK, Fulford RS, Rose KA, Boynton WR, Brady DC, Ciotti BJ, Diaz RJ, Friedland KD, Hagy JD, Hart DR, Hines AH, Houde ED, Kolesar SE, Nixon SW, Rice JA, Secor DH, Targett TE (2009) Nutrient enrichment and fisheries exploitation: interactive effects on estuarine living resources and their management. Hydrobiologia 629:31-47 Burger J, Gaines KF, Boring CS, Stephens WL, Snodgrass J, Gochfeld M (2001) Mercury and selenium in fish from the Savannah River: Species, trophic level, and locational differences. Environ Res 87:108-118 Claireaux G, Desaunay Y, Akcha F, Auperin B, Bocquene G, Budzinski H, Cravedi JP, Davoodi F, Galois R, Gilliers C, Goanvec C, Guerault D, Imbert N, Mazeas O, Nonnotte G, Nonnotte L, Prunet P, Sebert P, Vettier A (2004) Influence of oil exposure on the physiology and ecology of the common sole Solea solea: Experimental and field approaches. Aquat Living Resour/Ressour Vivantes Aquat 17:335-351 Clark BM (1997) Variation in surf-zone fish community structure across a wave-exposure gradient. Estuarine, Coastal and Shelf Science 44:659-674 Clements WH, Rohr JR (2009) Community responses to contaminants: Using basic ecological principles to predict ecotoxicological effects. Environmental Toxicology and Chemistry 28:1789-1800 Coleman N, Mobley M (1984) Diets of commercially exploited fish from Bass Strait and adjacent waters, Southeastern Australia. Marine and Freshwater Research 35:549-560

xxxv Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Collier TK, Stein JE, Sanborn HR, Hom T, Myers MS, Varanasi U (1992) Field studies of reproductive success and bioindicators of maternal contaminant exposure in English sole Parophrys-vetulus. Science of the Total Environment 116:169-185 Connell SD, Glasby TM (1999) Do urban structures influence local abundance and diversity of subtidal epibiota? A case study from Sydney Harbour, Australia. Marine Environmental Research 47:373-387 Costello, Mark J, Read, Paul (1994) Toxicity of sewage sludge to marine organisms: A review. Marine Environmental Research 37:23-46 Dafforn KA, Glasby TM, Johnston EL (2009) Links between estuarine condition and spatial distributions of marine invaders. Diversity and Distributions 15:807-821 Dallinger R, Prosi F, Segner H, Back H (1987) Contaminated food and uptake of heavy metals by fish: a review and a proposal for further research. Oecologia 73:91-98 Dorenbosch M, van Riel MC, Nagelkerken I, van der Velde G (2004) The relationship of reef fish densities to the proximity of mangrove and seagrass nurseries. Estuarine, Coastal and Shelf Science 60:37-48 DPI (2010) Survey of recreational fishing in New South Wales. Sydney: Department of Primary Industries Drinkwater KF, Frank KT (1994) Effects of river regulation and diversion on marine fish and invertebrates. Aquatic Conservation: Marine and Freshwater Ecosystems 4:135-151 Duke NC, Meynecke J-O, Dittmann S, Ellison AM, Anger K, Berger U, Cannicci S, Diele K, Ewel KC, Field CD, Koedam N, Lee SY, Marchand C, Nordhaus I, Dahdouh-Guebas F (2007) A world without mangroves? Science 317:41-42 Edgar GJ, Langhammer PF, Allen G, Brooks TM, Brodie J, Crosse W, De Silva N, Fishpool LDC, Foster MN, Knox DH, McCosker JE, McManus R, Millar AJK, Mugo R (2008) Key biodiversity areas as globally significant target sites for the conservation of marine biological diversity. Aquat Conserv: Mar Freshw Ecosyst 18:969-983 Edgar GJ, Shaw C (1995) The production and tropic ecology of shallow-water fish assemblages in southern Australia. III. General relationships between sediments, seagrasses, invertebrates and fishes. Journal of Experimental Marine Biology and Ecology 194:107- 131 Elliott M, Whitfield AK, Potter IC, Blaber SJM, Cyrus DP, Nordlie FG, Harrison TD (2007) The guild approach to categorizing estuarine fish assemblages: a global review. Fish and Fisheries 8:241-268 Folke C, Jsnsson A, Larsson J, Costanza R (1997) Ecosystem appropriation by cities. Ambio 26:167-172 Frank KT, Petrie B, Choi JS, Leggett WC (2005) Trophic cascades in a formerly cod-dominated ecosystem. Science 308:1621-1623 Froese R, Pauly D (2010) Fishbase. Available from www.fishbase.org (accessed January 25, 2010). Ganther HE, Goudie C, Sunde ML, Kopecky MJ, Wagner P, Oh S-H, Hoekstra WG (1972) Selenium: Relation to decreased toxicity of methylmercury added to diets containing tuna. Science 175:1122-1124 Glasby T, Connell S (1999) Urban structures as marine habitats. Ambio 28:595-598 Glasby T, Connell S, Holloway M, Hewitt C (2007) Nonindigenous biota on artificial structures: could habitat creation facilitate biological invasions? Mar Biol 151:887-895

xxxvi Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Gomon M, Bray D, Kuiter R (2008) Fishes of Australia's southern coast. Reed New Holland, Sydney Gray CA (1996) Intrusions of surface sewage plumes into continental shelf waters: interactions with larval and presettlement juvenile fishes. Mar Ecol Prog Ser 139:31-45 Gray CA (1997) Field assessment of numerical impacts of coastal sewage disposal on fish larvae relative to natural variability. Environ Biol Fishes 50:415-434 Gray CA, Otway NM, Laurenson FA, Miskiewicz AG, Pethebridge RL (1992) Distribution and abundance of marine fish larvae in relation to effluent plumes from sewage outfalls and depth of water. Mar Biol 113:549-559 Guo Y, Meng X-Z, Tang H-L, Zeng EY (2008) Tissue distribution of organochlorine pesticides in fish collected from the Pearl River Delta, China: Implications for fishery input source and bioaccumulation. Environmental Pollution 155:150-156 Haberl H, Wackernagel M, Krausmann F, Erb KHK-H, Monfreda C (2004) Ecological footprints and human appropriation of net primary production: a comparison. Land Use Policy 21:279-288 Hall BD, Bodaly RA, Fudge RJP, Rudd JWM, Rosenberg DM (1997) Food as the dominant pathway of methylmercury uptake by Fish. Water, Air, Soil Pollut 100:13-24 Henry GW, Lyle JM (2003) The national recreational and indigenous fishing survey. Commonwealth of Australia, Department of Agriculture, Fisheries, Forestry, Canberra Hu JY, Zhang ZB, Wei QW, Zhen HJ, Zhao YB, Peng H, Wan Y, Giesy JP, Li LX, Zhang B (2009) Malformations of the endangered Chinese sturgeon, Acipenser sinensis, and its causal agent. Proceedings of the National Academy of Sciences of the United States of America 106:9339-9344 Hylland K (2006) Biological effects in the management of chemicals in the marine environment. Marine Pollution Bulletin 53:614-619 Islam M, Tanaka M (2004) Impacts of pollution on coastal and marine ecosystems including coastal and marine fisheries and approach for management: a review and synthesis. Marine Pollution Bulletin 48:624-649 Isosaari P, Hallikainen A, Kiviranta H, Vuorinen PJ, Parmanne R, Koistinen J, Vartiainen T (2006) Polychlorinated dibenzo-p-dioxins, dibenzofurans, biphenyls, naphthalenes and polybrominated diphenyl ethers in the edible fish caught from the Baltic Sea and lakes in Finland. Environmental Pollution 141:213-225 IUCN (2010) Marine protected areas. International Union for the Conservation of Nature, Gland, Available from http://iucn.org/about/work/programmes/marine/marine_our_work/marine_mpas/ (Accessed October 31, 2010) Iverson RL (1990) Control of marine fish production. Limnology and Oceanography 35:1593- 1604 Iwata A, Hosoya S, Larson HK (2001) Paedogobius kimurai, a new genus and species of goby (Teleostei: Gobioidei: Gobiidae) from the West Pacific. Records of the Australian Museum 53:103-122 Jiao Y (2009) Regime shift in marine ecosystems and implications for fisheries management, a review. Rev Fish Biol Fish 19:177-191 Johnston E, Roberts DA (2009) Contaminants reduce the richness and evenness of marine communities: A review and meta-analysis. Environmental Pollution 157:1745-1752

xxxvii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Jones JC, Reynolds JD (1997) Effects of pollution on reproductive behaviour of fishes. Rev Fish Biol Fish 7:463-491 Karas P, Neuman E, Sandstrom O (1991) Effects of a pulp mill effluent on the population dynamics of perch perca-fluviatilis. Canadian Journal of Fisheries & Aquatic Sciences 48:28-34 Kelleher G, Bleakley C, Wells S (eds) (1995) A global representative system of marine protected areas: Antarctic, Arctic, Mediterranean, Northwest Atlantic, Northeast Atlantic and Baltic, Vol 1. Great Barrier Reef Marine Park Authority, World Conservation Union (IUCN), and World Bank, Washington, D.C. Kennish MJ (2002) Environmental threats and environmental future of estuaries. Environmental Conservation 29:78-107 Kingsford MJ, Suthers IM, Gray CA (1997) Exposure to sewage plumes and the incidence of deformities in larval fishes. Marine Pollution Bulletin 33:201-212 Knott NA, Aulbury J, Brown T, Johnston EL (2009) Contemporary ecological threats from historical pollution sources: impacts of large-scale resuspension of contaminated sediments on sessile invertebrate recruitment. Journal of Applied Ecology 46:770-781 Kojadinovic J, Potier M, Le Corre M, Cosson RP, Bustamante P (2007) Bioaccumulation of trace elements in pelagic fish from the Western Indian Ocean. Environmental Pollution 146:548-566 Last PR, Stevens PD (2009) Sharks and rays of Australia. CSIRO Publishing, Collingwood, Victoria Lotze HK, Lenihan HS, Bourque BJ, Bradbury RH, Cooke RG, Kay MC, Kidwell SM, Kirby MX, Peterson CH, Jackson JBC (2006) Depletion, degradation, and recovery potential of estuaries and coastal seas. Science 312:1806-1809 McGrouther M (2010) Dusky whaler juveniles in Sydney. Australia Museum, Melbourne. Available from http://australianmuseum.net.au/Dusky-Shark-juveniles-in-Sydney/ (Accessed September 10, 2010) Micheli F (1999) Eutrophication, fisheries, and consumer-resource dynamics in marine pelagic ecosystems. Science 21:1396-1398 Micheli F, Halpern BS, Botsford LW, Warner RR (2004) Trajectories and correlates of community change in no-take marine reserves. Ecological Applications 14:1709-1723 Miskiewicz AG, Gibbs PJ (1994) Organochlorine pesticides and hexachlorobenzene in tissues of fish and invertebrates caught near a sewage outfall. Environmental Pollution 84:269-277 MPA (2010) Marine Parks Authority of New South Wales, Sydney, Available from http://www.mpa.nsw.gov.au/ (Accessed September 15, 2010) Murphy CA, Rose KA, Alvarez M, Fuiman LA (2008) Modeling larval fish behavior: Scaling the sublethal effects of methylmercury to population-relevant endpoints. Aquatic Toxicology 86:470-484 Nixon S, Buckley B (2002) “A strikingly rich zone”—Nutrient enrichment and secondary production in coastal marine ecosystems. Estuaries and Coasts 25:782-796 Nixon SW (1995) Coastal marine eutrophication: A definition, social causes, and future concerns. Ophelia 41:199-219 NSW (1999) Marine parks (zoning plans) regulation of 1999 division marine park zones. New South Wales Government, Sydney, Available from www.austlii.edu.au/au/legis/nsw/consol_reg/mppr1999362/ (Accessed November 12, 2010)

xxxviii Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

NSWDNR (2010) Estuaries in New South Wales. Sydney: New South Wales Department of Natural Resources. Available from http://www.naturalresources.nsw.gov.au/estuaries/inventory/index_ns.shtml Accessed November 14, 2010. Oczkowski A, Nixon S (2008) Increasing nutrient concentrations and the rise and fall of a coastal fishery; a review of data from the Nile Delta, Egypt. Estuarine, Coastal and Shelf Science 77:309-319 Pauly D, Christensen V, Guenette S, Pitcher TJ, Sumaila UR, Walters CJ, Watson R, Zeller D (2002) Towards sustainability in world fisheries. Nature 418:689(687) Perry AL, Low PJ, Ellis JR, Reynolds JD (2005) Climate change and distribution shifts in marine fishes. Science 308:1912-1915 Piola RF, Johnston EL (2008) Pollution reduces native diversity and increases invader dominance in marine hard-substrate communities. Diversity and Distributions 14:329- 342 Potter IC, Hyndes GA (1999) Characteristics of the ichthyofaunas of southwestern Australian estuaries, including comparisons with holarctic estuaries and estuaries elsewhere in temperate Australia: A review. Aust J Ecol 24:395-421 Rakocinski CF, Baltz DM, Fleeger JW (1992) Correspondence between environmental gradients and the community structure of marsh-edge fishes in a Louisiana estuary. Mar Ecol Prog Ser 80:135-148 Ray GC (2004) Reconsidering ‘dangerous targets’ for marine protected areas. Aquat Conserv: Mar Freshw Ecosyst 14:211-215 Robertson A, Duke N (1987) Mangroves as nursery sites: comparisons of the abundance and species composition of fish and crustaceans in mangroves and other nearshore habitats in tropical Australia. Mar Biol 96:193-205 Robinet TT, Feunteun EE (2002) Sublethal effects of exposure to chemical compounds: A cause for the decline in Atlantic eels? Ecotoxicology 11:265-277 Rogers CE, Brabander DJ, Barbour MT, Hemond HF (2002) Use of physical, chemical, and biological indices to assess impacts of contaminants and physical habitat alteration in urban streams. Environmental Toxicology and Chemistry 21:1156-1167 Rose KA (2000) Why are quantitative relationships between environmental quality and fish populations so elusive? Ecol Appl 10:367-385 Roughgarden J, Gaines S, Possingham H (1988) Recruitment dynamics in complex life cycles. Science 241:1460-1466 Roy P, Williams R (2001) Structure and function of south-east Australian estuaries. Estuarine, Coastal and Shelf Science 53:351-384 Saintilan N, Williams RJ (1999) Mangrove transgression into saltmarsh environments in south- east Australia. Global Ecology and Biogeography 8:117-124 Scanes P (2010) NSW estuarine catchment disturbance ranks. NSW Department of Environment, Climate Change, and Water; Sydney Shane SH, Wells RS, Würsig B (1986) Ecology, behavior and social organization of the bottlenose dolphin: A review. Marine Mammal Science 2:34-63 Shears N, Babcock R (2002) Marine reserves demonstrate top-down control of community structure on temperate reefs. Oecologia 132:131-142

xxxix Assessing Ecological Changes to Fish Communities in Highly Disturbed Estuaries. McKinley 2012.

Taylor MD, Laffan SD, Fielder DS, Suthers IM (2006) Key habitat and home range of mulloway Argyrosomus japonicus in a south-east Australian estuary: finding the estuarine niche to optimise stocking. Marine Ecology Progress Series 328:237-247 Taylor MD, Palmer PJ, Fielder DS, Suthers IM (2005) Responsible estuarine finfish stock enhancement: An Australian perspective. Journal of Fish Biology 67:299-331 Tilman D, May RM, Lehman CL, Nowak MA (1994) Habitat destruction and the extinction debt. Nature 371:65-66 Tuya F, Sanchez-Jerez P, Dempster T, Boyra A, 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 Valiela I, Bowen JL, York JK (2001) Mangrove forests: One of the world's threatened major tropical environments. BioScience 51:807-815 van der Oost R, Beyer J, Vermeulen NPE (2003) Fish bioaccumulation and biomarkers in environmental risk assessment: a review. Environmental Toxicology and Pharmacology 13:57-149 Vetemaa, Markus, Forlin, Lars, Sandstrom, Olof (1997) Chemical industry effluent impacts on reproduction and biochemistry in a North Sea population of viviparous blenny (Zoarces viviparus). Journal of Aquatic Ecosystem Stress & Recovery 6:33-41 Walker DI, McComb AJ (1992) Seagrass degradation in Australian coastal waters. Marine Pollution Bulletin 25:191-195 Waring CP, Stagg RM, Fretwell K, McLay HA, Costello MJ (1996) The impact of sewage sludge exposure on the reproduction of the sand goby, Pomatoschistus minutus. Environmental Pollution 93:17-25 Waycott M, Duarte CM, Carruthers TJB, Orth RJ, Dennison WC, Olyarnik S, Calladine A, Fourqurean JW, Heck KL, Hughes AR, Kendrick GA, Kenworthy WJ, Short FT, Williams SL (2009) Accelerating loss of seagrasses across the globe threatens coastal ecosystems. Proceedings of the National Academy of Sciences 106:12377-12381 Whitfield AK, Elliott M (2002) Fishes as indicators of environmental and ecological changes within estuaries: a review of progress and some suggestions for the future. J Fish Biol 61:229-250 Wirgin, Isaac, Waldman, John R (1998) Altered gene expression and genetic damage in North American fish populations. Mutat Res 399:193-219 Wirgin I, Waldman JR (2004) Resistance to contaminants in North American fish populations. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis 552:73-100 Woodin SA (1979) Adult-larval interactions in dense infaunal assemblages: patterns of abundance. Journal of Marine Research 34:26-41 Xie L, Klerks PL (2004) Changes in cadmium accumulation as a mechanism for cadmium resistance in the least killifish Heterandria formosa. Aquatic Toxicology 66:73-81