Sydney's Designed Artificial Reef: the Recreational Fishery

Sydney's designed artificial reef: The recreational fishery

and movements of fish

Krystle Suzanne Keller

Evolution and Ecology Research Centre

School of Biological, Earth and Environmental Sciences

University of New South Wales

A thesis in fulfilment of the requirements for the degree of

Doctor of Philosophy

December 2016

PLEASE TYPE THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet Surname or Family name: Keller

First name: Krystle Other name/s: Suzanne Abbreviation for degree as given in University calendar: PhD

School: School of Biological, Earth and Environmental Sciences Faculty: Science Title: Sydney's designed artificial reef: The recreational fishery and movements of fish.

Abstract 350 words maximum: (PLEASE TYPE) Artificial reefs are believed to support increases in species diversity and abundance by providing food and refuge, as well as increased fishing opportunities for anglers. My thesis explores components of the ecosystem of a coastal designed artificial reef (AR) by estimating the recreational harvest and effort by anglers, and examining the activity, movements and residency of three co-occurring benthic species associated with this structure. Fishing effort was determined using a shore-based camera and was found to be low in the first two seasons since deployment. Recreational harvest was estimated by combining the effort data from June 2013- May 2014 with historical catch information and was calculated to be 700 ± 59 kg of fish in total, and 12,504 kg per km2 when standardised per unit area. Standardised effort and harvest (divided by area) were found to be higher at the AR compared to estuarine fisheries. Acoustic telemetry was used to study the activity patterns, influence of environmental parameters, movements and residency in the fiddler ray (Trygonorrhina fasciata), bluespotted flathead (Platycephalus caeruleopunctatus) and Port Jackson shark (Heterodontus portusjacksoni). Activity in fiddler rays was highest during the day, in contrast to Port Jackson sharks and bluespotted flatheads which were most active at night. Activity also increased with temperature in fiddler rays and bluespotted flatheads. The increase in activity is likely to be associated with foraging behaviour and differences in behaviour between these species may be related to resource partitioning. All three species exhibited low residency at the AR, however some individuals exhibited relatively high residency times at this site over the long-term monitoring period. Residency was highest at the AR compared to nearby natural reefs, but this was influenced by the original site of tagging. Frequent movements between the AR and nearby natural reefs by these species indicate strong connectivity between adjacent reefs. These results suggest that the AR has the potential to enhance recreational fisheries and may have altered the local distribution of these three species. The relatively low residency however indicates that only a fraction of their biomass production is likely to be derived from this AR.

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).

…………………………………………………….. …………………………………………………….. ……………………………………………………. 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:

iii

Cover image from Andrew Boomer

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 ……………………………………………......

‘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 ……………………………………………......

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 ……………………………………………...... Abstract

Artificial reefs are believed to support increases in species diversity and abundance by providing food and refuge, as well as increased fishing opportunities for anglers.

However it is currently unknown whether these reefs increase overall biomass production, or whether they merely attract and aggregate existing fish to a new location.

My thesis explores components of the ecosystem of a 700 m3 steel designed artificial reef (AR) located off coastal Sydney, Australia in 38 m depth of water by estimating the recreational fishing effort (Chapter 2) and harvest (Chapter 3) by anglers from this reef.

I also examine the activity patterns and influence of environmental parameters (Chapter

4), as well as movements and residency (Chapter 5) of three co-occurring benthic species at the AR and surrounding natural reefs, to determine the contribution of this reef to coastal ecosystems and recreational fisheries.

Recreational fishing effort was quantified at the Sydney AR using a shore-based camera over a two-year period. The effort estimates derived from the digital images were adjusted to account for visibility bias using information from a validation study. The levels of effort recorded in the first two seasons were low as the AR had been recently deployed and colonisation of the AR by sessile organisms and fishes was still occurring.

Effort intensity (calculated in units of fisher hours per square kilometre) at the Sydney

AR was compared with three South Australian ARs and 14 estuarine fisheries in New

South Wales (NSW) to provide context for the study. Effort intensity at the AR was found to be up to 12 times higher than that recorded from some estuarine fisheries in

NSW. Conversely, the levels of effort intensity at two South Australian ARs were much higher compared to those at the Sydney AR site in both survey years. Effort intensity

v comparisons showed that the relative levels of usage at Australian ARs were higher than those recorded from estuarine fisheries. These results indicate that the Sydney AR provides diverse fishing opportunities that may be concentrated in a small area. Camera- based technologies can provide a solution for cost-effective monitoring of AR sites, providing the accuracy of fishing effort information derived from camera images is validated. This study has broad implications for other recreational ARs, including many future deployments planned for eastern Australia.

The ability to assess recreational harvest is important for determining the effectiveness of AR deployments. Harvest estimation at AR fisheries poses many logistical and budgetary challenges. A pragmatic approach is used to estimate harvest at the Sydney

AR that combines existing datasets and a cost-effective sampling design. Total annual recreational harvest from the AR during June 2013- May 2014 was estimated to be

1,016 ± 82 fish by number, 700 ± 59 kg of fish by weight, and 12,504 kg per km2 when standardised per unit area. Harvest at the AR by number and by weight was relatively small, however when considered per unit area, this standardised harvest was very high

(2.3 - 43.6 times larger) compared to other fishery areas from which the fishable area is known. The harvest at the AR was dominated by 6 functional groups (ambush predators, leatherjackets, large to medium pelagic fish, small pelagic fish, medium demersal predators and large demersal predators), which accounted for 92% of the total annual harvest by number, and 95% of the total annual harvest by weight. Comparisons of standardised harvest between the AR and other fishery areas revealed two distinct groups, a) the AR and Swansea channel, a marine-dominated entrance to a large estuary, and b) all other fishery areas, which were grouped together. Future studies attempting to

vi estimate harvest at AR fisheries should consider an integrated methodology that combines existing datasets and cost-effective sampling designs.

Acoustic telemetry was used to study the activity patterns and influence of environmental parameters in three common benthic species around rocky reefs: the eastern fiddler ray (Trygonorrhina fasciata), eastern bluespotted flathead

(Platycephalus caeruleopunctatus), and Port Jackson shark (Heterodontus portusjacksoni). Fiddler rays were on average 28% more active during the day, in contrast to bluespotted flatheads and Port Jackson sharks which were on average 28% and 84% more active at night, respectively. Activity was not associated with fish length, tidal height or moon illumination, but increased with temperature in fiddler rays and bluespotted flatheads. The increase in activity is most likely to be associated with foraging behaviour, and a similarity in diet suggests that observed differences in movement behaviour may be due to resource partitioning between these species.

Different diel activity patterns have implications for conservation and fisheries management, whereby nocturnal species such as the bluespotted flathead and Port

Jackson shark are missed during visual surveys. The findings from this study improve the understanding of the ecology of co-existing benthic predators around coastal temperate reefs. Assessing the movements and residency of these species at both natural and ARs is needed to determine the role they play in structuring ecological communities, and for the management of marine coastal resources

Residency and connectivity of different species of fish associated with ARs and natural reefs is a key question to determine if ARs contribute to coastal ecosystems and fisheries. The movements and residency of 10 eastern fiddler rays (Trygonorrhina

vii fasciata), 17 Port Jackson sharks (Heterodontus portusjacksoni) and 23 bluespotted flatheads (Platycephalus caeruleopunctatus) were monitored using acoustic telemetry around a 12 x15 m designed AR in 38 m depth near Sydney, Australia. Residency at the

AR was low for all three species over the monitoring period, although some individual fiddler rays and bluespotted flathead exhibited relatively high residency times over the ~

20 month monitoring period, and were resident at least 50% of the time within the AR area. Fish tagged at the AR showed a higher degree of residency relative to those tagged near the natural reef. All three species moved frequently between the AR and 5-6 other reefs, indicating strong connectivity throughout the habitat mosaic. The relatively low residency shows that only 4- 32% of these species’ biomass production and contribution to total recreational harvest is likely to be derived from this AR.

viii

Acknowledgments

I would like to firstly thank all my supervisors, Iain Suthers, Mick Lowry and James

Smith, for all their advice, guidance, patience, support and assistance. Thank you for giving me the opportunity to do this research, encouraging me to pursue my own interests and believing in my abilities during the times when I lacked confidence in myself. I would like to extend a special thank you to Iain, for the constant motivation, support and friendship. Iain’s endless enthusiasm for marine science was a constant reminder for why I pursued this degree in the first place.

I am very grateful to Aldo Steffe, who so often gave up his time to help me with data analysis, interpretation, writing up and providing much needed advice for solving issues. You came to the rescue on so many occasions and for that I am so grateful.

Thank you for the constant support, patience and introducing me to the world of recreational fisheries. I am also grateful for the financial support of a number of funding bodies which made this research possible, including the NSW Recreational Fishing trust, NSW DPI fisheries and ARC linkage grant.

This research wouldn’t have been possible without the generous support of many volunteers, colleagues and friends. A big thank you to the Port Jackson marine rescue volunteers for assisting me with data collection, as well as the many volunteers and anglers who helped me directly with my fieldwork (and teaching me how to fish!).

Thanks to the FAMER lab for the friendship, support and encouragement over the years, especially Steph, Derrick and Ruan who assisted me with fieldwork and acoustic tagging procedures. A special thanks to Teagan for all the laughs, constant support, encouragement, travel adventures, and for generously assisting me in the field and

ix providing advice. Our shared passion for elasmobranchs and constant banter has inspired many of the ideas for this project. Thank you to the AATAMS guys: Andre,

James, Phil and Andrew Boomer, for all the assistance in the field and the receiver downloads. Thanks also to Matt Taylor and other colleagues from NSW DPI fisheries for the advice and assistance with data analysis.

To my family and friends, a huge thank you for your love and constant encouragement.

Thank you for supporting me on this journey and realising my dreams of studying marine science, which started from an early age.

Finally to my best friend and partner, Clint. Thank you for always being patient, understanding and supporting me through the ups and downs of my PhD adventure.

You have been my rock and a mountain of strength (not just literally!) and for that I am forever grateful.

Table of Contents

Chapter 1- Introduction ...... 1 1.1 Background ...... 1 1.2 The ‘attraction versus production’ debate ...... 3 1.3 Assessing attraction and production ...... 6 1.4 Study region ...... 7 1.5 Study Species ...... 10 1.6 Research objectives ...... 13 1.7 Thesis chapters ...... 15 Chapter 2- Monitoring boat-based recreational fishing effort at a nearshore artificial reef with a shore-based camera ...... 19 2.1 Abstract ...... 19 2.2 Introduction ...... 20 2.3 Materials and methods ...... 23 2.3.1 Artificial Reef description ...... 23 2.3.2 Camera imagery ...... 27 2.3.3 Validation of data derived from digital images ...... 28 2.3.4 Effort estimation from digital images ...... 29 2.3.5 Standardised comparisons of effort intensity ...... 32 2.3.6 Comparative coastal fishing effort data from the greater Sydney region ...... 36 2.4 Results ...... 38 2.5 Discussion ...... 45 2.6 Conclusions ...... 50 Chapter 3- Estimating the recreational harvest of fish from a nearshore artificial reef using a pragmatic approach ...... 52 3.1 Abstract ...... 52 3.2 Introduction ...... 53 3.3 Materials and methods ...... 59 3.3.1 The Artificial Reef ...... 59 3.3.2 Existing datasets used for inferring harvest composition and harvest rates at the AR ...... 60 3.3.3 Harvest rates, fishing effort and harvest estimation from the AR ...... 63 3.3.4 Standardised harvest comparisons ...... 64 3.3.5 Fish functional groups and data analysis ...... 67 3.4 Results ...... 69 3.5 Discussion ...... 75 3.5.1 A pragmatic approach for harvest estimation ...... 75 xi

3.5.2 Fish harvest at the AR ...... 78 3.6 Conclusions ...... 80 Chapter 4- A ray, a fish and a shark: the diel behaviour and activity of three co-occurring benthic species around temperate rocky reefs ...... 82 4.1 Abstract ...... 82 4.2 Introduction ...... 83 4.3 Material and methods ...... 87 4.3.1 Study area ...... 87 4.3.2 Acoustic tagging ...... 90 4.3.3 Environmental data ...... 98 4.3.4 Data analysis ...... 99 4.4 Results ...... 103 4.4.1 Diel activity ...... 103 4.4.2 Environmental variables ...... 107 4.5 Discussion ...... 113 4.5.1 Diel activity patterns ...... 113 4.5.2 Effect of temperature on activity ...... 115 4.5.3 Behavioural comparisons and implications ...... 117 4.6 Conclusions ...... 120 Chapter 5- Multispecies residency and connectivity around a designed artificial reef...... 123 5.1 Abstract ...... 123 5.2 Introduction ...... 124 5.2.1 Movement ecology and artificial reefs ...... 125 5.2.2 Objectives and rationale of study ...... 127 5.3 Materials and Methods ...... 129 5.3.1 Study area ...... 129 5.3.2 Acoustic tagging ...... 132 5.3.3 Site residency ...... 143 5.3.4 Site connectivity ...... 147 5.4 Results ...... 148 5.4.1 Residency ...... 148 5.4.2 Connectivity ...... 160 5.5 Discussion ...... 163 5.5.1 Residency ...... 163 5.5.2 Connectivity with natural reef ...... 167 5.5.3 Influence of the AR ...... 169 5.6 Conclusions ...... 171 Chapter 6- General discussion ...... 173 6.1 Summary ...... 173 6.2 Future research ...... 177

xii

References ...... 182 Appendix A. Effort Estimation Equations...... 197 Appendix B. Fish functional groups and species presence by site...... 202 Appendix C. Harvest Estimation Equations ...... 213 Appendix D. Estimated annual recreational harvest of the ten most commonly captured species from the AR ...... 217

xiii

List of Figures

Figure 1.1. Location of the designed artificial reef (AR) and Dunbar reef...... 9

Figure 1.2. Deployment of the AR on the 12th October 2011. Image: Justin Gilligan .... 9

Figure 1.3. The eastern fiddler ray, Trygonorrhina fasciata...... 11

Figure 1.4. The eastern bluespotted flathead, Platycephalus caeruleopunctatus...... 12

Figure 1.5. Port Jackson shark, Heterodontus portusjacksoni...... 12

Figure 1.6. Schematic diagram of subprojects for input into the AR ecosystem model. 14

Figure 2.1. Location of the Sydney Artificial Reef (AR) and the Old South Head Signal Station (indicated as camera)...... 24

Figure 2.2. Schematic of the 42-ton Sydney Artificial Reef (AR)...... 25

Figure 2.3. Camera view of the 0.056 km² monitored AR area (140 m (north to south) x 400 m (east to west)). ‘X’ indicates location of the AR. (Note the two different scale bars, the vertical scale bar (not to scale) indicates the increasing distance from the base of the monitored area)...... 26

Figure 2.4. Regression equation describing the relationship between the number of fishing events counted from camera digital images and from field validated observations at the AR. Number of overlying observations is provided above each data point. The dotted line denotes the y=x equation...... 39

Figure 2.5. Seasonal fishing effort (fisher hours ± SE) at the AR for each survey year (year 1=June 2012-May 2013, year 2=June 2013-May 2014)...... 40

Figure 2.6. Comparison of annual fishing effort by area (fisher hours/km2) between study locations. Sydney AR (this study); South Australia ARs (Gr= Grange, Gl= Glenelg, PN =Port Noarlunga), N. Lake Macquarie=Northern Lake Macquarie, S. Lake Macquarie=Southern Lake Macquarie, Swansea channel, PH estuary= Port Hacking estuary. For details of survey periods and sources refer to Table 2.2...... 44

Figure 2.7. Total seasonal fishing effort (angling trips ±SE) for line fishing in four greater Sydney coastal systems from March 2007 to February 2009 (Hawkesbury, Port Hacking, Botany Bay, Sydney Harbour)...... 47

Figure 3.1. Location of the Sydney Artificial Reef (AR), the Old South Head Signal Station (indicated as camera) and greater Sydney sites (Hawkesbury, Long Reef and Port Hacking)...... 61

Figure 3.2. Total annual harvest of fish functional groups by number (± SE) and weight (kg ± SE) at the AR from June 2013- May 2014 ( L*=Leatherjackets, M&G* =mullets and garfish; 41 taxa in total)...... 70

xiv

Figure 3.3. MDS (nMDS) of standardised annual fish harvest (kg) by site. Numbers indicate survey year, overlaid circles represent a similarity of 40% between groups. Refer to Table 3.3 for details of fishery areas...... 72

Figure 4.1. Location of receivers in greater Sydney. AR= Artificial reef, NR= Dunbar reef receivers (N: north, S: south) and ORS= Ocean Reference station...... 89

Figure 4.2. Average temporal activity (ms-2± SE) of fiddler ray (Trygonorrhina fasciata, n=7), bluespotted flathead (Platycephalus caeruleopunctatus, n=14) and Port Jackson shark (Heterodontus portusjacksoni, n=7) from June 2013 to June 2015...... 104

Figure 4.3. Contribution of the spline for hour to the fitted GAMM models of Activity (see Table 4.3), for a) fiddler rays b) bluespotted flatheads and c) Port Jackson sharks. Dotted lines represent 95% confidence intervals...... 105

Figure 4.4. Average hourly activity (ms-2) between a) the bluespotted flathead and fiddler ray, b) the Port Jackson shark and fiddler ray and c) the bluespotted flathead and Port Jackson shark during the monitoring period...... 106

Figure 4.5. Average activity (ms-2± SE) and temperature (°C) in the a) fiddler ray (Trygonorrhina fasciata), b) bluespotted flathead (Platycephalus caeruleopunctatus) and c) Port Jackson shark (Heterodontus portusjacksoni) from June 2013 to June 2015...... 109

Figure 4.6. Quantile boxplots of the detection frequency (detections per hour) for the control tag placed 200 m from the receiver. The detection frequency was significantly lower from 7-9 am (shaded grey; see Table 4.5)...... 110

Figure 5.1. Study area showing receiver locations. AR= Artificial Reef, NR= Dunbar reef, AM= Annie Miller reef, SG= Sydney harbour entrance, SH= Sydney Harbour, BSH= Between Bondi and South Head, BL= Bondi line, BCAR= Bronte-Coogee Aquatic reserve, SLB= South Long Bay, NAR= Narooma...... 130

Figure 5.2. Location of artificial reef (AR), nearby natural reef, and the two Dunbar receivers (north and south). Numbers are depths in metres. Bathymetry information is from acoustic surveys by the NSW Office of Environment and Heritage...... 131

Figure 5.3. Cumulative percentage (%) of detections of fiddler rays, bluespotted flatheads and Port Jackson sharks in study area during the monitoring period...... 145

Figure 5.4. Presence plot of fiddler rays monitored in the Sydney region from June 2013 to June 2015. Black dots indicate detection at the artificial reef (AR), colour symbols indicate detection at other receivers (SG= Sydney gate, NR=Dunbar, BCAR= Bronte- Coogee marine reserve, BL=Bondi line, BSH= between Bondi and South Head). Grey crosses indicates end of tag life...... 149

Figure 5.5. Presence plot of bluespotted flatheads monitored in the Sydney region from August 2013 to April 2015. Black dots indicate detection at the artificial reef (AR), colour symbols indicate detection at other receivers (SG= Sydney gate, NR=Dunbar, BL=Bondi line, BSH= between Bondi and South Head, SH= Sydney Harbour, SLB= South long Bay). Grey crosses indicates end of tag life...... 153 xv

Figure 5.6. Presence plot of Port Jackson sharks monitored in the Sydney region from August 2013 to December 2014. Black dots indicate detection at the artificial reef (AR), colour symbols indicate detection at other receivers (SG= Sydney gate, NR=Dunbar, BL=Bondi line, BSH= between Bondi and South Head, NAR= Narooma)...... 155

Figure 5.7. Proportion of total days that fiddler rays tagged from either the AR (artificial reef) or NR (Dunbar reef) were detected at receivers in the study area during the monitoring period. DBHN= north Dunbar reef, SG1-3= Sydney gate, SYT409=between Bondi and South Head, BL1-3= Bondi...... 157

Figure 5.8. Proportion of total days that bluespotted flatheads tagged from the either AR(artificial reef) or NR (Dunbar reef) were detected at receivers in the study area during the monitoring period. DBHN=north Dunbar reef, SG1-4= Sydney gate, SYT233= Sydney harbour, SYT409=between Bondi and South Head...... 158

Figure 5.9. Proportion of total days that Port Jackson sharks tagged from either the AR(artificial reef) or NR (Dunbar reef) were detected at receivers in the study area in a) greater Sydney area and b) south Sydney during the monitoring period. DBHN=north Dunbar reef, DBHS=south Dunbar reef, SG1-4= Sydney gate, N2-6= Narooma...... 159

Figure 5.10. Proportion of individuals per species detected and the corresponding distances from their tagging reef during the monitoring period...... 161

Figure 5.11. The average distance (km ± SE) detected from tagging reef by fiddler rays (n=10), bluespotted flatheads (n=23) and Port Jackson sharks (n=17) during the monitoring period. Grey dot indicates the median of distance travelled by all individuals...... 162

xvi

List of Tables

Table 2.1. The number of days sampled (n) and stratum sizes (N) within day-type, weekdays (WD) and weekend days (WE), during the survey period...... 30

Table 2.2. Study site survey periods, habitat types, distance from shore, depth, location (latitude and longitude) and measured areas (km2)...... 33

Table 2.3. Pairwise comparisons of effort between years and seasons over the two year survey period (NS: not significant, p>0.05; all other comparisons are significant, p<0.05)...... 41

Table 3.1. Comparison of sampling issues, ease of implementation and the relative cost associated with available sampling options that could be used to collect harvest information (N/A= not applicable) ...... 56

Table 3.2. Features of existing datasets and sources used to infer harvest composition and harvest rates for the fishery at the Artificial Reef...... 62

Table 3.3. Standardised harvest (kg/km2) and fishery details for the artificial reef (AR) and other fishery areas...... 65

Table 3.4. Number of taxa in each fish functional group at the artificial reef (AR) and other fishery areas used in the calculation of estimated harvest (refer to Table 3.3 for details of fishery areas)...... 68

Table 3.5. Percentage contribution of functional groups responsible for the average dissimilarity in standardised harvest (kg/km2) between the artificial reef and other fishery areas by survey year, identified using SIMPER analysis. Only metrics accounting for 90% of the similarity observed are shown here. Refer to Table 3.3 for details of fishery areas...... 73

Table 4.1. Summary of acoustically tagged animals. AR= artificial reef, NR= Dunbar reef, TL=total length; M= male; F= female; I=fish of indeterminate sex. L=low power, H=High power. Refer to Table 5.1 for the end of monitoring period of each tagged animal...... 92

Table 4.2. The top 2-3 model candidates based on AICc for GAMM analyses of activity (ms-2) for each species. s(Hour) is the spline of hour, Temp is the temperature variable (°C), Tide is the tidal level variable (m), TL is the total length, and df is the number of parameters in the model...... 108

Table 4.3. Summary statistics of the final GAMM model for activity with intercept, temperature and a spline of hour for each species. ‘edf’ is expected degrees of freedom...... 108

Table 4.4. Results from the Poisson GLM for the 50 m control tag...... 111

Table 4.5. Results from the Poisson GLM for the 200 m control tag...... 112

xvii

Table 5.1. Summary of acoustically tagged animals. AR= artificial reef, NR= Dunbar reef, AM= Annie Miller; TL=total length; M= male; F= female; I=fish of indeterminate sex. L=low power, H=High power; A= apparent survival, R= recaptured, U= undetected ...... 134

Table 5.2. Total residence time (tR, days), monitoring period (tM, days) and residency index (IR) of acoustically tagged animals at tagging reef. AR= artificial reef, NR= Dunbar reef...... 150

xviii

Chapter 1

Introduction

1.1 Background

Artificial reefs are deployed world-wide with the purpose of creating fish habitat and enhancing marine ecosystems (Bohnsack and Sutherland 1985; Baine 2001), as well as increasing recreational fishing opportunities (Branden et al. 1994; Carr and Hixon 1997;

Cenci et al. 2011). In the past, artificial reefs were constructed from a variety of waste materials such as rubble, car tyres, decommissioned ships and oil platforms among others, collectively referred to as ‘materials of opportunity’ (Bohnsack and Sutherland

1985; Brickhill et al. 2005). As awareness of the negative environmental impacts of this waste material increased, better reef developments became necessary. Extensive research eventually led to the development of designed and purpose-built, artificial reefs. These modern artificial reefs are built using design-specific materials which are structurally complex and include several internal spaces to mimic highly convoluted natural reefs (Ambrose and Swarbrick 1989; Branden et al. 1994).

Japan and Korea are considered to be the world leaders in artificial reef technology

(Kim et al. 2008). The design of these reefs permits water flow through the structure which provides an enhanced supply of nutrient and plankton to the artificial reef ecosystem to promote the growth of sessile organisms and resident fishes (Connell and

Anderson 1999; Redman and Szedlmayer 2009). The position of the artificial structure in relation to currents is also believed to influence the distribution of pelagic, demersal and benthic species whereby the highest abundance and density of species tend to be

1 attracted at the maximum current flow (Pickering and Whitmarsh 1997; Palardy and

Witman 2011; Champion et al. 2015).

Artificial reef development in Australia followed a similar pattern to artificial reef projects worldwide. These tended to occur in cycles where expenditure for reef developments was limited by an inability to assess the relative benefit of each initiative against objectives. The first reported artificial reef in Australia was constructed from concrete pipes and deployed in 1965 in Port Phillip Bay, a large heavily urbanised tide- dominated coastal bay on the southeast Australian coast (Kerr 1992). Over the next few years reefs were constructed primarily for recreational fishing and diving in estuarine and offshore locations of eastern and southern Australia (Kerr 1992; Branden et al.

1994). The inability to evaluate the projected benefits of these initial reef deployments resulted in the stalling of further artificial reef research in the mid-1980s. The failure of these early initiatives to develop as part of an integrated approach to fisheries enhancement was a result of inadequate knowledge regarding the design and deployment of artificial structures. This was primarily due to inadequate post- deployment assessment (Svane and Petersen 2001; Wilding and Sayer 2002).

A renewed interest in artificial reefs occurred in the state of New South Wales (NSW), where recreational fishing is a popular leisure activity with significant economic contributions for approximately 1 million people (17 % of the state’s population)

(Henry and Lyle 2003). The introduction of the recreational fishing license fee in NSW in 2001 provided funding for the NSW Department of Primary Industries-Fisheries to invest in the deployment and monitoring of artificial reefs as well as fish aggregating devices (FADs), as part of a fisheries enhancement program (Folpp and Lowry 2006;

Lowry et al. 2015). FADs have been deployed off the NSW coast since 2003 and consist of a single float, moored to the seabed at varying depths and provide a fixed location where fast-growing pelagic fish species such as dolphin fish (Coryphaena hippurus) can be targeted by recreational fishers (Folpp and Lowry 2006). FADs are unlikely to increase fish production; they simply attract fish to be more easily exploited by fishing.

A number of small multicomponent artificial reefs known as ‘Reef balls’ were also deployed in local estuaries of NSW in 2005. Since then other artificial reefs have been deployed in coastal estuarine areas of Australia (Branden et al. 1994; Folpp et al. 2011;

Folpp et al. 2013; Lowry et al. 2014). These reefs were quickly colonised by a variety of fish (Folpp et al. 2011). Following the success of the estuarine artificial reef program, design-specific artificial reefs typically made of steel or concrete were built and deployed in coastal NSW (Lowry et al. 2015). The potential for these artificial reefs is considered by the recreational fishing community as a high priority (Lowry and Folpp

2014). These reefs are increasingly designed according to the types of target species as different species of fish may respond to hard objects differently (Kim et al. 2008).

1.2 The ‘attraction versus production’ debate

Designed artificial reefs are referred to as management tools as they are believed to support increased species diversity and abundance by providing food and refuge, as well as increased fishing opportunities through the abundance of recreational fish (Branden et al. 1994; Carr and Hixon 1997; Cenci et al. 2011). There is an ongoing debate in the scientific community about whether artificial reefs increase overall production of a defined area, or whether they merely attract and aggregate existing fish to a new 3 location (Bohnsack and Sutherland 1985; Solonsky 1985; Bohnsack 1989; Carr and

Hixon 1997; Pickering and Whitmarsh 1997; Folpp et al. 2013; Smith et al. 2015). This has become known as the ‘attraction versus production’ issue. Attraction is defined as the net movement of individuals from natural to artificial habitats, which would otherwise have settled, survived and grown on natural habitats in its absence (Carr and

Hixon 1997; Brickhill et al. 2005). This is more likely to occur in locations where natural reef habitat is abundant and where species have a high fishing mortality, are recruitment limited, highly mobile, partially reef-dependent or opportunists (Bohnsack

1989). Production on the other hand is defined as an increase in total fish biomass and includes the growth, reproduction, and survival of fish (Bohnsack and Sutherland 1985;

Carr and Hixon 1997). Artificial reefs are expected to act as fish producers if the availability of habitat in an area is limited (Solonsky 1985; Carr and Hixon 1997;

Grossman et al. 1997).

In support of the production hypothesis, a study on the isotopic ratios of fishes confirmed that artificial reefs support production through an enhanced food supply

(Cresson et al. 2014). Previous studies also demonstrated that species richness, diversity and biomass density of fish is usually the same or higher on artificial reefs than natural reefs without notable effects on fishes dwelling in nearby non-reef habitats (eg.

Bohnsack and Sutherland 1985; Alevizon and Gorham 1989; Johnson et al. 1994;

Pondella et al. 2002). The presence of an artificial reef can also provide larger fishing catches compared to natural control reefs (Fabi and Fiorentini 1994; McGlennon and

Branden 1994; Carr and Hixon 1997; Santos and Monteiro 1998; Whitmarsh et al.

2008; Bortone et al. 2011; Leitão 2013). This creates concern however, as these reefs may negatively impact fish abundance by facilitating exploitation by fishermen through

4 aggregation (Bohnsack and Sutherland 1985; Solonsky 1985; Bohnsack 1989; Carr and

Hixon 1997; Grossman et al. 1997; Pickering and Whitmarsh 1997; Powers et al. 2003).

The possibility that both attraction and production occur simultaneously at artificial reefs rather than an either/or response, has been suggested by some studies, where the relative effect of these two processes may be influenced by various factors including species-specific characteristics, artificial structure design and placement (Wilson et al.

2001; Osenberg et al. 2002a). Furthermore, it is argued that while artificial reefs may simply attract and aggregate some species, they may promote the production of others and the situation is likely to lie between the two extremes (Bohnsack 1989). For example, attraction is considered more likely to be due to the abundance of large fish around an artificial reef within days to months after reef installation, whereas production can occur over several decades after the reef is established (i.e. due to increased biofouling communities) (Macreadie et al. 2011). The potential for fish production on artificial reefs needs to be evaluated so that fish biomass in terms of fish population growth must exceed losses of biomass through overfishing and/or other human related impacts (Carr and Hixon 1997; Pickering and Whitmarsh 1997; Shipley and Cowan Jr 2011).

In a designated artificial reef area heavily fished primarily by recreational anglers, limited movement can be detrimental to targeted species. In addition, if fishing mortality exceeds either productivity or recruitment and if fish production is not limited by the availability of habitat, then high fishing mortality rates may offset or diminish any net gains in productivity resulting from artificial reef construction (Strelcheck et al.

2007). Conversely, the location and proximity of artificial reefs to natural reefs can

5 positively influence fish abundance and diversity by increasing the ecological connectivity between habitats, thus facilitating species dispersal which reduces fishing- related mortality at these reefs (Cenci et al. 2011; Macreadie et al. 2011; Shipley and

Cowan Jr 2011; Smith et al. 2015). Movement by fish assemblages between natural and artificial reefs has been identified as important in ecological systems as it can be reliant upon prey availability, shelter and spawning opportunities (Topping and Szedlmayer

2011b; Topping and Szedlmayer 2011a).

1.3 Assessing attraction and production

The east coast of Australia has a highly urbanised coastline with a growing population and many people engaging in recreational fishing (Henry and Lyle 2003), highlighting the risk of increasing impacts on many marine species by recreational fisheries. Few studies have assessed recreational fishing effort at artificial reefs in relation to temporal fishing effort patterns within the region (Buchanan 1973; McGlennon and Branden

1994; Tinsman and Whitmore 2006). Furthermore, the impact of recreational catch and effort at an artificial reef in relation to local production is unknown. Understanding the level of catch and effort on fisheries by recreational anglers and how this changes through time and space is therefore critical for the sustainable management of fish stocks, as well as to maintain the societal and economic benefits to and from this sector

(Steffe et al. 2008; Griffiths 2012). This information is also important to assess the contribution of artificial reefs to recreational fisheries and to determine the economic benefit for implementing more of these reefs in the future. To better understand fishery aspects of artificial reefs, more emphasis is needed in artificial reef studies to examine and/or consider both aspects of production and concentration or attraction of fishes

(Shipley and Cowan Jr 2011). There is a need to focus on studying the residence time, 6 connectivity and residency of recreationally important benthic fish around artificial reefs and natural reefs. This can be achieved through acoustic telemetry studies of fish which are combined with the calculated recreational effort and harvest to determine the production potential of artificial reefs.

Acoustic telemetry involves the use of an electronic transmitter (or ‘tag’) which transmits data to a receiver apparatus moored at a fixed location that stores, displays or amplifies the information (Hussey et al. 2015). This technology has become a popular and effective method for remotely studying the movement, habitat preferences, behaviour and physiology of free-ranging aquatic animals (Hussey et al. 2015). Strong site residency of species at artificial structures has previously been reported for a number of species such as red snapper (Lutjanus campechanus) (Topping and

Szedlmayer 2011b; Piraino and Szedlmayer 2014), Atlantic cod (Gadus morhua)

(Reubens et al. 2013), Copper rockfish (Sebastes caurinus) and lingcod (Ophiodon enlongatus) (Reynolds et al. 2010). Assessing the movements and residency of fish at artificial reefs and nearby natural reefs via acoustic telemetry provides information regarding the suitability of artificial structures as habitat for fish, contribution to fisheries enhancement and therefore local production (Shipley and Cowan Jr 2011;

Leitão 2013).

1.4 Study region

This study focused on the fishing effort, harvest, and distribution of some key species around the Sydney (offshore) designed artificial reef (AR). The AR is Australia’s 7 largest single purpose-built artificial reef, constructed from 42 tonnes of steel and deployed in 38m of water off Sydney’s south head, NSW (Fig. 1.1). The reef was deployed with the primary purpose of enhancing recreational fishing opportunities. The structure was lowered into position on the morning of 12th October 2011 (Fig. 1.2) and was followed by the attachment of moorings and inspection by divers prior to commissioning on the 13th October 2011. A nearby natural reef (NR), Dunbar reef, was also monitored as part of this study. This site consists of a large outcrop of subtidal reef in 25-30 m of water and is located ~600-800 m from the AR (Fig. 1.1).

North head

South head

Sydney Harbour

Figure 1.1. Location of the designed artificial reef (AR) and Dunbar reef.

Figure 1.2. Deployment of the AR on the 12th October 2011. Image: Justin Gilligan 9

As part of the ongoing environmental monitoring of the AR by the NSW Department of

Primary Industries - Fisheries, the primary objectives involved assessing the species composition and residency times of the fish community associated with the AR, and comparing these with fish assemblages associated with natural (control) reefs in the immediate vicinity. Additional priorities included an assessment of the popularity of the

AR with recreational fishing groups to determine the economic benefit of the deployment of the structure (Lowry et al. 2015).

1.5 Study Species

Limited research has been undertaken on the ecology, residency and movement of benthic species associated with ARs, particularly in Australia. Understanding large and small-scale habitat use and movements are critical for the management of species that may be impacted either directly or indirectly by recreational fishing at the AR.

Investigating the association of these species with ARs is also important to determine the potential for these reefs in increasing the habitat availability in temperate rocky reef ecosystems. Part of the research (Chapters 3 and 4) presented here focuses on three common benthic species, the eastern fiddler ray (Trygonorrhina fasciata), eastern bluespotted flathead (Platycephalus caeruleopunctatus) and Port Jackson shark

(Heterodontus portusjacksoni). All three species are endemic to eastern Australia and inhabit soft substrate habitats in the vicinity of coastal rocky reefs (Hutchins and

Swainston 1986; Last and Stevens 1994; Powter and Gladstone 2008b; Moore et al.

2009). These species were selected for this study as they have previously been identified from routine Baited Remote Underwater Video (BRUV) and drop camera surveys at the

AR and surrounding natural reefs (Lowry and Folpp 2014; Scott et al. 2015).

The eastern fiddler ray (or banjo ray, hereafter referred to as ‘fiddler ray’, Fig. 1.3) occurs from southern Queensland to southern New South Wales in depths up to 100 m

(Last and Stevens 1994). This species was previously confused with the southern fiddler ray (T. dumerilii), which was referred to as T. fasciata then renamed as T. dumerilii

(White and Last 2012). Little is known about the life history and ecology of the fiddler ray, except that its reproductive biology is similar to that of the southern fiddler ray

(Marshall et al. 2007; Izzo and Gillanders 2008). The fiddler ray is reported to feed on crustaceans, fish, polychaetes and molluscs (Marshall et al. 2007; Izzo and Gillanders

2008).

Figure 1.3. The eastern fiddler ray, Trygonorrhina fasciata.

The eastern bluespotted flathead (hereafter referred to as ‘bluespotted flathead’, Fig.

1.4) occurs from southern Queensland to eastern Victoria to depths of approximately

100 m (Hutchins and Swainston 1986; Kuiter 2000; Rowling et al. 2010). The bluespotted flathead is an important commercial and recreational species, representing the primary species group harvested in New South Wales with an estimated harvest of between 320 and 450 tonnes (Henry and Lyle 2003; Rowling et al. 2010; Steffe and

Murphy 2011). Despite its recreational and commercial significance, little is known about the ecology and life history of this species. However reports indicate that the species grows to 90 cm, spawning occurs over an extended period from late winter to 11 summer, and its diet consists of crustaceans, fish, polychaetes and molluscs (Coleman and Mobley 1984; Hutchins and Swainston 1986; Moore et al. 2009; Barnes et al.

2011).

Figure 1.4. The eastern bluespotted flathead, Platycephalus caeruleopunctatus.

The Port Jackson shark occurs in southern Australian waters from southern Queensland to Tasmania, and west to the central coast of Western Australia, to depths of 275 m

(Last and Stevens 1994; Fig. 1.5). Extensive research into the species’ life history and ecology reveal that Port Jackson sharks are oviparous (Powter and Gladstone 2008c), and have a broad diet consisting of fish, echinoderms and decapod crustaceans (Powter et al. 2010). Tagging studies have shown that the Port Jackson Shark is a migratory species that exhibits large-scale movements associated with their breeding season

(McLaughlin and O'Gower 1971; O'Gower 1995). Both this species and the fiddler ray are not commercially targeted, but are taken as bycatch in fishery trawls (Jones et al.

2010; Rowling et al. 2010; Huveneers 2015; Huveneers and Simpfendorfer 2015).

Figure 1.5. Port Jackson shark, Heterodontus portusjacksoni.

1.6 Research objectives

The aim of this study was to address knowledge gaps relating to the contribution of designed ARs as a fisheries management strategy and more specifically for the role that the Sydney AR plays in the enhancement of NSW recreational fisheries. This was quantified by estimating the fishing effort and harvest of recreational species by fishers using the AR. The second component involves the use of acoustic telemetry to study the behaviour, movements and residency of benthic species associated with the AR and adjacent natural reefs, using a ray, a teleost and a shark as example study species. In addition to studying the movements and residency of these benthic species, this research also examines the activity patterns and influence of environmental parameters on these three species.

This study stems from a collaborative effort to determine the pelagic-benthic trophic ecosystem of the design-specific AR (Fig. 1.6). The specific aim of the overall project is to determine if coastal currents supply sufficient nutrients for the local production of fish through designing an ecosystem model. The estimated fishing effort and harvest

(subproject 3, Fig. 1.6) from the AR together with fish residency (i.e. proportion of fish derived from the AR, subproject 4) from this study will be used as inputs into an ecosystem model (Ecopath with Ecosim, EwE), along with two other components

(subproject 1: coastal currents and nutrients at the AR, and subproject 2: recruitment and growth of benthic food, Fig. 1.6), as part of a larger project. EwE creates a mass- balanced model of trophic flows among the functional groups in an ecosystem (in this case, a shallow coastal ecosystem including both natural and AR habitat), which is used to structure dynamic simulations of the biomass fluxes (with Ecosim) in response to certain forcings, including fish residency and harvest (Christensen and Pauly 1992; 13

Pauly et al. 2000). The impact on the biomass of all functional groups from deploying additional ARs can be investigated using Ecospace (the spatially-explicit component of

EwE). Using this model for the purpose of the overall project aim will improve our understanding of how much of the fish biomass is derived from the AR productivity and whether the AR has a net production or attraction effect, as well as the cost-benefit of additional AR deployments.

Subproject 3: Fishing effort and fish harvest from security camera Subproject 1: Flow * [plankton] = supply of food to benthos

Subproject 2: Recruitment and growth of benthos; food quality

Subproject 4: Fish residency –Proportion of fish derived from AR

Figure 1.6. Schematic diagram of subprojects for input into the AR ecosystem model.

The specific research aims of this study were to:

1) Estimate recreational boat-based fishing effort at the Sydney AR using a shore-

based camera;

2) Estimate the recreational harvest of fish from the AR fishery using a

combination of existing datasets;

3) Determine the diel behaviour and environmental drivers on the activity of three

benthic species around the AR; and,

4) Assess the site residency and movements of these three species at the AR and

surrounding natural reefs.

1.7 Thesis chapters

In Chapter 2, the boat-based recreational fishing effort at the AR over a two-year period is estimated using a shore-based camera. Monitoring recreational fishing effort is important to determine whether this AR enhances recreational fishing opportunities; since fishing effort is positively correlated to the levels of fishing-related mortality

(Polovina 1989). A validation study was developed by using volunteer observer counts of fishing events to adjust effort estimates derived from the digital images to account for visibility bias. Seasonal fishing effort in units of boat hours at the AR was converted to units of fisher hours and compared between the two survey years. To compare the level of effort at the AR with other Australian ARs and estuarine fisheries, fishing effort was standardised as intensity per unit area and calculated as effort intensity per square kilometre (fisher hours/ km2). Comparisons with other studies were limited to those for which information on the fishing effort (in fisher hours) and size of the fishery were available, therefore comparisons were made with three South Australian ARs and estuarine fisheries. These comparisons are important for determining the usage of the 15

AR by recreational anglers and for assisting managers to establish the economic benefits of implementing more ARs in Australia. This chapter was published in

Fisheries Research.

In Chapter 3, the recreational harvest at the AR using existing datasets to infer the harvest composition and harvest rates of recreational fishers using the reef was estimated. Estimating harvest is a challenge faced by researchers worldwide due to the many sampling options available which vary in their ability to deliver unbiased information about the fishery and their relative costs of implementation (Pollock et al.

1994; Smallwood et al. 2012; Hartill and Edwards 2015) . This study uses a combination of multiple datasets to obtain a list of taxa harvested by recreational fishers within the AR area and data from a previous probability-based survey in coastal Sydney waters which were used to obtain estimates of harvest rates for these taxa. Harvest was estimated by multiplying these harvest rates together with fishing effort derived from the validated digital images calculated in chapter 2. This information is presented as a case study in this chapter as it uses a robust and cost-effective sampling method to estimate harvest for these species. Harvest for each species was grouped into 10 fish functional groups for analysis. Standardised comparisons of harvest per unit area

(kg/km2) were made between the AR and 7 estuarine fishery areas for which the fishable area was known to provide context for the relative size of the recreational harvest from the AR fishery. This chapter was recently published in Fisheries Research.

Chapter 4 examines the activity patterns and influence of environmental parameters in the fiddler ray, bluespotted flathead and Port Jackson shark using acoustic telemetry.

Various environmental and biological factors are known to influence the movements

16 and distribution of fish, particularly in dynamic environments (Vianna et al. 2013;

Henderson et al. 2014; Schlaff et al. 2014; Gannon et al. 2015; Kruse et al. 2015). Thus these factors need to be considered when assessing the fish assemblages at a particular habitat such as an AR. As part of this research, accelerometer transmitters were implanted in 28 animals which were caught and released at the AR, and 9 animals which were caught and released at the NR. Diel activity patterns and the effect of temperature, luminosity and tide on the activity of all three species were examined using data from receivers at the AR and nearby natural reefs, as well as from a large integrated network of receivers in the greater Sydney area.

In Chapter 5, the residency and connectivity patterns of the fiddler ray, bluespotted flathead and Port Jackson were examined at the AR and surrounding natural reefs.

Understanding species’ residency and connectivity is needed to determine the potential contribution of an AR to production as well as the risk of fishing mortality (Szedlmayer and Schroepfer 2005; Cenci et al. 2011; Shipley and Cowan Jr 2011). As part of this research, a combination of acoustic transmitters was used and includes individuals which were tagged with accelerometer transmitters (from Chapter 4 above). A total of

43 animals were caught, tagged and released at the AR, 10 animals were caught, tagged and released at the NR and an additional 9 fish were caught, tagged and released at a nearby reef, the Annie Miller. Connectivity of the AR with adjacent natural reefs was determined by comparing data from the AR receiver with those from the integrated network of receivers across greater Sydney. The residency and connectivity information is used to infer the effect of this AR on the local distribution of these species. This chapter is currently in review in Marine and Freshwater Research.

Chapter 6 provides a synopsis of the main results presented in the individual chapters and discusses the potential for production and fisheries implications at ARs, using the

Sydney AR as a prime example. This chapter also explores future research directions.

The four data chapters of this thesis were written as individual publications for an international audience, and have been either submitted for publication, or prepared for submission. Each chapter is co-authored, with co-authors contributing to the data analysis and interpretation, or assisted in the preparation of manuscripts. The details of each prepared publication, or submission, and co-author contributions are listed at the end of each chapter.

Chapter 2

Monitoring boat-based recreational fishing effort at a nearshore artificial reef with a shore-based camera

2.1 Abstract

Recreational fishing effort was quantified at a 700 m3 steel artificial reef (AR) off coastal Sydney with a shore-based camera (06:00- 18:00) over a two-year period.

Stratified random sampling was used to select days for analysis of fishing effort from digital images. Fishing effort estimates derived from the digital images were adjusted to account for visibility bias using information from a validation study. The levels of effort recorded in the first two seasons were low as the AR had been recently deployed and colonization of the AR by sessile organisms and fishes was still occurring. The effort intensity (fisher hours per square kilometre) at the Sydney AR was compared with three

South Australian ARs and 14 estuarine fisheries in New South Wales (NSW) to provide context for the study. Effort intensity at the AR was found to be up to 12 times higher than that recorded from some estuarine fisheries in NSW. Conversely, the levels of effort intensity at two South Australian ARs were much higher compared to those at the

Sydney AR site in both survey years. Effort intensity comparisons showed that the relative levels of usage at Australian ARs were higher than those recorded from estuarine fisheries. The Sydney AR provides diverse fishing opportunities that may be concentrated in a small area. Camera-based technologies can provide a solution for cost- effective monitoring of AR sites, providing the accuracy of fishing effort information derived from camera images is validated. Our study has broad implications for other recreational ARs, including many future deployments planned for eastern Australia.

2.2 Introduction

A recent advance in technology for monitoring recreational fishing effort involves the use of shore-based camera imagery. This Internet Protocol (IP) webcam system can be used to monitor fishing effort at well-defined access points to a fishery (e.g. boat ramps, choke points) and fixed areas such as jetties, wharves, rock groynes (Ames and

Schlindler 2009; Smallwood et al. 2011; Smallwood et al. 2012) or the surface area above an artificial reef (AR). Camera imagery can reduce long-term program and monitoring costs and can provide a permanent record of activity which can be accessed and processed after the sampling period is complete (Ames and Schlindler 2009;

Smallwood et al. 2011; Smallwood et al. 2012). Shore-based cameras are becoming widely used for assessing nearshore recreational fishing. In the study by Ames and

Schlindler (2009), two digital cameras were mounted to two separate towers, one adjacent to a jetty and the other at the jetty entrance to capture images of passing boats in Newport, Oregon. Smallwood et al. (2011 and 2012) fixed cameras to four large groynes to observe recreational fishers in Perth, Western Australia. Similarly Van

Poorten et al. (2015) attached cameras to trees and other high stable permanent structures to record observations of angling effort in 49 small rural lakes in central

British Columbia, Canada.

Monitoring recreational fishing effort at ARs is important for determining their effectiveness for enhancing recreational fishing. These structures are built from various materials to create fish habitat and enhance marine ecosystems as well as increase recreational fishing opportunities (Branden et al. 1994; Carr and Hixon 1997; Cenci et al. 2011). An understanding of temporal patterns of fishing at ARs is also important because fishing effort is positively correlated to the levels of fishing-related mortality 20

(i.e. harvest and release-induced mortality). The success of an AR depends on its ability to enhance fishing opportunities (i.e. increasing the number of fish that are available for capture) while also maintaining the community equilibrium. Few studies have assessed recreational fishing effort at ARs in relation to temporal fishing effort patterns within the region (Buchanan 1973; McGlennon and Branden 1994; Tinsman and Whitmore

2006).

Whether ARs enhance the production of fish biomass, or simply attract and aggregate fish is a controversial topic (Bohnsack and Sutherland 1985; Solonsky 1985; Bohnsack

1989; Carr and Hixon 1997; Folpp et al. 2013; Smith et al. 2015). However, ARs have been viewed as an important resource for preventing localised overfishing by reducing fishing pressure on nearby natural reefs (Bohnsack and Sutherland 1985; Pickering and

Whitmarsh 1997; Santos and Monteiro 1998; Folpp et al. 2013; Smith et al. 2015).

Furthermore, studies have demonstrated that ARs may provide larger fishing catches compared to natural control reefs (Fabi and Fiorentini 1994; Carr and Hixon 1997;

Santos and Monteiro 1998; Whitmarsh et al. 2008; Bortone et al. 2011; Leitão 2013).

The AR was deployed 1.5 km off the Sydney coast in October 2011, to enhance recreational fishing opportunities. The 42-ton steel structure was designed with many open void spaces and towers that are particularly attractive to fish. The design allows water flows that provide an enhanced supply of nutrient and plankton to the AR ecosystem and can promote the growth of sessile organisms and resident fishes (Connell and Anderson 1999; Redman and Szedlmayer 2009). Similar ARs are being deployed elsewhere around Australia, and the need for cost-effective solutions to monitor recreational fishing at these locations is imperative.

The main aim of this study is to estimate recreational boat-based fishing effort at the

Sydney AR using a shore-based camera. We also quantify patterns of recreational boat- based fishing effort at the AR for each season over a two year period, and standardise the fishing effort by area at the AR to allow comparisons with other Australian ARs and estuarine fisheries. These comparisons are important for determining the usage by recreational anglers and for assisting managers to determine the economic benefits of implementing more ARs in Australia.

2.3 Materials and methods

2.3.1 Artificial Reef description

The Sydney AR (33°50.797'S, 151°17.988'E) was deployed in October 2011 in 38 m depth of water, approximately 1.2 km east of ‘The Gap’, near the southern headland of

Sydney Harbour, New South Wales (NSW), Australia (Fig. 2.1). The steel structure is

12 x 15 m (a footprint of 180 m2) and 12 m high with two 8 m tall pillars, resulting in a reef volume of 700 m3 (Fig. 2.2). The reef is moored at each corner with chain and a 60 ton concrete block. The AR area monitored was calculated using time and position data with a boat Global Position System (GPS). We defined the effective fishing area at and adjacent to the AR so that it included the AR structure, its footprint and an adjacent area that would be used by anglers to target fishes associated with the AR. The effective fishing area adjacent to the AR was determined by considering the direction of prevailing currents and the fishing practices (e.g. drifting, trolling, berleying) used by recreational anglers in the AR area. This area was calculated to be 140 m (north to south) x 400 m (east to west), providing coverage of recreational fishing activity for about 0.056 km² (Fig. 2.3). This calculated area also allowed for the ease of visual comparisons with the independent counts of fishing events made by observers, as part of a validation study (see section 2.3.3 for further details).

Figure 2.1. Location of the Sydney Artificial Reef (AR) and the Old South Head Signal

Station (indicated as camera).

Figure 2.2. Schematic of the 42-ton Sydney Artificial Reef (AR).

Figure 2.3. Camera view of the 0.056 km² monitored AR area (140 m (north to south) x

400 m (east to west)). ‘X’ indicates location of the AR. (Note the two different scale bars, the vertical scale bar (not to scale) indicates the increasing distance from the base of the monitored area).

2.3.2 Camera imagery

Recreational fishing effort at the AR was assessed by using a shore-based camera system. A mobotix M24, 3 megapixel (2048 x 1536 pixel resolution, 8 x digital zoom,

45° horizontal lens, 8 mm focal length, 2.0 aperture) twin-head IP camera

(www.anso.com.au) was fixed to a vantage point at 85 m above sea level at the Old

South Head Signal Station - a lighthouse (33°51'1.47"S, 151°17'12.41"E, Fig. 2.1), to capture digital images of the AR area, at a distance of 1.3 km from the reef.

Photographic stills were recorded continuously every minute during a defined period of daylight (06:00 to 18:00) over two years from 1 June 2012 to 30 May 2014 (total of

188,370 images). Images (400-800 KB each) were downloaded in JPEG format and stored on a 32 GB Micro-SD internal card. Images were downloaded on site via a laptop.

A pilot study was conducted prior to the main study to determine the time it took for a non-fishing vessel to transit the AR area. From this study we classified any vessel that remained in the AR area for 5 minutes or longer to be fishing, whether they were drift fishing or trolling. A fishing event was therefore defined as a vessel remaining in the

AR area for at least 5 minutes (i.e. 5 frames). All types of vessels, regardless of size, were counted and included in the estimation of fishing effort if they remained within the vicinity of the AR to ensure that all vessels fishing were included. The fishing effort data generated from the digital images was in units of fishing events and boat hours (i.e. the number of hours of boat-based fishing in the AR area). Two people were involved in the analysis of the digital images using Microsoft office picture manager. A reference set of digital images was used to train and standardise the image interpretation of the two readers. 27

2.3.3 Validation of data derived from digital images

We investigated the potential bias in the data derived from digital images by comparing them with independent counts of fishing events made by observers. A group of marine rescue volunteers based at the old signal station were trained to observe and record the number of fishing events during daylight hours (06:00 to 18:00) in the AR monitored area. All volunteers used a pair of standard binoculars (7x50 Tasco marine series, model

222YRZ) to observe fishing events. Volunteers were trained to observe and count the number of all vessels that remained at least 5 minutes within the same monitored area.

Observers also recorded the time each vessel arrived and departed. The location of the

AR and boundary of the monitored area was easily observed from the old signal station due to calibrated markings (previously calculated using GPS distance) on the window facing directly east. The observations of fishing effort reported by the volunteer observers were regarded as accurate measures of fishing effort in the AR area. Observer counts were regularly quality controlled by comparing current observations with those of the volunteers during onsite visits to download the data from the camera. These volunteer observations were used to calculate fishing events only. Volunteer observations of fishing effort were done on 55 randomly selected days in the period

November 2012 to June 2014 that covered all types of weather conditions. Volunteers also noted when no vessels were observed.

A linear regression forced through the origin was fitted using digital image estimates of daily fishing events on the y-axis and the observer-validated number of daily fishing events on the x-axis. This regression analysis can provide evidence of bias in the digital image generated estimates of fishing events if the slope of the regression line differs significantly from 1.0. The volunteer observers were assumed to be measured without 28 error, since they were trained to count the number of boats within the fishing area.

Forcing the validation regression through the origin was required otherwise the correction factor will be biased towards a higher number of fishing events which could result in an overestimate of fishing effort.

To determine whether there was any significant bias, a two-tailed t-test was used to test whether the sample value of b (i.e. slope) was different from the expected value of 1

(i.e. when there is no bias) (Sokal and Rohlf 1981). When bias is detected, it is possible to derive a correction factor using the regression coefficient and its variance derived from the “variance of a quotient” equation (Blumenfeld 2001). This correction factor was used to adjust the estimates of fishing effort (fishing events) for all strata. In addition, the number of fishing events on days when rainfall (≥0.4 mm) was recorded during the validation study were compared between the validated observations and digital images data (n=9) using a two-tailed t-test. No significant difference was found

(two-tailed t-test, t= 1.51, df= 8, P>0.05). See Appendix A for detailed calculations.

2.3.4 Effort estimation from digital images

Stratified random sampling methods were used to select a sample of daily digital images for processing. This method also takes into account both good and poor weather days.

Days were the primary sampling unit for all strata. By definition, a survey day started at

6 am and ended at 6 pm. Each year was stratified into austral seasons and day-types within seasons (weekdays and weekend days). Public holidays were classified as weekend days. Sample sizes for each base level stratum are given in Table 2.1.

Whenever possible, each month within a season was allocated an equal number of each day-type. Mean daily fishing effort values (i.e. daily mean of the mean number of 29 fishing events per day) and variances for each day-type stratum within each season and within each survey year were calculated. The first survey year spans from June 2012

(when the reef was 8 months old) to May 2013 and the second survey year spans from

June 2013 to May 2014.

Table 2.1. The number of days sampled (n) and stratum sizes (N) within

day-type, weekdays (WD) and weekend days (WE), during the survey

period.

Day- Days Number of days Season/ Year Type sampled (n) in stratum (N) WD 10 64 Winter 2012 WE 10 28

WD 15 64 Spring 2012 WE 15 27

WD 15 60 Summer 2012-13 WE 15 30

WD 15 63 Autumn 2013 WE 15 29

WD 15 64 Winter 2013 WE 15 28

WD 15 64 Spring 2013 WE 15 27

WD 15 61 Summer 2013-14 WE 15 29

WD 15 62 Autumn 2014 WE 15 30

The total fishing effort for each day-type stratum was estimated using a direct expansion method to account for the unsampled fraction of the stratum (Cochran 1977; Pollock et 30 al. 1994; Steffe and Chapman 2003). This was calculated by multiplying the mean daily fishing effort values by the number of possible sample days (N) in each day-type stratum. Day-type stratum totals were added together to obtain seasonal totals and the seasonal totals were summed to obtain annual estimates.

The variances for the base level strata are independent estimates therefore they were summed to obtain seasonal variances. Annual variances were obtained by adding seasonal variances together (which were already weighted by day type and season- see

Appendix A). A correction factor developed during the validation study was applied to the effort and variance estimates to adjust for the under-estimation of fishing effort caused by weather related visibility issues (Blumenfeld 2001; refer to Appendix A for further details of equations used). This correction factor was applied in units of fishing events.

Fishing effort was converted from fishing events into boat hours and finally into fisher hours, for comparisons with other studies (Appendix A). We multiplied the estimated total stratum effort (i.e. number of fishing events) independently for each stratum by the daily mean of the mean number of boat hours per fishing event for that stratum (derived from digital images) to obtain estimates of fishing effort in units of boat hours. We then used data from a survey of coastal marine fishing outside the Port Hacking estuary during the period March 2008 to February 2009 (Steffe and Murphy 2011) to obtain estimates of the daily mean of the mean number of fishers per boat (Appendix A). The conversion of fishing effort from boat hours to fisher hours was done by multiplying the boat hour estimates to the daily mean of the mean number of fishers per boat within

31 each base level stratum. Variances were also adjusted for this unit conversion

(Appendix A).

Pairwise comparisons of fishing effort (fisher hours) were made between seasons and years. We used the standard method recommended by Schenker and Gentleman (2001) to calculate an interval for each pairwise comparison. The standard method works by calculating a new interval for each pairwise comparison based on the nominal 95% confidence intervals of the point estimates being compared. The null hypothesis is tested at the convention of P=0.05 and is rejected if and only if the interval does not contain 0 (Schenker and Gentleman 2001).

2.3.5 Standardised comparisons of effort intensity

Standardised comparisons of effort intensity per unit of area were made between the AR and three ARs in the coastal waters of South Australia (McGlennon and Branden 1994;

Table 2.2), and 14 estuarine fisheries in NSW (Steffe et al. 1996; Table 2.2; Steffe et al.

2005a; Steffe et al. 2005b; Bucher 2006; Steffe and Murphy 2011).

Table 2.2. Study site survey periods, habitat types, distance from shore, depth, location (latitude and longitude) and measured areas (km2).

Distance Area Survey location Survey period Habitat type from shore & Latitude Longitude Source (km²) depth

Untreated Sydney artificial Jun 2012-May 2013; 1.2 km, steel 33°50.80'S 151°17.99'E 0.06 This study reef Jun 2013-May 2014 designed reef 38 m depth

Grange artificial Tyre modules 4.3 km, McGlennon and Sep 1990-Aug 1991 34°55.1'S 138°24'E 0.08 reef (1,200) 15 m depth Branden (1994)

Tyre modules Glenelg artificial 5 km, McGlennon and Sep 1990-Aug 1991 (900)/ sunken 34°58.8'S 138°26.4'E 0.19 reef 18 m depth Branden (1994) vessels (2)

Port Noarlunga Tyre modules 2.5 km, McGlennon and Sep 1990-Aug 1991 35°05.2'S 138°26.5'E 0.07 artificial reef (650) 18 m depth Branden (1994)

Northern Lake Mar 1999-Feb 2000; All estuarine - 33°02.0'S 151°37.0'E 60.73 Steffe et al. (2005b) Macquarie Dec 2003-Nov 2004 habitats

Southern Lake Mar 1999-Feb 2000; All estuarine - 33°06.0'S 151°35.0’E 43.10 Steffe et al. (2005b) Macquarie Dec 2003-Nov 2004 habitats

Table 2.2 (Cont.)

Distance Area Survey location Survey period Habitat type from shore & Latitude Longitude Source (km²) depth

Mar 1999-Feb 2000; All estuarine Swansea channel - 33°04.35’S 151°38.40'E 3.23 Steffe et al. (2005b) Dec 2003-Nov 2004 habitats

All estuarine Tweed River Mar 1994-Feb 1995 - 153°32.42'E 20.25 Steffe et al. (1996) habitats 28°14.38'S

All estuarine Richmond River Mar 1994-Feb 1995 - 28°52.24'S 153°32.7'E 25.85 Steffe et al. (1996) habitats

All estuarine Clarence River Mar 1994-Feb 1995 - 29°27.35'S 153° 9.39'E 101.37 Steffe et al. (1996) habitats

All estuarine Brunswick River Mar 1994-Feb 1995 - 28°31.95'S 153°32.0'E 1.58 Steffe et al. (1996) habitats

All estuarine Sandon River Mar 1994-Feb 1995 - 29°41.05'S 153°18.17'E 1.49 Steffe et al. (1996) habitats

Table 2.2 (Cont.)

Distance Area Survey location Survey period Habitat type from shore & Latitude Longitude Source (km²) depth

All estuarine Wooli River Mar 1994-Feb 1995 - 29°57.31'S 153°09.09'E 2.17 Steffe et al. (1996) habitats

All estuarine Mooball Creek Mar 1994-Feb 1995 - 28°25.75'S 153°33.33'E 0.40 Steffe et al. (1996) habitats

Mar 1999-Feb 2000; Dec All estuarine Tuross estuary - 36°03.80'S 150° 6.08'E 14.47 Steffe et al. (2005a) 2003- Nov 2004 habitats

Mar 2007-Feb 2008; Mar All estuarine Steffe and Murphy Hawkesbury reach - 33°33.0’S 151°20.15’E 120.81 2008-Feb 2009 habitats (2011)

Port Hacking Mar 2007-Feb 2008; Mar All estuarine Steffe and Murphy - 34°04.31’S 151°09.30’E 11.51 estuary 2008-Feb 2009 habitats (2011)

Mar 2007-Feb 2008; Mar All estuarine Manning River - 31°53'14’S 152°39'13’E 25.35 Bucher (2006) 2008-Feb 2009 habitats

We calculated standardised values of effort intensity for each fishery by dividing the total effort (in fisher hours) by area in square kilometres. The point estimates of effort intensity for the boat-based fishery at the AR were used to benchmark this fishery against these other recreational fisheries. These standardised comparisons provide relative measures of recreational usage across different fisheries, with the assumption that the patterns of fishing effort and the average number of fishers per boat per day within the study area had not changed. When necessary, survey areas were calculated in

ArcGIS (i.e. Steffe et al. 1996; Steffe et al. 2005a; Steffe et al. 2005b; Steffe and

Murphy 2011)

2.3.6 Comparative coastal fishing effort data from the greater Sydney region

We wanted to compare the seasonal fishing effort estimates from the Sydney AR in our study to the coastal fishing effort from the Sydney region. Total coastal fishing effort

(angling trips ± SE) was calculated using unpublished data from a survey of coastal marine fishing originating from four large waterways in the Sydney area (Hawkesbury,

Port Hacking, Botany Bay, Sydney Harbour) from March 2007 to February 2009 (Steffe and Murphy 2011). Counts of boats returning from the sea were made by observers located on the headlands to these sites. All counts started one hour after sunrise and ended at sunset.

The trailer boat data used were corrected to account and remove non-fishing, spearfishing and diving trips (Steffe and Murphy, unpublished data). Random stratified sampling was used. Each survey year was stratified into seasons and day-types within season (weekend and weekday strata). Public holidays were regarded as part of the weekend day stratum. Days were the primary sampling unit. Sampling was done on 9 36 weekdays and 9 weekend days within each season at each site. Further analyses of these data were conducted for the present study to describe the seasonal pattern of coastal fishing effort within the Sydney region.

2.4 Results

The validation study provided evidence that the data derived from the camera images was significantly biased (two-tailed t-test, t= -3.174, df= 53, P<0.01; Fig. 2.4).

Therefore effort estimates derived from camera images were adjusted to compensate for the underestimation caused by weather-related changes in detectability of vessels. We found that the visibility bias resulted in an underestimate of about 7.5% in the levels of fishing effort. All effort estimates and their measures of precision have been adjusted to correct this bias.

We estimated about 1,765 and 2,460 fisher hours of fishing effort were expended annually at the AR site during the two survey years (Fig.2.5). These annual estimates of effort did not differ significantly between years (P >0.05, Table 2.3). However, we found some seasonal differences in fishing effort (Table 2.3). The level of fishing effort in spring year 2 was significantly greater (P <0.05) than that of spring year 1. All other same season comparisons between years were not significantly different (Table 2.3).

The first survey year was characterized by low levels of fishing effort in the winter and spring seasons (Fig. 2.5). We found that in this first survey year that the fishing effort levels recorded during the autumn and summer seasons were significantly greater than those recorded in the winter and spring (Table 2.3). We found no significant differences in fishing effort among seasons in the second survey year (Table 2.3).

Figure 2.5. Seasonal fishing effort (fisher hours ± SE) at the AR for each survey year

(year 1=June 2012-May 2013, year 2=June 2013-May 2014).

Table 2.3. Pairwise comparisons of effort between years and seasons over the two year survey period (NS: not significant, p>0.05; all other comparisons are significant, p<0.05).

Comparisons Interval Result

Total effort (year 1 and 2) -539.30 - 1949.44 NS

Winter (year 1 and 2) -34.53 - 569.14 NS

Spring (year 1 and 2) 156.57 - 1117.22 Spring Y2>Spring Y1

Summer (year 1 and 2) -442.22 - 752.21 NS

Autumn (year 1 and 2) -1286.84 - 578.61 NS

Winter and spring (year 1) -135.85 - 120.92 NS

Winter and summer (year 1) 33.31 - 840.92 Summer Y1>Winter Y1

Winter and autumn (year 1) 42.82 - 1719.54 Autumn Y1>Winter Y1

Spring and autumn (year 1) 53.17 - 1724.13 Autumn Y1>Spring Y1

Summer and autumn (year 1) -474.96 - 1363.08 NS

Spring and summer (year 1) 46.80 - 842.37 Summer Y1>Spring Y1

Winter and spring (year 2) -190.44 - 914.69 NS

Winter and summer (year 2) -208.77 - 858.39 NS

Winter and autumn (year 2) -248.40 - 767.92 NS

Spring and autumn (year 2) -736.92 - 532.19 NS

Summer and autumn (year 2) -683.15 - 553.04 NS

Spring and summer (year 2) -692.40 - 617.77 NS

The Sydney AR received 31,525 and 44,116 fisher hours per square kilometre during the two survey years respectively (Fig. 2.6). Effort intensity comparisons between the three South Australian ARs (McGlennon and Branden 1994) and the Sydney AR showed that effort intensity was 2.1 times higher at the Grange AR than at the Sydney

AR in the first survey year and 1.5 times higher than at the Sydney AR in the second survey year. Effort intensity was also 1.8 times higher at the Glenelg AR than at the

Sydney AR in the first survey year and 1.3 times higher than at the Sydney AR in the second survey year. However, effort intensity was higher at the Sydney AR than at the

Port Noarlunga AR during both survey years; fishing intensity at the Sydney AR was

1.7 times and 2.4 times higher than at the Port Noarlunga AR in the first and second survey years respectively.

Annual effort intensity was higher at the Sydney AR compared to most estuarine fisheries (Fig. 2.6). Effort intensity at the AR during both survey years was between 5.8 and 9.1 times higher than effort intensity at northern Lake Macquarie during both survey years and between 4.1 and 6.8 times higher than at southern Lake Macquarie during both survey years (Steffe et al. 2005b). Similarly, effort intensity at the AR in survey year one and two was 2.9 and 4.1 times higher respectively than at the Tweed

River; 5.7 and 8 times higher respectively than at the Richmond River; 8.3 and 11.7 times higher respectively than at the Clarence River; 6.7 and 9.4 times higher respectively than at the Sandon River; 2.3 and 3.3 times higher respectively than at the

Wooli River; 2.4 and 3.3 times higher respectively than at the Mooball Creek (Steffe et al. 1996), and 5.5 and 7.7 times higher respectively than at the Manning River (Bucher

2006).

Effort intensity at the AR during both survey years was also between 5.5 and 9.5 times higher than at the Tuross estuary during both survey years (Steffe et al. 2005a); between

7.3 and 10.4 times higher than at the Hawkesbury estuary during both survey years

(Steffe and Murphy 2011) and between 3.8 and 5.6 times higher than at the Port

Hacking estuary during both survey years (Steffe and Murphy 2011). In contrast, effort intensity was similar between the Lake Macquarie entrance (Swansea channel),

Brunswick River and the AR. Effort intensity at the Lake Macquarie entrance in both survey years was 1.1 times higher than at the AR in the first survey year (Steffe et al.

2005b). However effort intensity at the AR in survey year two was between 1.3 and 1.4 times higher than at the Lake Macquarie entrance during both survey years (Steffe et al.

2005b). Similarly, effort intensity at the Brunswick River was 1.1 times higher than at the AR in survey year one, although effort intensity at the AR in survey two was 1.4 times higher than effort intensity at the Brunswick River (Steffe et al. 1996).

Figure 2.6. Comparison of annual fishing effort by area (fisher hours/km2) between study locations. Sydney AR (this study); South Australia

ARs (Gr= Grange, Gl= Glenelg, PN =Port Noarlunga), N. Lake Macquarie=Northern Lake Macquarie, S. Lake Macquarie=Southern Lake

Macquarie, Swansea channel, PH estuary= Port Hacking estuary. For details of survey periods and sources refer to Table 2.2.

2.5 Discussion

This study demonstrates that shore-based camera systems are effective for monitoring changes in fishing effort in a recreational fishery that is likely to be enhanced by creating an AR. We have shown the importance of validating data derived from digital camera images because changes in weather conditions (i.e. fog, rainfall, wind speed, sun glare), can affect the detection of vessels, leading to bias in the data. A similar study which used the same camera system as the one used in this study recorded a peak in boating activity at the same AR during the early morning period (Wood et al. 2016).

This period coincided with observations of high sun glare in our study, which may explain the discrepancy between the number of fishing events generated from the digital imagery and those from the validated observations. Similarly, Smallwood et al. (2011 and 2012) indicated that the fishing effort estimates from their camera study may have been underestimated due to a visibility bias when calculating fishing activity, since they were unable to identify all people fishing. By incorporating a validation study, we found that the visibility bias inherent in our digital image data would have produced an underestimate of about 7.5% in the level of fishing effort if not corrected. Therefore, future studies that rely on camera technologies to capture effort information for recreational fisheries should routinely include a validation component. The minimal additional cost required for the inclusion of a validation study far outweighs the potential risk of basing important management decisions on less accurate information.

The utility of camera systems for monitoring recreational fishing effort can be expected to increase as technological advances occur. Higher quality lenses may become available at lower cost thereby providing enhanced digital image quality with adjustable

ISO settings and better resolution of fishing events, particularly during peak activity 45 periods in variable-light conditions. Future surveys and monitoring programs may be able to more accurately record daytime fishing and possibly even extend coverage into the night, which was not possible to be quantified with the camera system in this study.

The seasonal pattern of fishing effort observed during the two year survey period was influenced by (a) the length of elapsed time since the deployment of the AR structure; and (b) normally occurring seasonal patterns in the activity of recreational fishers in this region. The Sydney AR was deployed in October 2011 and the monitoring of fishing effort at this site commenced just 8 months later. It is known that fish colonization on

ARs can occur rapidly after initial deployment and that this can continue to increase for a period of about 5 years until all elements of the reef ecosystem are established

(Bohnsack and Sutherland 1985; Bohnsack et al. 1994; Scott et al. 2015).

Adult fish recruitment at ARs generally occurs within the first year post-deployment and species richness is highest in the austral summer and autumn (Walsh 1985; Folpp et al. 2011; Lowry et al. 2014). Hence, it is likely that a resident assemblage of fish was not fully established at the AR site during the first two seasons of the monitoring period.

This may have either discouraged recreational fishers initially or they were unaware of the AR, which could have contributed to the relatively low levels of fishing effort recorded at the AR site during the first winter and spring seasons (11% of the maximum). The pattern of seasonal fishing effort recorded after this initial period at the

AR site closely resembled the known pattern of coastal fishing effort within this region.

That is, fishing effort levels tend to be lowest in the winter season and higher in the spring, summer and autumn seasons (Fig. 2.7).

Figure 2.7. Total seasonal fishing effort (angling trips ±SE) for line fishing in four greater Sydney coastal systems from March 2007 to February 2009 (Hawkesbury, Port

Hacking, Botany Bay, Sydney Harbour).

We found that the effort intensity recorded at the AR site was 31,525 and 44,116 fisher hours per square kilometre for years 1 and 2 respectively. This level of usage was up to

12 times more than that recorded from many estuarine fisheries in NSW (Fig. 2.6) and of similar magnitude to the estuarine fisheries in the Brunswick River and the Lake

Macquarie Channel, a relatively shallow but productive area that connects the coastal lagoon to the ocean. In comparison to the effort intensity at three ARs in South

Australia (McGlennon and Branden 1994, Fig. 2.6), the level of effort intensity at the

Sydney AR was higher than that reported for the Port Noarlunga AR (18,310 fisher hours per square kilometre). However, the two ARs at Grange (66,046 fisher hours per square kilometre) and Glenelg (57,505 fisher hours per square kilometre) had much higher levels of effort intensity than those observed at the Sydney AR site. This indicates that ARs may concentrate effort in a small area. Fish density is often much higher on ARs compared to natural reefs (Bohnsack and Sutherland 1985; Ambrose and

Swarbrick 1989). Therefore, ARs can be expected to concentrate fishing effort in the vicinity as anglers target the fish assemblages near them.

It is important to note that fishing effort is not homogenous in estuarine fisheries and can also be concentrated in certain areas. Our standardised estimates of effort in these fisheries are calculated on the whole fishery because we did not have data for smaller spatial scales. It is likely that some of the disparity in the fishing intensity comparisons among fisheries can be attributed to these spatial differences, as well as differences in habitat type (Table 2.2).

The relatively high levels of recreational fishing effort per square kilometre observed in both survey years at the AR site indicates that recreational fishing opportunities are likely to have been enhanced at this site. This may be due to a variety of factors which include a) an increase in biomass as a consequence of additional food being provided by the AR substrate; b) fish attraction and/or c) fish movements from adjacent habitats

(Bohnsack 1989; Cresson et al. 2014). Similarly, Santos and Monteiro (1997) found that the local fishing yields were higher at ARs and that fish biomass was enhanced, especially at protection reefs which provide shelter for fish.

ARs have been shown to provide a diverse range of opportunities for recreational fisheries (Bohnsack 1989; Milon 1989; Whitmarsh et al. 2008). AR deployments are currently being planned for eastern Australia and are expected to increase in the future, thus making it increasingly important to monitor the activities of recreational fishers at these sites. Camera-based technologies provide a solution for cost-effectively monitoring these AR fisheries.

2.6 Conclusions

Management of coastal resources at ARs especially in the vicinity of cities needs long- term data on resource use to determine their relative success for enhancing recreational fisheries. As digital imagery improves and costs decline, camera systems can contribute to this long-term data. Better image analysis technologies are needed to provide cost- effective solutions for monitoring. ARs are popular with recreational fishers and are likely to concentrate effort in a small area. Effort intensity comparisons revealed that

ARs received higher levels of recreational usage than many natural estuarine fisheries.

Camera-based technologies provide a solution for cost-effective monitoring of AR sites, however it is essential to validate the accuracy of data derived from digital images. Our study has broad application to many other recreational ARs around the world.

Publication details:

Keller, K., Steffe, A.S., Lowry, M., Murphy, J.J., and Suthers, I.M. (2016) Monitoring boat-based recreational fishing effort at an artificial reef with a shore based remotely- operated camera. Fisheries Research 181:84-92.

NSW DPI

Chapter 3

Estimating the recreational harvest of fish from a nearshore artificial reef using a pragmatic approach

3.1 Abstract

Artificial reefs (ARs) are deployed for various purposes including the enhancement of recreational fisheries. The ability to assess recreational harvest is important for determining the effectiveness of AR deployments. Harvest estimation at AR fisheries pose many logistical and budgetary challenges. We present a pragmatic approach to estimate harvest at an AR that combines existing datasets and a cost-effective sampling design. Total annual recreational harvest from the AR during June 2013- May 2014 was estimated to be 1,016 ± 82 fish by number, 700 ± 59 kg of fish by weight, and 12,504 kg per km2 when standardised per unit area. Harvest at the AR by number and by weight was relatively small, however when considered per unit area, this standardised harvest was very high (2.3 - 43.6 times larger) compared to other fishery areas from which the fishable area is known. The harvest at the AR was dominated by 6 functional groups

(ambush predators, leatherjackets, large to medium pelagic fish, small pelagic fish, medium demersal predators and large demersal predators), which accounted for 92% of the total annual harvest by number, and 95% of the total annual harvest by weight.

Comparisons of standardised harvest between the AR and other fishery areas revealed two distinct groups, a) the AR and Swansea channel, a marine-dominated entrance to a large estuary, and b) all other fishery areas that were grouped together. We recommend that future studies attempting to estimate harvest at AR fisheries consider an integrated methodology that combines existing datasets and cost-effective sampling designs.

3.2 Introduction

Recreational fishing is a popular leisure activity worldwide (Cooke and Cowx 2004;

Lewin et al. 2006) and within Australia (Henry and Lyle 2003). Recreational fishing is known to have substantial biological impacts, and high economic and social values

(Henry and Lyle 2003). The 2000-01 National Recreational and Indigenous Fishing

Survey (NRIFS) estimated that about 3.36 million residents aged 5 years or older had participated at least once in recreational fishing during the annual survey period (Henry and Lyle 2003). This equates to a national participation rate of about 19.5% (Henry and

Lyle 2003). The number of recreational fishers residing within New South Wales

(NSW) was estimated to be about 999,000 and these fishers undertook about 7.67 million fishing events (Henry and Lyle 2003). The NRIFS also found that the majority of the recreational fishers within NSW resided in the Sydney area (about 48%) and that more than 40% of their fishing events occurred in the coastal marine environment

(Henry and Lyle 2003).

Artificial reefs have been deployed worldwide to create recreational fishing opportunities, or to enhance and restore degraded habitats (Bohnsack and Sutherland

1985; Baine 2001). These artificial reefs are believed to provide many benefits to fish populations and fisheries in their immediate vicinity (Bohnsack and Sutherland 1985;

Fabi and Fiorentini 1994; Claisse et al. 2014; Ajemian et al. 2015; Scott et al. 2015).

These benefits include: reducing fishing pressure on nearby natural reefs and mitigating localized rates of fishing-related mortality (Pickering and Whitmarsh 1997; Zalmon et al. 2002; Cresson et al. 2014); providing a food source and shelter in an area where no reef habitat had previously existed (Fabi and Fiorentini 1994; Claisse et al. 2014;

Ajemian et al. 2015; Scott et al. 2015); and increasing fish density and biomass in the 53 vicinity of the artificial reef that in turn leads to increases in catch yields (Bohnsack and

Sutherland 1985; Fabi and Fiorentini 1994; Carr and Hixon 1997; Santos and Monteiro

1998; Zalmon et al. 2002; Whitmarsh et al. 2008; Bortone et al. 2011; Leitão 2013).

Whether artificial reefs actually enhance the production of fish biomass, or simply attract and aggregate fish leading to an increased risk of overfishing, is an ongoing debate (Bohnsack and Sutherland 1985; Solonsky 1985; Bohnsack 1989; Carr and

Hixon 1997; Folpp et al. 2013; Smith et al. 2015).

In NSW, artificial reefs have been deployed in both estuarine and coastal marine areas for the primary purpose of recreational fisheries enhancement (Folpp et al. 2011; Folpp et al. 2013; Lowry and Folpp 2014; Lowry et al. 2014). These man-made structures are purposely built for providing bottom structure in selected areas of the marine environment thereby increasing the availability of fish for recreational anglers

(McGlennon and Branden 1994; Lowry et al. 2014). The development and implementation of artificial reefs in NSW is considered a high priority by the recreational fishing community but the contribution of these reefs to recreational fisheries and local production is not well understood. Information describing the harvest composition and harvest of recreational anglers at these artificial reefs is needed to address these knowledge gaps and to provide a realistic evidence-based context for modelling studies. Simulation of realistic recreational harvest scenarios (e.g. using

Ecopath with Ecosim (EwE); Christensen and Pauly 1992; Pauly et al. 2000) may provide insights regarding the production potential of the system and hence the cost- benefit of additional artificial reef deployments.

Globally, researchers needing to estimate the harvest taken by recreational fishers from various artificial reef fisheries face a common problem. They must select and implement an appropriate sampling program to directly measure the harvest or use existing datasets to infer the harvest from the artificial reef fishery. There are many sampling options available but these vary greatly in their ability to deliver unbiased information about the fishery and their relative costs of implementation (Table 3.1;

Pollock et al. 1994; Smallwood et al. 2011; Smallwood et al. 2012; Hartill and Edwards

2015). All sampling options, including probability-based survey methods, can be subject to multiple forms of bias, which can be difficult to detect and quantify (Hartill and Edwards 2015). The cheaper options (i.e. fisher logbooks, web-based data) provide information that has many known biases and are not representative of the fishery (Table

3.1; Pollock et al. 1994; Connelly and Brown 1996; Bray and Schramm 2001; Conron and Bridge 2004; Smallwood et al. 2011; Smallwood et al. 2012; Hartill and Edwards

2015).

Alternatively, the statistical rigour of a well-designed probability-based sample survey comes at a prohibitive cost because of the difficulty of selecting an unbiased sample of fishers that use the artificial reef fishery from the many thousands of private and public access points (i.e. on-site surveys) or from the massive urban population that reside in the Sydney area (i.e. off-site surveys - Table 3.1, Fig. 3.1). This sampling problem is the equivalent of looking for micro-needles in a mega-haystack. Hence it is necessary to consider the use of existing datasets to infer the harvest composition and harvest rates of recreational fishers using the artificial reef. Of course, the use of any existing datasets requires the adoption of various assumptions about the representativeness of the data used.

Access Smartphone Scientific fishing Telephone Mail Roving point Fisher Fishing Web-based (standardised survey (off- survey survey survey Sampling issues logbooks application data gear) site) (off-site) (on-site) (on-site)

Probability-based sample of No No No No Yes Yes Yes Yes target fishery

Data representative of target No No No Unknown Yes Yes Yes Yes fishery

Coverage of target No No No No Yes Yes Yes Yes population known

Contact rates with eligible Low Low Low N/A Low Low High High fishers

Diffuse access to fishery (multiple private and public N/A N/A N/A N/A N/A N/A Yes Yes access points)

Safety issues working at sea N/A N/A N/A Yes N/A N/A Yes No

Fish identification bias High High High Low High High Low Low

Table 3.1 (Cont.)

Non-response bias High High High N/A Medium Medium Low Low

Avidity bias High High High N/A Low Low Low Low

Recall bias High High High N/A Medium Medium Low Low

Reporting bias High High High N/A Medium Medium Low Low

Prestige bias High High High N/A Low Low Low Low

Rounding bias High High High Low Low Low Low Low

Easy to implement Yes Yes Yes No No No No No

Relative cost of Low Low Low Medium High High High High implementation

Pollock et al. (1994), Connelly and Brown (1996), Bray and Schramm (2001), Hartill and Edwards (2015), Sources Smallwood et al. (2011; 2012).

We present a case study from a nearshore artificial reef in Sydney, Australia that uses existing datasets to obtain: (a) a list of species taken by recreational fishers from the vicinity of the artificial reef; and (b) the harvest rates of those species. We estimate the recreational harvest of fish from the artificial reef fishery by number and weight. These harvest estimates are standardised per unit area to provide context for the relative size of the recreational harvest from the artificial reef fishery. We also use the standardised harvest data to make comparisons with other recreational fishery areas from which the fishable area is known.

3.3 Materials and methods

3.3.1 The Artificial Reef

The Sydney Artificial Reef (AR) is a large purpose-built individual artificial structure that was deployed in October 2011 for the purposes of enhancing recreational fishing. It is located approximately 1.2 km east of ‘The Gap’, at South Head, Sydney, New South

Wales, Australia (33°50.797'S, 151°17.988'E, Fig. 3.1) in 38m depth of water. The steel structure is 12 x 15 m and 12 m high with two 8 m tall pillars and is moored at each corner with chain and a 60 ton concrete block, resulting in a reef volume of 700 m3

(Champion et al. 2015; Scott et al. 2015). The reef was designed with many open void spaces and towers that allow water flows through the structure. This water flow is important for supplying nutrient and plankton to the AR ecosystem and to promote the growth of sessile organisms and resident fishes (Connell and Anderson 1999; Redman and Szedlmayer 2009).

Information describing the recreational fishery in the vicinity of the AR was obtained by: (a) direct observation of fishing trips using binoculars and from analyses of digital images that were used to quantify fishing effort at the reef (Keller et al. 2016) and (b) discussions with some recreational anglers that had visited the AR. The available information indicates that the fishery at the AR is mainly a drift and trolling fishery.

Anglers are wary of anchoring near the reef because of the risk of fouling their fishing gear and anchor.

This fishery mainly targets baitfish (e.g. yellowtail scad, Trachurus novaezelandiae and slimy mackerel, Scomber australasicus), inshore pelagic fish (e.g. yellowtail kingfish,

Seriola lalandi) and demersal species (e.g. snapper, Chrysophrys auratus; blue morwong, Nemadactylus douglasii and flathead, Platycephalus spp.) that occur around the edges of the reef.

3.3.2 Existing datasets used for inferring harvest composition and harvest rates at

the AR

A variety of existing datasets and information sources were used to estimate the harvest of fish by recreational anglers in the vicinity of the AR (Table 3.2). A series of probability-based surveys of coastal boat-based fishing in the Greater Sydney region were carried out during the two year period March 2007 to February 2009 inclusive

(Steffe and Murphy 2011). Stratified random sampling protocols were used. A total of

72 days were sampled at each of 8 survey sites within each survey year but we only used data from three sites closest to the AR (Fig. 3.1). These survey data account for variation in species catchability, and were used to compile a comprehensive list of species and their harvest rates for estimating the recreational harvest in the vicinity of the AR. The use of the entire species list to represent the taxa harvested from the AR fishery would have resulted in an overestimation of harvest because many taxa would not occur at the AR. Thus, we used all of the other available data sources (e.g. underwater camera observations, diver surveys, anecdotal reports, expert opinion; Table

3.2, Appendix B) to confirm the presence of fish species in the vicinity of the AR. This procedure enabled us to develop a list of taxa that were regarded as being an accurate representation of those taxa taken by anglers in the vicinity of the AR.

Figure 3.1. Location of the Sydney Artificial Reef (AR), the Old South Head Signal

Station (indicated as camera) and greater Sydney sites (Hawkesbury, Long Reef and

Port Hacking).

Table 3.2. Features of existing datasets and sources used to infer harvest composition and harvest rates for the fishery at the Artificial Reef.

Data Description Source Associated Harvest Harvest Relative composition rate cost Bias

Probability- Data from historic fishing surveys Steffe and Murphy (2011). Low Yes Yes Low based fishing providing coverage of the greater survey Sydney region. Requires assumptions about representativeness of the data.

Baited remote Data from a concurrent monitoring Smith unpub. data, Lowry and High Yes No Low underwater program to assess the pelagic and Folpp (2014), Scott et al. (2015). video (BRUV) benthic fish assemblage associated with and unbaited the AR. drop cameras

Diver surveys Visual surveys of fish abundance made Reef Life Suvey database High Yes No Low by SCUBA divers using standardised (www.reeflifesurvey.com); scientific methods from 2009 onwards. Edgar and Stuart-Smith, (2014).

Published and Reports from anglers, information from Fishraider website Medium Yes No Low anecdotal online fishing sites, and published (www.fishraider.com.au); Otway reports of research within the AR area. et al (1996), Folpp and Lowry species (2006), Dempster (2005), observed/taken Annese and Kingsford (2005), from fishery Lowry and Folpp (2014).

Expert opinion Opinion from scientists that have Pers. comm. DPI staff, UNSW Low Yes No Low worked in the region. staff.

3.3.3 Harvest rates, fishing effort and harvest estimation from the AR

We used harvest rate data from the available survey (Steffe and Murphy 2011) from the three survey sites closest to the AR, these being Hawkesbury and Long Reef located to the north, and Port Hacking located south of the AR (Fig. 3.1). These sites were also selected since they are the main access points in greater Sydney where fishers return to.

Interviews with fishing parties were of completed coastal marine fishing trips conducted at boat ramps in these locations. Thus the calculation of daily harvest rates at each site was done using the ‘ratio of means’ estimator (Pollock et al. 1994; Appendix C; Pollock et al. 1997). The amount of recreational boat traffic was assumed to be the same in each site on each sampled day. Therefore we allocated equal-weighting to each survey site and to each of the two survey years before calculating a weighted daily mean harvest rate for each species within each of 8 base level strata (i.e. two day-type strata: weekend days and public holidays, weekdays; within four austral seasons: autumn, winter, spring, summer).

The fishing effort in the vicinity of the AR (an area of 0.056 km2) was assessed from digital images taken by a shore-based camera (www.anso.com.au) securely fixed to a vantage point at the Old South Head Signal Station - a lighthouse (33°51'1.47"S,

151°17'12.41"E, Fig. 3.1; Keller et al. 2016). All types of vessels were included if they remained within the defined fishable area for 5 minutes or more and this time criterion allowed vessels in transit to be excluded from counts of fishing effort (Keller et al.

2016). Effort from the AR was estimated in units of boat hours. Data were analysed over the period from 1 June 2013 to 30 May 2014. Stratified random sampling protocols were used. Effort estimation was done for each base level stratum. Effort estimation was validated via independent observer counts and a correction factor was applied to 63 account for the under-estimation of fishing effort caused by weather-related visibility issues (Keller et al. 2016). Harvest was estimated for each species by multiplying weighted daily harvest rates (fish per boat hour, derived from Steffe and Murphy 2011) with estimates of fishing effort at the AR (boat hours, derived from Keller et al. 2016).

Harvest estimation was done for each base level stratum. Harvest was estimated initially in units of the number of fish taken by recreational anglers. We converted this harvest estimate into units of weight by multiplying the harvest estimate (number of fish) by the site-weighted mean weight for each species within each base level stratum (Appendix

C). The mean weights used were obtained from the historic survey data (Steffe and

Murphy 2011).

3.3.4 Standardised harvest comparisons

We standardised annual harvest in kilograms per unit of area to enable comparisons with other fishery areas and to provide context for the study. Annual harvest per square kilometre was calculated for all taxa at the AR in this study and seven estuarine fishery areas: Lake Macquarie (north estuary, south estuary and Swansea channel; Steffe et al.

2005b), Port Hacking estuary, Pittwater/Broken Bay estuary, Hawkesbury reach (Steffe and Murphy 2011) and Tuross estuary (Steffe et al. 2005a; Table 3.3). Survey areas were calculated in ArcGIS. The point estimates of harvest for the boat-based fishery at the AR were used to benchmark this fishery against these other recreational fisheries.

These standardised comparisons provide relative measures of recreational fish harvest across different fishery areas and years, with the assumption that the background pattern of recreational fishing (i.e. harvest composition and harvest numbers) was similar within the study area. Comparative data quantifying annual recreational harvest and the fishery area were only available for some estuarine fisheries (Table 3.3). 64

Table 3.3. Standardised harvest (kg/km2) and fishery details for the artificial reef (AR) and other fishery areas.

Latitude Standardised Area Fishery area Survey period Habitat type Source Longitude harvest (kg/km2) (km²)

Untreated steel 33°50.80'S Sydney AR Jun 2013-May 2014 12,504 0.06 This study designed reef 151°17.99'E

Mar 1999-Feb 2000; All estuarine 33°02.0'S 627 Northern Lake Macquarie 60.73 Steffe et al., 2005b Dec 2003-Nov 2004 habitats 151°37.0'E 657

Mar 1999-Feb 2000; All estuarine 33°06.0'S 287 Southern Lake Macquarie 43.10 Steffe et al., 2005b Dec 2003-Nov 2004 habitats 151°35.0’E 1,445

Mar 1999-Feb 2000; All estuarine 33°04.4’S 3,182 Swansea channel 3.23 Steffe et al., 2005b Dec 2003-Nov 2004 habitats 151°38.4'E 5,464

Mar 2007-Feb 2008; All estuarine 34°04.3’S 981 Steffe and Murphy, Port Hacking estuary 11.51 Mar 2008-Feb 2009 habitats 151°09.3’E 815 2011

Pittwater/Broken Bay Mar 2007-Feb 2008; All estuarine 33°33.0’S 501 Steffe and Murphy, 35.88 estuary Mar 2008-Feb 2009 habitats 151°20.2’E 426 2011

Mar 2007-Feb 2008; All estuarine 33°32.1’S 466 Steffe and Murphy, Hawkesbury reach 84.93 Mar 2008-Feb 2009 habitats 151°14.3’E 462 2011

Table 3.3 (Cont.)

Latitude Standardised Area Fishery area Survey period Habitat type Source Longitude harvest (kg/km2) (km²)

Mar 1999-Feb 2000; All estuarine 36°03.8'S 668 Tuross estuary 14.47 Steffe et al., 2005a Dec 2003- Nov 2004 habitats 150° 6.1'E 813

3.3.5 Fish functional groups and data analysis

Taxa were grouped into 10 fish functional groups to enable comparisons of annual harvest by number and weight at the AR, and standardised harvest (harvest divided by fishery area- kg/km2) among different recreational fishery areas (Table 3.4; Appendix

B). Taxa were assigned into functional groups based on their ecological role and maximum adult weight (State Pollution Control Commission 1981; Rowling et al.

2010).

Standardised harvest data were square root transformed and a non-metric Multi-

Dimensional Scaling (MDS) analysis was performed on the Bray-Curtis similarity matrix. An MDS analysis was calculated for each fishery area by survey year to compare the annual harvest of functional groups from the AR with the standardised annual harvest of functional groups from the seven estuarine fishery areas. A one-way

Similarity of Percentages (SIMPER) analysis was conducted using the standardised data to identify functional groups that typified fishery areas and those which contributed to the average Bray-Curtis dissimilarity between areas. A 90% cut off for low contributions was selected in the analysis. All multivariate analyses were conducted in

PRIMER V6.1.14 (Plymouth Marine Laboratories, UK).

Pittwater/ Port N. Lake S. Lake Swansea Broken Hawkesbury Tuross Functional group AR Hacking Macquarie Macquarie channel Bay reach estuary estuary estuary Large-medium pelagic fish* 5 4 3 4 6 8 4 2 Small pelagic fish* 3 2 2 1 6 6 3 0 Mid-water planktivores 2 1 1 0 1 1 1 0 Omnivores (L*) 5 5 5 5 4 5 3 2 Omnivores (M &G*) 0 5 5 5 3 2 2 5 Ambush predators* 7 6 6 5 7 10 10 4 Large demersal predators* 6 2 2 2 3 4 5 1 Medium demersal predators* 10 5 4 2 8 10 10 6 Small demersal predators* 3 1 2 2 4 2 2 0 Herbivores 0 1 1 1 2 1 1 1 Total taxa 41 32 31 27 44 49 41 21 *Large-medium pelagic fish >3 kg; Small pelagic fish <3 kg; Large demersal predators > 5.0 kg; Medium demersal predators 1.0 to

5.0 kg; Small demersal predators < 1.0 kg, L= Leatherjackets, M &G=Mullets and garfish, Ambush predators= Ambush/cryptic predators, usually non-schooling. 68

3.4 Results

We estimated that 1,016 (± 82 SE) fish by number and 700 (± 59 SE) kg fish by weight were recreationally harvested from the AR during June 2013- May 2014. The standardised total annual harvest by area at the AR was 12,504 kg per km2 (Table 3.3).

The standardised annual harvest at the AR was between 2.3 to 43.6 times larger than the standardised harvest values recorded from the other fishery areas (Table 3.3).

The six most important functional groups (Fig. 3.2) harvested by recreational fishers at the AR were ambush predators, (27% by number and 24% by weight), omnivores

(leatherjackets) (21% by number and 12% by weight), medium demersal predators

(15% by both number and weight), large demersal predators (9% by number and 20% by weight), large to medium pelagic fish (7% by number and 20% by weight), and small pelagic fish (13% by number and 6% by weight). In total these six functional groups accounted for 92% of the total annual harvest by number, and 95% of the total annual harvest by weight. Two functional groups provided minor contributions to the recreational harvest, these being mid-water planktivores (5% by number and 4% by weight) and small demersal predators (3% by number and 1% by weight). In total these two functional groups accounted for 8% of the total annual harvest by number, and 5% of the total annual harvest by weight. There were no herbivores or omnivores (mullets and garfish) recorded in the recreational harvest taken from the AR.

Figure 3.2. Total annual harvest of fish functional groups by number (± SE) and weight (kg ± SE) at the AR from June 2013- May 2014 ( L*=Leatherjackets, M&G*

=mullets and garfish; 41 taxa in total).

Two separate groups could be identified by the MDS analysis of standardised annual harvest between the 8 fishery areas using a 40% similarity criterion (stress= 0.08, Fig.

3.3). The AR fishery was grouped with the Swansea channel fishery area. All other estuarine fishery areas were grouped together and separated from the AR and Swansea channel fishery area (Fig. 3.3). These two groupings were consistent across different survey years. SIMPER analysis identified the functional groups that made the greatest contribution to the dissimilarity between the standardised harvest (kg/km2) from the AR and estuarine fishery areas (Table 3.5). Large demersal predators (15.5-19.8%), large- medium pelagic fish (12.8–18%), ambush predators (8.5-18%), omnivores

(leatherjackets) (9.8-15.1%), medium demersal predators (10.1-13.9%) and small pelagic fish (8.5- 11.5%) made the greatest contribution to the average dissimilarity and were discriminators between all fishery areas and survey years. The mid-water planktivores (7.1-9.4%) was also a prominent functional group in the analysis. Two other functional groups contributed to the average dissimilarity between the AR and the

Swansea channel fishery area in both survey years, these were herbivores (6.24-14.58%) and omnivores (mullets and garfish) (5.2 -10.4%, Table 3.5). In addition the small demersal predators functional group was a minor contributor to the average dissimilarity between the AR and both the Port Hacking and Tuross estuary fishery areas (4.43-

9.01%). These 10 functional groups together contributed to an average dissimilarity of

48.7- 81.0 % between the AR and estuarine fishery areas.

Figure 3.3. MDS (nMDS) of standardised annual fish harvest (kg) by site. Numbers indicate survey year, overlaid circles represent a similarity of 40% between groups.

Refer to Table 3.3 for details of fishery areas.

Lake Macquarie estuary Hawkesbury estuary North South Swansea Pittwater- broken Hawkesbury Functional group Port Hacking Tuross Lake Lake channel estuary bay reach estuary Survey year 1* 2* 1* 2* 1* 2* 1* 2* 1* 2* 1* 2* 1* 2*

Large-medium 15.9 15.6 18.0 12.8 17.1 13.9 16.0 16.2 15.3 16.2 17.6 17.7 16.1 16.9 pelagic fish

Small pelagic 10.7 10.5 10.5 11.5 11.3 10.3 - 8.5 9.9 9.3 10.6 10.2 9.8 10.5 fish

Mid-water 8.1 8.3 7.8 9.4 7.6 8.2 7.9 7.1 8.0 7.8 8.1 7.9 7.2 7.8 planktivores

Omnivores 12.6 14.4 12.2 15.1 9.8 13.8 12.9 14.0 13.7 13.3 13.3 13.9 11.7 14.1 (leatherjackets)

Omnivores - - - - 10.4 5.2 ------(mullets/garfish)

Table 3.5 (Cont.)

Lake Macquarie estuary Hawkesbury estuary Tuross North South Swansea Pittwater- Hawkesbury Functional group Port Hacking estuary Lake Lake channel estuary broken bay reach Survey year 1* 2* 1* 2* 1* 2* 1* 2* 1* 2* 1* 2* 1* 2*

Ambush predators 17.2 16.4 16.8 14.3 12.6 8.5 17.1 18.0 17.6 17.2 17.0 16.6 13.7 12.4

Large demersal 14.4 17.6 17.4 19.8 17.9 18.6 18.1 17.7 17.8 17.6 16.1 15.5 17.3 18.0 predators

Med demersal 12.6 11.8 14.2 13.9 - - 10.1 10.5 12.6 13.2 11.5 11.6 11.3 10.4 predators

Small demersal ------9.01 - - - - - 4.43 - predators

Herbivores - - - - 6.24 14.58 ------

Average dissimilarity 70.4 68.2 78.8 56.9 58.1 48.7 61.9 63.6 69.9 71.2 73.0 74.1 81.0 75.1 *Lake Macquarie survey year 1= Mar 1999-Feb 2000, 2=Dec 2003-Nov 2004 *Port Hacking estuary survey year 1= Mar 2007-Feb 2008, 2=Mar 2008-

Feb 2009 *Hawkesbury estuary survey year; 1= Mar 2007-Feb 2008, 2=Mar 2008-Feb 2009 *Tuross estuary survey year 1= Mar 1999-Feb 2000, 2=

Dec 2003- Nov 2004

3.5 Discussion

3.5.1 A pragmatic approach for harvest estimation

The unbiased estimation of recreational harvest from a nearshore AR fishery adjacent to a large urban population is difficult to achieve. There are many data collection and sampling options available but these vary in their relative cost and the quality of information that they can provide (Table 3.1). At one end of the spectrum lie methods such as fisher logbooks and web-based surveys that rely on self-reported data (Pollock et al. 1994; Connelly and Brown 1996; Bray and Schramm 2001; Conron and Bridge

2004). The use of these methods is appealing because they are relatively easy to implement and cheaper than probability-based alternatives but they cannot provide data that are unbiased and representative of the fishery. Probability-based sample surveys, such as offsite and onsite survey methods (Pollock et al. 1994; Smallwood et al. 2011;

Smallwood et al. 2012; Hartill and Edwards 2015), lie at the far end of the sampling spectrum. A well designed and implemented probability-based sample survey can provide the unbiased and representative data needed for estimation of harvest. However, the relatively high cost of these methods can be prohibitive, particularly when trying to assess harvest from small, discrete fisheries located near large population centres and cities.

We present a case study that: (a) considers the strengths and weaknesses of these various data collection and sampling options; (b) assesses the utility of existing datasets; and (c) shows how existing datasets can be used in conjunction with the direct monitoring of fishing effort to enable a cost-effective solution for harvest estimation at the AR fishery. We did not have the resources needed to specifically design and

75 implement a probability-based survey to estimate fishing effort and harvest at the AR fishery. Instead, we used camera images from a fixed vantage point overlooking the AR area to obtain estimates of fishing effort (Keller et al. 2016). The data derived from the digital images was validated by direct observation (Keller et al. 2016). We then integrated multiple datasets to obtain a list of taxa that are harvested by recreational fishers within the AR area. Then we used data from a previous probability-based survey to obtain estimates of harvest rates for these taxa. Harvest was estimated by multiplying fishing effort and harvest rates together. This approach provided a feasible and cost- effective solution for addressing the trade-offs between cost, available resources and the reliability of harvest estimation (i.e. biases, accuracy and precision).

The use of existing datasets from different time periods to represent current conditions for a fishery has some issues. We used many different datasets to compile a list of harvested taxa at the AR. Each method is known to have different potential biases

(Table 3.2). For example, a) baited remote underwater video (BRUV) observations are known to be biased towards taxa that are strongly attracted to bait (Wraith et al. 2013); and b) visual diver surveys may not detect the presence of cryptic and nocturnal taxa

(Jennings and Polunin 1995; Edgar et al. 2004). Also it is well known that recreational fisheries are subject to inter-annual differences in harvest composition and harvest rates due to changes in the availability of some taxa (Steffe et al. 1996; Steffe and Murphy

2011; Carter et al. 2015). This source of variability is partially addressed because we use seasonal averages from two consecutive years of sampling from a well replicated, probability-based fishery survey to estimate the harvest rates. A final example of methodological bias is evident in the estimation of harvest for small pelagic baitfish (i.e. slimy mackerel and yellowtail scad). The harvest of these baitfish species from the AR

76 will be underestimated when using harvest rates derived from landed catches at boat ramps (i.e. a probability-based survey done at boat ramps) because the data do not include those fish that were used for bait during those trips (Lowry et al. 2006).

The use of probability-based fishing surveys requires a number of assumptions regarding the representativeness of the data. This includes assuming that fishing behaviour at public access points is the same as that at private access points. Hence onsite surveys (i.e. roving or access point surveys) are best suited to fisheries with few access points that are located away from large population centres. As in the case of the

AR fishery in this study, it is difficult to avoid potential biases in future surveys due to the proximity of this reef to the large Sydney population. Thus the use of multiple datasets to infer the current harvest composition at the AR will be preferable to relying on any single method alone.

We have used a pragmatic approach for estimating the harvest from the AR that provides a balanced way of addressing issues related to data quality, potential biases and cost. We make a number of assumptions regarding the representativeness of the harvest composition and harvest rate data used (see discussion below). We contend that our approach is preferable to the alternatives of: (a) using cheaper self-reported data collection strategies that are known to be biased, unrepresentative of the target population and require the acceptance of many assumptions regarding the quality of the data that are untenable; and (b) attempting to implement complex probability-based survey methods for which we do not have adequate funding and thus cannot be done appropriately. We urge readers to treat our reported harvest values as indicative of the

AR fishery and the estimates of precision as being approximate values.

3.5.2 Fish harvest at the AR

We estimated that about 1,016 fish and 700 kg by weight of annual harvest in total and

12,504 kg per km2 of standardised harvest were derived from the AR. The harvest estimates from the AR appear to be relatively small when considered in absolute terms without regard for the size of the area being fished. However, this perception changes when harvest is standardised per unit area. The estimated harvest per unit area at the AR was 2.3 to 43.6 times greater than comparative values from other fishery areas (Table

3.3). This indicates that the AR is supporting a viable recreational fishery that attracts recreational anglers thereby resulting in the concentration of fishing effort in a small area (Keller et al. 2016). Previous studies have also shown that catch rates of species were higher at ARs compared to natural reefs (Fabi and Fiorentini 1994; McGlennon and Branden 1994; Santos and Monteiro 1998; Whitmarsh et al. 2008; Bortone et al.

2011; Leitão 2013). ARs are believed to sustain both fish populations and fisheries in their immediate vicinity (Bohnsack and Sutherland 1985; Fabi and Fiorentini 1994;

Claisse et al. 2014; Ajemian et al. 2015; Scott et al. 2015). Therefore, it is likely that the calculated harvest estimates from the AR in this study are also conservative, because they are based on unenhanced boat catch rates prior to the implementation of the AR.

Fishing is known to be a selective process (Jennings and Kaiser 1998; Hsieh et al.

2006). Recreational fishing methods commonly used at the AR (e.g. trolling and drift fishing) are aimed mainly at predatory taxa (Jennings and Kaiser 1998). Thus, it is not surprising that functional groups containing leatherjackets, ambush predators, demersal predators and pelagic fish contributed greatly to the recreational harvest at the AR. Of the top ten most commonly captured species from the AR, the eastern bluespotted flathead (Platycephalus caeruleopunctatus), an ambush predator, was estimated to have 78 the highest recreational harvest, both by number and weight (Appendix D). Catches of ecologically similar species have also been reported at other ARs in the world (e.g. Fabi and Fiorentini 1994; McGlennon and Branden 1994; Santos and Monteiro 1997; Santos and Monteiro 1998; Zalmon et al. 2002; Santos and Monteiro 2007). Herbivores (e.g.

Luderick) and omnivores (e.g. mullets and garfish) were not represented in the harvest at the AR.

The comparisons of annual standardised harvest per unit area separated the fishery areas into two distinct groups (Fig. 3.3). The AR fishery area was most similar to the Swansea channel fishery area. The Swansea channel is a shallow and productive area that connects the coastal lagoon (Lake Macquarie) to the ocean (Table 3.3; Steffe et al.

2005b). This channel area has a predominately marine environment that is similar to the

AR. All other estuarine fishery areas had much lower levels of harvest per unit area and were grouped together. The SIMPER analysis revealed that the ambush predators functional group was most responsible for the similarity between the AR and Swansea channel fishery areas. In contrast, both the large demersal predators and large-medium pelagic fish were responsible for the differences between the AR and other fishery areas, particularly the Tuross estuary which was most dissimilar to the AR. The differences between the AR and other areas that we measured may be overstated because it is known that fishing effort is not spatially homogeneous (Keller et al. 2016).

Furthermore catch rates were unavailable from the AR and were instead taken from three coastal sites from the greater Sydney area to estimate harvest from this reef (Fig.

3.1). A better understanding of fine-scale fishing intensity within the larger estuarine fishery areas is needed to provide further insights into this issue. Even so, the fishery area comparisons do highlight the potential for recreational fishers to exploit localized

79 fish assemblages. The estimates of recreational harvest at the AR are vital for providing a better understanding of the production potential of these structures. This can be achieved by incorporating harvest information into simulation models (e.g. Ecopath with Ecosim) and may inform the current debate on the cost-effectiveness of additional

AR deployments being planned for eastern Australia.

3.6 Conclusions

The pragmatic approach used to estimate harvest at the AR provides a solution to a difficult sampling problem. Harvest at the AR by number and by weight was relatively small, but when considered per unit area, harvest was very high. Functional groups consisting of ambush predators, leatherjackets, pelagic fish and demersal predators were found to contribute the most to the recreational harvest at the AR. Comparisons of standardised harvest between the AR and other fishery areas revealed two distinct groups, where the AR was most similar to Swansea channel and the other fishery areas were grouped together. The ability to assess recreational harvest and the potential for localised production are important for determining the cost-effectiveness of future AR deployments. We recommend that future studies attempting to estimate harvest at AR fisheries consider an integrated methodology that combines existing datasets and cost- effective sampling designs.

Publication details:

Keller, K., Steffe, A.S., Lowry, M., Murphy, J.J., Smith, J. and Suthers, I.M. (2017)

Estimating the recreational harvest of fish from a nearshore artificial reef using a pragmatic approach. Fisheries Research. 187: 158-167.

A. Boomer

Chapter 4

A ray, a fish and a shark: the diel behaviour and activity of three co-occurring benthic species around temperate rocky reefs

4.1 Abstract

(Platycephalus caeruleopunctatus), and Port Jackson shark (Heterodontus portusjacksoni). Fiddler rays were on average 28% more active during the day, in contrast to bluespotted flatheads and Port Jackson sharks which were on average 28% and 84% more active at night respectively. Activity was not associated with fish length, tidal height or moon illumination, but increased with temperature in fiddler rays and bluespotted flatheads. The increase in activity is most likely to be associated with foraging behaviour, and a similarity in diet suggests that observed differences in movement behaviour may be due to resource partitioning between these species.

Different diel activity patterns have implications for conservation and fisheries management, whereby nocturnal species such as the bluespotted flathead and Port

Jackson shark are missed during visual surveys. The findings from this study improve the understanding of the ecology of co-existing benthic predators around coastal temperate reefs.

4.2 Introduction

Understanding behavioural patterns and habitat use of fish are important for sustainable fisheries management and conservation because these characteristics provide evidence for developing management zones such as sanctuary areas (Espinoza et al. 2015c).

Measuring activity in particular enhances our understanding of individual-level patterns in habitat use (Schlaff et al. 2014), which is essential information for species management. Activity varies due to differences in individual behaviour, season, and structural characteristics of the habitat (Santos et al. 2002; Schlaff et al. 2014). Another factor contributing to the variation in reef fish assemblages is the diel cycle in activity

(Santos et al. 2002). Diel activity and movements can bias visual surveys of assemblage composition, or influence a species’ vulnerability to directed harvest, but can also aid management and conservation programs, as well as stock assessments (Stoner 2004;

Parsley et al. 2008; Stehfest et al. 2015). For example, species which are more active during daylight periods are more likely to be observed during fishery surveys so are included in fisheries management plans, compared to cryptic nocturnal species which hide during the day.

Activity in marine animals is influenced by various factors, such as photoperiod, temperature, tide, (Stoner 2004; Brownscombe et al. 2014; Schlaff et al. 2014), dissolved oxygen, lunar phase, barometric pressure, (Stoner 2004; Schlaff et al. 2014), salinity (Stoner 2004; Schlaff et al. 2014; Gannon et al. 2015), and pH (Schlaff et al.

2014). Combinations of these factors have been shown to influence fish assemblages and movements in various environments including estuaries, coral reefs and seagrass habitats (Vianna et al. 2013; Henderson et al. 2014; Gannon et al. 2015; Kruse et al.

2015). Photoperiod is known to influence activity as it is linked to feeding, predator 83 avoidance, aggregation and reproduction in many fish species (Stoner 2004). This is because the level of light affects vision and contrast which for some species drives increased activity levels during the day (Stoner 2004). Changes in the lunar phase, tidal cycle and temperature can also trigger fish behaviour such as spawning and migration

(Claydon et al. 2014). The role of temperature is particularly evident in temperate regions which experience strong seasonal variation (Vianna et al. 2013; Espinoza et al.

2015b; Stocks et al. 2015). Temperature not only affects fish activity patterns, but also metabolic processes and feeding motivation (Stoner 2004). As marine species face an array of anthropogenic threats especially in coastal ecosystems with a growing urban population, understanding how species respond to changes in their environment is becoming increasingly important (Schlaff et al. 2014).

Studies of the distribution and abundance of fish in coastal environments of Australia are typically conducted through visual surveys, such as drop cameras, baited remote underwater video surveys (BRUVS) and diver surveys (e.g. Edgar and Stuart-Smith

2014; Champion et al. 2015; Scott et al. 2015). However, these methods are poorly suited for measuring activity and behaviour, which instead requires acoustic telemetry

(Hussey et al. 2015). Acoustic telemetry has become an important tool in studying the movements and behaviour of aquatic animals over a variety of spatial and temporal scales (Hussey et al. 2015). A relatively recent advance in this technology includes the incorporation of sensors such as triaxial accelerometers which measure body activity to directly relate behaviour and physiology to stimuli in the environment (Hussey et al.

2015). This has improved the understanding of the relationship between activity and environmental variables in many marine species such as the giant cuttlefish (Sepia apama) (Payne et al. 2011), dusky flathead (Platycephalus fuscus) (Gannon et al.

2014), sandy flathead (Platycephalus bassensis) (Stehfest et al. 2015), as well as luderick (Girella tricuspidata), mulloway (Argyrosomus japonicas), and sand whiting

(Sillago ciliata) (Payne et al. 2016). The use of accelerometers has also been used to determine the diel activity patterns in species including whitetip reef sharks

(Triaenodon obesus) (Whitney et al. 2007), and bonefish (Albula vulpes)

(Brownscombe et al. 2014). Without determining diel behaviour through acoustic telemetry studies, assessment of the ecological role of these species is influenced by visual surveys only.

In this study, passive acoustic telemetry was used to quantify the activity and behaviour of three common benthic species: the eastern fiddler ray (Trygonorrhina fasciata), eastern bluespotted flathead (Platycephalus caeruleopunctatus), and Port Jackson shark

(Heterodontus portusjacksoni). All three species are endemic to eastern Australia and inhabit soft substrate habitats (Hutchins and Swainston 1986; Last and Stevens 1994;

Moore et al. 2009). The Port Jackson shark also inhabit rocky reef environments

(Powter and Gladstone 2008b), and all three species have been observed in surrounding rocky reefs from BRUVS (Lowry and Folpp 2014) and diver surveys

(www.reeflifesurvey.com; Edgar and Stuart-Smith 2014). The distribution of the eastern fiddler ray extends from southern Queensland to southern New South Wales in depths up to 100 m (Last and Stevens 1994). The fiddler ray feeds on crustaceans, fish, polychaetes and molluscs (Marshall et al. 2007; Izzo and Gillanders 2008). Little is known about the life history and ecology of the fiddler ray, except that it is similar to the southern fiddler ray (T. dumerilii; Marshall et al. 2007; Izzo and Gillanders 2008).

Similarly, few studies have examined the ecology and life history of the bluespotted flathead, which is an important commercial and recreational species with an estimated

85 annual harvest between 320 and 450 tonnes (Henry and Lyle 2003; Rowling et al. 2010;

Steffe and Murphy 2011). This species occurs from southern Queensland to eastern

Victoria in depths up to 100 m, and grows to 90 cm (Hutchins and Swainston 1986;

Kuiter 2000; Rowling et al. 2010). Similar to the fiddler ray, the diet of the bluespotted flathead is also composed of crustaceans, fish, polychaetes and molluscs (Coleman and

Mobley 1984; Hutchins and Swainston 1986; Moore et al. 2009; Barnes et al. 2011).

The distribution of the Port Jackson shark extends from southern Queensland to

Tasmania and west to the central coast of Western Australia to depths of 275 m (Last and Stevens 1994). It has a broader diet than the fiddler ray and bluespotted flathead, consisting of fish, echinoderms and decapod crustaceans (Powter et al. 2010). Nocturnal activity has previously been observed in Port Jackson sharks (McLaughlin and O'Gower

1971; O'Gower 1995; Powter and Gladstone 2009; Powter et al. 2010). Furthermore, conventional tagging studies have showed that individuals of this species are capable of large-scale movements associated with the breeding season (McLaughlin and O'Gower

1971; O'Gower 1995).

The effect of environmental variables and the diel cycle on the activity of the fiddler ray, bluespotted flathead and Port Jackson shark are not well understood. This lack of information means that these species could be overlooked in the planning of marine park zones and fisheries management plans. A better understanding of activity and how it relates to environmental conditions will help clarify the ecological function of these three abundant benthic species. The specific aims of this study were to: 1) determine and compare the diel pattern of activity in these species, and 2) assess the effect of temperature, luminosity and tide on their activity.

4.3 Material and methods

4.3.1 Study area

Individuals from all three species were captured and tagged at two main reefs, the

Sydney artificial reef (AR, 33°50.797'S, 151°17.988'E) and a nearby natural reef (NR),

Dunbar reef (33°51.10’73"S, 151°17.19’36"E), located ~2 km south-east of South

Head, Sydney, Australia (Fig. 4.1). The steel structure is 12 x 15 m x 12 m high including two 8 m tall pillars, at a depth of 38 m. It is moored at each corner with chain and a 60 ton concrete block. The substrate surrounding the AR is mostly flat and sandy, whereas the NR is a larger outcrop of subtidal reef in 25-30 m depth, ~600-800 m from the AR.

A VR2W receiver (Vemco Ltd, Nova Scotia, Canada) was deployed on the AR from

2011 to 2013 and was replaced with a VR4 receiver (Vemco Ltd, Nova Scotia, Canada) from 2013 onwards, providing remote uploading capability. These receivers were tethered to a cross-beam on the artificial structure approximately 8 m from the seafloor.

Two VR2W receivers were also deployed near the NR, one north (33°50'47.76"S,

151°17'27.60"E; Fig. 4.1) and another south of the reef (33°51'4.32"S, 151°17'27.60"E;

Fig. 4.1) at approximately 25 m depth. All receivers were coated with a copper-based antifouling paint to prevent possible signal occlusion due to biofouling (Heupel et al.

2008).

Receivers were downloaded every 3 to 6 months over a period of 2 years. Detections from the greater Sydney area (e.g. Sydney Harbour, Bronte-Coogee Aquatic reserve,

Bondi; refer to Fig. 5.1, Chapter 5) were downloaded from the Integrated Marine

Observing System-Australian Animal Tagging and Monitoring System (IMOS-

AATAMS) fisheries acoustic telemetry array, which is publically available online

(https://aatams.emii.org.au/aatams).

Figure 4.1. Location of receivers in greater Sydney. AR= Artificial reef, NR= Dunbar reef receivers (N: north, S: south) and ORS= Ocean Reference station.

4.3.2 Acoustic tagging

A number of recreationally important species were initially planned to be incorporated in this study (e.g. blue morwong, Nemadactylus douglasii; silver trevally, Pseudocaranx dentex; pink snapper, Chrysophrys auratus and ocean leatherjacket, Nelusetta ayraud).

However the effects of barotrauma (e.g. Hughes and Stewart 2013; Curtis et al. 2015) and difficulty of capture prevented the tagging and subsequent monitoring of these species. The fiddler ray, bluespotted flathead and Port Jackson shark however were previously identified as abundant from routine Baited Remote Underwater Video

(BRUV) and drop camera surveys at the AR and NR (Lowry and Folpp 2014; Scott et al. 2015). Therefore these benthic species were selected for this study due to the ease of capture and abundance at these reefs. Six fiddler rays, 13 bluespotted flatheads and 9

Port Jackson sharks were caught at the AR, and 1 fiddler ray, 1 bluespotted flathead, and 7 Port Jackson sharks were caught from the NR (Table 4.1). All animals were captured using various fishing methods, including rod and reel, commercial fishing trap, or modified long-line. Individuals collected by rod and reel were captured with circle- style hooks on monofilament line, baited with pilchard or squid. Modified long-lines comprised of a 10 m bottom set 8 mm mainline rope weighted on each end (3 kg), with one end attached to 20-30 m (depending on depth) 8 mm nylon float line tied to a surface buoy. Each long-line had 3 gangions which consisted of 1 m of monofilament line and a circle hook (6/0) baited with pilchard or squid, attached 2 m apart with a shark clip. Two long-lines were usually deployed simultaneously, depending on weather conditions and soaked for approximately 30 min.

Upon capture, individuals were placed into a 20-30 L tub containing fresh seawater that was continuously aerated, and tagging was conducted either on the boat at the site of 90 capture, or transported to a nearby private jetty. Prior to surgery, captured animals were anaesthetised with 0.5 ml Aqui-S (AQUI-S New Zealand Ltd) per kg of fish, except for

Port Jackson sharks which were inverted to induce tonic immobility and operated on without anaesthetic (Henningsen 1994; Holland et al. 1999; Wells et al. 2005). All captured individuals were measured to the nearest cm (total length – TL), sexed, and surgically implanted with a 69 kHz Vemco acoustic accelerometer transmitter (Vemco

Ltd, NovaScotia, Canada) into the peritoneal cavity. Each transmitter records dynamic body acceleration (ms-2), which is calculated as a root mean square value from tri-axial acceleration, these data are transmitted to the receiver after mean delay periods

(specified in Table 4.1). The type of transmitter used (V9A-2H, V9A-2L, or V9AP-2L) depended on animal size and availability (Table 4.1).

Fish ID TL Approx. tag Min Max Power Location Species Sex Tag type Date tagged code (mm) life (d) delay (s) delay (s) output tagged

fiddler ray B2 840 M 61 190 290 V9A-2H H 28/06/2013 AR

fiddler ray B3 750 M 61 190 290 V9A-2H H 28/06/2013 AR

fiddler ray B4 790 M 61 190 290 V9A-2H H 28/06/2013 AR

fiddler ray B5 764 M 156 25 25 V9A-2L L 25/07/2013 AR

fiddler ray B8 760 M 602 220 500 V9AP-2L L 27/08/2013 AR

fiddler ray B9 820 M 602 220 500 V9AP-2L L 24/09/2013 AR

fiddler ray B10 820 M 602 220 500 V9AP-2L L 25/02/2014 NR

bluespotted F1 415 I 156 25 25 V9A-2x L 14/08/2013 NR flathead

bluespotted F3 500 I 61 190 290 V9A-2x H 14/08/2013 AR flathead

Table 4.1 (Cont.)

Fish ID TL Approx. tag Min Max Power Location Species Sex Tag type Date tagged code (mm) life (d) delay (s) delay (s) output tagged

bluespotted F4 365 I 61 190 290 V9A-2x H 14/08/2013 AR flathead bluespotted F6 410 I 602 220 500 V9AP-2L L 22/08/2013 AR flathead bluespotted F7 387 I 602 220 500 V9AP-2L L 27/08/2013 AR flathead bluespotted F8 450 I 602 220 500 V9AP-2L L 27/08/2013 AR flathead bluespotted F9 520 I 602 220 500 V9AP-2L L 27/08/2013 AR flathead bluespotted F10 550 I 602 220 500 V9AP-2L L 27/08/2013 AR flathead bluespotted F11 420 I 602 220 500 V9AP-2L L 24/09/2013 AR flathead bluespotted F13 425 I 602 220 500 V9AP-2L L 31/10/2013 AR flathead

Table 4.1 (Cont.) Min Fish ID TL Approx. tag Max Power Location Species Sex delay Tag type Date tagged code (mm) life (d) delay (s) output tagged (s)

bluespotted F14 400 I 602 220 500 V9AP-2L L 31/10/2013 AR flathead bluespotted F15 490 I 602 220 500 V9AP-2L L 1/05/2014 AR flathead bluespotted F20 300 I 602 220 500 V9AP-2L L 12/09/2014 AR flathead bluespotted F23 520 I 602 220 500 V9AP-2L L 12/09/2014 AR flathead

Port Jackson PJ1 930 M 602 220 500 V9AP-2L L 22/08/2013 NR shark

Port Jackson PJ2 950 M 602 220 500 V9AP-2L L 22/08/2013 NR shark

Port Jackson PJ3 950 M 602 220 500 V9AP-2L L 22/08/2013 NR shark

Port Jackson PJ4 950 M 602 220 500 V9AP-2L L 22/08/2013 AR shark

Table 4.1 (Cont.)

Fish ID TL Approx. tag Min Max Power Location Species Sex Tag type Date tagged code (mm) life (d) delay (s) delay (s) output tagged

Port Jackson PJ5 873 M 602 220 500 V9AP-2L L 27/08/2013 AR shark

Port Jackson PJ6 640 M 602 220 500 V9AP-2L L 27/08/2013 AR shark

Port Jackson PJ7 900 M 602 220 500 V9AP-2L L 27/08/2013 NR shark

Port Jackson PJ8 660 M 602 220 500 V9AP-2L L 27/08/2013 AR shark

Port Jackson PJ9 820 M 602 220 500 V9AP-2L L 27/08/2013 AR shark

Port Jackson PJ10 910 M 602 220 500 V9AP-2L L 27/08/2013 AR shark

Port Jackson PJ11 900 M 602 220 500 V9AP-2L L 6/09/2013 NR shark

Port Jackson PJ12 985 M 602 220 500 V9AP-2L L 6/09/2013 NR shark

Table 4.1 (Cont.)

Fish ID TL Approx. tag Min Max Power Location Species Sex Tag type Date tagged code (mm) life (d) delay (s) delay (s) output tagged

Port Jackson PJ13 880 M 602 220 500 V9AP-2L L 24/09/2013 NR shark

Port Jackson PJ14 1150 F 602 220 500 V9AP-2L L 24/09/2013 AR shark

Port Jackson PJ15 960 M 602 220 500 V9AP-2L L 24/09/2013 AR shark

Port Jackson PJ16 660 M 602 220 500 V9AP-2L L 24/09/2013 AR shark

The approximately 3 cm-long incision was sutured with one or two stitches using non- absorbable polyamide monofilament with a curved cutting needle. Each individual was injected with Engemycin 100 (50 mg kg-1 body weight) to prevent infection and allow fast recovery. All flatheads were individually tagged with externally visible T-bar anchor tags (Hallprint, Hindmarsh Valley, SA, Australia, http://www.hallprint.com/) in the dorsal fin musculature which were marked with contact details in case of recapture by fishermen. Sterile surgical methods and Betadine were used throughout the procedure.

After surgery, individuals were returned to the tub containing fresh seawater and transported by boat back to the site of capture for release. Fish were released only after full recovery and normal signs of activity were observed. In order to avoid mortality after release, or monitoring of biased movement patterns due to injuries caused during the tag implantation, individuals which did not recover fully within 20 min were euthanized and removed from analysis. Fiddler rays and Port Jackson sharks were tagged between June 2013 and February 2014, and bluespotted flatheads were tagged between August 2013 and September 2014 (Table 4.1). All surgical procedures were conducted following protocols approved by University of New South Wales Animal

Care and Ethics Committee (Approval Number 12/111A).

Environmental variables are known to affect the probability of a transmitter being detected by the acoustic receivers and thus can influence the inference of animal behaviour (Payne et al. 2010; Gjelland and Hedger 2013). To determine the variability in tag detection with distance from receiver, two stationary V9-2L transmitters (with a random delay of 170-310 s, power output: low; battery life ~365 days; hereafter referred

97 to as “control tags”) were deployed at two fixed distances from the AR (50 m north of the AR and the other was placed 200 m south of the AR) on 1st December 2013, to determine the level of environmental and/or biological noise affecting the detectability of the animal tags. Each tag was placed in a mesh cotton bag and covered with thick plastic mesh, then attached to a 2 m rope at 1 m off the bottom substrate. The rope was anchored by a weight with the top attached to a float.

4.3.3 Environmental data

The effect of water temperature, tidal height and lunar phase on activity of the three benthic species was assessed. The VR4 receiver logged water temperature data (°C) once daily at the AR from December 2013 onwards. For detections at receivers outside the AR and NR, temperature from the Sydney Water ORS station (Ocean Reference

Station, 33°53'52.80"S, 151°18'54.00"E; Fig. 4.1) at 36 m depth was used and a HOBO temperature logger was attached to the North Head receiver at 25 m depth

(33°49'32.16"S, 151°18'0.00"E). A temperature logger was also attached to the south

NR receiver (33°51'4.32"S, 151°17'27.60"E) at 25 m depth and two temperature loggers were deployed near the AR, one north 530 m from the AR at approximately 35 m depth

(33°50'28.08"S, 151°18'3.00"E) and another 500 m south of the AR (33°51'1.44"S,

151°18'2.16"E) at 45m depth. Each logger deployed near the AR was secured to a 2.5 m rope at 1.4 m from the seafloor. One end was attached to a weight and an eight inch float was attached at the top of the rope. Temperature from the ORS station was downloaded from August 2013 to June 2015 and data from the loggers were downloaded from 10th June to 19th October 2013. Data were averaged daily where necessary.

Tidal level (in metres) was calculated from daily tide recordings made at Fort Denison in Sydney Harbour (33°49’31.56"S, 151°15’30.72"E). Tide data from June 2013 to June

2015 were provided courtesy of Manly Hydraulics Laboratory, Office of Environment and Heritage NSW. Luminosity was calculated according to percentage of illumination, where a new moon had 0% illumination and a full moon had 100% illumination. Data were obtained from the United States Naval Observatory Astronomical Applications

Department (USA Astronomical Application Department website. Available: http://aa.usno.navy.mil/data/docs/MoonFraction.php, accessed 14th July 2015).

4.3.4 Data analysis

Data spanned over approximately 2 years, where receivers were downloaded from the

AR (June 2013 to June 2015) and NR (June 2013 to July 2014). Detections within the first 24 hours after release were excluded to reduce the monitoring of behaviour influenced by the tagging procedure. Data were filtered to remove potential spurious detections; single transmitter detections were considered false detections and removed from the analyses (Reubens et al. 2013). Activity (ms-2) was adjusted from raw values for each transmitter detection, where the minimum activity value was subtracted from each activity value per detection, so that the minimum activity was zero (ie. no activity,

Payne et al. 2016). Data were converted to Eastern Standard Time (EST). Activity for all tagged fish and across the entire sampling period was binned (averaged) in 1 hour intervals and was calculated for each hour in the 24 hour day period. Activity data were also averaged for every 0.1°C increment of water temperature. Diel periods were analysed as hourly periods with the day period (06:00 to 18:00) and the night period

(18:00 to 06:00). 99

All fiddler rays tagged in this study were classified as mature males, with TL greater than 740mm (Izzo and Gillanders 2008). Tagged fiddler rays measured 750-840 mm so only TL was used in analysis, no females or juveniles were caught. Similarly, only 3 male Port Jackson sharks which were tagged were classified as maturing subadults, with a TL between 500 and 750 mm (Powter and Gladstone 2008c; Powter et al. 2010). 12 sexually mature males and one sexually mature female were tagged (males ≥750 mm

TL; females ≥900 mm TL; Powter and Gladstone 2008c; Powter et al. 2010), no juveniles were caught or tagged. All tagged flatheads measured 250 mm or more in total length and were also classified as mature (Rowling et al. 2010). Sex was unable to be determined in this species.

A generalized additive mixed-effects model (GAMM) was used to examine the effect of time of day (Hour; 0-23), total length (TL; mm), moon illumination (Luminosity; %), tide height (Tide; m), and water temperature (Temp °C) on activity. Fish identity

(FishID) was included as a random effect to account for any dependency of observations within individuals. Austral season, calendar month and day were not included as variables as they were correlated with temperature. Only one adult female

Port Jackson shark was tagged and no female fiddler rays were caught and tagged, so no sex interaction was tested. Similarly, maturity stage was not tested due to the lack of juveniles and/or subadults, and due to the correlation with TL. Site was not included in the model due to the low numbers of individuals tagged at the NR compared to the AR.

Additional environmental variables, such as rain and wind speed were also not included in the model to avoid over-parameterisation given the limited number of animals

100 tagged. Therefore the full model available for model selection for fiddler rays, bluespotted flatheads and Port Jackson sharks was (eqn 1):

Activity= s(Hour) + Temp+ TL + Tide + Luminosity+ FishIDrandom

where s represents a cyclic penalised thin plate regression spline. GAMMs used a basis dimension k = 6 to avoid unrealistic smoothers. Splines for Tide, Luminosity and Temp were also investigated but these did not improve model fit so were not included in the final model.

A model selection process was used to find the most parsimonious combination of these parameters, based on corrected Akaike information criterion (AICc). The differences in

AICc (Δ AICc) between models in a set were calculated to compare the relative weights of these models, and to identify the probability that each model is the best model in the set (Anderson 2008). Average activity in relation to these variables was analysed for all individuals within species. GAMMs were run using the ‘gamm4’ package (Wood and

Scheipl 2015), and model selection was done using ‘MuMIn’ package (Barton 2015) in

R (R Core Team, 2015).

Generalised linear modelling with Poisson family was used to test the effect of environmental factors on the detection frequency (detections per hour) of the two control tags. A month of hourly detection data from the 4th of December 2013 was used, which gave approximately 720 hourly detections. The factors included in the

GLM were hour of the day (0-23, fixed), wave height (m, continuous), water temperature (°C, continuous), and rainfall (mm). Wave height was obtained from the

101

Office of Environment and Heritage’s offshore Sydney Waverider buoy (33°45'56”S,

151°24'39”E), rainfall was obtained from the BOM Dover Heights weather station

(33°52'12"S, 151°16'48"E; BOM 2015), and water temperature was collected from the

VR4 receiver at the AR.

102

4.4 Results

4.4.1 Diel activity

All three species exhibited diel activity patterns throughout the study period. Each monitored individual displayed various levels of activity, indicating that mortality did not occur during the study period. On average fiddler rays were 28% more active during the day, compared to Port Jackson sharks which were on average 84% more active at night (Figs. 4.2 and 4.3). Average activity was lowest in the bluespotted flathead, but was on average 28% higher at night (Figs 4.2 and 4.3). The weak diel activity patterns observed in the bluespotted flathead were due to the large amount of variation in activity across all hourly periods (Fig. 4.3b). There was a negative correlation in activity between the fiddler ray and the bluespotted flathead, and the fiddler ray and Port

Jackson shark (Figs. 4.4a and b). In contrast there was a positive correlation in activity between the bluespotted flathead and the Port Jackson shark (Fig. 4.4c).

103

Figure 4.2. Average temporal activity (ms-2± SE) of fiddler ray (Trygonorrhina fasciata, n=7), bluespotted flathead (Platycephalus caeruleopunctatus, n=14) and Port

Jackson shark (Heterodontus portusjacksoni, n=7) from June 2013 to June 2015. The narrow standard error bars are based on the high number of detections per hour per individual.

104

Figure 4.3. Contribution of the spline for hour to the fitted GAMM models of Activity

(see Table 4.3), for a) fiddler rays b) bluespotted flatheads and c) Port Jackson sharks.

Dotted lines represent 95% confidence intervals.

105

Figure 4.4. Average hourly activity (ms-2) between a) the bluespotted flathead and fiddler ray, b) the Port Jackson shark and fiddler ray and c) the bluespotted flathead and

Port Jackson shark during the monitoring period.

106

4.4.2 Environmental variables

Model selection of the GAMMs showed that the best model for fiddler ray and bluespotted flathead activity was one including the hour spline and temperature (Table

4.2), both of which were significant (Table 4.3). Temperature had little effect on Port

Jackson shark activity and the best model was a spline of hour (Fig. 4.5c; Table 4.2).

The model weights for each best model were approaching 1, so model averaging was considered unnecessary. Average activity was found to increase with increasing temperature for both fiddler rays and bluespotted flatheads (Figs. 4.5a and b; Table 4.3).

There was no effect of time of day (Hour), Water Temperature, Rainfall, or Wave

Height on the detection frequency of the 50 m control tag (Table 4.4), but there was an effect of Hour (P = 0.007) and Wave Height (P = 0.002) on the detection frequency of the 200 m tag (Table 4.5). The detection frequency was significantly lower from 7-9 am

(Fig. 4.6), and was also reduced with increasing wave height. Model results reveal that the detection frequency of the 200 m tag declined by 20-30% between 7-9 am compared to the detection frequency at midnight, and there was a 7% decline in detection frequency for every 1 m increase in wave height (Table 4.5).

107

Table 4.2. The top 2-3 model candidates based on AICc for GAMM analyses of activity

(ms-2) for each species. s(Hour) is the spline of hour, Temp is the temperature variable

(°C), Tide is the tidal level variable (m), TL is the total length, and df is the number of

parameters in the model.

Species df Model AIC ΔAIC Model c c Weight 5 Activity ~ s(Hour) + Temp 3432.00 0.00 0.97 fiddler rays 6 Activity ~ s(Hour) + Temp + Tide 3440.30 8.30 0.02

bluespotted 5 Activity ~ s(Hour) + Temp 3408.30 0.00 0.75 flatheads 6 Activity ~ s(Hour) + Temp + Tide 3410.70 2.40 0.23

Port Jackson 4 Activity ~ s(Hour) 1927.40 0.00 0.84 sharks 5 Activity ~ s(Hour) + TL 1932.00 4.60 0.08 5 Activity ~ s(Hour) + TL + Tide 1933.30 5.60 0.05

Table 4.3. Summary statistics of the final GAMM model for activity with intercept, temperature and a spline of hour for each species. ‘edf’ is expected degrees of freedom.

Species Estimate SE t-value edf F-statistic P fiddler rays Intercept 0.095 0.046 2.063 <0.05 Temp 0.020 0.002 9.156 <0.0001 s(Hour) 5.571 471.90 <0.0001 bluespotted Intercept -0.022 0.063 -0.350 0.726 flatheads Temp 0.016 0.003 4.795 <0.0001 s(Hour) 3.910 45.27 <0.0001

Port Jackson Intercept 0.365 0.031 11.880 <0.0001 sharks s(Hour) 6.574 551.80 <0.0001

108

Figure 4.5. Average activity (ms-2± SE) and temperature (°C) in the a) fiddler ray

(Trygonorrhina fasciata), b) bluespotted flathead (Platycephalus caeruleopunctatus) and c) Port Jackson shark (Heterodontus portusjacksoni) from June 2013 to June 2015.

109

110

Table 4.4. Results from the Poisson GLM for the 50 m control tag.

Std. Estimate z value P value Error Intercept 2.360 0.185 12.724 < 0.01 hour _1 -0.038 0.071 -0.529 0.597 hour_2 0.028 0.070 0.393 0.695 hour_3 0.017 0.072 0.232 0.816 hour_4 -0.017 0.071 -0.242 0.809 hour_5 0.008 0.070 0.107 0.915 hour_6 -0.043 0.071 -0.606 0.545 hour_7 -0.099 0.072 -1.373 0.170 hour_8 -0.121 0.073 -1.664 0.096 hour_9 -0.051 0.072 -0.715 0.475 hour_10 -0.046 0.071 -0.639 0.523 hour_11 -0.067 0.072 -0.935 0.350 hour_12 0.004 0.071 0.050 0.960 hour_13 0.002 0.071 0.035 0.972 hour_14 -0.025 0.071 -0.356 0.722 hour_15 0.040 0.069 0.571 0.568 hour_16 -0.003 0.070 -0.042 0.967 hour_17 0.007 0.070 0.097 0.923 hour_18 -0.013 0.070 -0.181 0.857 hour_19 0.039 0.071 0.556 0.579 hour_20 0.041 0.070 0.592 0.554 hour_21 0.010 0.070 0.139 0.890 hour_22 0.029 0.070 0.414 0.679 hour_23 0.010 0.070 0.140 0.889 Temp 0.011 0.009 1.229 0.219 Rainfall 0.002 0.002 0.657 0.511 Height 0.014 0.020 0.684 0.494

111

Table 4.5. Results from the Poisson GLM for the 200 m control tag.

Std. Estimate z value P value Error Intercept 2.393 0.200 11.988 < 0.01 hour_1 0.004 0.074 0.048 0.961 hour_2 0.009 0.074 0.116 0.908 hour_3 -0.007 0.076 -0.098 0.922 hour_4 -0.033 0.075 -0.440 0.660 hour_5 -0.096 0.076 -1.259 0.208 hour_6 -0.078 0.076 -1.027 0.305 hour_7 -0.264 0.080 -3.291 0.001 hour_8 -0.196 0.079 -2.464 0.014 hour_9 -0.187 0.078 -2.404 0.016 hour_10 -0.106 0.076 -1.387 0.165 hour_11 -0.122 0.077 -1.599 0.110 hour_12 -0.016 0.075 -0.215 0.830 hour_13 -0.020 0.074 -0.264 0.792 hour_14 -0.022 0.075 -0.300 0.765 hour_15 0.017 0.073 0.234 0.815 hour_16 -0.013 0.074 -0.171 0.864 hour_17 0.008 0.073 0.114 0.909 hour_18 0.017 0.073 0.230 0.818 hour_19 -0.021 0.075 -0.275 0.783 hour_20 -0.104 0.076 -1.371 0.170 hour_21 0.013 0.074 0.183 0.855 hour_22 -0.061 0.075 -0.810 0.418 hour_23 -0.009 0.074 -0.117 0.907 Temp 0.011 0.010 1.103 0.270 Rainfall -0.001 0.003 -0.380 0.704 Height -0.067 0.022 -3.038 0.002

112

4.5 Discussion

4.5.1 Diel activity patterns

Activity of three species was analysed from the AR, NR and several temperate coastal reef sites in greater Sydney, and a distinct diel pattern was found for two of the three species during the study period. Fiddler rays sustained increased activity level throughout the day period which was opposite to the increased nocturnal activity in Port

Jackson sharks. Bluespotted flatheads also showed a nocturnal activity pattern, but with much less difference between day and night average activity. Diel movement behaviours have been reported in a range of reef fishes (Santos et al. 2002; Topping and

Szedlmayer 2011a; Topping and Szedlmayer 2011b; Lee et al. 2015a; Stocks et al.

2015), as well as elasmobranchs (Vianna et al. 2013; Espinoza et al. 2015b). It is plausible that light intensity is the factor responsible for the difference in activity patterns, and similar results have been described previously (e.g. Abecasis et al. 2013;

Henderson et al. 2014).

High activity sustained during the day has also been observed in other temperate reef fish including the eastern blue groper (Achoerodus viridis) (Lee et al. 2015a) and the rock blackfish (Girella elevata) (Stocks et al. 2015). Previous studies that observed similar activity patterns in other species suggest that high daytime activity is related to foraging, whereas low activity patterns are attributed to sheltering, energy conservation and predator avoidance (Santos et al. 2002; Leitão et al. 2007; Abecasis et al. 2013;

Henderson et al. 2014; Lee et al. 2015a). Diel movement between habitats can also be related to daytime foraging and night time resting (Topping and Szedlmayer 2011b;

Topping and Szedlmayer 2011a; Villegas-Ríos et al. 2013). Daytime group resting and

113 night time foraging have been observed previously in Port Jackson sharks, whereby individuals spent significant amounts of time inactive during the day and feed on small invertebrates that are buried in the sediment at night (McLaughlin and O'Gower 1971;

Powter and Gladstone 2009; Powter et al. 2010). Nocturnal activity is a common behaviour exhibited by elasmobranchs which increase foraging opportunities to take advantage of prey that is available at night (Abecasis et al. 2013; Vianna et al. 2013;

Kessel et al. 2014).

Average activity in bluespotted flatheads followed a fairly irregular pattern and was lower in comparison to activity in both fiddler rays and Port Jackson sharks. The bluespotted flathead is a “sit and wait” ambush predator (Coleman and Mobley 1984;

Moore et al. 2009; Barnes et al. 2011). The dusky flathead (P. fuscus) is also an ambush predator which was observed to be active during crepuscular and diurnal periods

(Gannon 2016), and is thought to be foraging during these periods. Similarly the dottyback (Pseudochromis fuscus), which is a small coral reef predator, was observed to use a combination of feeding modes: an ambush or a pursuit mode, but was most active in the middle of the day (Feeney et al. 2012). The bluespotted flathead is also likely to be an opportunistic hunter that feeds throughout the day with bursts of active movements, explaining some of the variability in activity and also because this species is frequently caught by anglers during the day (Rowling et al. 2010). However the slight increase in activity during the night may indicate larger scale movements and sustained periods of swimming, but may also be in response to predation risk (Metcalfe et al.

1999; Koeck et al. 2013).

114

It is important to note that the control tag results in this study indicate that there is a reduced chance of detecting a tagged fish at a moderate distance from the reef (~200 m away) when there are large swells, and particularly in the early morning. These diel differences in detections were considered to be minor and unlikely to influence the results in this study, since accelerometry tags were used to directly measure activity and a reduced likelihood of detection would not impact these results. However this pattern would be of much greater importance for studies examining behavioural or movement patterns on finer spatial or temporal scales.

4.5.2 Effect of temperature on activity

In this study, TL, tide and luminosity were found to have no influence on activity in any species, however activity increased significantly with increasing water temperature in both fiddler rays and bluespotted flatheads during the monitoring period. Temperature is well known to have an effect on the metabolic, physiological and reproductive functions of ectotherms (Schlaff et al. 2014). In a previous study, both activity and metabolic rates in the sand flathead (P. bassensis) were observed to increase with increasing temperature up to 22°C (Stehfest et al. 2015). Many species also display a behavioural thermoregulation strategy by maximising foraging success during warmer water temperatures and minimising energetic costs (Charnov 1976; Schlaff et al. 2014). The grey reef shark (Carcharhinus amblyrhynchos) for example is known to adopt behavioural strategies to maintain optimum body temperature (Vianna et al. 2013).

In many fish species, activity peaks at an optimum temperature and then decreases as temperatures continue to rise (i.e. ‘pejus’ thresholds; Gannon et al. 2014; Henderson et al. 2014; Payne et al. 2016). Activity of the dusky flathead (P. fuscus) was found to 115 increase with temperature up to ~23°C at which point activity levels rapidly declined

(Gannon et al. 2014). The distribution of this species overlaps with that of the bluespotted flathead, however its range extends further north to Cairns in Queensland

(Gray et al. 2002). A multi-species comparison of temperate fish (with P. fuscus included) established that optimal ecological performance of these species reflects the highest environmental temperatures encountered in their range while maintaining a safety buffer from the adverse effects of upper critical temperatures (Payne et al. 2016).

Activity data for bluespotted flatheads in temperatures greater than ~23°C were unable to be collected but these results suggest that activity of this species increases to at least

~22 °C. Research suggests that the activity of the bluespotted flathead may be particularly sensitive to temperatures higher than 22-23 °C due to its more contracted range leading to a narrower thermal tolerance (Payne et al. 2016). The same may be true for fiddler rays, which have a similar distribution to the bluespotted flathead and therefore experience the same variation in temperatures.

Temperature was found to have little effect on Port Jackson shark activity, since it was not included in the best model with only hour explaining activity in this species. This may be due to the migratory behaviour by Port Jackson sharks, which alter their thermal habitat through seasonal movements (McLaughlin and O'Gower 1971; Powter and

Gladstone 2008a; Powter and Gladstone 2009). Adult male and female Port Jackson sharks are known to migrate annually during the breeding season from July to

November, from deep offshore waters to shallow coastal rocky reefs (Powter and

Gladstone 2008a). In Sydney waters this breeding season, which occurs during the austral winter-spring, is correlated with low water temperatures where sea temperatures are usually below 18 °C (McLaughlin and O'Gower 1971). Hence temperature may

116 function as a cue for migratory behaviour in this species. However, other factors may be more important in driving activity within this species and could relate to the size, sex and/or maturity state of individuals, which were limited in this study. Similarly, these factors may also be responsible for driving the activity patterns in fiddler rays, since only males were tagged in this study. A study on the yellowfin bream (Acanthopagrus australis) in an estuarine system, for instance, found that activity in this species was not related to temperature or lunar phase, but both fish size and conductivity (Gannon et al.

2015). Future studies should therefore investigate other factors which may influence activity as these may determine the distribution and movements (including site residency- see chapter 5) of the Port Jackson shark and similar species.

4.5.3 Behavioural comparisons and implications

This study demonstrates that diel activity patterns varied between the three species: a ray, a teleost and a shark; and that temperature influenced the activity in two of these species. The bluespotted flathead and Port Jackson shark had similar nocturnal activity patterns which produced a positive relationship between these two species. In contrast, both these species had a negative relationship in activity with the fiddler ray, which was active during the day. Opposite behavioural patterns in ecologically similar species have been previously observed in other studies and are suggested as strategies to reduce resource competition (Koeck et al. 2013; Espinoza et al. 2015a; Espinoza et al. 2015b).

Species are able to adapt their behaviour in various ways, such as changing the timing of seasonal movements and diet to optimise foraging in their ecological niche (Charnov

1976; Koeck et al. 2013; Espinoza et al. 2015a; Espinoza et al. 2015b). A previous study on four co-occurring elasmobranch species, which included the Port Jackson shark for example, found that all four species had different diets (Sommerville et al. 117

2011). They suggested that inter- and intraspecific variations in dietary composition reduced the potential for competition between and within the elasmobranch species in south-western Australian waters.

The fiddler ray, bluespotted flathead and Port Jackson shark are reported to have similar diets (Hutchins and Swainston 1986; Marshall et al. 2007; Izzo and Gillanders 2008;

Powter et al. 2010). It is possible that the fiddler ray may compete for resources with the bluespotted flathead and Port Jackson shark, and other ecologically similar species such as snapper (C. auratus) and the blue morwong (N. douglasii). Therefore the different diel activity patterns by these species may be one strategy to reduce interspecific resource competition. Investigating the diets of these three species and other members of the epibenthic ecosystem, and measuring changes in activity patterns in systems with different species composition, would help quantify the extent to which interspecific competition influences diel activity patterns.

High activity during the day, as well as movements in response to environmental variables, can have relevance for species management. Fiddler rays and other species that exhibit high activity during daylight periods are more susceptible to fishing at these times as a negative relationship generally exists between recreational fishing effort and the time of day (Vianna et al. 2013; Villegas-Ríos et al. 2013). An increase in activity with increased temperature can also lead to higher susceptibility to fishing mortality in fiddler rays during warmer periods, as well as for bluespotted flatheads because this species is targeted by anglers (Steffe et al. 1996; Henry and Lyle 2003; Steffe and

Murphy 2011). Similarly, catches of the recreationally-targeted sand flathead were found to vary seasonally and declined at lower temperatures (Stehfest et al. 2015). This

118 was suggested to be due to a lower probability of encountering bait by this species as well as lower feeding motivation due to a lower metabolic debt. However both the bluespotted flathead and the Port Jackson shark are nocturnal species and are less likely to encounter recreational fishing gears as long as minimal horizontal diel movements are made and fishing is temporally restricted (Parsley et al. 2008).

119

4.6 Conclusions

This study assessed behaviour and activity of three species across a number of sites in a temperate coastal reef ecosystem, and found different diel patterns in three co-existing benthic species, as well as an effect of temperature on the activity of fiddler rays and bluespotted flatheads. An increase in activity is likely to be associated with foraging behaviour, with fiddler rays feeding during the day compared to bluespotted flatheads and Port Jackson sharks which feed at night. Conversely, resting by all three species occurs during non-activity periods. Furthermore, differences in activity plus an overlapping diet suggest that resource partitioning may occur whereby these species forage at different times of the day. Species which are active during the day can be at a higher risk of fishing mortality, and the activity patterns of the targeted bluespotted flathead in this study suggest it is at most risk from recreational harvest during crepuscular periods and in warmer months.

Some of the observed diel activity patterns in these species may be influenced by the presence of the AR. Previous studies have found that habitat type can influence individual temporal behaviour (Santos et al. 2002; Abecasis et al. 2013; Koeck et al.

2013), ARs in particular have the potential to modify individual behavioural patterns of fishes as well as seascape connectivity (Koeck et al. 2013). The strong diel pattern of white seabream (Diplodus sargus) for instance was found to be associated with feeding at the AR and sheltering at nearby natural reefs (Leitão et al. 2007; Abecasis et al.

2013). Similar diel trends have also been observed in other species that were resident around ARs, where fish were present in higher densities during the day (Santos et al.

2002). The selection of preferred habitats (i.e. natural or artificial) by a species can have significant implications on population dynamics, including life-history, mortality, and 120 reproductive success (Koeck et al. 2013). Future studies should therefore assess the movements and residency of the fiddler ray, bluespotted flathead and Port Jackson shark at both natural and ARs to determine the role these habitats play in movement behaviour of benthic predators.

Publication details:

Keller, K., Smith, J., Lowry, M, and Suthers, I.M. (In prep). A ray, a fish and a shark: the diel behaviour and activity of three co-occurring benthic species around temperate rocky reefs.

121

122

Chapter 5

Multispecies residency and connectivity around a designed artificial reef

5.1 Abstract

Designed artificial reefs (ARs) may increase fishing opportunities by enhancing fish abundance and species diversity via providing food and refuge. Residency and connectivity of different species of fish associated with ARs and natural reefs is a key question to determine if ARs contribute to coastal ecosystems and fisheries. The movements and residency of 10 eastern fiddler rays (Trygonorrhina fasciata), 17 Port

Jackson sharks (Heterodontus portusjacksoni) and 23 bluespotted flatheads

(Platycephalus caeruleopunctatus) were monitored using acoustic telemetry around a

12 x15 m designed AR in 38 m depth near Sydney, Australia. Residency at the AR was low for all three species over the monitoring period, although some individual fiddler rays and bluespotted flathead exhibited relatively high residency times over the ~ 20 month monitoring period, and were resident at least 50% of the time within the AR area.

Fish tagged at the AR showed a higher degree of residency relative to those tagged near the natural reef. All three species moved frequently between the AR and 5-6 other reefs indicating strong connectivity throughout the habitat mosaic. The relatively low residency shows that only 4- 32% of these species’ biomass production and contribution to total recreational harvest is likely to be derived from this AR.

123

5.2 Introduction

Designed artificial reefs are widely used to support increased recreational fishing opportunities (Branden et al. 1994; Carr and Hixon 1997; Cenci et al. 2011). These modern artificial reefs are not constructed from materials of opportunity (rubble, scuttled vessels), but built from steel or concrete to a design that incorporates vertical relief, void space and water movement (Ambrose and Swarbrick 1989; Branden et al.

1994). The design and structure of artificial reefs are believed to support fisheries enhancement by providing food and refuge to fish communities in an area where the availability of habitat is limited (Carr and Hixon 1997; Grossman et al. 1997).

Whether an increase in fish biomass leads to production or existing biomass is attracted to these reefs is still debated (Bohnsack and Sutherland 1985; Bohnsack 1989; Carr and

Hixon 1997; Grossman et al. 1997; Pickering and Whitmarsh 1997; Powers et al. 2003;

Brickhill et al. 2005; Smith et al. 2015). Attraction is considered to be detrimental to fish abundance by facilitating exploitation by fishermen through aggregation (Carr and

Hixon 1997; Grossman et al. 1997; Powers et al. 2003). Previous studies suggesting high fish production can occur on artificial reefs demonstrate that species richness, diversity, and biomass density of fish were usually the same or higher on artificial reefs compared to natural reefs (Bohnsack and Sutherland 1985; Pondella et al. 2002). In fact decommissioned oil rigs are suggested to be the most productive marine habitats for fish

(Claisse et al. 2014). It is likely that both production and attraction occur and depends on species-specific characteristics (Wilson et al. 2001; Osenberg et al. 2002b) and their connectivity with natural reef. Estimating the residency of fish associated with artificial reefs is essential for understanding the production versus attraction issue as the more resident a species is, the more its production can be attributed to the artificial reef 124

(Smith et al. 2015; Smith et al. 2016). On the other hand, reduced residency implies that a population is less vulnerable to over-exploitation at the artificial reef (Smith et al.

2015).

5.2.1 Movement ecology and artificial reefs

The temporal and spatial scales at which fish interact with artificial and natural reefs are related to their ecological function (Bohnsack 1989; Shipley and Cowan Jr 2011). Fine- scale movements of fish at artificial habitats using kernel utilisation distributions are often used to reveal habitat use and home range (Topping and Szedlmayer 2011a;

Abecasis et al. 2013; Piraino and Szedlmayer 2014). Fish that show limited movements, with a narrow home range and high residency at artificial reefs, are more likely to benefit from these reefs (Bohnsack 1989; Shipley and Cowan Jr 2011). Site residency or fidelity is defined as the continuous presence by a species over time and can be measured as the degree to which an animal returns to a specific site (Schroepfer and

Szedlmayer 2006; Reubens et al. 2013). Fidelity of fishes to artificial structures often arises due to increased prey availability, shelter and potential spawning opportunities

(Topping and Szedlmayer 2011b). Knowledge of species residency and movements are important for assessing the risk of exposure to fishing, since site residency is proportional to the risk of capture at a given location (Parsley et al. 2008). However various factors can influence overall site residency, such as seasonal migration whereby a species may only be partial residents (e.g. Reubens et al. 2013; Espinoza et al. 2015a;

Espinoza et al. 2015b).

125

Red snapper (Lutjanus campechanus) in the north-eastern Gulf of Mexico showed high residency over extended periods on artificial reefs compared to natural reefs, indicating that artificial reefs provide suitable habitat for this species (Szedlmayer and Schroepfer

2005; Schroepfer and Szedlmayer 2006; Topping and Szedlmayer 2011b; Piraino and

Szedlmayer 2014). Likewise, Atlantic cod (Gadus morhua) displayed high residency at artificial benthic habitats associated with an offshore wind farm (Reubens et al. 2013).

Copper rockfish (Sebastes caurinus) and lingcod (Ophiodon enlongatus) also exhibited long periods of residency in an artificial reef in Alaska, implying that this reef provides suitable habitat (Reynolds et al. 2010). Similarly, residency in white seabream

(Diplodus sargus) was found to be high on both natural and artificial reefs in the northwestern Mediterranean Sea (Koeck et al. 2013). In Australia, studies of fish presence at artificial reefs have been limited to coastal estuarine environments in New South Wales, and indicate that a number of species, including sparids and carangids, frequently occur at design-specific artificial reefs (Burchmore et al. 1985; Folpp et al. 2011; Folpp et al.

2013; Lowry et al. 2014).

Connectivity between artificial and natural reefs is central to understanding the effect of an artificial reef within a coastal ecosystem. The location, size and proximity of artificial reefs to other reefs can facilitate the dispersion and recruitment of species to newly deployed artificial reefs (Airoldi et al. 2005; Cenci et al. 2011; Shipley and

Cowan Jr 2011; Smith et al. 2015). These structures can function as ecological stepping stones by increasing the connectivity between existing habitats (Airoldi et al. 2005).

Connectivity can also contribute to overall production of an artificial reef by increasing total fish abundance and biomass (Fernandez et al. 2008; Koeck et al. 2013). Studying the movements of fish between artificial and natural habitats is therefore useful for

126 assessing the suitability of artificial habitats, developing fisheries management zones

(e.g. marine parks), as well as the potential for contribution to overall fisheries productivity.

5.2.2 Objectives and rationale of study

The objective of this study is to examine the site residency, connectivity and movements of the fiddler ray (Trygonorrhina fasciata), eastern bluespotted flathead

(Platycephalus caeruleopunctatus), and Port Jackson shark (Heterodontus portusjacksoni) at a coastal artificial reef off Sydney, NSW, Australia, and throughout the adjacent coastal ecosystem. These species are endemic to the eastern coast of

Australia and inhabit soft substrates near coastal rocky reef environments (Hutchins and

Swainston 1986; Last and Stevens 1994; Powter and Gladstone 2008b; Moore et al.

2009). The bluespotted flathead was a focal species for this study, as it supports a statewide recreational harvest of between 320 and 450 tonnes (Henry and Lyle 2003;

Rowling et al. 2010; Steffe and Murphy 2011). This species was also found to be the most commonly captured species at the Sydney designed artificial reef, with an estimated annual recreational harvest of 144 kg from the AR (Appendix D). The fiddler ray and Port Jackson shark were also selected as they are common benthic predators which are taken as bycatch in commercial trawls (Jones et al. 2010; Rowling et al.

2010). These two species are often incidentally captured by recreational fishers, with an estimated state-wide recreational catch of over 3 tonnes for Port Jackson shark (West et al. 2015), and an estimated annual harvest of at least 2 kg for the fiddler ray from the

Sydney artificial reef alone (see Chapter 3). All three species have a similar diet, composed of crustaceans, fish, polychaetes and molluscs (Coleman and Mobley 1984;

Hutchins and Swainston 1986; Barnes et al. 2011). In addition Port Jackson Shark may 127 play an important role as an ecosystem regulator through their predation on echinoderms and fish (Powter et al. 2010). Large-scale seasonal movements and high site residency at natural reefs have previously been documented in Port Jackson sharks

(McLaughlin and O'Gower 1971; O'Gower 1995; Powter and Gladstone 2009). Little is known about the movements in the fiddler ray and bluespotted flathead, other than their diurnal and nocturnal activity patterns respectively (Chapter 4).

Acoustic telemetry is widely used to study the movements and behaviour of aquatic animals and has greatly improved our understanding of temporal and spatial movements in fish (Payne et al. 2011; Hussey et al. 2015; Payne et al. 2016). The specific aims of this study were to: 1) determine the site residency of three benthic species at a designed artificial reef and at nearby natural reefs using acoustic telemetry; 2) determine the connectivity between the artificial reef and natural reefs for these species, and 3) use this information to infer the effect of this artificial reef on the local distribution of these species.

128

5.3 Materials and Methods

5.3.1 Study area

This study was conducted at the Sydney artificial reef (AR), which was deployed in

October 2011 in 38 m depth of water near South Head of Sydney Harbour, Australia

(33°50'47.82"S 151°17'59.28"E, Fig. 5.1). The steel structure has a footprint of 12 x 15 m and 12 m high with two 8 m tall pillars extending from the top and is attached by chain at each corner to a 60 ton concrete block (Champion et al. 2015; Scott et al.

2015). The substrate surrounding the AR is mostly flat and sandy. A nearby natural reef

(NR), Dunbar reef (33°51’10.73"S, 151°17’19.36"E) was also monitored (Fig. 5.2).

This site consists of a significant outcrop of subtidal reef that gets as close as 550 m to the AR.

129

Figure 5.1. Study area showing receiver locations. AR= Artificial Reef, NR= Dunbar reef, AM= Annie Miller reef, SG= Sydney harbour entrance, SH= Sydney Harbour,

BSH= Between Bondi and South Head, BL= Bondi line, BCAR= Bronte-Coogee

Aquatic reserve, SLB= South Long Bay, NAR= Narooma.

130

(Note: Dunbar reef (NR) is the area of natural reef that projects out from the coast between the two receivers).

131

A VR2W receiver (Vemco Ltd, NovaScotia, Canada) was deployed on the AR from

2011 to 2013, it was then replaced by a VR4 receiver (Vemco Ltd, NovaScotia, Canada) to provide remote uploading capability. The receiver was tethered to a cross-beam on the artificial structure approximately 8 m from the seafloor. Two VR2W receivers were also deployed on areas of scattered reef near the NR, one north (33°50'47.76"S,

151°17'27.60"E) and another south (33°51'4.32"S, 151°17'27.60"E) at approximately 25 m depth (Fig. 5.2). All receivers were coated with a copper-based antifouling paint to prevent possible signal occlusion due to biofouling (Heupel et al. 2008). Receivers were downloaded every 3 to 6 months over a period of 2 years. Broad-scale detections of tagged individuals detected in the greater Sydney area (Fig. 5.1), and further afield, were downloaded from the Integrated Marine Observing System – Australian Animal

Tagging and Monitoring System (IMOS– AATAMS) fisheries acoustic telemetry array.

This array consists of over 350 receivers in the region, which is publically available online (https://aatams.emii.org.au/aatams).

5.3.2 Acoustic tagging

The fiddler ray (Trygonorrhina fasciata), bluespotted flathead (Platycephalus caeruleopunctatus) and Port Jackson shark (Heterodontus portusjacksoni), are three large-bodied species that were identified as abundant (prior to tagging) using Baited

Remote Underwater Video (BRUV) at both the AR and NR (Lowry and Folpp 2014;

Scott et al. 2015). These three species were selected for tagging, captured using various fishing methods (Chapter 4), and were implanted with a variety of available acoustic transmitters (Table 5.1). A total of 9 fiddler rays, 25 bluespotted flatheads and 9 Port

Jackson sharks were tagged at the AR; and 1 fiddler ray, 1 bluespotted flathead and 8

Port Jackson sharks were tagged at the NR. An additional 9 bluespotted flatheads were 132 tagged at Annie Miller (AM), an artificial reef consisting of a wreck in 45 m depth of water, located approximately 3 km south of the AR (33°52'26.1"S, 151°17'55.8"E; Fig.

5.1). All surgical procedures were conducted following protocols approved by

University of New South Wales Animal Care and Ethics Committee (Approval Number

12/111A).

133

Table 5.1. Summary of acoustically tagged animals. AR= artificial reef, NR= Dunbar reef, AM= Annie Miller; TL=total length; M= male; F= female;

I=fish of indeterminate sex. L=low power, H=High power; A= apparent survival, R= recaptured, U= undetected

Animal Fish TL Sex Estimated Min Max Tag type Power Date tagged End of Location Status ID (mm) tag life (d) delay delay output monitoring tagged & code (s) (s) period released fiddler ray B1 770 M 1021 400 800 V9-1x L 28/06/2013 15/06/2015 AR A fiddler ray B2 840 M 61 190 290 V9A-2x H 28/06/2013 1/09/2013 AR A fiddler ray B3 750 M 61 190 290 V9A-2x H 28/06/2013 1/09/2013 AR A fiddler ray B4 790 M 61 190 290 V9A-2x H 28/06/2013 1/09/2013 AR A fiddler ray B5 764 M 156 25 25 V9A-2x L 25/07/2013 28/12/2013 AR A fiddler ray B6 800 M 1021 400 800 V9-1x L 14/08/2013 15/06/2015 AR A fiddler ray B7 742 M 821 170 310 V9-2x L 14/08/2013 15/06/2015 AR A fiddler ray B8 760 M 602 220 500 V9AP-2L L 27/08/2013 26/04/2015 AR A fiddler ray B9 820 M 602 220 500 V9AP-2L L 24/09/2013 19/05/2015 AR A fiddler ray B10 820 M 602 220 500 V9AP-2L L 25/02/2014 15/06/2015 NR A

134

Table 5.1 (Cont.)

Animal Fish TL Sex Estimated Min Max Tag type Power Date tagged End of Location Status ID (mm) tag life (d) delay delay output monitoring tagged & code (s) (s) period released

bluespotted F1 415 I 156 25 25 V9A-2x L 14/08/2013 18/01/2014 NR A flathead bluespotted F2 383 I 156 25 25 V9A-2x L 14/08/2013 18/01/2014 AR U flathead bluespotted F3 500 I 61 190 290 V9A-2x H 14/08/2013 14/10/2013 AR R flathead bluespotted F4 500 I 61 190 290 V9A-2x H 14/08/2013 14/10/2013 AR A flathead bluespotted F5 365 I 821 170 310 V9-2x L 17/08/2013 15/06/2015 AR A flathead bluespotted F6 410 I 602 220 500 V9AP-2L L 22/08/2013 16/04/2015 AR A flathead bluespotted F7 387 I 602 220 500 V9AP-2L L 27/08/2013 21/04/2015 AR A flathead

135

Table 5.1 (Cont.)

Animal Fish TL Sex Estimated Min Max Tag type Power Date tagged End of Location Status ID (mm) tag life (d) delay delay output monitoring tagged & code (s) (s) period released

bluespotted F8 450 I 602 220 500 V9AP-2L L 27/08/2013 21/04/2015 AR A flathead bluespotted F9 520 I 602 220 500 V9AP-2L L 27/08/2013 3/11/2013 AR R flathead bluespotted F10 550 I 602 220 500 V9AP-2L L 27/08/2013 21/04/2015 AR A flathead bluespotted F11 420 I 602 220 500 V9AP-2L L 24/09/2013 19/05/2015 AR A flathead bluespotted F12 425 I 602 220 500 V9AP-2L L 24/09/2013 19/05/2015 AR U flathead bluespotted F13 425 I 602 220 500 V9AP-2L L 31/10/2013 15/06/2015 AR A flathead bluespotted F14 400 I 602 220 500 V9AP-2L L 31/10/2013 15/06/2015 AR A flathead

136

Table 5.1 (Cont.)

Animal Fish TL Sex Estimated Min Max Tag type Power Date tagged End of Location Status ID (mm) tag life (d) delay delay output monitoring tagged & code (s) (s) period released

bluespotted F15 490 I 602 220 500 V9AP-2L L 1/05/2014 15/06/2015 AR A flathead bluespotted F16 270 I 270 130 230 V7-4x L 25/08/2014 22/05/2015 AR A flathead bluespotted F17 270 I 270 130 230 V7-4x L 25/08/2014 22/05/2015 AR A flathead bluespotted F18 250 I 270 130 230 V7-4x L 12/09/2014 9/06/2015 AR A flathead bluespotted F19 250 I 270 130 230 V7-4x L 12/09/2014 9/06/2015 AR U flathead bluespotted F20 300 I 602 220 500 V9AP-2L L 12/09/2014 15/06/2015 AR A flathead bluespotted F21 280 I 270 130 230 V7-4x L 12/09/2014 9/06/2015 AR A flathead

137

Table 5.1 (Cont.)

Animal Fish TL Sex Estimated Min Max Tag type Power Date tagged End of Location Status ID (mm) tag life (d) delay delay output monitoring tagged & code (s) (s) period released

bluespotted F22 430 I 602 220 500 V9AP-2L L 12/09/2014 15/06/2015 AR A flathead bluespotted F23 520 I 602 220 500 V9AP-2L L 12/09/2014 15/06/2015 AR A flathead bluespotted F24 260 I 270 130 230 V7-4x L 12/09/2014 9/06/2015 AR A flathead bluespotted F25 255 I 270 130 230 V7-4x L 12/09/2014 9/06/2015 AR A flathead bluespotted F26 265 I 270 130 230 V7-4x L 12/09/2014 9/06/2015 AR A flathead bluespotted F27 433 I 61 190 290 V9A-2x H 24/09/2014 24/11/2014 AM U flathead bluespotted F28 420 I 270 130 230 V7-4x L 24/09/2014 15/06/2015 AM A flathead

138

Table 5.1 (Cont.)

Animal Fish TL Sex Estimated Min Max Tag type Power Date tagged End of Location Status ID (mm) tag life (d) delay delay output monitoring tagged & code (s) (s) period released

bluespotted F29 420 I 270 130 230 V7-4x L 24/09/2014 15/06/2015 AM A flathead bluespotted F30 415 I 270 130 230 V7-4x L 24/09/2014 15/06/2015 AM U flathead bluespotted F31 290 I 270 130 230 V7-4x L 24/09/2014 15/06/2015 AM U flathead bluespotted F32 265 I 270 130 230 V7-4x L 24/09/2014 15/06/2015 AM U flathead bluespotted F33 260 I 270 130 230 V7-4x L 24/09/2014 15/06/2015 AM U flathead bluespotted F34 277 I 270 130 230 V7-4x L 24/09/2014 15/06/2015 AM U flathead bluespotted F35 262 I 270 130 230 V7-4x L 24/09/2014 15/06/2015 AM U flathead

139

Table 5.1 (Cont.)

Animal Fish TL Sex Estimated Min Max Tag type Power Date tagged End of Location Status ID (mm) tag life (d) delay delay output monitoring tagged & code (s) (s) period released

Port Jackson PJ1 930 M 602 220 500 V9AP-2L L 22/08/2013 16/04/2015 NR A shark

Port Jackson PJ2 950 M 602 220 500 V9AP-2L L 22/08/2013 16/04/2015 NR A shark

Port Jackson PJ3 950 M 602 220 500 V9AP-2L L 22/08/2013 16/04/2015 NR A shark

Port Jackson PJ4 950 M 602 220 500 V9AP-2L L 22/08/2013 16/04/2015 AR A shark

Port Jackson PJ5 873 M 602 220 500 V9AP-2L L 27/08/2013 21/04/2015 AR A shark

Port Jackson PJ6 640 M 602 220 500 V9AP-2L L 27/08/2013 21/04/2015 AR A shark

Port Jackson PJ7 900 M 602 220 500 V9AP-2L L 27/08/2013 21/04/2015 NR A shark

140

Table 5.1 (Cont.)

Animal Fish TL Sex Estimated Min Max Tag type Power Date tagged End of Location Status ID (mm) tag life (d) delay delay output monitoring tagged & code (s) (s) period released

Port Jackson PJ8 660 M 602 220 500 V9AP-2L L 27/08/2013 21/04/2015 AR A shark

Port Jackson PJ9 820 M 602 220 500 V9AP-2L L 27/08/2013 21/04/2015 AR A shark

Port Jackson PJ10 910 M 602 220 500 V9AP-2L L 27/08/2013 21/04/2015 AR A shark

Port Jackson PJ11 900 M 602 220 500 V9AP-2L L 6/09/2013 1/05/2015 NR A shark

Port Jackson PJ12 985 M 602 220 500 V9AP-2L L 6/09/2013 1/05/2015 NR A shark

Port Jackson PJ13 880 M 602 220 500 V9AP-2L L 24/09/2013 19/05/2015 NR A shark

Port Jackson PJ14 1150 F 602 220 500 V9AP-2L L 24/09/2013 19/05/2015 AR A shark

141

Table 5.1 (Cont.)

Animal Fish TL Sex Estimated Min Max Tag type Power Date tagged End of Location Status ID (mm) tag life (d) delay delay output monitoring tagged & code (s) (s) period released

Port Jackson PJ15 960 M 602 220 500 V9AP-2L L 24/09/2013 19/05/2015 AR A shark

Port Jackson PJ16 660 M 602 220 500 V9AP-2L L 24/09/2013 19/05/2015 AR A shark

Port Jackson PJ17 1900 F 1212 140 240 V13-1x L 31/10/2013 15/06/2015 NR A shark

142

5.3.3 Site residency

Data collection spanned approximately 2 years, with receiver downloads from June

2013 to June 2015 at the AR, and from June 2013 to November 2014 at the NR.

Detections within the first 24 h after release of tagged individuals were excluded to ensure that behaviour was not influenced by the tagging procedure and data were filtered to remove potential spurious detections. Single transmitter detections were considered false detections and removed from the analyses (Reubens et al. 2013). Data were converted to Eastern Standard Time (EST). Heavy rainfall events (>30 mm) from the Dover Heights weather station (BOM 2015), were compared with detections from each species, however no relationship between individual absence/presence at the AR or

NR was observed during the monitoring period.

Total residence time was calculated for every tagged individual and was defined as the total number of days that an individual was detected at a specific receiver. Total residence time was often the sum of multiple residence periods. A residence period for an individual began when a minimum of two detections on a specific receiver were recorded within a 24 h period (Campbell et al. 2012). Non-residence time was calculated as the duration between consecutive residence periods. A residence period ended when that individual was either detected at another receiver or was not detected within 24-48 h after the last detection, depending on the minimum number of days that an individual was considered resident. In this study, the minimum number of days that an individual was considered resident was determined by calculating a cumulative percentage of detections, as days between consecutive detections for each species in the study area during the monitoring period (Fig. 5.3). We found that 85% of the total detections of bluespotted flatheads were recorded with 3 or fewer days between them, 143 whereas 97% and 93% of total detections of fiddler rays and Port Jackson sharks were recorded with 2 or fewer days between them, respectively (Fig. 5.3). Hence, fiddler rays and Port Jackson sharks were considered resident if there was no more than 1 day between consecutive detections, and bluespotted flatheads were considered resident if there were no more than 2 days between consecutive detections.

These definitions of residency for each species were considered to be appropriate since increasing the number of days between detections made negligible difference to the overall residency calculations. The rationale behind the 24-48 hour residency ‘grace period’ is that receivers can have a fairly short range (assumed here as a 200 m radius from the receiver; this was a conservative estimate based on a maximum detection range of 500 m (e.g. Abecasis et al. 2013; Villegas-Ríos et al. 2013; Espinoza et al. 2015a;

Espinoza et al. 2015b; Lee et al. 2015a)). I considered this range less than the home range of a typical resident fish, so allowed this period of absence within a residency period. All residency analyses were conducted using the ‘VTrack’ R package (Campbell et al. 2012; Dwyer et al. 2015) with the R 2.15.1 software (www.rproject.org).

144

Figure 5.3. Cumulative percentage (%) of detections of fiddler rays, bluespotted flatheads and Port Jackson sharks in study area during the monitoring period.

145

A residency index (IR) was determined for each individual for each tagging reef (AR or

NR). IR was calculated by dividing the total residence time at a tagging reef by the total monitoring period (Abecasis et al. 2013; Reubens et al. 2013; Villegas-Ríos et al. 2013;

Espinoza et al. 2015b):

tR IR = tM

where tR is the total number of days an individual was detected during the monitoring period (total residence time) and tM is the total monitoring period for each fish, from the first day an individual was detected till the last day of the monitoring period

(monitoring time). The monitoring period depended on the battery life of the transmitters, known captures of tagged fish, and the date of last receiver download (June

2015), which meant some variation in monitoring periods among individuals (Table

5.1). Due to these various monitoring periods which can potentially bias the IR when comparing between individuals and species, tR was also compared between all individuals. IR = 0 indicates no residency and IR = 1 indicates permanent residency.

The proportion of total days spent at the AR, NR and other sites in the study area (Fig.

5.1) by all tagged individuals was determined to examine the effect of tagging location on these individuals. The number of days detected at each receiver was summed together for all individuals and the proportion of total days detected at that receiver was examined separately for individuals tagged from either the AR or the NR. Receivers in the study area were classified into 6 main sites (Fig.5.1): AR (artificial reef receiver),

NR (north and south Dunbar receivers), Sydney Harbour (SH- SYT233 receiver),

Sydney gate (4 receivers at entrance to Sydney harbour– SG1-4), Bondi (Bondi line

146 receivers– BL1-3) and between Bondi and South Head (BSH-SYT409 receiver). A seventh site was also included; Narooma (5 receivers– N2-6), located south of Sydney.

5.3.4 Site connectivity

Connectivity between the AR and natural reefs was determined for each species, by comparing the proportion of individuals detected at each receiver in the study area

(Fig.5.1). A distance matrix (in metres) between all sites in the network of receivers was calculated using a 200 m radius for each receiver with the ‘VTrack’ R package (Dwyer et al. 2015). Mean and median distance travelled from the tagging reef (either the AR or

NR) were determined for each species by calculating the proportion of individuals per species detected within 8 km, at 1 km intervals, and those detected at distances greater than 8 km.

147

5.4 Results

5.4.1 Residency

Fiddler ray

Of the 10 fiddler rays tagged, 9 were tagged at the AR, and these individuals were detected at this reef for more than 2 days (Fig. 5.4). One fiddler ray (B5) was detected at the AR and between Bondi and South Head, but only had a few detections so was removed from further analysis. Three fiddler rays (B2, B3 and B4) were tagged with short-term tags (61 day tag life; Table 5.1) so long-term residency was unable to be determined for these individuals. Over the monitoring period, individual fiddler rays exhibited a range of residency from IR=0.01 to IR =0.77 and were detected from 2 to 402 days at the AR (Table 5.2). Two individuals (B8 and B9) showed the highest residency at the AR (IR ≥ 0.48, Table 5.2). The average residency of all individuals at the AR was

IR = 0.32, with a median IR = 0.28 and the average long-term residency at the AR for all individuals with a total monitoring period greater than 61 days was IR = 0.26. Only one individual was tagged at the NR (B10), which had an IR = 0.06 and was detected at this site for a total of 28 days (Table 5.2). Three individuals (B2, B6 and B9) were only detected at the AR during the monitoring period (Fig. 5.4). All other individuals were detected at up to 5 additional sites.

148

Figure 5.4. Presence plot of fiddler rays monitored in the Sydney region from June

2013 to June 2015. Black dots indicate detection at the artificial reef (AR), colour symbols indicate detection at other receivers (SG= Sydney gate, NR=Dunbar, BCAR=

Bronte-Coogee marine reserve, BL=Bondi line, BSH= between Bondi and South Head).

Grey crosses indicates end of tag life.

149

Table 5.2. Total residence time (tR, days), monitoring period (tM, days) and residency index (IR) of acoustically tagged animals at tagging reef. AR= artificial reef, NR= Dunbar reef.

Fish ID Tagging Animal t t I code R M R reef

fiddler ray B1 7 714 0.01 AR

fiddler ray B2 32 65 0.49 AR

fiddler ray B3 2 65 0.03 AR

fiddler ray B4 50 65 0.77 AR

fiddler ray B6 32 662 0.05 AR

fiddler ray B7 53 670 0.08 AR

fiddler ray B8 402 601 0.67 AR

fiddler ray B9 286 593 0.48 AR

fiddler ray B10 28 473 0.06 NR

bluespotted flathead F1 15 157 0.10 NR

bluespotted flathead F3 3 61 0.05 AR

bluespotted flathead F4 2 61 0.03 AR

bluespotted flathead F6 64 602 0.11 AR

bluespotted flathead F7 1 601 0 AR

bluespotted flathead F8 0 593 0 AR

bluespotted flathead F9 12 66 0.18 AR

bluespotted flathead F10 1 600 0 AR

bluespotted flathead F11 2 602 0 AR

bluespotted flathead F13 19 592 0.03 AR

bluespotted flathead F14 6 592 0.01 AR

bluespotted flathead F15 16 410 0.04 AR

bluespotted flathead F16 6 270 0.02 AR 150

Table 5.2 (Cont.)

Fish ID Tagging Animal t t I code R M R reef bluespotted flathead F17 2 270 0.01 AR bluespotted flathead F18 0 270 0 AR bluespotted flathead F20 130 276 0.47 AR bluespotted flathead F21 35 270 0.13 AR bluespotted flathead F23 24 276 0.09 AR bluespotted flathead F24 39 270 0.14 AR bluespotted flathead F25 24 266 0.09 AR bluespotted flathead F26 23 270 0.09 AR

Port Jackson shark PJ1 19 602 0.03 NR

Port Jackson shark PJ2 15 602 0.02 NR

Port Jackson shark PJ3 11 602 0.02 NR

Port Jackson shark PJ4 37 601 0.06 AR

Port Jackson shark PJ5 32 602 0.05 AR

Port Jackson shark PJ6 13 602 0.02 AR

Port Jackson shark PJ7 6 602 0.01 NR

Port Jackson shark PJ8 77 600 0.13 AR

Port Jackson shark PJ9 29 600 0.05 AR

Port Jackson shark PJ10 4 599 0.01 AR

Port Jackson shark PJ12 9 602 0.01 NR

Port Jackson shark PJ13 1 602 0 NR

Port Jackson shark PJ14 11 599 0.02 AR

Port Jackson shark PJ15 7 602 0.01 AR

Port Jackson shark PJ16 19 602 0.03 AR

151

Bluespotted flathead

Of the 25 bluespotted flatheads which were tagged at the AR, 17 individuals (68%) were detected at the AR for 2 or more days (Fig. 5.5). Three individuals (F2, F12 and

F19) were tagged and released at the AR but were undetected during the monitoring period. In addition two individuals (F5 and F22) were detected at the AR and between

Bondi and South Head but only had a few detections so these individuals were removed from further analysis. Three individual bluespotted flatheads (F1, F3, and F4) were tagged with short term tags (61 or 157 day tag life; Table 5.1) and two individuals (F3 and F9) were recaptured by anglers. Residency of bluespotted flatheads at the AR ranged from IR =0 to IR =0.47 and individuals were detected at this reef from 0 to 130 days. One individual (F20) which was tagged at the AR exhibited the highest residency at this reef over the monitoring period (IR =0.47, Table 5.2). Both the average residency of all individuals at the AR and the average long-term residency for individuals with a total monitoring period greater than 157 days at the AR was IR = 0.07 respectively. The median residency of these individuals was IR = 0.04. Only one individual (F1) was tagged at the NR, had an IR =0.1 and was detected at this site for 15 days (Table 5.2). Of the 9 individual bluespotted flatheads that were tagged at the AM, only two individuals

(F28 and F29) were detected at Bondi and Long Bay (2-10 km further south from this site). These individuals only had a few detections so were removed from further analysis. Of all individual bluespotted flatheads that were analysed, 11 individuals

(52%) were only detected at the AR during the monitoring period (Fig. 5.5). In contrast, one bluespotted flathead (F8) which was tagged at the AR, was only detected at Bondi and Sydney Harbour. The other individuals (43%) were detected at the AR and up to 6 additional sites (Fig. 5.5).

152

Figure 5.5. Presence plot of bluespotted flatheads monitored in the Sydney region from

August 2013 to April 2015. Black dots indicate detection at the artificial reef (AR), colour symbols indicate detection at other receivers (SG= Sydney gate, NR=Dunbar,

BL=Bondi line, BSH= between Bondi and South Head, SH= Sydney Harbour, SLB=

South long Bay). Grey crosses indicates end of tag life.

153

Port Jackson shark

Of the total 17 Port Jackson sharks tagged, 10 individuals (59%) were detected at the

AR for more than 2 days (Fig. 5.6). Two individuals (PJ11 and PJ17) were tagged at the

NR but had few detections so were removed from further analysis. Residency of Port

Jackson sharks at the AR ranged from IR =0.01 to IR =0.13 and individuals were detected from 4 to 77 days at this reef, with one individual (PJ8) exhibiting the highest residency at this reef over the monitoring period (Table 5.2). The average residency of all individuals at the AR was IR = 0.04, with a median IR = 0.03. Residency of Port

Jackson sharks at the NR ranged from IR=0 to IR =0.03 and individuals were detected from 1 to 19 days at this reef, with one individual (PJ1) exhibiting the highest residency over the monitoring period (Table 5.2). Both the average and median residency of all individuals at the NR was IR = 0.02. All tagged Port Jackson sharks were detected at 1 to 5 additional sites from their original tagging reef (Fig. 5.6). The majority of Port

Jackson sharks (87%) were detected at their tagging reefs irregularly from August 2013 when they were first tagged and detections ceased after October 2013. Five individuals

(33%) were detected again at their tagging reefs between 282 to 334 days after the last date of detection (Fig. 5.6).

154

Figure 5.6. Presence plot of Port Jackson sharks monitored in the Sydney region from

August 2013 to December 2014. Black dots indicate detection at the artificial reef (AR), colour symbols indicate detection at other receivers (SG= Sydney gate, NR=Dunbar,

BL=Bondi line, BSH= between Bondi and South Head, NAR= Narooma).

155

The proportion of total days detected at the AR and NR was highest by individuals that were tagged from each corresponding reef (Figs. 5.7, 5.8 and 5.9). In total, 96%, 99% and 98% of days spent at the AR were by fiddler rays, bluespotted flatheads and Port

Jackson sharks which were tagged near this reef respectively. Similarly, 99% and 92% of days spent near the north NR receiver were by fiddler rays and bluespotted flatheads which were tagged near this reef respectively, and 85% of days spent near the north NR receiver and 79% of days spent at the south NR receiver were by Port Jackson sharks tagged near the NR. In contrast, only 4%, 1% and 2% of days spent at the AR were by fiddler rays, bluespotted flatheads and Port Jackson sharks which were tagged near the

NR respectively. The total number of days detected at the AR by all three species was highest at this reef compared to other sites. However 1% of days spent at the SG2 receiver, and 3% of days spent at the SG1 were by fiddler rays and Port Jackson sharks tagged from the NR respectively. Individuals tagged from the AR also spent time at other sites in the study area during the monitoring period. The total number of days spent at these sites was between 1 to 9 days for fiddler rays, 1 to 37 days for bluespotted flatheads and 1 to 106 days for Port Jackson sharks. In addition, Port Jackson sharks tagged from the AR spent between 2 to 72 days over summer and autumn near

Narooma, located approximately 280 km further south (Fig. 5.9b).

156

Figure 5.7. Proportion of total days that fiddler rays tagged from either the AR

(artificial reef) or NR (Dunbar reef) were detected at receivers in the study area during the monitoring period. DBHN= north Dunbar reef, SG1-3= Sydney gate,

SYT409=between Bondi and South Head, BL1-3= Bondi.

157

Figure 5.8. Proportion of total days that bluespotted flatheads tagged from the either

AR (artificial reef) or NR (Dunbar reef) were detected at receivers in the study area during the monitoring period. DBHN=north Dunbar reef, SG1-4= Sydney gate,

SYT233= Sydney harbour, SYT409=between Bondi and South Head.

158

159

5.4.2 Connectivity

The distance from tagging reef at which the majority of fiddler rays, bluespotted flatheads and Port Jackson sharks were detected was up to 2 km, with 62%, 81% and

71% of individuals detected respectively (Fig. 5.10). The maximum distance that individuals were detected from their tagging reef was 7 to 8 km for fiddler rays (8% of individuals), 5 to 6 km for bluespotted flatheads (6% of individuals) and >8 km for Port

Jackson sharks (14% of individuals). The average distance detected from the original tagging reef was similar between fiddler rays and bluespotted flatheads, being 2.5 (±1.4

S.E.) km and 1.4 (±1.2) km respectively (Fig. 5.11). In contrast, the average distance detected from the tagging reef by Port Jackson sharks was 30.6 (±1.4) km. The median distances detected from the tagging reef by all three species however was similar, both fiddler rays and Port Jackson sharks had a median distance of 1.4 km and the median distance for bluespotted flatheads was 1.2 km.

160

Figure 5.10. Proportion of individuals per species detected and the corresponding distances from their tagging reef during the monitoring period.

161

162

5.5 Discussion

5.5.1 Residency

The three species examined in this study all displayed low residency, with average residency indices <0.5. However the total residence time by some individuals was relatively high over the long-term monitoring periods. Two individual fiddler rays and one bluespotted flathead had a moderate degree of residency at the AR and were detected for at least half of the total monitoring period, which ran up to ~20 months depending on the tagged individual. Presence at an estuarine AR by a similar benthic predator, the pink snapper (Chrysophrys auratus) was also observed to occur over a long period, possibly due to the increased access to food (Ross et al. 2007; Lowry et al.

2014). Other individuals also displayed high residence times at the AR in this study, but for short periods at the beginning of the monitoring period. Similarly, the fidelity of tagged red snapper (Lutjanus campechanus) to petroleum platforms in the Northern

Gulf of Mexico was found to be high in the short-term and lower over a long-term period of more than 3 years (Westmeyer et al. 2007). This was hypothesised to be due to the proximity and abundance of available habitat nearby, which provides alternative foraging and/or shelter opportunities for red snapper. Thus the residency of individuals at the AR may be attributed to foraging in the vicinity of this reef or other factors, which can potentially contribute to some biomass production at the AR.

It was apparent that the site of tagging influenced the residency of all three species, where the proportion of days spent at either the AR or NR was highest by individuals tagged from those reefs. This indicates site fidelity over small spatial scales, at least over short-term periods. In this study only one fiddler ray and one bluespotted flathead

163 were tagged at the NR. Conversely, several Port Jackson sharks were tagged from the

NR and all these individuals spent more time at this reef than any other site, including the AR. Likewise, the giant trevally (Caranx ignobilis) was observed to exhibit high reef fidelity to their tagging reef with limited movement to other reefs or regions in the

Great Barrier Reef (Lédée et al. 2015). Both the grey reef shark (Carcharhinus amblyrhynchos) and silvertip shark (C. albimarginatus) also spent most time present at or near their tagging reef (Vianna et al. 2013; Espinoza et al. 2015a; Espinoza et al.

2015b; Espinoza et al. 2015c). The majority of individuals in this study however were tagged at the AR and were mostly present at this reef. Red snapper (L. campechanus) tagged at ARs in a previous study were also found to be most resident at these reefs during the monitoring period (Szedlmayer and Schroepfer 2005). Further tagging studies are needed to evaluate the influence of the tagging reef by comparing more individuals tagged from the NR with those from the AR. It is also interesting that the

AR was deployed only 3 years earlier, yet many individuals > 3 years old (e.g. Port

Jackson sharks, Tovar-Ávila et al. 2009; Izzo and Rodda 2012) had incorporated this new structure into their home range and developed a degree of site fidelity.

A number of factors may determine the residency and distribution of species at ARs and adjacent natural reefs. These can include food availability, and habitat preference for shelter and/or spawning opportunities (Santos and Monteiro 1997; Santos and Monteiro

1998; Strelcheck et al. 2007). ARs may provide varying levels of these resources, depending on characteristics of the artificial structure, available habitat and fish behaviour (Topping and Szedlmayer 2011a; Topping and Szedlmayer 2011b). The AR in this study is unlikely to provide much benefit in terms of habitat for the fiddler ray and bluespotted flathead, since these are both sand-affiliated species (Hutchins and

164

Swainston 1986; Last and Stevens 1994; Rowling et al. 2010), but could possibly increase benthic food production at a very localised scale (Barros et al. 2001; Danovaro et al. 2002; Leitão et al. 2007). It is possible that these species used the sandy habitat that was available prior to the deployment of the AR, but take advantage of the altered substrate, as the bluespotted flathead is known to use both reef and sand-inundated reef

(Moore et al. 2009). Port Jackson sharks are more likely to associate closely with the

AR, and have been observed in the structure itself (Smith et al. 2016). This species is known to have a preference for shallow coastal rocky reefs for mating and oviposition

(Powter and Gladstone 2008b). Structured reef habitat is also suggested to facilitate social interactions between individual Port Jackson sharks, as well as provide protection against predation (Powter and Gladstone 2009). Likewise the NR may provide a heterogeneous environment favouring the presence of Port Jackson sharks and other species characteristic of soft and hard bottoms (Santos and Monteiro 2007). This is because this species has a preference for the sand/reef interface on the lee side of reefs which may act as a refuge against strong water movements (Hutchins and Swainston

1986; Powter and Gladstone 2008b).

Other factors which determine the site residency and movements of species include body size, life history, as well as environmental and biological drivers (Vianna et al.

2013; Henderson et al. 2014; Schlaff et al. 2014; Espinoza et al. 2015a; Espinoza et al.

2015b). Port Jackson sharks have seasonal movements related to the breeding season

(Powter and Gladstone 2008c), which did affect their residency. Most individuals were only detected at their tagging reefs at the beginning of the monitoring period (from

August to October 2013) and the majority were not detected until the following year.

Adult male and female Port Jackson sharks are known to migrate annually from deep

165 offshore waters to shallow coastal rocky reefs during the breeding season from July to

November (Powter and Gladstone 2008a). The majority of individuals tagged in this study were classified as sexually mature adult males (TL > 750 mm; McLaughlin and

O'Gower 1971; Powter and Gladstone 2008c; Powter et al. 2010) and possibly emigrated from offshore reefs to the AR, NR or nearby coastal reefs, to use these reefs during the breeding season. Furthermore a small proportion of individuals were detected again at the AR and/or NR approximately a year since the last detection at these reefs.

Many sharks are known to undertake seasonal migrations and display inter-annual site residency, returning to the same area every year (e.g. Vianna et al. 2013; Lee et al.

2015b). In previous studies Port Jackson sharks have also been resighted at the same reefs after multiple seasons indicating that they have a highly-developed spatial memory of refuge and oviposition locations (McLaughlin and O'Gower 1971; O'Gower 1995;

Powter and Gladstone 2009).

In this study, other individual Port Jackson sharks were either not detected after leaving their tagging reef or were detected at other sites, including Narooma. Reoccurrence at the same tagging reef by this species can sometimes occur up to 2 years (McLaughlin and O'Gower 1971). The majority of individual Port Jackson sharks that were tagged in this study were male and thus we were unable to quantify the differences between sexes.

Breeding females are reported to exhibit a male avoidance strategy during the reproductive season, whereas males tend to congregate together in deep water reefs during this period (Powter and Gladstone 2009). These findings suggest that the AR may provide suitable habitat for Port Jackson sharks, particularly males, even for a limited period. The potential for AR habitat preference by Port Jackson sharks should thus be taken into consideration when deploying these reefs.

166

5.5.2 Connectivity with natural reef

Connectivity was evident between the AR and nearby natural reefs, with all species exhibiting movements 5 km or more from their tagging reefs and visiting up to 5 or 6 other reef areas during the monitoring period. Most individuals however remained within 2 km from their tagging reefs. Similarly, the majority of tagged red snapper individuals were reported to have moved 2 km or less from the ARs where they were released (Strelcheck et al. 2007), although a separate study reported that the same species moved up to 8 km away to other sites (Topping and Szedlmayer 2011b). Fish assemblages associated with this AR have previously been observed at increased densities on a localised scale of less than 30m (Scott et al. 2015). This is suggested to be due to the close proximity of the AR to nearby natural reefs, thus promoting inter-reef connectivity. Closely-spaced habitats can promote the dispersal of mobile species, particularly those which have the ability to move over relatively large distances

(Workman et al. 2002; Chin et al. 2013; Espinoza et al. 2015c). Certain Port Jackson shark individuals in this study for instance were found to travel over 8 km from their tagging reefs, and the average distance travelled by all individuals was 30 km from the original tagging reef.

The furthest location that this species was detected at was Narooma, (located ~280 km south from the AR and NR). Port Jackson sharks travel several hundred kilometres during their migration between coastal and offshore habitats (McLaughlin and O'Gower

1971). This indicates that large-scale movements and high dispersal frequency can occur between reef areas by this species. In contrast, for more resident species with limited movements, sand can act as a barrier to movement, thus reducing inter-reef 167 connectivity (e.g. Fernandez et al. 2008; Koeck et al. 2013; Lee et al. 2015a). ARs have the potential to alter the seascape connectivity and habitat use of individuals by facilitating or restricting movements between habitat patches (Koeck et al. 2013).

Regular movements by all three species in this study however indicate that connectivity between this AR and natural reefs can occur on a localised scale.

A number of individual fiddler rays and bluespotted flatheads tagged at the AR were only detected at this reef during the monitoring period. Furthermore, a large number of individuals left the study site completely or were absent for long intervals during the monitoring period. Absence of individuals is not uncommon in acoustic telemetry studies and may be due to various factors, including transmitter failure, natural or fishing mortality, acoustic coverage, reproduction and movement to unmonitored areas of a reef and/or to unmonitored reefs within the study area (Heupel et al. 2006;

Espinoza et al. 2011a; Espinoza et al. 2015b; Hussey et al. 2015). The risk of fishing mortality is likely to be high for bluespotted flatheads since this species is targeted by recreational anglers (Henry and Lyle 2003; Rowling et al. 2010). In this study, two individual bluespotted flatheads were recaptured by anglers during the monitoring period. In contrast, individuals tagged with accelerometry tags (ie. V9A or V9AP) displayed various levels of activity throughout this study, indicating that mortality was unlikely (Fig. 4.2, chapter 4). However, the results from this study may have been compromised by a number of issues including 1) the lack of an acoustic array at the AR,

2) lack of data due to lost receivers, 3) the number of unmonitored reefs (i.e. nearby reefs without acoustic coverage), 4) possible fouling of the AR receiver, and 5) transmitter failure.

168

Movements to unmonitored areas could explain the absence of some individuals in this study. Wide-ranging species such as Port Jackson sharks for instance can leave an area without returning and/or spend long periods outside the detection range of receivers even if individuals remain within the study site (Espinoza et al. 2015c). This may also be the case for some bluespotted flatheads which exhibit strong localised movements in unmonitored areas such as the AM reef. However a lack of detections could also be as a result of transmitter failure or fishing mortality.

5.5.3 Influence of the AR

The residency by all three species at the AR suggests that only 4- 32% of their biomass production is likely to be derived from this AR, and this would likely decline further based on how these species interact with the reef itself. The local fish production at this

AR was recently estimated to be 350 g m-2 y-1 (Smith et al. 2016). This ascribed to the potential for this AR to support the production of reef-resident zooplanktivorous fish such as mado (Atypichthys strigatus) (Champion et al. 2015). The contribution to local benthic fish production however is expected to be low due to the low overall residency, but moderate residence time by the species in this study. This is because the fiddler ray and bluespotted flathead are likely to spend much of their time both near as well as away from the AR, whereas the Port Jackson shark may only occasionally occur inside the AR for short periods of time.

The combination of low residency and frequent movements by all three species suggest that the deployment of this AR may have altered the local distribution of these species by increasing the connectivity between adjacent habitats, therefore contributing to the distribution of their biomass. The strength of their association with the AR means that 169 these species would only contribute to a small fraction of the total estimated harvest from the AR (Chapter 3), since more dispersed species can be harder to catch (Smith et al. 2015). The AR may also aid in the dispersion of a range of recreationally-targeted species like the bluespotted flathead which has localised movements.

Comparisons with multiple ARs of similar design and location using an acoustic array need to be made. Future purpose-built reefs are currently being planned for coastal

NSW waters and the proximity between AR structures and natural reefs is a key consideration. Scott et al. (2015) suggest that multiple AR units could be deployed as close as 60 m to ensure reef connectivity while avoiding overlapping distributions of most associated fish species. The findings from this study indicate that ARs should be deployed no more than 2 km from adjacent reefs to promote connectivity for species recruitment and dispersion of biomass.

170

5.6 Conclusions

Movements of benthic species at ARs and nearby natural reefs using acoustic telemetry can improve the understanding of the AR ecosystem. The low residency at the AR by all individuals in this study indicates that at most, 4- 32% of their biomass production could be derived from this reef. Localised and large-scale movements by all species to 5 or 6 other reefs demonstrate that a degree of connectivity exists between the AR and nearby natural reefs. The distribution of these species is likely to have been altered by this AR through increased inter-reef connectivity, which highlights the need for telemetry studies around ARs. When planning future ARs, managers should consider the location and proximity of these structures to natural reef habitats to increase the likelihood of fish recruitment, as well as dispersal to reduce the risk of AR targeted fishing mortality.

Publication details:

Keller, K., Smith, J., Lowry, M, Taylor, M. and Suthers, I.M. (In review) Multispecies presence and connectivity around a designed artificial reef. Marine and Freshwater

Research.

171

172

Chapter 6

General discussion

6.1 Summary

The research presented here focused on monitoring the first offshore deployment of a large designed artificial reef (AR) in Australia. The assessment involved determining the utility of artificial structures as a means of enhancing recreational fishing quality as well as broader ecosystem effects such as the potential for ARs to contribute to secondary production. The research program included the estimated recreational fishing effort and harvest of fish by anglers from the AR fishery (Chapters 2 and 3) and examining the behaviour, movements and residency of 3 benthic species which co-exist in the vicinity of the AR (Chapters 4 and 5). The findings have been used by relevant regulatory authorities to understand the cost-benefit of future AR deployments for fisheries enhancement and how AR development is incorporated into a broader strategy for the management of fisheries resources. The results of this thesis are derived from an

AR in a temperate region of southeast Australia, but the methods and approaches are relevant for other AR and natural reef ecosystems, as well as the monitoring of other recreational fisheries.

These findings indicate that this AR has the potential to enhance recreational fisheries.

This is demonstrated through the comparisons of effort intensity and estimated harvest between the AR and estuarine fisheries, which revealed that ARs received higher levels of recreational usage and harvest than many natural estuarine fisheries (Chapter 2 and

3). Effort was lowest when the AR had just recently been deployed, thus effort can be

173 expected to increase over time with ongoing colonisation by sessile organisms and fishes at the structure (Bohnsack and Sutherland 1985; Bohnsack et al. 1994; Scott et al.

2015). Observations of a 10% increase in effort at this reef have already been observed since the initial two-year monitoring period, through the ongoing monitoring of the reef using camera imagery (K. Keller pers. obs.). An increase in effort however is not directly related to an increase in catch or harvest rates since this depends on both the catchability and the availability of the target species (Stoner 2004; Stehfest et al. 2015).

Catchability depends on various factors including fishing gear used, fisher behaviour, environmental variables and fish behaviour, whereas availability varies with local fish abundance (Stoner 2004; Stehfest et al. 2015).

Regardless, recent model simulations of this AR indicated that there is a potential for the overexploitation of current fish stocks at this reef (Smith et al. 2016). Furthermore, this research indicates that the estimated fishing effort and harvest may be concentrated at this fishery. The estimates of annual effort (Chapter 2) and harvest (700 kg) combined with the fish functional groups targeted at the AR (Chapter 3) indicate that ongoing monitoring at this fishery is necessary for sustainable fisheries management.

Camera-based technologies provide a cost-effective solution for monitoring fishing effort (e.g. Ames and Schlindler 2009; Smallwood et al. 2011; Smallwood et al. 2012;

Hartill et al. 2016), as long as the information derived from camera images is validated with some independent observations. Likewise, multiple datasets can be used to estimate recreational harvest, but only after all available resources, trade-offs between cost, and the reliability of harvest estimates have been considered, including potential biases. These methods have broad application to many other recreational fisheries around the world.

174

It was important to assess the diel activity, movements and residency of benthic fish to infer the local production around the AR. The findings from this research component revealed that both the bluespotted flathead and Port Jackson shark are nocturnal, and therefore may be at less risk of fishing, although night time fishing at the AR is unknown. In contrast, the fiddler ray was most active during the day (Chapter 4; Vianna et al. 2013; Villegas-Ríos et al. 2013), and it may be at a higher risk of capture by recreational fishers since it has a similar diet and habitat preference as the recreationally-targeted bluespotted flathead (Hutchins and Swainston 1986; Marshall et al. 2007). All three species exhibited relatively low residency and frequent movements between the AR to nearby natural reefs (Chapter 5) which reduces their vulnerability to the negative effects of attraction to the AR (Bohnsack 1989; Smith et al. 2015).

The current distribution of these species is likely to have been altered by this AR through the increased inter-reef connectivity which aids in the dispersion of their biomass. This is an important consideration when planning future AR deployments, as it indicates that this AR may potentially contribute to fisheries management by minimising the effects of fishing on these species (Milon 1989; Pickering and

Whitmarsh 1997; Whitmarsh et al. 2008; Bortone et al. 2011). The AR may also contribute to the dispersion of other similar benthic species that are recreationally targeted in NSW and have been observed at the AR, such as the pink snapper (C. auratus), blue morwong (N. douglasii) and silver trevally (P. dentex) (Appendix B;

Rowling et al. 2010; Lowry et al. 2015; Scott et al. 2015). Furthermore, the AR may influence the distribution of pelagic fish, including the yellowtail kingfish (S. lalandi), a transient species which has been observed in large numbers at the AR (pers. obs; Scott et al. 2015). This species was observed to exhibit a much lower residency at the AR

175 structure compared to the benthic species in this study, with an average residency of just

22 hours for nine individuals that were monitored over one year (S. Brodie pers. comm.).

The results from this research further the understanding of the role this AR plays in a temperate marine ecosystem. The low residency of the three benthic species at the AR in this study suggests that 4 to 32% of their biomass production may be derived from this reef (Chapter 5). Fish production is often dependent upon the critical density of

ARs which includes fish movement off, or between, nearby reefs during foraging

(Shipley and Cowan Jr 2011). Likewise the size and spacing of these reefs can alter fish production through growth rates, site residency, and population dynamics of reef fishes

(Strelcheck et al. 2007; Champion et al. 2015).

Interestingly, a simulation model on this AR estimated that only ~5% of ‘new’ fish production, that is the production that occurred as a result of the deployment of this AR, would have occurred relative to the existing local production (Smith et al. 2016).

However the estimated annual flux of biomass across this reef was very large, ~360 times greater than the standing-stock biomass. This suggests that this designed reef contributes to fish production at a larger scale than natural reef, and similar to oil platforms, can be considered as one of the most productive marine fish habitats (Claisse et al. 2014; Smith et al. 2016). This may be due to the large food production at the AR as a result of the structure’s exposure to ocean currents, as well as the association of zooplanktivores and large target pelagic fish (Champion et al. 2015; Scott et al. 2015).

Hence the low estimated biomass production from my findings on these three species suggests that this AR has a minor role in supporting local benthic fish communities.

176

This reef may instead have a more important role for other species which are more reef- associated, such as mado (Atypichthys strigatus) and yellowtail scad (Trachurus novaezelandiae) (Champion et al. 2015).

6.2 Future research

The research presented here was limited to three benthic predators that were associated with an AR, however initially a number of other recreationally important species were planned to also be incorporated. These included the blue morwong (N. douglasii), silver trevally (P. dentex), pink snapper (C. auratus), sixspine leatherjacket (Meuschenia freycineti) and ocean leatherjacket (Nelusetta ayraud), which have all been observed in large numbers at the AR (Lowry et al. 2015; Scott et al. 2015). Movements and residency behaviour of the pink snapper have previously been examined in some acoustic tagging studies (e.g. Egli and Babcock 2004; Lowry et al. 2014). With improved capture techniques and prevention of barotrauma, future studies will be able to incorporate these species into monitoring studies at the AR to determine the degree at which these species interact with the structure. Identification of the fish-AR relationship

(i.e. spawning, feeding, refuge and attraction) for these recreationally important species can help to further assess the economic value and contribution of this AR to fisheries enhancement and local production.

Other non-recreationally targeted species may also use the AR and may be just as important in contributing to the local production at this reef. One such species is the blind shark (Brachaelurus waddi), an abundant cryptic rocky reef species endemic to the east coast of Australia (Last and Stevens 1994). To date, little information is available regarding this species’ current population trends, biology or ecology, except 177 anecdoctal reports suggest that it is a relatively sessile and nocturnal species and may be a significant bycatch component of the inshore trap fishery (Kuiter 2000; Macbeth et al.

2009; Pedersen 2012). The blind shark was tagged as part of a separate research project to investigate its movements and residency patterns in temperate coastal rocky reefs in

Sydney, and some individuals were detected in the vicinity of the AR during the montioring period. Preliminary results indicate that, similar to the Port Jackson shark, the blind shark and other cryptic species may be an important component of the temperate rocky reef ecosystem and may occasionally use the AR habitat. This highlights the need for further tagging studies as this could provide critical information to direct conservation planning, and also the ecosystem effects that these nocturnal and uncensused benthic predators may have.

Further studies examining the fine-scale movements and habitat use of fish associated with ARs is needed to determine whether these reefs provide suitable habitat for these fish assemblages. The Vemco VR2W Positioning System (VPS) can be useful for providing high resolution data of animal movements and habitat use in these environments (How and de Lestang 2012; Vemco 2016). The VPS involves 3 or more precisely located and synchronized receivers to provide an instantaneous estimate of a tagged animals’ location with up to 95% accuracy. The difficulty of depth and precisely fine-tuning the re-location of receivers prevented the deployment of a VPS at the AR.

VPS tracking has been used to determine a species’ habitat preference in marine environments (e.g. Heupel et al. 2006; Espinoza et al. 2011a; Espinoza et al. 2011b).

Likewise, increasing the acoustic coverage across the study region will assist in determining the scale of connectivity between this AR and nearby unmonitored natural reefs, including the nearby Annie Miller wreck (AM), which will better inform

178 managers on the placement of ARs in the future. Interestingly we tagged nine bluespotted flathead at this wreck but they were not recorded at the AR just 3 km further north.

There was only one AR in this study, therefore comparisons with multiple ARs of similar design and location need to be made to validate the observed patterns in fish residency and distribution, as these patterns may be site-specific. A number of concrete designed multiple-AR clusters have been recently deployed >2 km off the Shoalhaven

River, and Port Macquarie in NSW since the deployment of the Sydney AR, and more are being planned for deployment off Botany Bay in the near future (Folpp 2015; NSW

DPI 2015). Future research of these reef deployments will benefit from quantifying the distribution of fish around these multiple reef units, particularly as production is suggested to increase with the addition of reefs, and depends on the size and shape of these structures (Brickhill et al. 2005; Leitão et al. 2007; Champion et al. 2015). Reef clusters could also act as nursery areas for economically important species by providing refuge from fishing (or if fishing is banned) (Brickhill et al. 2005).

Fishing effort from the AR in this study in comparison with other new AR fisheries with a less urbanised coast can further our understanding of the spatial and temporal changes in effort at these reefs. For example, an integrated survey design using a combination of aerial-roving surveys together with shore-based cameras (e.g. Smallwood et al. 2011;

Smallwood et al. 2012) can be used to record fishing activity at multiple ARs. The use of cameras can assist in identifying peak times of fishing activity to determine critical times for scheduling aerial flight surveys. This effort information is important to inform managers about which ARs are receiving high levels of fishing intensity, especially

179 after implementation of a sanctuary zone. Similarly, comparing catch rates between AR fisheries is needed to determine the benefit of these ARs to recreational fishers. The harvest rates used to estimate harvest from the AR in this study was collected prior to the deployment of this AR, therefore it is recommended that a creel survey (depending on available budget) is conducted prior to and after the deployment of future ARs. This

Before-After Control-Impact survey design can help managers measure any change in harvest to determine whether these reefs do enhance recreational fisheries.

The continuing advancement in camera technology provides a variety of options for the improved monitoring of effort. One such technology is the Gigapan camera system that can be deployed for extended periods to capture high resolution digital panoramas of small-scale temporal and spatial trends in recreational fishing effort (Lynch et al. 2015;

Wood et al. 2016). This system allows for precise geo-referencing of fishing vessels across multiple kilometre scales, identification of boat registrations, party size, user type and other identifying features by sampling at the rate of 1 image every 15 minutes.

These features can improve the monitoring of effort at nearshore fisheries including the

AR. Similarly, the development of technologies such as thermal imaging cameras and radar can increase the monitoring of recreational fishing activities in poor visibility conditions and during peak night time periods, which were limited by the camera used in this research. However these systems have certain limitations including relatively high equipment costs (e.g. Wood et al. 2016), which requires consideration for future monitoring studies. In contrast, computer software programs are becoming increasingly available which can reduce the image processing time by using algorithms to automate the boat counting procedure for monitoring effort (Shapiro and Stockman 2001; Hartill et al. 2015; Hartill et al. 2016). This enables counts to be provided in real time without

180 the need for any subsampling, thereby making such systems more cost-effective.

Monitoring recreational fishing effort and harvest is a priority for the management of

ARs and a combination of cost-effective methods and tools will assist managers in determining the economic benefit that these reefs provide to the fishing community as well as information to direct conservation and spatial planning.

181

References

Abecasis, D., Bentes, L., Lino, P.G., Santos, M.N., and Erzini, K. (2013) Residency, movements and habitat use of adult white seabream (Diplodus sargus) between natural and artificial reefs. Estuarine, Coastal and Shelf Science 118, 80-85.

Airoldi, L., Abbiati, M., Beck, M.W., Hawkins, S.J., Jonsson, P.R., Martin, D., Moschella, P.S., Sundelöf, A., Thompson, R.C., and Åberg, P. (2005) An ecological perspective on the deployment and design of low-crested and other hard coastal defence structures. Coastal Engineering 52(10–11), 1073-1087.

Ajemian, M.J., Wetz, J.J., Shipley-Lozano, B., Shively, J.D., and Stunz, G.W. (2015) An analysis of artificial reef fish community structure along the northwestern gulf of Mexico shelf: Potential impacts of “Rigs-to-Reefs” programs. PLoS ONE 10(5), e0126354.

Alevizon, W.S., and Gorham, J.C. (1989) Effects of artificial reef deployment on nearby resident fishes. Bulletin of Marine Science 44(2), 646-661.

Ambrose, R.F., and Swarbrick, S.L. (1989) Comparison of fish assemblages on artificial and natural reefs off the coast of southern California. Bulletin of Marine Science 44(2), 718-733.

Ames, R.T., and Schlindler, E. (2009) Video monitoring of ocean recreational fishing effort. Oregon, USA.

Anderson, D.R. (2008) 'Model based inference in the life sciences: A primer on evidence.' (Springer Science & Business Media: New York)

Annese, D.M., and Kingsford, M.J. (2005) Distribution, movements and diet of nocturnal fishes on temperate reefs. Environmental Biology of Fishes 72(2), 161-174.

Baine, M. (2001) Artificial reefs: a review of their design, application, management and performance. Ocean & Coastal Management 44(3-4), 241-259.

Barnes, L., Leclerc, M., Gray, C., and Williamson, J. (2011) Dietary niche differentiation of five sympatric species of Platycephalidae. Environmental Biology of Fishes 90(4), 429-441.

Barros, F., Underwood, A.J., and Lindegarth, M. (2001) The influence of rocky reefs on structure of benthic macrofauna in nearby soft-sediments. Estuarine, Coastal and Shelf Science 52(2), 191-199.

Barton, K. (2015) MuMIn: Multi-model inference. R package version 1.15.1. edn.

Blumenfeld, D. (2001) Means and variances. In Operations research calculations handbook. (Ed. D Blumenfeld). (CRC Press LLC: Boca Raton, Florida)

Bohnsack, J.A. (1989) Are high densities of fishes at artificial reefs the result of habitat limitation or behavioral preference? Bulletin of Marine Science 44(2), 631-645.

182

Bohnsack, J.A., Harper, D.E., McClellan, D.B., and Hulsbeck, M. (1994) Effects of reef size on colonization and assemblage structure of fishes at artificial reefs off southeastern Florida, USA. Bulletin of Marine Science 55(3), 796-823.

Bohnsack, J.A., and Sutherland, D.L. (1985) Artificial reef research: a review with recommendations for future priorities. Bulletin of Marine Science 37(1), 11-39.

BOM (2015) Climate Data Online: Rainfall. Vol. 2015. (Bureau of Meteorology)

Bortone, S.A., Brandini, F.P., and Otake, S. (2011) 'Artificial reefs in fisheries management ' (CRC Press: Boca Raton ) 332

Branden, K.L., Pollard, D.A., and Reimers, H.A. (1994) A review of recent artificial reef developments in Australia. Bulletin of Marine Science 55(3), 982-994.

Bray, G.S., and Schramm, H.L. (2001) Evaluation of a statewide volunteer angler diary program for use as a fishery assessment tool. North American Journal of Fisheries Management 21(3), 606-615.

Brickhill, M.J., Lee, S.Y., and Connolly, R.M. (2005) Fishes associated with artificial reefs: attributing changes to attraction or production using novel approaches. Journal of Fish Biology 67, 53-71.

Brownscombe, J.W., Gutowsky, L.F.G., Danylchuk, A.J., and Cooke, S.J. (2014) Foraging behaviour and activity of a marine benthivorous fish estimated using tri-axial accelerometer biologgers. Marine Ecology Progress Series 505, 241-251.

Buchanan, C.C. (1973) Effects of an artificial habitat on the marine sport fishery and economy of MurrellsInlet, South Carolina. Marine Fisheries Review 35(9), 15-22.

Bucher, D.J. (2006) Spatial and temporal patterns of recreational angling effort in a warm- temperate Australian estuary. Geographical Research 44(1), 87-94.

Burchmore, J.J., Pollard, D.A., Bell, J.D., Middleton, M.J., Pease, B.C., and Matthews, J. (1985) An ecological comparison of artificial and natural rocky reef fish communities in Botany Bay, New South Wales, Australia. Bulletin of Marine Science 37(1), 70-85.

Campbell, H.A., Watts, M.E., Dwyer, R.G., and Franklin, C.E. (2012) V-Track: software for analysing and visualising animal movement from acoustic telemetry detections. Marine and Freshwater Research 63(9), 815-820.

Carr, M.H., and Hixon, M.A. (1997) Artificial Reefs: The importance of comparisons with natural reefs. Fisheries 22(4), 28-33.

Carter, D.W., Crosson, S., and Liese, C. (2015) Nowcasting intraseasonal recreational fishing harvest with internet search volume. PLoS ONE 10(9), e0137752.

Cenci, E., Pizzolon, M., Chimento, N., and Mazzoldi, C. (2011) The influence of a new artificial structure on fish assemblages of adjacent hard substrata. Estuarine, Coastal and Shelf Science 91, 133-149.

183

Champion, C., Suthers, I.M., and Smith, J.A. (2015) Zooplanktivory is a key process for fish production on a coastal artificial reef. Marine Ecology Progress Series 541, 1-14.

Charnov, E.L. (1976) Optimal foraging, the marginal value theorem. Theoretical Population Biology 9(2), 129-36.

Chin, A., Heupel, M.R., Simpfendorfer, C.A., and Tobin, A.J. (2013) Ontogenetic movements of juvenile blacktip reef sharks: evidence of dispersal and connectivity between coastal habitats and coral reefs. Aquatic Conservation: Marine and Freshwater Ecosystems 23(3), 468-474.

Christensen, V., and Pauly, D. (1992) ECOPATH II — a software for balancing steady-state ecosystem models and calculating network characteristics. Ecological Modelling 61(3–4), 169- 185.

Claisse, J.T., Pondella, D.J., Love, M., Zahn, L.A., Williams, C.M., Williams, J.P., and Bull, A.S. (2014) Oil platforms off California are among the most productive marine fish habitats globally. Proceedings of the National Academy of Sciences 111(43), 15462-15467.

Claydon, J.A.B., McCormick, M.I., and Jones, G.P. (2014) Multispecies spawning sites for fishes on a low-latitude coral reef: spatial and temporal patterns. Journal of Fish Biology 84(4), 1136- 1163.

Cochran, W.G. (1977) 'Sampling techniques.' (John Wiley and Sons: New York)

Coleman, N., and Mobley, M. (1984) Diets of commercially exploited fish from Bass Strait and adjacent Victorian Waters, south-eastern Australia. Marine and Freshwater Research 35(5), 549-560.

Connell, S.D., and Anderson, M.J. (1999) Predation by fish on assemblages of intertidal epibiota: effects of predator size and patch size. Journal of Experimental Marine Biology and Ecology 241(1), 15-29.

Connelly, N.A., and Brown, T.L. (1996) Using diaries to estimate fishing effort and fish consumption: A contemporary assessment. Human Dimensions of Wildlife 1(1), 22-34.

Conron, S., and Bridge, N. (2004) Statewide angler diary program 1997 –2003. Fisheries Victoria.

Cooke, S.J., and Cowx, I.G. (2004) The role of recreational fishing in global fish crises. BioScience 54(9), 857-859.

Cresson, P., Ruitton, S., and Harmelin-Vivien, M. (2014) Artificial reefs do increase secondary biomass production: mechanisms evidenced by stable isotopes. Marine Ecology Progress Series 509, 15-26.

Curtis, J.M., Johnson, M.W., Diamond, S.L., and Stunz, G.W. (2015) Quantifying delayed mortality from barotrauma impairment in discarded red snapper using acoustic telemetry. Marine and Coastal Fisheries: Dynamics, Management, and Ecosystem Science 7, 434-449.

Danovaro, R., Gambi, C., Mazzola, A., and Mirto, S. (2002) Influence of artificial reefs on the surrounding infauna: analysis of meiofauna. ICES Journal of Marine Science 59, S356–S362. 184

Dempster, T. (2005) Temporal variability of pelagic fish assemblages around fish aggregation devices: biological and physical influences. Journal of Fish Biology 66(5), 1237-1260.

Dwyer, R.G., Watts, M.E., Campbell, H.A., and Franklin, C.E. (2015) VTrack: A collection of tools for the analysis of remote acoustic telemetry data. Vol. R package version 1.11. 2015 edn.

Edgar, G.J., Barrett, N.S., and Morton, A.J. (2004) Biases associated with the use of underwater visual census techniques to quantify the density and size-structure of fish populations. Journal of Experimental Marine Biology and Ecology 308(2), 269-290.

Edgar, G.J., and Stuart-Smith, R.D. (2014) Systematic global assessment of reef fish communities by the Reef Life Survey program. Scientific Data 1, 140007.

Egli, D.P., and Babcock, R.C. (2004) Ultrasonic tracking reveals multiple behavioural modes of snapper (Pagrus auratus) in a temperate no-take marine reserve. ICES Journal of Marine Science 61(7), 1137.

Espinoza, M., Farrugia, T.J., and Lowe, C.G. (2011a) Habitat use, movements and site fidelity of the gray smooth-hound shark (Mustelus californicus Gill 1863) in a newly restored southern California estuary. Journal of Experimental Marine Biology and Ecology 401(1-2), 63-74.

Espinoza, M., Farrugia, T.J., Webber, D.M., Smith, F., and Lowe, C.G. (2011b) Testing a new acoustic telemetry technique to quantify long-term, fine-scale movements of aquatic animals. Fisheries Research 108(2–3), 364-371.

Espinoza, M., Heupel, M.R., Tobin, A., and Simpfendorfer, C. (2015a) Residency patterns and movements of grey reef sharks (Carcharhinus amblyrhynchos) in semi-isolated coral reef habitats. Marine Biology 162(2), 343-358.

Espinoza, M., Heupel, M.R., Tobin, A.J., and Simpfendorfer, C.A. (2015b) Movement patterns of silvertip sharks (Carcharhinus albimarginatus) on coral reefs. Coral Reefs 34(3), 807-821.

Espinoza, M., Lédée, E.J.I., Simpfendorfer, C.A., Tobin, A.J., and Heupel, M.R. (2015c) Contrasting movements and connectivity of reef-associated sharks using acoustic telemetry: implications for management. Ecological Applications 25(8), 2101-2118.

Fabi, G., and Fiorentini, L. (1994) Comparison between an artificial reef and a control site in the Adriatic Sea: analysis of four years of monitoring. Bulletin of Marine Science, 55 2(3), 538-558.

Feeney, W.E., Lönnstedt, O.M., Bosiger, Y., Martin, J., Jones, G.P., Rowe, R.J., and McCormick, M.I. (2012) High rate of prey consumption in a small predatory fish on coral reefs. Coral Reefs 31(3), 909-918.

Fernandez, T.V., D'Anna, G., Badalamenti, F., and Pérez-Ruzafa, A. (2008) Habitat connectivity as a factor affecting fish assemblages in temperate reefs. Aquatic Biology 1, 239-248.

Folpp, H. (2015) Shoalhaven offshore artificial reef– Long term management plan. NSW Department of Primary Industries, NSW.

185

Folpp, H., and Lowry, M. (2006) Factors affecting recreational catch rates associated with a fish aggregating device (FAD) off the NSW coast, Australia. Bulletin of Marine Science 78(1), 185- 193.

Folpp, H., Lowry, M., Gregson, M., and Suthers, I.M. (2011) Colonization and community development of fish assemblages associated with estuarine artificial reefs. Brazilian Journal of Oceanography 59(Special Issue 1 (CARAH)), 55-67.

Folpp, H., Lowry, M., Gregson, M., and Suthers, I.M. (2013) Fish assemblages on estuarine artificial reefs: natural rocky-reef mimics or discrete assemblages? PLoS ONE 8(6), e63505.

Gannon, R. (2016) Habitat and spatial ecology of two iconic estuarine fishes using acoustic telemetry. Doctor of Philosophy Thesis, University of New South Wales, Sydney, NSW

Gannon, R., Payne, N.L., Suthers, I.M., Gray, C.A., Meulen, D.E., and Taylor, M.D. (2015) Fine- scale movements, site fidelity and habitat use of an estuarine dependent sparid. Environmental Biology of Fishes 98(6), 1599-1608.

Gannon, R., Taylor, M.D., Suthers, I.M., Gray, C.A., van der Meulen, D.E., Smith, J.A., and Payne, N.L. (2014) Thermal limitation of performance and biogeography in a free-ranging ectotherm: insights from accelerometry. The Journal of Experimental Biology 217, 3033-3037.

Gjelland, K.Ø., and Hedger, R.D. (2013) Environmental influence on transmitter detection probability in biotelemetry: developing a general model of acoustic transmission. Methods in Ecology and Evolution 4(7), 665-674.

Gray, C.A., Gale, V.J., Stringfellow, S.L., and Raines, L.P. (2002) Variations in sex, length and age compositions of commercial catches of Platycephalus fuscus (Pisces : Platycephalidae) in New South Wales, Australia. Marine and Freshwater Research 53(7), 1091-1100.

Griffiths, S.P. (2012) Recreational catch composition, catch rates, effort and expenditure in a specialised land-based pelagic game fish fishery. Fisheries Research 127–128, 40-44.

Grossman, G.D., Jones, G.P., and Seaman, W.J. (1997) Do artificial reefs increase regional fish production? A review of existing data. Fisheries 22(4), 17-23.

Hartill, B.W., and Edwards, C.T.T. (2015) Comparison of recreational harvest estimates provided by onsite and offsite surveys: detecting bias and corroborating estimates. Canadian Journal of Fisheries and Aquatic Sciences 72(9), 1379-1389.

Hartill, B.W., Payne, G., Rush, N., and Bian, R. (2016) Bridging the temporal gap: Continuous and cost-effective monitoring of dynamic recreational fisheries by web cameras and creel surveys. Fisheries Research 183, 488–497.

Hartill, B.W., Rush, N., Bian, R., Miller, A., Payne, G., and Armiger, H. (2015) Web camera and creel survey monitoring of recreational fisheries in FMAs 1, 8, and 9. Ministry for Primary Industries, Wellington, NZ.

Henderson, M.J., Fabrizio, M.C., and Lucy, J.A. (2014) Movement patterns of summer flounder near an artificial reef: Effects of fish size and environmental cues. Fisheries Research 153, 1-8.

186

Henningsen, A.D. (1994) Tonic immobility in 12 elasmobranchs: Use as an aid in captive husbandry. Zoo Biology 13(4), 325-332.

Henry, G.W., and Lyle, J.M. (Eds) (2003) 'The national recreational and indigenous fishing Survey.' Final report to the Fisheries Research and Development Corporation (Fisheries Research and Development Corporation: Canberra (Australia))

Heupel, M.R., Reiss, K.L., Yeiser, B.G., and Simpfendorfer, C.A. (2008) Effects of biofouling on performance of moored data logging acoustic receivers. Limnology and Oceanography: Methods 6(7), 327-335.

Heupel, M.R., Semmens, J.M., and Hobday, A.J. (2006) Automated acoustic tracking of aquatic animals: scales, design and deployment of listening station arrays. Marine and Freshwater Research 57(1), 1-13.

Holland, N.K., Wetherbee, M.B., Lowe, G.C., and Meyer, G.C. (1999) Movements of tiger sharks (Galeocerdo cuvier) in coastal Hawaiian waters. Marine Biology 134(4), 665-673.

How, J.R., and de Lestang, S. (2012) Acoustic tracking: issues affecting design, analysis and interpretation of data from movement studies. Marine and Freshwater Research 63(4), 312- 324.

Hsieh, C.-h., Reiss, C.S., Hunter, J.R., Beddington, J.R., May, R.M., and Sugihara, G. (2006) Fishing elevates variability in the abundance of exploited species. Nature 443(7113), 859-862.

Hughes, J.M., and Stewart, J. (2013) Assessment of barotrauma and its mitigation measures on the behaviour and survival of snapper and mulloway. NSW Department of Primary Industries Mosman NSW.

Hussey, N.E., Kessel, S.T., Aarestrup, K., Cooke, S.J., Cowley, P.D., Fisk, A.T., Harcourt, R.G., Holland, K.N., Iverson, S.J., Kocik, J.F., Mills Flemming, J.E., and Whoriskey, F.G. (2015) Aquatic animal telemetry: A panoramic window into the underwater world. Science 348(6240), 1-10.

Hutchins, B., and Swainston, R. (1986) 'Sea fishes of southern Australia. Complete field guide for anglers and divers.' (Swainston Publishing: Perth, Australia)

Huveneers, C. (2015) Trygonorrhina fasciata. The IUCN Red List of Threatened Species 2015. Vol. 2015.

Huveneers, C., and Simpfendorfer, C.A. (2015) Heterodontus portusjacksoni. The IUCN Red List of Threatened Species 2015:. Vol. 2015.

Izzo, C., and Gillanders, B.M. (2008) Initial assessment of age, growth and reproductive parameters of the southern fiddler ray Trygonorrhina fasciata (Müller & Henle, 1841) from South Australia. Pan-American Journal of Aquatic Sciences 3(3), 321-327.

Izzo, C., and Rodda, K.R. (2012) Comparative rates of growth of the Port Jackson shark throughout its southern Australian range. Marine and Freshwater Research 63(8), 687-694.

187

Jennings, S., and Kaiser, M.J. (1998) The Effects of Fishing on Marine Ecosystems. In Advances in Marine Biology. Vol. Volume 34. (Eds. AJS J.H.S. Blaxter and PA Tyler) pp. 201-352. (Academic Press)

Jennings, S., and Polunin, N.V.C. (1995) Biased underwater visual census biomass estimates for target-species in tropical reef fisheries. Journal of Fish Biology 47(4), 733-736.

Johnson, T.D., Barnett, A.M., DeMartini, E.E., Craft, L.L., Ambrose, R.F., and Purcell, L.J. (1994) Fish production and habitat utilization on a southern california artificial reef. Bulletin of Marine Science 55(2-3), 709-723.

Jones, A.A., Hall, N.G., and Potter, I.C. (2010) Species compositions of elasmobranchs caught by three different commercial fishing methods off southwestern Australia, and biological data for four abundant bycatch species. Fishery Bulletin 108(4), 365-380.

Keller, K., Steffe, A.S., Lowry, M., Murphy, J.J., and Suthers, I.M. (2016) Monitoring boat-based recreational fishing effort at a nearshore artificial reef with a shore-based camera. Fisheries Research 181, 84–92.

Kerr, S. (1992) 'Artificial reefs in Australia. Their construction, location and function.' (Bureau of Rural Resources: Canberra (Australia))

Kessel, S.T., Chapman, D.D., Franks, B.R., Gedamke, T., Gruber, S.H., Newman, J.M., White, E.R., and Perkins, R.G. (2014) Predictable temperature-regulated residency, movement and migration in a large, highly mobile marine predator (Negaprion brevirostris). Marine Ecology Progress Series 514, 175-190.

Kim, C.G., Kim, H.S., Baik, C.I., Kakimoto, H., Seaman, W., Nielsen, J.L., Dodson, J.J., Friedland, K., Hamon, T.R., and Musick, J. Design of artificial reefs and their effectiveness in the fisheries of eastern Asia. 2008,

Koeck, B., Alós, J., Caro, A., Neveu, R., Crec'hriou, R., Saragoni, G., and Lenfant, P. (2013) Contrasting fish behavior in artificial seascapes with implications for resources conservation. PLoS ONE 8(7), e69303.

Kruse, M., Taylor, M., Muhando, C.A., and Reuter, H. (2015) Lunar, diel, and tidal changes in fish assemblages in an east African marine reserve. Regional Studies in Marine Science 3, 49- 57.

Kuiter, R.H. (2000) 'Coastal fishes of south-eastern Australia.' (Gary Allen) 437

Last, P.R., and Stevens, J.D. (1994) 'Sharks and rays of Australia.' (CSIRO Division of Fisheries: Hobart, Tasmania, Australia)

Lédée, E.J., Heupel, M.R., Tobin, A.J., and Simpfendorfer, C.A. (2015) Movements and space use of giant trevally in coral reef habitats and the importance of environmental drivers. Animal Biotelemetry 3(1), 1-14.

Lee, K.A., Huveneers, C., Macdonald, T., and Harcourt, R.G. (2015a) Size isn't everything: movements, home range, and habitat preferences of eastern blue gropers (Achoerodus viridis)

188 demonstrate the efficacy of a small marine reserve. Aquatic Conservation: Marine and Freshwater Ecosystems 25(2), 174-186.

Lee, K.A., Huveneers, C., Peddemors, V., Boomer, A., and Harcourt, R. (2015b) Born to be free? Assessing the viability of releasing captive-bred wobbegongs to restock depleted populations. Frontiers in Marine Science 2(18).

Leitão, F. (2013) Artificial reefs: from ecological processes to fishing enhancement tools. Brazilian Journal of Oceanography 61, 77-81.

Leitão, F., Santos, M.N., and Monteiro, C.C. (2007) Contribution of artificial reefs to the diet of the white sea bream (Diplodus sargus). ICES Journal of Marine Science.

Lewin, W.C., Arlinghaus, R., and Mehner, T. (2006) Documented and potential biological impacts of recreational fishing: Insights for management and conservation. Reviews in Fisheries Science 14, 305-367.

Lowry, M., and Folpp, H. (2014) Sydney offshore artificial reef – Annual environmental monitoring report 2012-13. NSW Department of Primary Industries, Sydney.

Lowry, M., Folpp, H., and Becker, A. (2015) Sydney offshore artificial reef – Environmental monitoring final report. NSW Department of Primary Industries, Sydney.

Lowry, M., Steffe, A., and Williams, D. (2006) Relationships between bait collection, bait type and catch: A comparison of the NSW trailer-boat and gamefish-tournament fisheries. Fisheries Research 78(2-3), 266-275.

Lowry, M.B., Glasby, T.M., Boys, C.A., Folpp, H., Suthers, I., and Gregson, M. (2014) Response of fish communities to the deployment of estuarine artificial reefs for fisheries enhancement. Fisheries Management and Ecology 21(1), 42-56.

Lynch, T.P., Alderman, R., and Hobday, A.J. (2015) A high-resolution panorama camera system for monitoring colony-wide seabird nesting behaviour. Methods in Ecology and Evolution 6(5), 491-499.

Macbeth, W.G., Geraghty, P.T., Peddemors, V.M., and Gray, C.A. (2009) Observer-based study of targeted commercial fishing for large shark species in waters off northern New South Wales. Industry & Investment Cronulla, NSW, Australia.

Macreadie, P.I., Fowler, A.M., and Booth, D.J. (2011) Rigs-to-reefs: will the deep sea benefit from artificial habitat? Frontiers in Ecology and the Environment 9(8), 455-461.

Marshall, L.J., White, W.T., and Potter, I.C. (2007) Reproductive biology and diet of the southern fiddler ray, Trygonorrhina fasciata (Batoidea : Rhinobatidae), an important trawl bycatch species. Marine and Freshwater Research 58(1), 104-115.

McGlennon, D., and Branden, K.L. (1994) Comparison of catch and recreational anglers fishing on artificial reefs and natural seabed in Gulf St. Vincent, South Australia. Bulletin of Marine Science 55(2-3), 510-523.

189

McLaughlin, R.H., and O'Gower, A.K. (1971) Life history and underwater studies of a heterodont shark. Ecological Monographs 41(4), 271-289.

Metcalfe, N.B., Fraser, N.H.C., and Burns, M.D. (1999) Food availability and the nocturnal vs. diurnal foraging trade-off in juvenile salmon. Journal of Animal Ecology 68(2), 371-381.

Milon, J.W. (1989) Economic evaluation of artificial habitat for fisheries: Progress and challenges. Bulletin of Marine Science 44(2), 831-843.

Moore, C.H., Harvey, E.S., and Van Niel, K.P. (2009) Spatial prediction of demersal fish distributions: enhancing our understanding of species–environment relationships. ICES Journal of Marine Science: Journal du Conseil 66(9), 2068-2075.

NSW DPI (2015) Port Macquarie offshore artificial reef– Long term management plan. NSW Department of Primary Industries, NSW.

O'Gower, A. (1995) Speculations on a spatial memory for the Port Jackson shark (Heterodontus portusjacksoni) (Meyer) (Heterodontidae). Marine and Freshwater Research 46(5), 861-871.

Osenberg, C.W., St. Mary, C.M., Wilson, J.A., and Lindberg, W.J. (2002a) A quantitative framework to evaluate the attraction–production controversy. ICES Journal of Marine Science: Journal du Conseil 59(suppl), S214-S221.

Osenberg, C.W., St. Mary, C.M., Wilson, J.A., and Lindberg, W.J. (2002b) A quantitative framework to evaluate the attraction–production controversy. ICES Journal of Marine Science 59(suppl), S214.

Otway, N.M., Sullings, D.J., and Lenehan, N.W. (1996) Trophically-based assessment of the impacts of deepwater sewage disposal on a demersal fish community. Environmental Biology of Fishes 46(2), 167-183.

Palardy, J.E., and Witman, J.D. (2011) Water flow drives biodiversity by mediating rarity in marine benthic communities. Ecol Lett 14(1), 63-8.

Parsley, M.J., Popoff, N.D., Wright, C.D., and van der Leeuw, B.K. (2008) Seasonal and diel movements of white sturgeon in the lower Columbia River. Transactions of the American Fisheries Society 137(4), 1007-1017.

Pauly, D., Christensen, V., and Walters, C. (2000) Ecopath, Ecosim, and Ecospace as tools for evaluating ecosystem impact of fisheries. ICES Journal of Marine Science: Journal du Conseil 57(3), 697-706.

Payne, N.L., Gillanders, B.M., Seymour, R.S., Webber, D.M., Snelling, E.P., and Semmens, J.M. (2011) Accelerometry estimates field metabolic rate in giant Australian cuttlefish Sepia apama during breeding. Journal of Animal Ecology 80(2), 422-30.

Payne, N.L., Gillanders, B.M., Webber, D.M., and Semmens, J.M. (2010) Interpreting diel activity patterns from acoustic telemetry: the need for controls. Marine Ecology Progress Series 419, 295-301.

190

Payne, N.L., Smith, J.A., van der Meulen, D.E., Taylor, M.D., Watanabe, Y.Y., Takahashi, A., Marzullo, T.A., Gray, C.A., Cadiou, G., and Suthers, I.M. (2016) Temperature dependence of fish performance in the wild: links with species biogeography and physiological thermal tolerance. Functional Ecology, n/a-n/a.

Pedersen, A.F. (2012) Reproductive biology in the blind shark, Brachaelurus waddi, with notes on feeding biology and habitat use. Master's Thesis, Aarhus University, Denmark

Pickering, H., and Whitmarsh, D. (1997) Artificial reefs and fisheries exploitation: a review of the "attraction versus production" debate, the influence of design and its significance for policy. Fisheries Research 31(1-2), 39-59.

Piraino, M.N., and Szedlmayer, S.T. (2014) Fine-scale movements and home ranges of red snapper around artificial reefs in the northern Gulf of Mexico. Transactions of the American Fisheries Society 143(4), 988-998.

Pollock, K.H., Hoenig, J.M., Jones, C.M., Robson, D.S., and Greene, C.J. (1997) Catch rate estimation for roving and access point surveys. North American Journal of Fisheries Management 17(1), 11-19.

Pollock, K.H., Jones, C.M., and Brown, T.L. (1994) 'Angler survey methods and their applications in fisheries management.' (American Fisheries Society: Bethesda, Maryland, USA)

Polovina, J.J. (1989) Artificial reefs: nothing more than benthic fish aggregators. ColCOFl Report 30, 37-39

Pondella, D.J., Stephens Jr, J.S., and Craig, M.T. (2002) Fish production of a temperate artificial reef based on the density of embiotocids (Teleostei: Perciformes). ICES Journal of Marine Science 59(suppl), S88-S93.

Powers, S.P., Grabowski, J.H., Peterson, C.H., and Lindberg, W.J. (2003) Estimating enhancement of fish production by offshore artificial reefs: uncertainty exhibited by divergent scenarios. Marine Ecology Progress Series 264, 265-277.

Powter, D.M., and Gladstone, W. (2008a) Demographic analysis of the Port Jackson shark Heterodontus portusjacksoni in the coastal waters of eastern Australia. Marine and Freshwater Research 59(5), 444-455.

Powter, D.M., and Gladstone, W. (2008b) Habitat preferences of Port Jackson sharks, Heterodontus portusjacksoni, in the coastal waters of eastern Australia. Proceedings of the Linnean Society of New South Wales 129, 151-165.

Powter, D.M., and Gladstone, W. (2008c) The reproductive biology and ecology of the Port Jackson shark Heterodontus portusjacksoni in the coastal waters of eastern Australia. Journal of Fish Biology 72(10), 2615-2633.

Powter, D.M., and Gladstone, W. (2009) Habitat-mediated use of space by juvenile and mating adult Port Jackson sharks, Heterodontus portusjacksoni, in eastern Australia. Pacific Science 63(1), 1-14.

191

Powter, D.M., Gladstone, W., and Platell, M. (2010) The influence of sex and maturity on the diet, mouth morphology and dentition of the Port Jackson shark, Heterodontus portusjacksoni. Marine and Freshwater Research 61(1), 74-85.

Redman, R.A., and Szedlmayer, S.T. (2009) The effects of epibenthic communities on reef fishes in the northern Gulf of Mexico. Fisheries Management and Ecology 16(5), 360-367.

Reubens, J.T., Pasotti, F., Degraer, S., and Vincx, M. (2013) Residency, site fidelity and habitat use of Atlantic cod (Gadus morhua) at an offshore wind farm using acoustic telemetry. Marine Environmental Research 90, 128-135.

Reynolds, B.F., Powers, S.P., and Bishop, M.A. (2010) Application of acoustic telemetry to assess residency and movements of rockfish and lingcod at created and natural habitats in Prince William Sound. PLoS ONE 5(8), e12130.

Ross, P.M., Thrush, S.F., Montgomery, J.C., Walker, J.W., and Parsons, D.M. (2007) Habitat complexity and predation risk determine juvenile snapper (Pagrus auratus) and goatfish (Upeneichthys lineatus) behaviour and distribution. Marine and Freshwater Research 58(12), 1144-1151.

Rowling, K., Hegarty, A., and Ives, M. (2010) Status of fisheries resources in New South Wales 2008/09. Industry and Investment NSW, Cronulla.

Santos, M.N., and Monteiro, C.C. (1997) The Olhão artificial reef system (south Portugal): Fish assemblages and fishing yield. Fisheries Research 30(1–2), 33-41.

Santos, M.N., and Monteiro, C.C. (1998) Comparison of the catch and fishing yield from an artificial reef system and neighbouring areas off Faro (Algarve, south Portugal). Fisheries Research 39(1), 55-65.

Santos, M.N., and Monteiro, C.C. (2007) A fourteen-year overview of the fish assemblages and yield of the two oldest Algarve artificial reefs (southern Portugal). Hydrobiologia 580(1), 225- 231.

Santos, M.N., Monteiro, C.C., and Gaspar, M.B. (2002) Diurnal variations in the fish assemblage at an artificial reef. ICES Journal of Marine Science: Journal du Conseil 59(suppl), S32-S35.

Schenker, N., and Gentleman, J.F. (2001) On Judging the Significance of Differences by Examining the Overlap Between Confidence Intervals. The American Statistician 55(3), 182- 186.

Schlaff, A., Heupel, M., and Simpfendorfer, C. (2014) Influence of environmental factors on shark and ray movement, behaviour and habitat use: a review. Reviews in Fish Biology and Fisheries 24(4), 1089-1103.

Schroepfer, R.L., and Szedlmayer, S.T. (2006) Estimates of residence and site fidelity for red snapper Lutjanus campechanus on artificial reefs in the northeastern Gulf of Mexico. Bulletin of Marine Science 78(1), 93-101.

192

Scott, M.E., Smith, J.A., Lowry, M.B., Taylor, M.D., and Suthers, I.M. (2015) The influence of an offshore artificial reef on the abundance of fish in the surrounding pelagic environment. Marine and Freshwater Research 66, 429–437.

Shapiro, L.G., and Stockman, G.C. (2001) 'Computer Vision.' (Prentice Hall: New Jersey, USA) 608

Shipley, J.B., and Cowan Jr, J.H. (2011) Artificial reef placement: a red snapper, Lutjanus campechanus, ecosystem and fuzzy rule-based model. Fisheries Management and Ecology 18(2), 154-167.

Smallwood, C.B., Pollock, K.H., Wise, B.S., Hall, N.G., and Gaughan, D.J. (2011) Quantifying recreational fishing catch and effort: a pilot study of shore-based fishers in the Perth Metropolitan area. Department of Fisheries, No. Fisheries Research Report No. 216, Western Australia.

Smallwood, C.B., Pollock, K.H., Wise, B.S., Hall, N.G., and Gaughan, D.J. (2012) Expanding aerial–roving surveys to include counts of shore-based recreational fishers from remotely operated cameras: benefits, limitations, and cost effectiveness. North American Journal of Fisheries Management 32(6), 1265-1276.

Smith, J.A., Lowry, M., Champion, C., and Suthers, I.M. (2016) A designed artificial reef is among the most productive marine fish habitats: new metrics to address ‘production versus attraction’. Marine Biology 163, 188.

Smith, J.A., Lowry, M.B., and Suthers, I.M. (2015) Fish attraction to artificial reefs not always harmful: a simulation study. Ecology and Evolution 5, 4590-4602.

Sokal, R.R., and Rohlf, F.J. (1981) 'Biometry: the principles and practice of statistics in biological research.' 2nd edn. (W. H. Freeman and Co.: New York ) 859

Solonsky, A.C. (1985) Fish colonization and the effect of fishing activities on two artificial reefs in Monterey Bay, California. Bulletin of Marine Science 37(1), 336-347.

Sommerville, E., Platell, M.E., White, W.T., Jones, A.A., and Potter, I.C. (2011) Partitioning of food resources by four abundant, co-occurring elasmobranch species: relationships between diet and both body size and season. Marine and Freshwater Research 62(1), 54-65.

State Pollution Control Commission (1981) The ecology of fish in Botany bay-Biology of commercially and recreationally valuable species. Environmental control study of Botany Bay. State Pollution Control Commission, Sydney, Australia.

Steffe, A.S., and Chapman, D.J. (2003) A survey of daytime recreational fishing during the annual period, March 1999 to February 2000, in Lake Macquarie, New South Wales. New South Wales Fisheries, No. 52, Cronulla.

Steffe, A.S., and Murphy, J.J. (2011) Recreational fishing surveys in the Greater Sydney Region. New South Wales Department of Primary Industries, Cronulla.

Steffe, A.S., Murphy, J.J., Chapman, D.J., Barrett, G.P., and Gray, C.A. (2005a) An assessment of changes in the daytime, boat-based, recreational fishery of the Tuross Lake estuary following 193 the establishment of a ‘Recreational Fishing Haven’. New South Wales Department of Primary Industries, Cronulla.

Steffe, A.S., Murphy, J.J., Chapman, D.J., and Gray, C.A. (2005b) An assessment of changes in the recreational fishery of Lake Macquarie following the establishment of a ‘Recreational Fishing Haven’. New South Wales Department of Primary Industries, Cronulla.

Steffe, A.S., Murphy, J.J., Chapman, D.J., Tarlinton, B.E., Gordon, G.N.G., and Grinberg, A. (1996) An assessment of the impact of offshore recreational fishing in New South Wales on the management of commercial fisheries. New South Wales fisheries institute, No. null, Cronulla, NSW.

Steffe, A.S., Murphy, J.J., and Reid, D.D. (2008) Supplemented access point sampling designs: A cost-effective way of improving the accuracy and precision of fishing effort and harvest estimates derived from recreational fishing surveys. North American Journal of Fisheries Management 28(4), 1001-1008.

Stehfest, K.M., Lyle, J.M., and Semmens, J.M. (2015) The use of acoustic accelerometer tags to determine seasonal changes in activity and catchability of a recreationally caught marine teleost. ICES Journal of Marine Science: Journal du Conseil.

Stocks, J.R., Gray, C.A., and Taylor, M.D. (2015) Out in the wash: spatial ecology of a temperate marine shallow rocky-reef species derived using acoustic telemetry. Marine and Freshwater Research 66(6), 559-571.

Stoner, A.W. (2004) Effects of environmental variables on fish feeding ecology: implications for the performance of baited fishing gear and stock assessment. Journal of Fish Biology 65(6), 1445-1471.

Strelcheck, A.J., Cowan, J.H., and Patterson, W.F. (2007) Site fidelity, movement, and growth of red snapper: Implications for artificial reef management American fisheries society symposium 60, 147-162

Svane, I.B., and Petersen, J.K. (2001) On the problems of epibioses, fouling and artificial reefs, a review. Marine Ecology 22(3), 169-188.

Szedlmayer, S.T., and Schroepfer, R.L. (2005) Long-term residence of red snapper on artificial reefs in the northeastern gulf of mexico. Transactions of the American Fisheries Society 134(2), 315-325.

Tinsman, J.C., and Whitmore, W.H. (2006) Aerial flight methodology to estimate and monitor trends in fishing effort on Delaware artificial reef sites. Bulletin of Marine Science 78(1), 167- 176.

Topping, D.T., and Szedlmayer, S.T. (2011a) Home range and movement patterns of red snapper (Lutjanus campechanus) on artificial reefs. Fisheries Research 112(1–2), 77-84.

Topping, D.T., and Szedlmayer, S.T. (2011b) Site fidelity, residence time and movements of red snapper Lutjanus campechanus estimated with long-term acoustic monitoring. Marine Ecology Progress Series 437, 183-200.

194

Tovar-Ávila, J., Troynikov, V.S., Walker, T.I., and Day, R.W. (2009) Use of stochastic models to estimate the growth of the Port Jackson shark, Heterodontus portusjacksoni, off eastern Victoria, Australia. Fisheries Research 95(2–3), 230-235.

Van Poorten, B.T., Carruthers, T.R., Ward, H.G.M., and Varkey, D.A. (2015) Imputing recreational angling effort from time-lapse cameras using an hierarchical Bayesian model. Fisheries Research 172, 265-273.

Vemco (2016) VPS (VEMCO Positioning System) online.

Vianna, G.M.S., Meekan, M.G., Meeuwig, J.J., and Speed, C.W. (2013) Environmental influences on patterns of vertical movement and site fidelity of grey reef sharks (Carcharhinus amblyrhynchos) at aggregation sites. PLoS ONE 8(4), e60331.

Villegas-Ríos, D., Alós, J., March, D., Palmer, M., Mucientes, G., and Saborido-Rey, F. (2013) Home range and diel behavior of the ballan wrasse, Labrus bergylta, determined by acoustic telemetry. Journal of Sea Research 80, 61-71.

Walsh, W.J. (1985) Reef fish community dynamics on small artificial reefs: the influence of isolation, habitat structure, and biogeography. Bulletin of Marine Science 36(2), 357-376.

Wells, R.M.G., McNeil, H., and MacDonard, J.A. (2005) Fish hypnosis: Induction of an atonic immobility reflex. Marine and Freshwater Behaviour and Physiology 38(1), 71-78.

West, L.D., Stark, K.E., Murphy, J.J., Lyle, J.M., and Ochwada-Doyle, F.A. (2015) Survey of Recreational Fishing in New South Wales and the ACT, 2013/14. NSW Department of Primary Industries,

Westmeyer, M.P., Wilson III, C.A., and Nieland, D.L. (2007) Fidelity of red snapper to petroleum platforms in the northern Gulf of Mexico. In Red Snapper ecology and fisheries in the U.S. Gulf of Mexico. Symposium 6 edn. (Eds. WF Patterson III, JH Cowan Jr, GR Fitzhugh and DL Nieland) pp. 105–121. (American Fisheries Society: Bethesda, Maryland)

White, W.T., and Last, P.R. (2012) A review of the taxonomy of chondrichthyan fishes: a modern perspective. Journal of Fish Biology 80(5), 901-917.

Whitmarsh, D., Santos, M.N., Ramos, J., and Monteiro, C.C. (2008) Marine habitat modification through artificial reefs off the Algarve (southern Portugal): An economic analysis of the fisheries and the prospects for management. Ocean & Coastal Management 51(6), 463-468.

Whitney, N.M., Papastamatiou, Y.P., Holland, K.N., and Lowe, C.G. (2007) Use of an acceleration data logger to measure diel activity patterns in captive whitetip reef sharks, Triaenodon obesus. Aquatic Living Resources 20(04), 299-305.

Wilding, T.A., and Sayer, M.D.J. (2002) Evaluating artificial reef performance: approaches to pre- and post-deployment research. ICES Journal of Marine Science 59(suppl), S222-230.

Wilson, J., Osenberg, C.W., St. Mary, C.M., Watson, C.A., and Lindberg, W.J. (2001) Artificial reefs, the attraction-production issue, and density dependence in marine ornamental fishes. Aquarium Sciences and Conservation 3(1), 95-105.

195

Wood, G., Lynch, T., Devine, C., Keller, K., and Figueira, W. (2016) High-resolution photo- mosaic time-series imagery for monitoring human use of an artificial reef. Ecology and Evolution 6, 6963–6968.

Wood, S., and Scheipl, F. (2015) Gamm4: Generalized additive mixed models using mgcv and lme4. In R package version 0.2-3.

Workman, I., Shah, A., Foster, D., and Hataway, B. (2002) Habitat preferences and site fidelity of juvenile red snapper. ICES Journal of Marine Science 59, S43-S50.

Wraith, J., Lynch, T., Minchinton, T.E., Broad, A., and Davis, A.R. (2013) Bait type affects fish assemblages and feeding guilds observed at baited remote underwater video stations. Marine Ecology Progress Series 477, 189-199.

Zalmon, I.R., Novelli, R., Gomes, M.P., and Faria, V.V. (2002) Experimental results of an artificial reef programme on the Brazilian coast north of Rio de Janeiro. ICES Journal of Marine Science 59(suppl), 83-87.

196

Appendix A. Effort Estimation Equations

Basic notation

j=the stratum being considered (j=1….J)

J= the total number of strata i=sample day unit within stratum (i=1…N)

Nj=total population size (all possible sampling days in stratum) nj= sample size in stratum

The sample variance for stratum j: 2 𝑗𝑗 [ ( ) ]𝑠𝑠 𝑛𝑛 = 𝑗𝑗 2 𝑖𝑖𝑖𝑖 𝑗𝑗 2 ∑𝑖𝑖=1 𝑥𝑥 −𝑥𝑥̅ 𝑠𝑠𝑗𝑗 𝑛𝑛𝑗𝑗−1 =the value of the ith unit of stratum (i=1…j)

𝑖𝑖𝑖𝑖 𝑥𝑥 =the sample mean for stratum j

𝑗𝑗 𝑥𝑥 ̅

Fishing effort calculations with direct expansion

1. Fishing effort expansion for each base level stratum Ê :

𝑗𝑗 Ê = × (A. 1)

𝑗𝑗 𝑗𝑗 𝑗𝑗 = mean𝑁𝑁 dailyē fishing effort for stratum j

𝑗𝑗 ē

2. Variance of the mean daily estimates of fishing effort using the finite population correction factor Var(ēj):

= 2 × (A. 2) 𝑠𝑠𝑗𝑗 𝑁𝑁𝑗𝑗−𝑛𝑛𝑗𝑗 𝑉𝑉𝑉𝑉𝑉𝑉�𝑒𝑒𝑗𝑗̅ � 𝑛𝑛𝑗𝑗 𝑁𝑁𝑗𝑗 197

3. Variance of fishing effort for each base-level stratum Ê :

𝑉𝑉𝑉𝑉𝑉𝑉� 𝑗𝑗� Ê = × (A. 3) 2 𝑗𝑗 𝑗𝑗 𝑗𝑗 𝑉𝑉𝑉𝑉𝑉𝑉� � 𝑁𝑁 𝑉𝑉𝑉𝑉𝑉𝑉�𝑒𝑒̅ �

4. Total fishing effort for each season is calculated by adding day-type strata for

𝑇𝑇𝑇𝑇𝑇𝑇 seasonal totals and addition of seasons𝐸𝐸 �for annual totals:

= (A. 4) 𝐽𝐽 𝑇𝑇𝑇𝑇𝑇𝑇 𝐽𝐽=1 𝑗𝑗 𝐸𝐸� ∑ 𝐸𝐸�

5. Total variance Ê for total fishing effort is calculated by adding variances of

𝑇𝑇𝑇𝑇𝑇𝑇 base level strata: 𝑉𝑉𝑉𝑉𝑉𝑉� �

Ê = Ê (A. 5) 𝐽𝐽 𝑇𝑇𝑇𝑇𝑇𝑇 𝐽𝐽=1 𝑗𝑗 𝑉𝑉𝑉𝑉𝑉𝑉� � ∑ 𝑉𝑉𝑉𝑉𝑉𝑉� �

6. Standard error Ê for the total fishing effort:

𝑆𝑆𝑆𝑆� 𝑇𝑇𝑇𝑇𝑇𝑇� Ê = Var Ê (A. 6)

𝑆𝑆𝑆𝑆� 𝑇𝑇𝑇𝑇𝑇𝑇� � � 𝑇𝑇𝑇𝑇𝑇𝑇�

Fishing effort - Adjusting stratum totals and variances for visibility bias

1. The regression equation is forced through the origin:

= (A. 7)

Y=𝑌𝑌 daily𝑏𝑏𝑏𝑏 fishing events from digital images x= daily fishing events from direct observations b=slope of the regression

198

2. The correction factor Cf is the inverse of the slope from the regression equation:

= 1/ (A. 8)

3.𝐶𝐶𝐶𝐶 Variance𝑏𝑏 of the correction factor Var (Cf) is derived from the “variance of a quotient” equation (Blumenfeld, 2001):

( ) ar ( ) = (A. 9) Var 𝑏𝑏 4 Var𝑉𝑉 (b)=𝐶𝐶𝐶𝐶 variance𝑏𝑏 of the slope

4. Adjusted estimate of fishing effort Ê for stratum j (units= fishing events):

𝑎𝑎𝑎𝑎𝑎𝑎 Ê = Ê × (A. 10)

𝑎𝑎𝑎𝑎𝑎𝑎 𝑗𝑗 Ê = unadjusted𝐶𝐶𝐶𝐶 estimate of fishing effort for stratum j (units= fishing events)

𝑗𝑗

5. Adjusted variance of fishing effort Var Ê for stratum j (fishing events):

� 𝑎𝑎𝑎𝑎𝑎𝑎� Var Ê = Ê × Var ( ) + × Ê 2 2 � 𝑎𝑎𝑎𝑎𝑎𝑎� � 𝑗𝑗 𝐶𝐶𝐶𝐶 � �𝐶𝐶𝐶𝐶 𝑉𝑉𝑉𝑉𝑉𝑉� 𝑗𝑗�� Ê × Var ( ) (A. 11)

𝑗𝑗 Ê = unadjusted−�𝑉𝑉𝑉𝑉𝑉𝑉 variance� � of fishing𝐶𝐶𝐶𝐶 effort� for stratum j

𝑗𝑗 𝑉𝑉𝑉𝑉𝑉𝑉� �

Converting fishing effort units (fishing events to boat hours)

1. Calculate the mean number of boat hours per fishing event for each day f (all data

𝑖𝑖𝑖𝑖 derived from digital images) ̅

199

2. Calculate the daily mean of the mean boat hours for fishing event fj for stratum j : ̅ f = (A. 12) ∑ f̅𝑖𝑖𝑖𝑖 𝒋𝒋̅ 𝑛𝑛𝑗𝑗 3. Sample variance of the mean daily fishing effort for stratum j in boat hours Var (fj): ̅

� f = ∑ 𝑓𝑓𝑖𝑖𝑖𝑖 (A. 13) 𝑉𝑉𝑉𝑉𝑉𝑉� 𝑛𝑛𝑗𝑗 � 𝑉𝑉𝑉𝑉𝑉𝑉� 𝒋𝒋̅ � 𝑛𝑛𝑗𝑗 4. Adjusted estimate of fishing effort for stratum j in boat hours Ê :

𝑏𝑏𝑏𝑏 Ê = Ê × f (A. 14)

𝑏𝑏𝑏𝑏 𝑎𝑎𝑎𝑎𝑎𝑎 𝒋𝒋 ̅

5. Variance of adjusted fishing effort) for stratum j in boat hours Var Ê :

� 𝑏𝑏𝑏𝑏� Var Ê = Ê × f + f × Var Ê 2 2 � 𝑏𝑏𝑏𝑏� � 𝑎𝑎𝑎𝑎𝑎𝑎 𝑉𝑉𝑉𝑉𝑉𝑉� 𝒋𝒋̅ �� 𝒋𝒋̅ � 𝑎𝑎𝑎𝑎𝑎𝑎�� Var Ê × f (A. 15)

𝑎𝑎𝑎𝑎𝑎𝑎 𝒋𝒋 −� � � 𝑉𝑉𝑉𝑉𝑉𝑉� ̅ ��

Converting fishing effort units (boat hours to fisher hours)

1. Calculate the mean number of fishers per boat trip for each day (data derived

𝑖𝑖𝑖𝑖 from coastal marine fishing outside the Port Hacking estuary- Steffeℎ� and Murphy,

unpublished data).

2. Calculate the daily mean of the mean fishers per boat trip j for stratum j:

ℎ� h = (A. 16) ∑ ℎ�𝑖𝑖𝑖𝑖 �𝒋𝒋 𝑛𝑛𝑗𝑗

200

3. Sample variance of the mean daily fishing effort for stratum j in fisher hours Var

( ):

ℎ�𝑗𝑗

� = ∑ ℎ𝑖𝑖𝑖𝑖 (A. 17) 𝑉𝑉𝑉𝑉𝑉𝑉� 𝑛𝑛𝑗𝑗 � 𝑉𝑉𝑉𝑉𝑉𝑉�ℎ�𝑗𝑗� 𝑛𝑛𝑗𝑗

4. Adjusted estimate of fishing effort for stratum j in fisher hours Ê

� 𝑓𝑓𝑓𝑓� Ê = Ê × h (A. 18)

𝑓𝑓𝑓𝑓 𝑏𝑏𝑏𝑏 𝒋𝒋 �

5. Variance of adjusted fishing effort) for stratum j in fisher hours Var Ê :

� 𝑓𝑓𝑓𝑓� Var Ê = Ê × + × Var Ê 2 2 � 𝑓𝑓𝑓𝑓� � 𝑏𝑏𝑏𝑏 𝑉𝑉𝑉𝑉𝑉𝑉�ℎ�𝑗𝑗�� ℎ�𝑗𝑗 � 𝑏𝑏𝑏𝑏�� Var Ê × (A. 19)

𝑏𝑏𝑏𝑏 𝑗𝑗 −� � � 𝑉𝑉𝑉𝑉𝑉𝑉�ℎ� ��

201

Appendix B. Fish functional groups and species presence by site.

South Total no. Presence by site Functional group / head of obs. at Name Family Species AR1 LM2 PH3 PB3 HR3 Tuross4 RLS5 AR6

Large-Medium Pelagic fish

Australian salmon Arripidae Arripis trutta x x x x x x 0 0

Australian bonito Scombridae Sarda australis x x x 0 1

Leaping bonito Scombridae Cybiosarda elegans x x 0 0

Yellowtail kingfish Carangidae Seriola lalandi x x x x x 0 30

Cobia Rachycentridae Rachycentron x 0 0 canadum

Dolphinfish Coryphaenidae Coryphaena hippurus x 0 0

Tailor Pomatomidae Pomatomus saltatrix x x x x x x 0 0

Samsonfish Carangidae Seriola hippos x x 0 0

202

Appendix B (Cont.)

Presence by site South Total no. Functional group / head of obs. at Name Family Species AR1 LM2 PH3 PB3 HR3 Tuross4 RLS5 AR6

Amberjack Carangidae Seriola dumerili x 0 0

Australian spotted Scombridae Scomberomorus x 0 0 mackerel munroi

Smooth hammerhead Sphyrnidae Sphyrna zygaena x 0 0

Short-fin mako Lamnidae Isurus oxyrinchus x 0 1

Small Pelagic fish

Yellowtail scad Carangidae Trachurus x x x x x 8 59 novaezelandiae

Australian sardine Clupeidae Sardinops x 0 0 neopilchardus

Frigate tuna Scombridae Auxis thazard x x 0 0

Jack mackerel Carangidae Trachurus declivis x 0 0

203

Appendix B (Cont.)

Presence by site South Total no. Functional group / head of obs. at Name Family Species AR1 LM2 PH3 PB3 HR3 Tuross4 RLS5 AR6

Striped seapike Sphyraenidae Sphyraena obtusata x x 0 0

Stout longtom Belonidae Tylosurus gavialoides x 0 0

Longfin pike Dinolestidae Dinolestes lewini x x x x x 10 22

Slimy mackarel Scombridae Scomber australasicus x x x x x 0 1

Mid-water planktivores

Silver sweep Kyphosidae Scorpis lineolata x x x x x 11 20

Mado Kyphosidae Atypichthys strigatus x 13 47

204

Appendix B (Cont.)

Presence by site Total South no. of Functional group / head obs. at Name Family Species AR1 LM2 PH3 PB3 HR3 Tuross4 RLS5 AR6

Omnivores (leatherjackets)

Sixspine leatherjacket Monacanthidae Meuschenia freycineti x x x x x x 4 8

Yellowfin leatherjacket Monacanthidae Meuschenia trachylepis x x x x x x 6 0

Fan-bellied Monacanthidae Monacanthus chinensis x x x x x 7 0 Leatherjacket*

Velvet leatherjacket* Monacanthidae Meuschenia scaber x 0 6

Mosaic leatherjacket Monacanthidae Eubalichthys mosaicus x x 2 2

Rough leatherjacket Monacanthidae Scobinichthys x x x x 6 1 granulatus

Ocean jacket Monacanthidae Nelusetta ayraudi x x x x x 1 61

205

Appendix B (Cont.)

Presence by site South Total no. Functional group / head of obs. at Name Family Species AR1 LM2 PH3 PB3 HR3 Tuross4 RLS5 AR6

Omnivores (mullets & garfish)

Flat-tail mullet Mugilidae Liza argentea x 0 0

Sand mullet Mugilidae Myxus elongatus x x x x 0 0

Sea mullet Mugilidae Mugil cephalus x x x x 0 0

Sea garfish Hemiramphidae Hyporhamphus x x x x 0 0 australis

River garfish Hemiramphidae Hyporhamphus x x x 0 0 regularis

Yellow-eye mullet Mugilidae Aldrichetta forsteri x 0 0

Ambush/cryptic predators

Marbled flathead Platycephalidae Platycephalus x x x x 0 0 marmoratus

206

Appendix B (Cont.)

Presence by site South Total no.

Functional group / 1 2 3 3 3 4 head of obs. at Name Family Species AR LM PH PB HR Tuross RLS5 AR6

Bluespotted flathead Platycephalidae Platycephalus x x x x x 0 24 caeruleopunctatus

Dusky flathead Platycephalidae Platycephalus fuscus x x x x x 0 0

Southern sand flathead Platycephalidae Platycephalus bassensis x x 0 0

Long-spined flathead Platycephalidae Platycephalus x 0 0 longispinis

Northern-sand flathead Platycephalidae Platycephalus arenarius x x x 0 0

Tiger flathead Platycephalidae Platycephalus x 0 0 richardsoni

Eastern red scorpionfish Scorpaenidae Scorpaena cardinalis x x x x x 2 1

Sergeant baker Aulopidae Latropiscis x x x x 0 4 purpurissatus

207

Appendix B (Cont.)

Presence by site South Total no. Functional group / head of obs. at Name Family Species AR1 LM2 PH3 PB3 HR3 Tuross4 RLS5 AR6

Painted grinner Synodontidae Trachinocephalus x x 0 0 myops

Smalltooth flounder Paralichthyidae Pseudorhombus x x x x x x 0 1 jenynsii

Largetooth flounder Paralichthyidae Pseudorhombus arsius x x x x x x 0 0

John Dory Zeidae Zeus faber x x 0 1

Large demersal predators

Snapper Sparidae Chrysophrys auratus x x x x x x 10 10

Eastern blue groper Labridae Achoerodus viridis x 10 1

Bronze whaler Carcharhinidae Carcharhinus x x x 0 0 brachyurus

Gummy shark Triakidae Mustelus antarcticus x x x 0 0

208

Appendix B (Cont.)

Presence by site South Total no. Functional group / head of obs. at Name Family Species AR1 LM2 PH3 PB3 HR3 Tuross4 RLS5 AR6

Australian angel shark Squatinidae Squatina australis x 0 0

Eastern fiddler Ray Rhinobatidae Trygonorrhina fasciata x x x 0 16

Port Jackson shark* Heterodontidae Heterodontus spp. x 0 6

Jewfish Sciaenidae Argyrosomus japonicas x x x x 0 0

Teraglin Sciaenidae Atractoscion aequidens x 0 0

Medium demersal predators

Yellowfin bream Sparidae Acanthopagrus australis x x x x x x 10 1

Black bream Sparidae Acanthopagrus butcheri x x 0 0

Tarwhine Sparidae Rhabdosargus sarba x x x x x x 0 3

209

Appendix B (Cont.)

Presence by site South Total no. Functional group / head of obs. at Name Family Species AR1 LM2 PH3 PB3 HR3 Tuross4 RLS5 AR6

Silver trevally Carangidae Pseudocaranx dentex x x x x x x 0 36

Red morwong Cheilodactylidae Cheilodactylus fuscus x x x 12 1

Blue morwong Cheilodactylidae Nemadactylus x x 1 20 douglasii

Eastern shovelnose ray Rhinobatidae Aptychotrema rostrata x x x x 0 0

Crimsonbanded wrasse Labridae Notolabrus x x x x x 15 2 gymnogenis

Estuary perch Percichthyidae Macquaria x x 0 0 colonorum

Bastard trumpeter Latridae Latridopsis forsteri x 0 0

Eastern wirrah Serranidae Acanthistius ocellatus x x x 0 0

Red gurnard Triglidae Chelidonichthys kumu x x x x x x 0 0

210

Appendix B (Cont.)

Presence by site South Total no. Functional group / head of obs. at Name Family Species AR1 LM2 PH3 PB3 HR3 Tuross4 RLS5 AR6

Redfish Berycidae Centroberyx affinis x 0 1

Sand whiting Sillaginidae Sillago ciliata x x x x x 0 0

Small demersal predators

Bluestriped goatfish Mullidae Upeneichthys lineatus x 14 18

Trumpeter whiting Sillaginidae Sillago maculata x x x x 0 0

Eastern school whiting Sillaginidae Sillago flindersi x 0 0

Maori wrasse Labridae Ophthalmolepis x x x x 15 12 lineolatus

Comb wrasse Labridae Coris picta x x 0 1

Black sole Soleidae Brachirus nigra x 0 0

211

Appendix B (Cont.)

Presence by site South Total no. Functional group / head of obs. at Name Family Species AR1 LM2 PH3 PB3 HR3 Tuross4 RLS5 AR6

Herbivores

Silver drummer Kyphosidae Kyphosus sydneyanus x 3 0

Luderick Kyphosidae Girella tricuspidata x x x x x 0 0

Other (unindentified) 0 33 * Species for which harvest by number and/or weight were unable to be calculated

1Observations at the artificial reef (AR)

2LM= Lake Macquarie estuary (Steffe and Chapman, 2003; Steffe et al., 2005a)

3PH= Port Hacking estuary; PB= Pittwater/ Broken Bay estuary, HR=Hawkesbury reach (Steffe and Murphy, 2011)

4Tuross estuary (Steffe et al., 2005b)

5Reef life survey (RLS) diver observations from South head, Sydney (2009- 2015, http://reeflifesurvey.com)

6 Total number of observations (rounded off to nearest whole number) from drop cameras and Baited Remote Underwater Videos (BRUVs) at the AR from 2011- 2014 (Lowry et al., unpublished data; Scott et al., 2015; Smith, unpublished data)

212

Appendix C. Harvest Estimation Equations

Basic notation

j= the stratum being considered (j=1….J)

J= the total number of strata i= sample day unit within stratum (i=1…N)

Nj= total population size (all possible sampling days in stratum) nj= sample size in stratum

𝑖𝑖𝑖𝑖 𝑥𝑥 = the sample mean for stratum j

𝑗𝑗 𝑥𝑥 ̅

Harvest calculations

1. Harvest rate estimation using ratio of means Ȓij (number of fish per boat hour):

= 𝑛𝑛𝑛𝑛 (C. 1) ∑𝑖𝑖=1 𝐻𝐻𝑖𝑖𝑖𝑖 𝑖𝑖𝑖𝑖 𝑛𝑛𝑛𝑛 Ȓ ∑𝑖𝑖=1 𝐿𝐿𝑖𝑖𝑖𝑖

Hij=the complete harvest for the ith unit of stratum (i=1…j)

Lij= the complete trip length for the ith unit of stratum (i=1…j)

213

2. Mean daily harvest rates in each stratum j for each site (number of fish per boat

𝑗𝑗 hour): 𝑅𝑅�

= (C. 2) ∑ Ȓ𝑖𝑖𝑖𝑖 𝑅𝑅�𝑗𝑗 𝑛𝑛𝑗𝑗

3. Variance of the mean daily harvest rates in each stratum j for each site:

𝑉𝑉𝑉𝑉𝑉𝑉 𝑅𝑅�𝑗𝑗 = 2 (C. 3) 𝑠𝑠𝑗𝑗 𝑉𝑉𝑉𝑉𝑉𝑉 𝑅𝑅�𝑗𝑗 𝑛𝑛𝑗𝑗

4. Site-weighted mean daily harvest rates, ( ) for each stratum j (day-type within

𝑗𝑗 season): 𝑠𝑠𝑠𝑠 𝑅𝑅�

( ) ( ) ( ) ( ) = + + (C. 4) 𝑅𝑅�𝑗𝑗 𝑙𝑙𝑙𝑙 𝑅𝑅�𝑗𝑗 𝑃𝑃𝑃𝑃 𝑅𝑅�𝑗𝑗 𝐻𝐻𝐻𝐻 �𝑗𝑗 𝑠𝑠𝑠𝑠 𝑅𝑅 � 3 3 3 �

LR= Long Reef

PH= Port Hacking

HK= Hawkesbury

Note: 3 sites were weighted equally

5. Site-weighted variance for mean daily harvest rates, ( )for each stratum j

𝑗𝑗 (day-type within season): 𝑠𝑠𝑠𝑠 𝑉𝑉𝑉𝑉𝑉𝑉 𝑅𝑅�

1 1 ( ) = × + × 3 2 ( ) 3 2 ( ) 𝑠𝑠𝑠𝑠 𝑉𝑉𝑉𝑉𝑉𝑉 𝑅𝑅�𝑗𝑗 �� 𝑉𝑉𝑉𝑉𝑉𝑉 𝑅𝑅�𝑗𝑗 𝐿𝐿𝐿𝐿 � �� 𝑉𝑉𝑉𝑉𝑉𝑉 𝑅𝑅�𝑗𝑗 𝑃𝑃𝑃𝑃 �

+ × ( ) (C. 5) 1 2 ��3� 𝑉𝑉𝑉𝑉𝑉𝑉 𝑅𝑅�𝑗𝑗 𝐻𝐻𝐻𝐻 �

214

6. Harvest estimation ( ) for each stratum j (units= number of fish per boat hour)

𝐻𝐻�𝑗𝑗 𝑛𝑛𝑛𝑛 ( ) = ( ) × (C. 6)

𝑗𝑗 𝑛𝑛𝑛𝑛 𝑗𝑗 𝑗𝑗 𝐻𝐻� 𝑠𝑠𝑠𝑠 𝑅𝑅� 𝐸𝐸�

= Total fishing effort for the stratum in boat hours (Keller et al. 2016)

𝑗𝑗 𝐸𝐸�

7. Variance of harvest ( )at the AR ( number of fish per boat hour):

𝑉𝑉𝑉𝑉𝑉𝑉 𝐻𝐻�𝑗𝑗 𝑛𝑛𝑛𝑛 ( ) = ( ) × + × ( ) 2 2 𝑉𝑉𝑉𝑉𝑉𝑉 𝐻𝐻�𝑗𝑗 𝑛𝑛𝑛𝑛 �𝑠𝑠𝑠𝑠 𝑅𝑅�𝑗𝑗 𝑉𝑉𝑉𝑉𝑉𝑉�𝐸𝐸�𝑗𝑗�� 𝐸𝐸�𝑗𝑗 𝑠𝑠𝑠𝑠 𝑉𝑉𝑉𝑉𝑉𝑉 𝑅𝑅�𝑗𝑗 � − [ ( ) × ( )] (C. 7)

𝑗𝑗 𝑗𝑗 𝑉𝑉𝑉𝑉𝑉𝑉 𝐸𝐸� 𝑠𝑠𝑠𝑠 𝑉𝑉𝑉𝑉𝑉𝑉 𝑅𝑅�

= Variance of fishing effort for each base-level stratum j (Keller et al. 2016)

𝑗𝑗 𝑉𝑉𝑉𝑉𝑉𝑉 �𝐸𝐸� �

8. Standard error of harvest ( ) is calculated using the square root formula: � 𝑆𝑆𝑆𝑆 𝐻𝐻𝑗𝑗 𝑛𝑛𝑛𝑛 ( ) = ( ) (C. 8)

𝑆𝑆𝑆𝑆 � 𝑗𝑗 𝑛𝑛𝑛𝑛 � � 𝑗𝑗 𝑛𝑛𝑛𝑛 𝐻𝐻 𝑉𝑉𝑉𝑉𝑉𝑉 𝐻𝐻

Converting harvest from number of fish per boat hour to weight (kg)

1. Harvest estimation in fish weight ( ) for each stratum j (weight in kg):

𝐻𝐻�𝑗𝑗 𝑛𝑛𝑛𝑛 ( ) = ( ) × (C. 9)

𝑗𝑗 𝑘𝑘𝑘𝑘 𝑗𝑗 𝑛𝑛𝑛𝑛 𝑗𝑗 𝐻𝐻� 𝐻𝐻� 𝑊𝑊�

= Site-weighted mean daily weight (kg) in each stratum j

𝑗𝑗 Note:𝑊𝑊� Harvest was converted to fish weight for all taxa for which suitable length to

weight conversion equations were available 215

2. Approximate variance for derived from proportional allocation (fish weight in kg):

( ) Var H ( ) = × H ( ) C.10 𝑉𝑉𝑉𝑉𝑉𝑉 𝐻𝐻(�𝑗𝑗 𝑛𝑛𝑛𝑛) �j kg � 𝐻𝐻�𝑗𝑗 𝑛𝑛𝑛𝑛 � �j kg

216

Appendix D. Estimated annual recreational harvest of the ten most commonly captured species from the AR Common name Species Harvest (number Harvest by weight of fish) ± SE (Kg) ± SE Bluespotted flathead Platycephalus 238.70 ±49.24 143.96 ±38.40 caeruleopunctatus

Ocean jacket Nelusetta ayraudi 202.04 ±47.02 71.11 ±27.32

Silver trevally Pseudocaranx 101.84 ±29.63 62.05 ±22.35 dentex

Snapper Chrysophrys 83.79 ±17.56 73.08 ±15.85 auratus

Yellowtail scad Trachurus 62.66 ±15.28 14.59 ±7.40 novaezelandiae

Silver sweep Scorpis lineolata 51.44 ±12.02 23.77 ±8.22

Maori wrasse Ophthalmolepis 30.94 ±7.17 8.56 ±3.75 lineolatus

Yellowtail kingfish Seriola lalandi 20.16 ±5.04 65.63 ±9.02

Blue morwong Nemadactylus 20.07 ±4.08 22.41 ±4.34 douglasii

Yellowfin bream Acanthopagrus 12.13 ±3.23 9.15 ±2.88 australis

217

218