Characteristics of the fish fauna of the Mandurah artificial

Submitted by

Lachlan A. W. Ramm

This thesis is presented for the degree of Bachelor of Science Honours

College of Science, Health, Engineering & Education, of Murdoch University, 2020

Declaration

I declare that this thesis is my own account of my research and contains as its main content work which has not previously been submitted for a degree at any tertiary education institution

Lachlan Ramm

i Acknowledgements

I would like to thank a number of people who have helped made the submission of this thesis possible. Firstly, a massive thanks to my primary supervisor, Dr James Tweedley, for his endless support and advice. Advances in my knowledge throughout this thesis is a huge credit to yourself and for that I am very grateful. Thank you to my secondary supervisor James Florisson and Steph Watts from Recfishwest for their advice, particularly in Chapter two, as well as Alistair Becker for his contribution to that chapter also. A big thank you to my parents, Julie and Ash, who have supported me financially and emotionally throughout this journey, this one’s for you. Many thanks to my partner Nicola who has been very patient and helpful throughout the stressful milestones of this thesis and for providing transport, food and emotional support when I needed it most. Credit is due for the Reef Vision volunteers that made BRUV deployments on the Mandurah at the start of the study, thank you for your efforts. I would also like to thank the many volunteers who assisted my own deployments of BRUV units throughout the 17 months of the study period, specifically Ash, Nathan, Will, Jethro, Luke and Emilie for their efforts on multiple occasions. Thank you Tackle World Miami for donating the bait used in the BRUV units throughout the duration of this study. Finally, it should be acknowledged that funding was provided by Recfishwest and the Department of Primary Industries & Regional Development through the Recreational Fishing Initiatives Fund and by Murdoch University.

ii Abstract

Artificial reefs have been deployed throughout the world for a range of different purposes and thus differ in their characteristics. To better understand how their designs and implementation differ spatially, temporally and according to their stated purpose, an extensive collation of information on artificial reefs is needed. Furthermore, in Australia and particularly Western Australia, artificial reefs are increasingly being deployed to enhance recreational fishing experiences and catch rates of key species. To address these topics, this thesis aimed to conduct a systematic literature review on existing artificial reefs around the world and describe their characteristics. It also aimed to compare and describe the characteristics of the fish fauna on the recently deployed

Mandurah artificial reef (south-western Australia) to those on nearby natural habitat. Finally, a comparison of the fish community of the three ‘Fish Box’ artificial reefs in Mandurah, Bunbury and Dunsborough was conducted. The literature review identified 1,074 unique artificial reefs from 71 countries, with 89% located in the northern hemisphere, but with an equal distribution between east and west. A more in-depth investigation of artificial reefs that were intentionally deployed found that these deployments increased markedly after 1965 and were most commonly sunk between depths of 10 and 30 m. Most reefs were designed to enhance faunal communities and/or fisheries. More affluent countries have monitored artificial reefs using more technologically advanced methods including video systems over recent decades. However, many reef deployments have not had an associated monitoring program and those that do are generally too short to detect long-term temporal changes. Fish and the associated faunal assemblages were monitored on an artificial reef and natural (control) site using Baited Remote Underwater Video (BRUVs) between February 2019 and July 2020. A far greater number of species and individuals were observed on the artificial reef than the control site. Additionally, three key recreational fishing species, i.e. Chrysophrys auratus, Seriola hippos and Pseudocaranx dentex, were also found in far greater abundances on the artificial reef. Pelagic fish were also more

iii common on the artificial habitat, likely due to the upward flow diversion produced by the modules and the subsequent increased availability of plankton. While community composition was mainly driven by differences between the artificial reef and control site, the faunas of both sites did change temporally, albeit to a lesser extent. There was a general decrease in the number of species of fish in winter and increase in summer. An aim of this study was to develop a standardized protocol for describing artificial reefs, which is hoped to be adopted globally. If successful, this will allow comparisons between artificial reefs and their outcomes to be made thus increasing our understanding of these reefs and improving their designs. The study also demonstrates that the Mandurah artificial reef supports a distinct faunal community compared with an adjacent natural habitat with no reef, including species that are of recreational importance. A comparison of three ‘Fish Box’ artificial reefs demonstrated that the fauna on the Mandurah reef was the most speciose and harboured the greater number of fish, indicating that site selection has a major influence on the resultant fish community composition and thus should be considered in future reef deployments.

iv Table of contents

Acknowledgements ...... ii Abstract ...... iii Publication list ...... vi

Chapter 1: General introduction ...... 1 1.1. Artificial reefs ...... 1 1.2. Artificial reefs in Western Australia ...... 4 1.3. Recreational fishing ...... 8 1.4. Aims ...... 9

Chapter 2: Artificial reefs in the Anthropocene: A review of geographical and historical trends in their design, purpose and monitoring ...... 10 2.0. Abstract ...... 10 2.1. Introduction ...... 11 2.2. Materials and methods ...... 14 2.2.1. Collation and analysis of publication data for fisheries enhancement terms ...... 14 2.2.2. Literature search ...... 15 2.2.3. Data extraction and classification ...... 16 2.3. Results and discussion ...... 19 2.3.1. Number of publications ...... 19 2.3.2. Overview ...... 20 2.3.3. Geographical distribution and abundance ...... 23 2.3.4. Trends between the categories of artificial reefs ...... 26 2.3.5. Category C reefs ...... 28 2.4. Conclusions and guidelines for artificial reef description ...... 40

Chapter 3: Comparisons between the fish faunas of the Mandurah artificial reef and adjacent unstructured habitat ...... 43 3.0. Abstract ...... 43 3.1. Introduction ...... 44 3.2. Materials and methods ...... 46 3.2.1. Study region and artificial reef ...... 46 3.2.2. Sampling regime ...... 48 3.2.3. Data extraction ...... 50 3.2.4. Statistical analysis ...... 52 3.3. Results ...... 55 3.3.1. Preliminary analysis ...... 55 3.3.2. Primary analysis ...... 59 3.4. Discussion ...... 77 3.4.1. Number of species and total MaxN ...... 77 3.4.2. Differences among habitat and feeding guilds ...... 81 3.4.3. Faunal composition ...... 82 3.4.4. Study limitations and future directions ...... 83

Chapter 4: General conclusion ...... 86 4.1. Global trends in artificial reef construction and deployment...... 86 4.2. Fish fauna of the Mandurah artificial reef ...... 86 4.3. Comparisons of the fish faunas of the Mandurah artificial to reefs of the same design in Geographe Bay ...... 87 4.4. Future directions ...... 92

References ...... 93 Appendices ...... 104

v Publication list

Journal articles Ramm, L.A.W., Florisson, J.H., Watts, S.L., Becker, A. & Tweedley, J.R. (BMS2020-0046 – major revisions). Artificial reefs in the anthropocene: a review of spatial and temporal trends in their design, purpose and monitoring. Bulletin of Marine Science. This publication was developed from my literature review and has been included verbatim from the submission to Bulletin of Marine Science at Chapter 2. According to the chief editor, this manuscript was reviewed by two experts in artificial reefs and is acceptable for publication following modification. Contributions. LR: conceptualisation, methodology, investigation, writing - original draft, writing- review and editing, visualization - production of figures and tables. JF: writing- review and editing, development of concept. SW: writing- review and editing, development of concept. JT: conceptualisation, methodology, formal analysis, writing - original draft, visualization - production of figures and tables, supervision, writing- review and editing.

Conference presentation Ramm, L.A.W., Florisson, J.H., Watts, S.L. & Tweedley, J.R. (2019). Global trends in artificial reefs: how has design, purpose, location and monitoring method changed over the last 50 years. 10th Florida State University-Mote International Symposium on Fisheries Ecology and 6th International Symposium on Stock Enhancement and Sea Ranching. Florida, United States of America. Contributions. LR: investigation, conceptualisation, writing- review and editing. JF: conceptualisation. SW: conceptualisation. JT: conceptualisation, writing - original draft, visualization - production of figures and tables, supervision, writing- review and editing, presenting. A copy of this document is provided in Appendix 1.

Conference poster Ramm, L.A.W., Florisson, J.H., Walker, T.H.E., Watts, S.L. & Tweedley, J.R. (2019). Scientifically robust citizen science monitoring of artificial reefs. 10th Florida State University-Mote International Symposium on Fisheries Ecology and 6th International Symposium on Stock Enhancement and Sea Ranching. Florida, United States of America. Contributions. LR: investigation, conceptualisation, writing – original draft, writing- review and editing. JF: conceptualisation. TW: investigation. SW: conceptualisation. JT: conceptualisation, writing - original draft, visualization - production of figures, supervision, writing- review and editing. A copy of this document is provided in Appendix 2.

vi Chapter 1: General introduction

1.1. Artificial reefs Artificial reefs, which can be broadly defined as “human-made structures installed in aquatic habitats that serve as a substrate and/or shelter for organisms” (Lima et al.,

2019), have been used by humans for at least 3,000 years (Riggio et al., 2000). With increasing degradation of the marine environment and overexploitation of its fauna, particularly in recent decades (Pauly et al., 2002; Halpern et al., 2015), artificial reefs are increasingly being used as one of a suite of mitigation tools. For example, these structures have been used to compensate for overfishing (Bombace, 1989), habitat loss (Dupont, 2008) and to prevent illegal trawling (Jensen, 2002; Iannibelli and Musmarra, 2008). Moreover, the recreational fishing and tourism (e.g. ) sectors also utilize artificial reefs, providing social and economic benefits (Sutton and Bushnell, 2007; Becker et al., 2017). For instance, Ex-HMAS Brisbane, which was sunk as a dive site in Queensland (Australia) in 2005 at a cost of AU$4.75 million with ongoing maintenance costs of AU$77,900 per year, was estimated to have generated in excess of AU$17 million in local revenue between 2005 and 2009 (Schaffer, 2011). If done carefully, the disposal of oil and gas structures as artificial reefs can be cheaper, whilst also providing environmental benefits (Kaiser and Pulsipher, 2005). Furthermore, the sinking of decommissioned ships has been shown to reduce fishing and scuba diving on nearby natural reefs (Leeworthy et al., 2006). Artificial reefs can be assigned to two broad types, i.e. those constructed from materials of opportunity and purpose-built structures (Bateman, 2015). Materials of opportunity are objects that are no longer needed for their primary function and can make the construction of an artificial reef very cost effective. Common materials used include car tyres, concrete and rock rubble, whitegoods, oil and gas structures and vehicles/vessels (e.g. Branden et al., 1994; Bortone et al., 1997). Purpose-built structures are those constructed for the exclusive purpose of an artificial reef. They are most commonly made from concrete and/or steel and vary significantly in size and complexity, often in coherence to their individual purposes (Baine, 2001). The Convention on the Prevention of Marine Pollution by Dumping of Wastes and Other

1 Matter was signed by 87 countries and ratified in 1975 and following revision in 2006 is now known as the London Protocol (London Convention and Protocol/UNEP, 2009). One of its objectives is to prevent pollution of the sea by dumping of wastes and other matter and thus outlaws the construction of artificial reefs from materials of opportunity without approval. In Australia, which is one of the signatories, the protocol has been incorporated into the Environment Protection (Sea Dumping) Act 1981 (Department of Agriculture, Water and th Environment, 2020b). Due to the diversity in the morphology of artificial reefs and range of materials used in their construction, they are used for a range of purposes (Baine, 2001). These include, but are not limited to, restoration and enhancement of commercial, recreational and artisanal fishing (Polovina and Sakai, 1989; Becker et al., 2017; Lima et al., 2019), aquaculture and sea ranching (Qin et al., 2020), provision and restoration of habitat (Clark and Edwards, 1999; Jensen, 2002; Silva et al., 2016), coastal defence (Hegde, 2010), tourism and recreation, e.g. scuba diving (Stolk et al., 2007) and surfing (Evans and Ranasinghe, 2001), and science (Sherman et al., 2002). There are also a range of structures that were not constructed as artificial reefs per se but do provide habitat for fish, and includes active oil and gas drilling platforms (Ajemian et al., 2015), undersea pipelines (Kaiser, 2006), breakwaters (Hegde, 2010), marinas and jetties (Rilov and Benayahu, 2000) and offshore wind farms (Langhamer, 2012). Despite many studies showing that artificial reefs are colonised by fish and invertebrates and can support a greater diversity than unstructured habitats and similar or enhanced communities to natural reefs (e.g. Fabi and Fiorentini, 1994), the use of artificial reefs as a fisheries management tool has long been controversial. Much of this debate focuses on whether reefs produce biomass, in particularly fish, or simply attract them from surrounding natural reef (e.g. Pickering and Whitmarsh, 1997). The production hypothesis states that artificial reefs function independently in that they only recruit larvae that otherwise would not have settled on natural reef (Osenberg et al., 2002). In contrast, under the attraction hypothesis, larval settlement is redirected and/or already settled fish are attracted from natural reef with no net increase in abundance (Bohnsack and Sutherland, 1985). Determining whether an artificial reef is

2 producing or attracting fish involves consideration of many processes, including reef design, size and proximity to natural reef (Osenberg et al., 2002, Brickhill et al., 2005). Assessments as to whether artificial reefs are producing or attracting fish have been made by a range of methods including; studying the length of fish over time (Szedlmayer and Shipp, 1994), the movement of fish with acoustics (Keller et al., 2017), age comparisons between fish and reef (Syc and Szedlmayer, 2012) and numerical modelling (Smith et al., 2017). Previous studies investigating these effects have supported both the attraction (Simon et al., 2011) and production (Szedlmayer and Shipp, 1994) hypotheses, indicating that both may occur under different conditions/situations and thus the results are context dependent on the design of the reef and the biological characteristics of the target species. To further confuse the situation, it appears that attraction and production both occur in majority of artificial reef cases, described by Osenberg et al. (2002) as “… end points on a continuum” rather than discrete categories. It has been accepted by most that attraction is often detrimental to an ecosystem as it removes fish from natural reefs and can make them more accessible to be removed by commercial and recreational fishers. However, some studies have suggested that attraction may simply disperse fish biomass and make them harder for fisherman to catch (Smith et al., 2015). It is clear that more research is required to better understand the attraction/ production hypotheses. The setting of goals for artificial reefs and their subsequent monitoring has progressively been recognised as a crucial factor for their success (Becker et al., 2017). Critical evaluation (ideally quantitative) of the effect that a reef has on ecological and social dimensions is important to improve their design and increase our knowledge. The ecology of the reef community, comparisons to natural habitat, site fidelity and movements/migration of key species and shifts in fishing pressure and catch rates are amongst the most commonly studied aspects during monitoring studies of artificial reefs and their fauna (Leeworthy et al., 2006; Schroepfer and Szedlmayer, 2006; Becker et al., 2017). Although the choice of monitoring method is best determined by comparing the aim of the study against the benefits and limitations of a range of possible methods,

3 some of the more common ones employed are Underwater Visual Census (UVC) (Abelson and Shlesinger, 2002), Baited Remote Underwater Video (BRUV; Florisson et al., 2018b), line fishing (Feigenbaum et al., 1989), netting (Coll et al., 2009) and Remotely

Operated Video (ROV; Harris et al., 2019). Despite recognition of the value of setting goals and targets for artificial reefs, a meta-analysis undertaken by Becker et al. (2017) found that 62% of artificial reef studies failed to mention the goals that the reef of interest aimed to achieve and those that did used almost exclusively qualitative goals. Moreover, when monitoring was undertaken the duration of studies was also short (mean = 2.4 years) and inadequate to incorporate small-scale temporal variability to allow for ecological communities to stabilize and mature (Becket et al., 2017).

1.2. Artificial reefs in Western Australia Significant investment has been made by industry, state government and community groups for the deployment of artificial reefs in Western Australia (Florisson et al., 2018a). The subsequent text details the history of deployments in WA and future plans for reef construction. The first recorded artificial reef was created in 1971 using 80 tyres off in an attempt to provide habitat for Western Rock Lobster Panulirus cygnus (Sanders, 1974). While the structure was seeded with P. cygnus, individuals were not found during subsequent surveys (Florrison et al., 2018a). The first major artificial reef in WA, which was constructed from 34,000 disused vehicle tyres and deployed in 20 m of water in Geographe Bay (near Busselton) in 1987, was based on similar reefs in South Australia (Branden et al., 1994). The reef, which was partially funded by the WA Tourist Commission, was colonised “rapidly” by fauna with three surveys conducted up to 16 months post deployment by the then Department of Fisheries finding 25 species and yielded “excellent catch rates” of recreationally targeted species (Branden et al., 1994). A subsequent survey in 2012, however, showed that many of the tyres had broken apart and lay scattered on the seabed in a 300 x 300 m area, albeit some of the pyramid and row structures were still intact (Lewis, 2015). Moreover, despite 37 fish species being recorded, only three of those were of

4 recreational importance, leading the author to conclude that “the suitability of using old tyres as an artificial habitat type for a recreational fishing reef is in question”. Another artificial reef was deployed in Geographe Bay in 1997 when the 113 m long decommissioned warship Ex-HMAS Swan was sunk to enhance and tourism (Dowling and Nichol, 2001). The site proved popular with scuba divers and, over the following 18 months, the economic impact was estimated at US$1.39 million (Dowling and Nichol, 2001). Similarly, the Ex-HMAS Perth was scuttled in King George Sound (near Albany) in 2001 (Macleod et al., 2004). In 2006, rock and recycled concrete sleepers were submerged over an area of 6,000 m2 off the coast of Dampier as an environmental offset for a community that had been damaged by land reclamation from oil and gas activities (Blakeway et al., 2013). One of the world’s first documented artificial reefs specifically designed to improve surfing conditions was deployed in 1999 (Pattiaratchi, 2003) just north of the mouth of the Swan-Canning Estuary. The surf break was constructed by adding a boomerang-shaped rock structure to an already existing limestone reef and was considered successful in achieving its goals, providing surfable waves on 142 days of 1999 (Pattiaratchi, 2003). In 2011 the first artificial reef for sea ranching, i.e. the release of cultured juveniles into unenclosed marine and estuarine environments for harvest at a larger size in “put, grow, and take” operations (Bell et al., 2008) was deployed in Flinders Bay (near Augusta, south-west Australia). The reefs (called Abitats), which were designed for Greenlip Abalone Haliotis laevigata, are 0.9 m high, constructed of concrete shaped in a pyramid, have a total surface area of 10 m2 and of 900 kg (Greenwell et al., 2019a). As of March 2019, 10,000 Abitats have been deployed in Flinders Bay and a second sea ranch (~ 400 Abitats) was established off Esperance, ~ 600 km to the east, in 2017 (Ocean Grown Abalone, 2019). The next suite of artificial reefs to be deployed in WA have all been related to recreational fishing. The first of these reefs were deployed in Geographe Bay off the coasts of Bunbury (33° 18.500'S 115° 35.900'E; 17 m deep) and Dunsborough (33° 33.962'S 115° 9.980'E; 27 m deep) in 2013 at a total project cost of $2.38 million (Recfishwest, 2020). Each reef comprises 30 ‘Fish Box’ modules, placed in six clusters of

5 five units, and deployed over a four-hectare area. The modules, which measure 3 m3 and weigh 10 tonnes, are constructed from steel-reinforced concrete with curved cross braces designed to promote localised ‘’, or flow diversion (Recfishwest, 2020;

Fig. 1.1a). These reefs were designed to increase the abundance of recreationally- important fish species, such as the sparid Pink Snapper Chrysophyrs auratus and the carangids Pseudocaranx spp. and Samson Fish Seriola hippos, and thus improve recreational fishing opportunities. The large modules on the Bunbury artificial reef were augmented in 2019, with the addition of 90 small-scale Appollo (1 m high x 1.3 m diameter and 930 kg) and large Abitat (1 m high, 2.8 m wide per side and 1,800 kg) concrete modules. The addition of these smaller modules increased the fishable area by 50% (Recfishwest, 2020). An essentially identical artificial reef to the Fish Box reefs in Geographe Bay prior to augmentation was deployed off Mandurah in 2016. This ‘Mandurah Artificial Reef’ (32° 31.59'S 115° 34.98'E; 25 m) is the topic of Chapter 3 in this thesis. Two 12.5 m tall, 10 m long, 7.8 m, wide and 50,000 kg Fish Towers became the state’s first steel artificial reefs, when they were deployed off Rottnest Island in 2017 (Tower 1: 32° 07’.527 S, 115° 27’.013 E; Tower 2: 32° 07’.461 S, 115° 26’.978 E; 45 m deep). The lattice-like steel upper part of the towers (Fig. 1.1b) aims to concentrate small schooling baitfish such as Yellowtail Scad Trachurus novaezelandiae, which, in turn, attract large predatory pelagic species such as Spanish Mackerel Scomberomorus commerson and Yellowtail Kingfish, Seriola lalandi (Recfishwest, 2020). The large area, vertical profile and differing types and shapes of the bottom part of the structure are designed to provide habitat for demersal species such as C. auratus and the highly sought-after and endemic Dhufish Glaucosoma hebraicum. The Exmouth Integrated Artificial Reef (King Reef; 21° 54.938'S 114° 11.235'E; 17 m deep; Fig.1.1c) was deployed in July 2018. This reef is the first in Australia to combine repurposed oil and gas infrastructure (i.e. six mid-depth buoys) with 49 purpose-built concrete reef modules and provides 27,000 m3 of habitat (Florisson et al., 2020). Within two years, over 90 species of fish have been observed, which is much greater than the sand habitat upon which the reef was installed.

6 The latest artificial reef to be installed in WA is the Esperance Artificial Reef (Cooper Reef; 33° 52.2638'S 121° 58.834'E; 32m deep) in 2019 (Recfishwest, 2020). The reef comprises 128 concrete modules, i.e. 96 Apollo, 16 large Abitat and 16 reef dome

(1.5 m high, 1.8 m in diameter and 2,000 kg; Fig. 1.1d), has a volume of 330 m3 and is spread out across 221 m x 116 m of sandy substrate. The reef was designed to enhance recreational fishing by creating a new, accessible and safe fishing location, with S. hippos, Pseudocaranx spp. and the Bight redfish Centroberyx gerrardi being target species (see Chapter 2). More artificial reefs are planned and in various stages of approval. The most advanced of these is the Northern Artificial Reef, located off the northern suburbs of Perth (the state capital of WA), near Ocean Reef Marina (Department of Primary Industries and Regional Development, 2019). This reef will have a minimum size of 1,000 m3 and the design is still being finalised. As part of a State Government initiative to boost the economy post COVID-19, $6 million was allocated to recreational fishing including artificial reefs for Albany and two other regional towns, one of which will be Carnarvon.

Thus, the majority of recently deployed and planned artificial reefs are designed to increase recreational fishing opportunities and catches.

Fig. 1.1. Photographs showing the different types of artificial reef modules used in Western Australia. (a) Fish Box modules using on Bunbury, Dunsborough and Mandurah; (b) Perth Fish Towers; (c) mid-buoys on the Exmouth Integrated Reef and (d) Apollo modules used in Esperance and in Bunbury following augmentation. Images kindly provided by Recfishwest.

7 1.3. Recreational fishing Recreational fishing, i.e. fishing that does not constitute the individual's primary resource to meet basic nutritional needs and are not generally sold or otherwise traded on markets (FAO, 2012) is a global activity and provides wide-ranging social and economic benefits (Arlinghaus et al., 2019). Across the industrialised world, 10.5% of the population fish recreationally, equating to an estimated 118 million people and at least 220 million worldwide (Arlinghaus et al., 2015). Fishing for recreation is an integral part of Australian culture, involving, as of 2000 - 2001, an estimated 3.4 million people or 16.8% of the population (Henry and Lyle, 2003). Participation in WA is greater than the national average at ~ 30% (612,000 people 95% CL = 535,000 to 690,000), albeit it has been declining slightly since the late 1990s (Department of Primary Industries and Regional Development, 2018). Despite the reduction in participation in recent years, the relative high proportion of fishers, combined with the state’s rapidly increasing population, has resulted in the estimated number of recreational fishers in WA more than doubling from 315,000 in 1989/90 to 711,000 25 years later (Ryan et al., 2015).

Among those recreational fishers using a boat, 57% of effort occurs in nearshore marine habitat where the majority of artificial reefs are located (see above), with lower proportions in inshore demersal (27%), estuaries (11%), pelagic (2%), offshore demersal (2%) and freshwaters (1%; Ryan et al., 2019). Across Australia, recreational fishing is estimated to contribute AU$1.8 billion annually and support 90,000 jobs (Department of Agriculture, Water and the Environment, 2020c). Among the residents of the various states and territories, fishers in WA had the second highest average expenditure, i.e. AU$706 fisher/y-1, behind Victoria ($721) and substantially greater than both New South Wales and Queensland ($555 and $407, respectively; Campbell and Murphy, 2005). When scaled for population, Campbell and Murphy (2005) estimated annual expenditure by recreational fishers in WA to be AU$338 million. Note this is vastly different to that calculated by Economic

Research Associates (McLeod and Lindner, 2018) of AU$2.4 billion, of which the greatest spend was for ‘food and drink’ at $605 million, which was not valued as an expense by the earlier study as it is not solely related to fishing. In any case, it is clear that

8 recreational fishing provides economic value, particularly for regional areas. Recreational fishers in WA, depending on their target species and whether or not they are fishing from a boat, are required to have a number of licences ranging from $ 40 to

$ 50 per license. A portion of revenue from the fishing from a boat licence is used to fund the Recreational Fishing Initiatives Fund, which has provided funds for several artificial reefs in the state including the construction, deployment and monitoring of the Mandurah artificial reef.

1.4. Aims This thesis has two main aims. 1. Conduct a systematic literature review on the characteristics of deployed artificial reefs around the world and determine how their attributes, e.g. location, materials of construction, purpose and any associated monitoring activities vary spatially and temporally (Chapter 2). As part of this, a standardised approach to describing artificial reefs is presented, which could form the basis for a formal database of future deployments and allow comparisons to enhance our understanding and evaluation of these structures. 2. Undertake monitoring to describe the fish fauna (which includes associated taxa such as cephalopods and marine mammals) present on the Mandurah artificial reef and a nearby unstructured site over a 17 month period between February 2019 and July 2020. This will provide some evidence as to whether the artificial reef is achieving its goals of increasing the abundance of certain recreationally important fish species, namely C.auratus, S.hippos and P.dentex (Chapter 3). As the Mandurah artificial reef is structurally identical to the Fish Box module reefs off Bunbury and Dunsborough, a subset of the resultant data from the study will be compared statistically to those collected by Walker (2016) and Florisson et al. (2018b) monthly over one year to determine whether the fish fauna of the three artificial reefs and the Mandurah control site differ (part of Chapter 4). This will provide valuable information about the performance of each reef and may be useful when deciding on the location of such reefs in the future.

9 Chapter 2: Artificial reefs in the Anthropocene: A review of geographical and historical trends in their design, purpose and monitoring

2.0. Abstract The long history of artificial reefs has stimulated diversity in their physical properties and deployment for a range of purposes. A systematic literature search yielded 804 scientific publications on artificial reefs. A database of their characteristics was constructed and used to investigate geographical and historical trends. A total of 1,074 unique artificial reefs from 71 countries were identified, with 89% located in the northern hemisphere, but equally distributed between east and west. Reefs were assigned to one of three categories; A: unintentional deployment, B: intentional deployment but unintentional reef, and C: intentional artificial reef. Category A reefs consisted predominantly of accidental shipwrecks. Category B reefs were primarily coastal defense structures in shallow waters and active oil and gas infrastructures at greater depths. The number of Category C reefs increased after 1965, with most in depths of 10-30 m. Most were constructed from concrete or steel, followed by rock and rubber. Usage of concrete as a material steadily increased, while those of steel and rubber decreased, coinciding with the transition from objects (materials) of opportunity to purpose-built reefs. Most reefs were deployed to enhance faunal communities or fisheries, particularly recreational fishing in North America and Australia. Monitoring was most often performed using underwater visual census but transitioned to more technologically-advanced methods, particularly in more affluent countries over recent decades. We present a standardized protocol for describing artificial reefs and urge authors to include all relevant data in their publications to allow future comparisons to enhance our understanding and evaluation of these structures.

10 2.1. Introduction The story of human history is intrinsically linked to the ocean. The earliest known fishing activities occurred 100,000 years ago (Walter et al., 2000; Pauly, 2018) and access to the goods and services provided by the ocean is thought to have facilitated coastal migrations and the peopling of the world (Pauly, 2018; Steneck and Pauly, 2019). During the Anthropocene, human population size has increased dramatically, concomitant with our use of the ocean. In 2008, it was estimated that no parts of the oceans were unaffected by human influence, with more recent work suggesting the magnitude of this deleterious impact has increased in 66% of the ocean and in 77% of all exclusive economic zones (Halpern et al., 2008; 2015). According to the FAO, annual production from commercial capture fisheries has plateaued since the 1980s (~ 80 million tonnes; FAO, 2018), but a rapidly growing global population, combined with increased demand and consumption of seafood, has led to declines in stock abundances and a greater percentage of stocks being fished at unsustainable levels. There is also increased recognition of the impact of the > 220 million recreational fishers on stocks (Cooke and Cowx, 2004; Arlinghaus et al., 2019). Reconstructed global recreational marine catches average ~ 900,000 tonnes per year, however, in some regions and for some taxa, they exceed those of commercial fishers (Freire et al., 2020). Anthropogenic influences have also resulted in the degradation of coastal habitats such as salt marshes, coral reefs, mangroves, seagrasses and shellfish reefs (Waycott et al., 2009; Polidoro et al., 2010; Beck et al., 2011; Deegan et al., 2012). These habitats act as nursery areas for many fish and invertebrate species, including some of those targeted by fisheries (Beck et al., 2001; Tweedley et al., 2016). Fish stocks and therefore fisheries productivity can be constrained by several factors, including recruitment, habitat, trophic and genetic bottlenecks (Becker et al., 2018). Fisheries enhancement, which includes both aquaculture-based and habitat- based enhancements (Taylor et al., 2017), can potentially alleviate these constraints. Aquaculture-based enhancement, which includes stock enhancement, restocking and sea ranching, can address issues of poor recruitment and potentially also low genetic

11 diversity by rearing and releasing individuals of target species (Crisp et al., 2018; Kitada, 2018). The provision of new artificial structures (e.g. an artificial reef) or restoration of a habitat (e.g. oyster reef), can provide space for colonization by sessile biota and alter water currents causing localized upwelling, thus increasing productivity and providing prey for consumers (Bailey-Brock, 1989; Okano et al., 2011; Wu et al., 2018). Moreover, these structures provide complex physical habitat for benthic invertebrates (Wu et al., 2019) and overtime fish assemblages develop and can provide increased fisheries productivity, particularly for habitat-limited species (Jan et al., 2003; Leitao et al., 2009). At a broad level, artificial reefs are defined as “human-made structures installed in aquatic habitats that serve as a substrate and/or shelter for organisms” (Lima et al., 2019). The use of artificial reefs dates back at least 3,000 years in the Mediterranean when rocks used to anchor tuna nets were discarded and the accumulated rocky habitat attracted benthic fauna and fish (Riggio et al., 2000). The success of these rubble reefs, in part, led to the ruins of ancient Greek temples and old ships being submerged. Similarly, in Japan, fishers sank vessels, rocks and bamboo with the earliest documented in 1650 (Thierry, 1988; Ito, 2011). The first documented purpose-built artificial reefs were deployed in Japan in 1952 as part of a plan to improve commercial fishing (Lee et al., 2018). This was expanded considerably in 1974, and, by 1989, 9% of the coastal shelf (< 200 m depth) had been affected. By 2001, purpose-built reefs had been deployed at ~ 20,000 sites (Bohnsack and Sutherland, 1985; Seaman and Lindberg, 2009; Ito, 2011). The concept of combining aquaculture-based enhancement with artificial reefs was first conceived in China in 1979 and termed marine ranching (Qin et al., 2020). Between 2000 and 2016, 42 demonstration areas comprising 200 marine ranches with ~ 61 million m3 of artificial reef units (covering an area of 853 km2) were created, generating substantial direct economic and ecological benefits (Zhou et al., 2019). On a far smaller scale, the seeding of cultured juvenile abalone onto artificial reefs has been done successfully in Australia (Greenwell et al., 2019a;b).

Artificial reefs have been constructed from a wide variety of materials, which can be broadly categorized as either objects (materials) of opportunity, e.g. rock/rubble, tires, wood, plastic, old vehicles/vessels and even whitegoods, or purpose-built, e.g.

12 fabricated concrete modules and steel structures (Pickering and Whitmarsh, 1997; Baine, 2001; Florisson et al., 2018a). The impetus for many such reefs has been based around the observations of increased abundances of fish species on and around these structures, and thus many reefs have been deployed to improve fishing (Riggio et al., 2000; Ito, 2011; Florisson et al., 2018b;). However, as technology and our understanding of artificial reefs has increased over time, so has the range of purposes for which they have been deployed, including recreation and tourism, coastal/habitat protection, habitat restoration and as experimental sites for scientific and technological innovation (Baine, 2001; Lima et al., 2019). Moreover, there is evidence that objects deployed underwater for other purposes (e.g. oil and gas platforms) can support productive fish populations (Claisse et al., 2014). Despite the wide range of reef types and purposes for which they are used, only 62% of publications on artificial reefs described the objective of the reef deployment (Becker et al., 2018). Furthermore, where goals were provided, they were typically qualitative, and few studies conducted enough monitoring to determine whether those stated objectives had been met.

The scarcity of reporting on the purpose and achievement of goals also extends to information on the characteristics of those reefs. As pointed out by Seaman Jr (2002), there is currently no global database for artificial reefs. Therefore, the overarching aim of this study was to use information from peer-reviewed sources to develop such a census of artificial reefs around the world and describe how their attributes, such as location, materials, purpose and monitoring activities vary geographically and historically. Based on the approach taken in collating data for this review, a standardized approach to describing artificial reefs is presented. This framework could form the basis for a formal database of future deployments and allow global comparisons to enhance our understanding and evaluation of these structures.

13 2.2. Materials and methods 2.2.1. Collation and analysis of publication data for fisheries enhancement terms The Scopus database (Elsevier), which contains > 1.4 billion records dating back to 1788 was used to calculate the annual number of peer-reviewed publications produced on various types of fisheries enhancements. Note that only Scopus was used to generate data for this analysis as the Web of Science database (Clarivate Analytics) only searches for terms in titles of publications produced before 1990, but expands this to titles, keywords and abstracts for subsequent publications, which has led some authors to make erroneous statements about publication rates (Pautasso, 2014; Calver et al., 2017;). Searches were conducted for the following keywords “habitat enhancement”, “artificial reef”, “fish aggregating device”, “aquaculture enhancement”, “stock enhancement”, “restocking” and “sea ranching” in the title, abstract or keywords. The search term “fish” was included in each of those above to distinguish enhancements for the purpose of fish and fisheries, from more generic types of enhancement. As the search was conducted in October 2019, only data up until 2018 was included in this analysis. Separate one-tailed Pearson’s correlation tests were conducted in SPSS v24 to determine whether the annual number of publications containing each of the six terms increased over time (p = < 0.05). The annual publication data for the most widely employed terms for each type of fisheries enhancement, i.e. artificial reef and stock enhancement (see Results and discussion) were subjected to Analysis of Covariance (ANCOVA) in SPSS v24. This test was used to determine whether the annual number of publications containing each term increased over time (i.e. the 50 years between 1969 and 2018) and if they differed between the two terms (p = < 0.05). The starting year was chosen as 1969, as papers containing both terms were not published frequently before this date (i.e. in two and zero years for artificial reef and stock enhancement, respectively). This test was repeated for the two types of habitat enhancement, i.e. artificial reef and fish aggregating device. Before each ANCOVA analysis data were subjected to Levene’s Test for Equal Variance. As these returned non-significant values (p = 0.484 and 0.300, respectively), no transformations were required.

14 2.2.2. Literature search A systematic literature search (Okoli and Schabram, 2010) was conducted to find scientific publications on artificial reefs. As the various databases and search engines can yield different results (e.g. Calver et al., 2017) the search for “artificial reef” & “fish” was conducted separately in Scopus, Web of Science Core Collection (formerly known as Web of Knowledge) and Google Scholar. The citation of each publication from each search was downloaded and used to create a reference library. As in Lima et al. (2019), publications were only included if they passed the inclusion criteria adapted from Moher et al. (2010). These were; i) the publication was published in a journal, ii) duplicate publications were removed and considered as a single document, and iii) publications that were available online or via inter-library loan. A total of 1,367 citations were extracted from the three searches. These were reduced to 1,027 when duplicate publications were removed (25%), with 804 of these (78%) being accessible and had at least the abstract written in English. This approach, like that for any systematic literature search, does have some limitations. Firstly, Bohnsack and Sutherland (1985) found that scientific papers made up only 32% of the 413 publications they reviewed on artificial reefs. However, we followed Becker et al. (2018) and Lima et al. (2019) in excluding, books, thesis, dissertations, technical reports, magazine articles and pamphlets. While we recognize the grey literature contains significant amounts of information, these publications are often difficult to access and the majority have not been peer-reviewed. Moreover, the information in some may also have been published in journals at a later date. We also feel there is a greater emphasis to publish in scientific journals in recent decades compared to 1983 when Bohnsack and Sutherland (1985) conducted their search. Secondly, the journals indexed in Scopus and Web of Science are predominantly written in English and therefore papers written in languages such as Chinese or Japanese would not have been included. Lastly, many reefs are deployed with no record being published in any form of accessible literature, particularly those used by artisanal fishers, commercial companies (e.g. fishing corporations) or in parts of the world where publishing in the scientific literature occurs less frequently. As an example of these

15 issues, the current study contains only a small proportion of the reported 28,700 and ~ 20,000 artificial reefs deployed in China and Japan, respectively (Ito, 2011; Si, 2012). Despite these caveats, the analyses undertaken below are considered to provide a representative account of the attributes of the 1,074 artificial reefs documented in the scientific literature and represent the most comprehensive global dataset compiled to date (Seaman Jr, 2002).

2.2.3. Data extraction and classification A suite of data on the characteristics of the artificial reefs described, and if and how they were monitored were extracted from each of the 804 publications (Table 2.1). These included the name and location, general properties (e.g. benthic area covered, depth, date of deployment), physical properties (e.g. construction material, single or multiple types of components), purpose and monitoring (e.g. monitoring type, frequency and duration). Note that the reefs were not always well described (Seaman Jr, 2002) and thus data were not always available for all characteristics of every reef

(Table 2.1).

16 Table 2.1. List of the artificial reef characteristics (bold) and, when applicable, their component categories extracted from the published peer-reviewed literature. The percentage number of papers that contained relevant information is provided for each characteristic in parentheses. Note those with an * were assigned by the authors of the current paper.

Reef name (100%*) Dominant construction material (81.6%) Latitude (100%*) Ash Longitude (100%*) Brick Country (100%*) Ceramic Continent (100%*) Concrete Africa Dead coral Antarctica Fibreglass Asia Mixed components Australia Other metal (e.g. Aluminium) Europe Plastic North America Rock (e.g. rubble, limestone) South America Rubber (e.g. tyres) World Bank income group (100%*) Shell Low income Steel Lower middle income Wood Upper middle income Purpose (75.5%) High income Faunal enhancement (i.e. increase animal abundance, usually fish, but not specifically related to a fishery) Depth (m) (64.5%) Artisanal fishing Year deployed (62.7%) Commercial fishing Location habitat (94.7%*) Recreational fishing Inshore (≤ 10 nm from shore) Mixed fishing (e.g. any combination of commercial, recreation and/or artisanal fishing) Offshore (> 10 nm from shore) Habitat enhancement (increase quantify of available habitat) Substrate (45.0%) Habitat rehabilitation Soft sediment (e.g. silt, clay, mud) Anti-trawling (prevent trawling activities) Sand Tourism (e.g. scuba diving) Rock (e.g. rock, limestone, rubble) Science (e.g. experimental reef) Coral (dead or alive) Other Macrophytes (e.g. seagrass, kelp) Monitoring method (61.1%) Reef category (97.1%*) Photography A (accidental deployment) Underwater Visual Census (UVC; i.e. visual survey performed using or scuba, with or without a recording device) B (intentional deployment but Underwater video (e.g. Baited Remote unintentional artificial reef) Underwater Video) C (intentional artificial reef) Remotely Operated underwater Vehicle (ROV) Design (79.4%) Line fishing (e.g. rod and line, longline) Single (one type of Netting/trawling (e.g. gillnetting, beam module/component) trawling) Trapping mixed (comprised of different components) Catch records Material type (82.1%) Tagging Material of opportunity (materials Poison (e.g. Rotenone) once used for other purposes, e.g. Acoustics tires) Other Purpose-built (built using materials Monitoring frequency (year-1) (46.6%) sourced exclusively for artificial reef Monitoring duration (months) (56.8%) construction)

17 The initial definition of an artificial reef used in this study, that of Lima et al. (2019), is quite broad to compensate for their large range of applications, purposes and types. Therefore, each reef was assigned to one of three categories. A: Accidental deployments. This includes, for example, ships and other vessels that have sunk as a result of conflict or accident (e.g. Lechanteur and Griffiths, 2001). B: Intentional deployment but unintentional artificial reef. Structures that have been deliberately deployed, but not for the purpose of constructing an artificial reef. Typically, these have been for economic purposes (e.g. oil and gas extraction and offshore power generation (Reubens et al., 2014), or physical structures such as jetties and breakwaters (Burt et al., 2013). C: Intentional artificial reef. This category includes structures deliberately deployed to mimic some characteristics of a natural reef (Baine, 2001). Artificial reefs constructed from repurposed oil and gas structures that are removed from their original location (e.g. mid-depth buoys on the Exmouth Integrated Artificial Reef; Florisson et al., 2020) are included as Category C, whereas artificial reefs comprised of in situ oil and gas structures (such as a platform) are included in Category B.

Maps and descriptions from each publication were used in conjunction with Google Earth to determine the approximate geographical location for each artificial reef and the measure function used to calculate their distance ‘as the crow flies’ from the nearest large landmass. Those located ≤ 18.5 km (10 nautical miles) from the coast were termed inshore and those > 18,5 km (10 nautical miles) away were classified as offshore. The geographical coordinates were used to determine the country and continent within which the reef lies. The income group for each country was extracted from the World Bank Database (https://databank.worldbank.org/home.aspx). This metric assigns the world's economies to one of four income groups using gross nation income per capita (current US$) calculated using the Atlas method. These groups are low income (≤ $996), lower-middle income ($996 - 3,895), upper-middle income ($3896 - 12,055) and high income (> $12,055).

In order to be as objective as possible, information on characteristics, such as construction materials and reef purpose, each of which can be multifactorial, all were initially recorded as stated in the source material. The ‘responses’ were subjected to

18 content analysis by two independent researchers to create categories of responses that reflected the source material but reduced the overall number of categories (Obregón et al., 2020). In the case of the year of deployment, each year was aggregated to a five- year period (year category) and decade. Finally, in cases where multiple publications were produced on the same artificial reef, data for duplicate records were merged. For example, one publication may describe the physical properties of the reef, while another detailed the results of biological monitoring. In total, the 804 publications, yielded 1,074 unique artificial reefs.

2.3. Results and discussion 2.3.1. Number of publications Of the seven terms relating to fisheries enhancement, artificial reef (980) and stock enhancement (852) were used in by far the greatest number of publications, followed by habitat enhancement (605), aquaculture enhancement (387) and restocking (392). Fish aggregating devices and sea ranching were the least used terms, with 268 and 119 publications, respectively (Fig. 1). The oldest aquaculture-based enhancement publication on restocking was published in 1942 and that for habitat enhancement on artificial reefs twenty years later in 1962. There was a significant positive correlation between the number of publications using each of the seven terms with year (all p = < 0.001). Component r values ranged from 0.810 to 0.917 for all terms except sea ranching (r = 0.573; Fig. 1). ANCOVA detected a significant difference in the number of publications using the terms artificial reef and stock enhancement between years (p = < 0.001), but not between the two terms (p = 0.174), indicating that their usage in the scientific literature was increasing at a similar rate. When the same test was run using the publication counts for artificial reef and fish aggregating devices, while they both increased over time (p = < 0.001), there were more for the former term (p = < 0.001; Fig. 1).

Despite the general increase in the number of publications over time, there were years where particularly high numbers were recorded for artificial reef and, to a lesser extent, stock enhancement (Fig. 1). These dates coincide with the proceedings from the

19 International Conference on Artificial Reef & Related Aquatic Habitats and the International Symposium on Stock Enhancement & Sea Ranching being published (Bortone, 2015). However, the extent of the disparity between years with published proceedings and neighboring years without has declined markedly since 1997, reflecting a general increase in the number of publications produced. This is true of many areas of science (e.g. Pautasso, 2012; Bornmann and Mutz, 2015), and for artificial reefs could reflect the increased interest in these fields as reefs were deployed in more countries, with research being conducted by a greater number of workers. For example, early research focused on the evaluation of different types of materials used to construct reefs and comparisons of faunal communities on artificial vs natural reefs, while more recently, monitoring was performed more frequently as a legislative requirement.

Fig. 2.1. Number of publications produced annually over the 50 years between 1969 and 2018 for the various terms of fisheries enhancement. Linear trend lines (dashed lines) and regression equations are provided for the most commonly used terms, i.e. artificial reef (df = 49; y=1.10x -10.96; R2 = 0.75) and stock enhancement (df = 49; y=1.06x -7.48; R2 = 0.75). * and # denote years in which a conference proceeding from the International Conference on Artificial Reefs & Related Aquatic Habitats and the International Symposium on Stock Enhancement & Sea Ranching, respectively, were published.

2.3.2. Overview A total of 1,074 unique artificial reefs (subsequently referred to as reefs unless otherwise stated) were identified from the 804 publications found in the systematic

20 literature search. These reefs were present in the waters of 71 different countries from all continents except Antarctica; with a geographical distribution extending from the Troms og Finnmark in Norway (70° N) and Iceland (66° N) in the north to southern

Argentina and Chile in the south (~ 42° S; Fig. 2a). A greater percentage of reefs were found in the northern (89%) than the southern hemisphere (11%; 952 vs 122), however, there was an even distribution between the eastern and western hemispheres (~ 50% each; 534 vs 538). Areas with relatively large numbers of reefs included the eastern seaboard of the USA, Western Europe, the Mediterranean Sea and the coasts of Japan, Korea and China (Fig. 2a). Reefs in the southern hemisphere were mainly located in southern Brazil and around the entire coast of Australia. The oldest reef identified was the SS Dunraven, a cargo ship that collided with a coral reef in Sha'ab Mahmoud (Egypt) in 1876 and is now a popular tourist attraction for scuba divers. The most recent reef identified from the literature was constructed following the decommissioning of two oil and gas platforms in Sarawak (Malaysia) in 2017, although artificial reefs continued to be deployed regularly.

Reefs varied markedly in size from one in the Bahamas at only 0.75 m2 that comprised of 10 cinder blocks (Allgeier et al., 2013) to a mound of dredged material in the Gulf of Mexico that was 4,800 m wide x 12,400 m long and 6 m high, with a volume of ~ 14 million m3 (Clarke et al., 1988). The median size was 643 m2. Reefs were deployed in a range of depths from some mangrove roots in Panama at < 1 m deep to an oil and gas platform off Brazil situated in 1,070 m of water. While the average depth was 39 m, the median was 21 m. In addition to being deployed in fairly shallow waters, 85% were located in inshore waters and far fewer in deeper, difficult to access remote offshore waters. Most reefs were deployed on unvegetated and structurally simple habitats e.g. sand (73%) and mud (15%), with only 5% deployed on rock, 5% in macrophytes beds and the remaining 2% on coral. A total of 656 of the reef deployments (61%) were accompanied by a monitoring program, although it is likely some of the remaining 39% were monitored, but the results of which were not described in the source publication or have not been published elsewhere. Of these programs, 18% involved the collection of data on a single sampling

21 occasion, with seasonal, monthly and annual monitoring undertaken by 31, 18 and 10% of studies, respectively. The duration of the monitoring varied with 12% occurring once, 25% lasting for a year, 17% for two years and 19% between four and 10 years. Long- term monitoring was rare, with only 0.65% of programs running for over 10 years.

Fig. 2.2. Location of (a) all artificial reefs identified from the systematic literature search and those classified in (b) Category A, (c) Category B and (d) Category C.

22 2.3.3. Geographical distribution and abundance Of the 1,074 artificial reefs recorded in the study, 43 (~ 4%) belonged to Category A, i.e. accidental deployments. While such reefs were present on all continents except

Australia (Fig. 2b), they were most common in Europe. Reefs in this category tended to be either ships that were either accidentally sunk (e.g. ran aground) or involved in combat. Category B reefs, i.e. those intentionally deployed but unintentional artificial reefs, comprised 7.4% of all reefs (n = 77). Almost two-thirds (64%) were actively involved in the extraction of oil and gas, with 11% being used for coastal protection (e.g. breakwaters). The majority of those used by the petroleum industry were in the Gulf of Mexico (USA), with other hot spots including the North Sea (UK) and North West Shelf of Australia (Fig. 2c). It is noteworthy that fewer of the oil platforms in the North Sea, which numbered 1,212 in 2009, were referred to as artificial reefs. This likely reflects that, in the USA, the conversion of obsolete offshore oil and gas structures into artificial reefs instead of onshore disposal has been used since the late 1980s (Kaiser and

Pulsipher, 2005). A total of 420 structures have been donated, with 120 and 73 platforms in Louisiana and Texas, being used to produce 83 and 35 reefs, respectively (Kaiser and Pulsipher, 2005; Techera and Chandler, 2015). In Europe, however, protests over the disposal of the Brent Spar (e.g. Dickson and McCulloch, 1996) led the OSPAR Commission excluding Rigs-to-Reefs as a viable decommissioning option (Jørgensen, 2012). Intentional artificial reefs (Category C), of which 921 were recorded, were by far the most common (88.5%). Such reefs were present in the waters of 60 countries from all continents, with 14 countries having deployed > 10 (Fig. 2d). Countries with relatively large numbers included the USA (360), Japan (68), South Korea (68), Australia and the Philippines (54), France (39) and Brazil (34). Almost 43% of the artificial reefs recorded were deployed in North America, followed by Asia (27%), Europe (17%) and Australia (6%; Fig. 3a). The marked abundance of reefs in North America mirrors the results of Lima et al. (2019), with the majority being Category C, reflecting the continent’s (particularly the USA) large focus on recreational fishing (Hughes, 2015) and the numerous studies on various aspects of

23 artificial reef design and the effects on fish populations (Bohnsack and Sutherland, 1985; Baine, 2001). It is also worth noting that in the USA, the permitting process for deploying a reef is relatively cheap and easy, allowing deployments to be made by small organizations, e.g. local fishing clubs, in addition to state government programs. For example, reefs in Alabama only require a US$31 inspection/permit fee for each individual structure before deployment within specified zones (Alabama Department of Conservation and Natural Resources, 2017), whereas in Australian waters the placement of an artificial reef requires a permit to be issued following the Environmental Protection (Sea Dumping) Act 1981. This act fulfills the international obligations under the London Protocol to regulate the ‘placement for a purpose, that purpose being an Artificial Reef, and the purpose being not contrary to the aims of the protocol’. An artificial reef permit application must be completed together with a payment of AU$10,000 (~ US$6,500) for the assessment of the application (Department of Agriculture Water and the Environment, 2020). While this process is administered by the federal government, further exemptions and approvals may be required depending on the relevant state and territory policies and legislation. Baine (2001) found that 38% of publications submitted to six journal volumes on global artificial reef research were from authors in the USA, compared to 29% from Europe and 15% from Asia. While the true number of reefs in all countries is underestimated, this effect is particularly more likely in Asia, due to limited access to the scientific literature and that publications written in languages other than English were excluded. However, despite the underrepresentation, the 286 reefs recorded in the current study do provide information on their characteristics. Very few reefs were recorded in South America (48) and Africa (16; Fig. 3a), which is consistent with the findings of Baine (2001) and Lima et al. (2019), who also found that publications from these continents represented only 1% and <0.5% of their total datasets, respectively. With the exception of South Africa (Arlinghaus et al., 2015) expenditure on marine recreational fishing was lowest in Africa compared to the rest of the world (Cisneros- Montemayor and Sumaila, 2010), as many reefs are deployed to enhance these recreational fisheries, the lack of deployments in Africa is unsurprising.

24

Fig. 2.3. Percentage contribution of artificial reefs in each of the three categories (A, B, C) in each (a) continent, (b) year category and (c) depth category (m). (d) Percentage contribution of various construction materials used to produce artificial reefs in each category. The total number of samples is provided above each column.

25 2.3.4. Trends between the categories of artificial reefs Category C reefs dominated (75-96%) the numbers of reefs in each continent except Africa where these reefs comprised only 50%, due to the relatively large proportion of Category A reefs (Fig. 3a). The number of artificial reefs ‘deployed’ was very low (< 2) in each of the five-year periods between 1875 and 1950, except for 1945, where multiple vessels were sunk during World War 2 (Fig. 3b). Generally, earlier reefs fell into Category A, but this changed from 1965 onwards when increasingly large numbers of Category C reefs were deployed. Deployments of these types of reefs increased from 2 in 1960 to 10 in the following five-year period to 95 in 1985 and remained at around this level until 2005. Thus, this type of reef was mainly responsible for the increasing number of reefs overall during this time. Category B reefs were constructed at a relatively similar pace (i.e. 2-10) in each five-year period. The dramatic increase in Category C reefs follows the initiation of a large Japanese artificial reef program starting in 1952 (Ito, 2011; Lee et al., 2018) and the passing of federal and state legislation, such as the US National Fishing Enhancement

Act (1984) and Artificial Reefs Act (1989) in Texas, the latter of which was designed to ‘promote, develop, maintain, monitor, and enhance the artificial reef potential in state waters’ (Stephan, 1990). As many reefs are deployed to attempt to increase the abundance of fauna (including particularly fishery species; see later), the implementation of artificial reef programs could be due to the increasing recognition that fish stocks were becoming overexploited (Myers and Worm 2003) and to enhance recreational fishing experiences. It should be noted, however, that there is debate over to what extent the fish present on and around artificial reefs were attracted from nearby areas, in which case their removal through fishing would facilitate over-exploitation, or produced, with this continuum likely to vary for different species, with reefs probably most important for habitat-limited species (Bohnsack, 1989; Layman and Allgeier, 2020). In Australia, the increased number of reefs being deployed is mainly due to recreational fishers paying license fees (e.g. New South Wales, Victoria and Western Australia), with a portion of those funds being used to enhance fishing, including constructing and deploying artificial reefs (e.g. Becker et al., 2017; Florisson et al., 2018a).

26 The depths in which the reefs identified in the literature search were deployed ranged from < 1 to 1,074 m, with 76% located in waters between 5 and 40 m deep (Fig. 3c). Depth patterns differed according to category, with Category C reefs having the lowest median depth (20 m), followed by Category A (27 m) and Category B (43 m). Waters between 10 and 30 m deep contained 55% of all Category C reefs, with < 5% of them deployed in waters > 50 m deep and with 87% in inshore waters. On the other hand, Category B reefs were relatively evenly spread throughout the depth categories (i.e. 5-16% in each category) and with 40% in offshore locations. This marked difference in depth patterns between these two categories echoes their usages, with the shallower Category C reefs easily accessible from small recreational fishing vessels and also cheaper to deploy and visit. For example, in both Western Australia and New South Wales the distance needed to travel by recreational fishers from a boat launching facility to the reefs is considered during the site selection process. For reefs used in scientific trials, proximity to shore reduces travel costs and shallow depths also allow sampling using snorkel and scuba rather than more expensive ROV or submersible equipment.

The uniform depth distribution of Category B reefs is due to structures in shallow waters (< 10 m) being used primarily for coastal defense and those deeper environments (> 30 m) utilized for oil and gas extraction. The materials used to construct reefs in each category differed. Unsurprisingly, 80% of Category A reefs were constructed from steel, as most were ships that accidentally sunk or were involved in combat (Fig. 3d). Steel was also a major component used in the construction of Category B reefs (67%), with concrete and rock also commonly used (13 and 10%, respectively). The first of these materials is readily used to produce oil and gas platforms and pipelines (Bull and Love, 2019), while rock and concrete are used in coastal defenses (Burt et al., 2013). By contrast, a wide range of materials have been used to construct Category C reefs due to their diverse range of purposes, many of which require different materials. The historical shift from producing reefs from objects of opportunity (e.g. rubber tires) to purpose-built materials progressed as technology improved and the environmental implications became more evident and international guidelines (e.g. London Convention and Protocol/UNEP, 2009)

27 were created to help develop legislation. Category C reefs were most commonly constructed from concrete, as it is cheap to produce, can be molded into complex structures and is considered one of the most effective materials for encouraging colonization and succession of fouling organisms due to its porous structure (Fitzhardinge and Bailey-Brock, 1989).

2.3.5. Category C reefs While the broad trends between reefs identified in the literature search from the three categories are described above, given the intended purposes of Category C reefs, more detail on their locations, construction, purpose and monitoring was undertaken. We hope this more detailed analysis could provide some guidance into the development of a standardized approach to describing future artificial reef initiatives and the creation of an online database of these structures.

Trends across income groups Countries in the progressively more affluent World Bank income groups (low through to high) had sequentially greater numbers of reefs (1, 69, 99 and 742, respectively) and average reef area (25, 41, 34,069 and 36,618 m2, respectively). This reflects the fact that countries such as China (upper-middle) and the USA, Japan, South Korea and Australia (all high income) have established artificial reef programs (e.g. Polovina and Sakai, 1989; Kim et al., 1994; Kaiser and Pulsipher, 2005; Si, 2012). Moreover, they have the resources to deploy larger structures or a greater number of modules, increasing the size of the reef. Note that only a single reef was recorded from a country in the low- income group, which is almost certainly an underestimation due to the literature search utilizing only scientific publications. The materials used to construct reefs in low-middle income countries were proportionally different from those in higher income countries (Fig. 4a). The rubber category (essentially repurposed tires) represented 41% of reefs in low-middle, but only 4 and 7% in the upper-middle and high groups, respectively. Conversely, a higher proportion of reefs in the latter two groups were made from steel (23 and 26%,

28 respectively) than in low-middle income countries (1%). Concrete reefs were used in similar proportions by all groups, except low income (i.e. 25-52%; Fig. 4b). The use of rubber over steel in reefs deployed in low-middle income countries could be due to cost, with used tires being a cheap option and were once suggested as a suitable reef material (Stone et al., 1975) before the environmental impacts of such reefs were realized (Collins et al., 1995). While tires were used extensively in the USA, restrictions or bans were brought in some states (Collins et al., 2002) and, in Japan, government funding for constructing reefs was not available for reefs constructed from waste material (Bohnsack and Sutherland, 1985).

Fig. 2.4. Percentage contribution of artificial reefs (a) constructed from different dominant materials, (b) deployed for different purposes and (c) monitored using different methods in countries in each World Bank income group. Full category names given in Table 2.1. The total number of samples is provided above each column.

29 Regardless of income group, reefs were deployed for the purposes of faunal and habitat enhancement, mixed fishing and science (Fig. 4b). Some purposes were, however, largely restricted to reefs in countries within a single group, namely, recreational fishing and anti-trawling. Recreational fishing reefs were only deployed in high-income countries where they represented 20% of all reefs. Fishing for recreation is generally practiced in higher-income countries (Arlinghaus et al., 2015) and is a key cultural feature of countries such as the USA and Australia (Henry and Lyle, 2003; Steinback et al., 2004). Reefs placed to prevent illegal trawling represented 25% of those deployed in upper-middle income countries such as Brazil which lacks the recourses to patrol its large coastline (Brandini, 2014). These reefs were also present in high-income countries where they comprise 5% of the reefs. Such reefs were exclusively deployed in the Mediterranean, i.e. southern France, Italy, Spain and Cyprus. They are recognized as effective actions to prevent illegal trawling and help to protect seagrass meadows (Sánchez-Jerez and Ramos-Esplá, 2000) and ensure the continuation of local small-scale fisheries, which are crucial from a social and economic perspective (Harmelin, 2000; Fabi et al., 2011). In terms of monitoring, underwater visual census (UVC) was the dominant method in each income group (49-87%, excluding low income; Fig. 4c). This is not surprising given it is very simple and cost-effective (Murphy and Jenkins, 2010) and many reefs were deployed in shallow depths. Cost appeared to play an important role in determining the monitoring method chosen, with the prevalence of UVC declining in higher-income countries, being replaced by visual methods employing more expensive technology, such as Remotely Operated Vehicles (ROV) and Baited Remote Underwater Video (BRUV; Fig. 4c). It is also worth noting the relatively high use of netting/trawling in the upper-middle income group (16.4%) compared to countries in the high income group (6.3%). Countries in the latter group utilized a range of monitoring methods, including expensive equipment, and therefore relied less on nets and trawl gear.

30 Geographical trends in characteristics Among the reefs identified in the literature search, the percentage contribution made by reefs built using components of a single type (e.g. concrete blocks), as opposed to different components (e.g. a selection of waste materials) was fairly consistent in

Asia, Europe and Australia (49-66%), but lower in South America (37%) and higher in North America and Africa (77 and 100%, respectively). The proportion of purpose-built reefs was more variable ranging from only 22% in North America and 39% in Australia to 57 and 71% in Europe and Asia, respectively. This reflects the deployment of a range of waste materials in both the USA and Australia in the 1970s, 1980s and 1990s (Bohnsack and Sutherland, 1985; Kerr, 1992). Concrete was the most frequently used material overall (38%), being used in all continents except Africa. Its contribution ranged from 26% in Australia, 30% in North America and as high as 58 and 67% in South America and Europe, respectively (Fig. 5a). This material is popular due to its relatively cheap cost, compared to steel, and the ability to be molded into complex structures. Moreover, a meta-analysis of 39 artificial reefs indicated that higher densities of fish were recorded on concrete than natural reefs, but that artificial reefs comprised of other materials hosted similar amounts of fish to natural reefs (Paxton et al., 2020). The lower proportional usage of this material in Asia (particularly the Philippines) and Australia is due to their relatively high use of rubber (22 and 38%, respectively). In the Philippines repurposed tires are used due to their low costs (relative to materials such as concrete) despite the productivity of tire reefs being low compared to natural habitats (Delos Reyes and Martens, 1996). While Australia does have a high number of tire reefs, this type of reef was rarely deployed after the 1980s (Kerr, 1992; Florisson et al., 2018a). Excluding the low number of reefs in Africa, the deployment of reefs constructed from steel was greatest (34%) in North America due to large numbers of oil and gas platforms decommissioned for use as artificial reefs through the Rigs-to-Reef program (Dauterive, 2000; Bull and Love, 2019). It appears that the choice of material when constructing Category C artificial reefs is heavily dependent on cost and availability.

31 The rationale for deploying artificial reefs differs substantially between continents, with the proportion of all varying, the only exception being for science. The prevalence of reefs deployed for purely scientific purposes may well be overestimated as the scientific literature was utilized as the sole data source, however, it does highlight our increasing knowledge in this relatively new field of biology (Fig. 1) and the improving technology to facilitate more effective and productive reefs. Although enhancing faunal communities was a fairly common purpose across all continents, the use of reefs for this purpose was most pronounced in Asia (58%). In many parts of this large continent, coastal environments have been degraded and marine life and fisheries production depleted (Pomeroy, 2012), and these environments are crucial for the provision of social and economic services (Stobutzki et al., 2006). As such, reefs are seen as a method of helping to manage these environments and increase their productivity (Zhou et al., 2019). Reefs deployed for recreational fishing were mainly present in Australia (64%) and North America (28%). As mentioned previously, fishing is an important component of leisure in these countries, with Australia having one of the highest participation rates among industrialized countries (Arlinghaus et al., 2015). Moreover, recreational fishing bodies actively encourage participation and fund schemes to improve fishing experiences, such as fish stocking and artificial reefs (e.g. Becker et al., 2017; Broadley et al., 2017) as they are popular with fishers (Obregón et al., 2020). Anti-trawling artificial reefs are common in Europe and South America and are recognized as important tools for managing illegal trawling activities (Sánchez-Jerez and Ramos-Esplá, 2000; Brandini, 2014). Monitoring of artificial reefs was most commonly conducted using UVC, both overall (54%) and in each continent (45-100%; Fig. 5c), likely due to its ease of use and limited cost of equipment (Hill and Wilkinson, 2004; Murphy and Jenkins, 2010). A relatively large number of artificial reefs in Australia have been monitored by both governmental and citizen scientists using BRUVs (e.g. Becker et al., 2017; Florisson et al., 2018b), reflecting a broader trend in Australia where this equipment is used to monitor a range of marine habitats (Whitmarsh et al., 2017; Langlois et al., 2018).

32 Destructive sampling using netting or trapping was rarely reported in Australia and North America (0 and 3%, respectively), but was relatively common in Asia, Europe and South America (11, 13 and 24%, respectively). These methods are also cost-effective and can be used in areas where the water can be turbid or have limited clarity due to eutrophication (Wu et al., 2019) and also involve local artisanal or commercial fishers.

Fig. 2.5. Percentage contribution of artificial reefs (a) constructed from different dominant materials, (b) deployed for different purposes and (c) monitored using different methods in each continent. Full category names given in Table 2.1. The total number of samples is provided above each column.

Reef characteristics among purposes The majority of reefs in the literature search (81%) were deployed in depths between 5 and 40 m deep, with reefs situated between 10-20 m being on average the

33 largest and declining with increasing depth (Fig. 6a). Reefs in shallow waters (< 5 m) were typically small in area and volume, possibly for navigation purposes for vessels, particularly during low , while those in deeper waters are less accessible by end- users and more expensive to build as they are often taller and so more costly to deploy. Most reefs were deployed over sandy substrates (77%) and would add habitat complexity to an otherwise barren seascape. Reefs were deployed on more structurally complex macrophyte habitats (e.g. seagrass and kelp), but only in waters < 30 m and, even then, mainly in very shallow depths (25% of all reefs in water < 5 m deep; Fig. 6b). The proportion of reefs deployed over mud increased with depth, likely as this type of habitat is more common at deeper depths (Harris, 2012). It should be noted, however, that reefs have been known to ‘sink’ into soft sediments and become completely buried and so ineffective (Wu et al., 2019). In terms of construction materials, the usage of plastic and concrete reefs declined with increasing depth, with those of steel increased (Fig. 6c), which could reflect the decommissioning of oil and gas platforms (Kaiser and Pulsipher, 2005; Bull and Love, 2019) and purpose-built reefs such as the Sydney

Offshore Artificial Reef (Becker et al., 2017). The proportion of reefs constructed using different modules (i.e. a mixed design) rather than a single module type, declined progressively with depth from 50% (0-10 m deep), to 35 (20- 30 m) and 25% (30-50 m) to only 15% in waters > 50 m deep (Fig. 6c). In contrast, there was no change in the proportion of reefs built using objects of opportunity as opposed to those that are purpose-built, with on average 56% being purpose-built. Of those publications that listed a purpose for reef deployment, science dominated the shallower depth categories, reflecting the ease at which these reefs would be readily and cheaply assessed (Fig. 6d). Most reefs designed to increase fish catch rates were in < 40 m, which would increase access for recreational, artisanal and small-scale commercial fishers and provide safer inshore fishing opportunities. Reefs designed to enhance habitat were more prevalent at greater depths perhaps as these environments are, by their nature, less complex (Fig. 6d). Studies have shown that the

34 biomass of fish and other marine biota in these deeper areas can be increased without negatively impacting shallower ecosystems (Szedlmayer and Shipp, 1994).

Fig. 2.6. (a) Mean area (1,000 m2) of artificial reefs and percentage contribution of artificial reefs (a) deployed in different substrates, (b) constructed from different dominant materials and (c) deployed for different purposed in each depth category. Full category names are given in Table 2.1. The total number of samples is provided above each column. Note no Category C reefs were deployed in depths between 200 and 500 m.

35 In terms of the purpose of reefs deployed on various substrates, science dominated those on coral and macrophytes (Fig. 7a). These studies generally investigated the potential for artificial reefs as restoration tools for coral (Ferse, 2008), fish (Beets and

Hixon, 1994) and other biota. Reefs positioned on sandy and soft habitats were deployed for a large variety of purposes, probably driven by the significant number of reefs in these habitat types, as well as the opportunity to increase the complexity of these three-dimensionally structurally simple habitats. Approximately half (53%) of all artificial reefs found were constructed from objects of opportunity, with the highest proportions for the purpose of tourism (89%) and recreational fishing (84%), with very low proportions (8%) of anti-trawling reefs. Many of the reefs deployed for tourism purposes were designed for scuba diving and involve the scuttling of vehicles (e.g. planes) and vessels, which provide interesting and novel dive sites (Shani et al., 2012) and so are constructed from steel (Fig. 7b). Recreational fishing reefs tended to be composed of rock (e.g. rubble) and rubber, which accounts for the large proportion of objects of opportunity (Fig. 7b). Such materials are easy for fishing clubs to obtain and deploy and the construction of reefs along the Gulf coast of the USA has been encouraged (Stephan, 1990; Alabama Department of Conservation and Natural Resources, 2017). The dominant material used for anti- trawling artificial reefs was concrete, often in combination with other materials e.g. protruding steel bars, as they are heavy enough not to be displaced by trawlers and are effective at snagging nets (Muñoz-Pérez, 2008; Giansante et al., 2010). Reefs constructed for artisanal fishing were mainly composed from objects of opportunity (62%), e.g. wood, rock and concrete rubble, all of which are readily available and cheap to produce (d'Cruz et al., 1994).

36

Fig. 2.7. Percentage contribution of artificial reefs in each purpose category deployed (a) on different substrates and (b) constructed from different dominant materials. Full category names are given in Table 2.1. The total number of samples is provided above each column.

Historical trends in characteristics Among the reefs identified from the literature search, there has been a transition from Category C reefs being constructed from objects of opportunity to those that are purpose-built over time. The proportion of purpose-built reefs increased from 0% in the 1940s and 1960s to 18% (1970s) and 46% (1980s) to over 60% in the 1990s and 2000s. The timing and magnitude of this shift differs among countries, even those with the financial resources to deploy more expensive purpose-built reefs. For example, the shift to a greater proportion of purpose-built reefs in the USA (52%) occurred in the 2000s, being 0 and 13% in the 1960s and 1970s and 28 and 23% in the following two decades respectively. In contrast, this shift was more marked in Australia, with > 89% of the reefs deployed being purpose-built in since the 2000s, from a base of 0-18% in the previous four decades. The proportions of deployments of single and mixed artificial reefs have not changed since the 1960s at between ~ 30-45% being comprised of mixed modules.

37 While steel and concrete were the most commonly used materials, comprising 42 to 66% of reefs deployed in each decade from the 1960s, their individual proportions were inversely correlated (Fig. 8a). The use of steel decreased by around half from 20-

33% in the 1960s to 1980s to 11 to 15% between 1990s and 2010s, whereas the uses of concrete increased sequentially from 8% in 1960s to 56% in 2000s. This highlights the shift from objects of opportunity to purpose-built structures, likely driven by the London Convention (London Convention and Protocol/UNEP, 2009) in which 89 countries signed, encapsulating 89% of the reefs in the current study. The increased usage of concrete reflects its ability to form complex structures and that, modern marine-grade mixtures are pH balanced and facilitate colonization by fouling organisms. The proportion of reefs constructed from rubber (mainly disused tires) declined markedly from 43 and 33% in the 1960s and 1970s to ~15% in the next two decades and 0% thereafter (Fig. 8b). This is due to the London protocol, the well-publicized negative effects of existing tire reefs (e.g. Osborne in Florida, USA; Morley et al., 2008) and research on the leaching of heavy metals into the marine environment (Collins et al.,

2002; Sherman and Spieler, 2006). In all decades, the majority of Category C reefs have been deployed to increase the abundance of fauna (including those of fishery species), for scientific purposes and to enhance habitat (Fig. 8b), with their relative proportions varying. In general, the deployment of reefs for recreational fishing has decreased between the 1960s and 2000s but increased again in the 2010s (Fig. 8b). This upturn is driven by deployments and research in Australia, where the availability of recreational fisher license funds has facilitated the deployment of numerous reefs and associated research (e.g. Becker et al., 2017; Florisson et al., 2018b). The increase in the proportion of reefs designed to enhance habitat could be due to recognition of the role of habitat complexity in increasing the abundance of associated fauna and also the impacts of anthropogenic degradation on coastal waters.

Underwater visual census methods were the mainstay (> 50%) for monitoring artificial reefs in all decades until the 2010s (Fig. 8c), reflecting their low cost and ease of use (Hill and Wilkinson, 2004; Murphy and Jenkins, 2010). The diversity of monitoring

38 methods has increased in recent decades, following advances in technology and mass production, improving the resolution of video and acoustic outputs and lowing the unit costs of electronic equipment. Electronic equipment, e.g. ROV has the additional advantage of being deployed in depths beyond those which humans can attain using or SCUBA (Ajemian et al., 2015). Moreover, comparative studies have shown that video methods such as BRUVs can provide more representative data on fish communities due to bait attracting rarer predatory fish and being unaffected by the potential attraction or repulsion of fish to divers as in UVC (Willis et al., 2000).

Fig. 2.8. Percentage contribution of artificial reefs (a) constructed from different dominant materials, (b) deployed for different purposes and (c) monitored using different methods in each decade. Note no Category C reefs were deployed in the 1950s. Full category names are given in Table 2.1. The total number of samples is provided above each column.

39 2.4. Conclusions and guidelines for artificial reef description This study conducted a systematic literature search of scientific peer-reviewed publications on artificial reefs to develop a database of artificial reefs around the world and describe how their characteristics such as location, materials, purpose and monitoring activities vary geographically and historically. A total of 804 scientific publications were obtained and data from 1,074 unique artificial reefs from 71 countries were extracted. Almost 90% of reefs were in the northern hemisphere, but an equal distribution existed between the east and west. Reefs were allocated into one of three categories; A: unintentional deployment, B: Intentional deployment but unintentional reef, and C: intentional artificial reef. Category A reefs consist predominantly of accidental shipwrecks at varying depths and locations. Category B was comprised of primarily coastal defense structures in shallow waters (<10 m) and active oil and gas infrastructure at depths >30 m. A rapid increase in Category C reefs since 1965 saw their dominance in waters between 10 and 30 m deep. Higher income countries deployed more reefs over a larger area with expensive materials and monitoring was performed using more technologically advanced methods. The majority of artificial reefs were made of concrete or steel, followed by rock and rubber. North America, which had the highest count of artificial reefs (43%), had the lowest number of purpose-built structures and the highest count of steel reefs, highly influenced by the Rigs-to-Reef program. Globally, the use of concrete as an artificial reef material has steadily increased as steel and rubber have decreased, coinciding with the transition from objects (materials) of opportunity to purpose-built reefs. Most were deployed with the aim to increase faunal abundance or for the benefit of fisheries, particularly recreational fishing in North America and Australia. Monitoring is most often performed using UVC but has transitioned to utilize more technologically advanced methods in more affluent countries over recent times. As pointed out by Seaman Jr (2002), there is no global database for artificial reefs that would facilitate the exchange of technology and information. The data collated for this study represents a potential starting point for the creation of such a database, however, important information is still lacking on key features for many reefs such as

40 their goals (Table 1; Becker et al., 2018). To address this knowledge gap, we propose a standardized approach to describing artificial reefs, which, if followed, could be used to facilitate the collation of such a database in the future. While such data could be written in the materials and methods section of a publication as text, pictorial examples of both a single and mixed module artificial reef are provided (Figs 9, 10). The following 15 attributes are essential for understanding the function of a reef and we recommend their inclusion in all artificial reef publications when applicable. Identifying information: i) name of artificial reef, ii) geographical location. Deployment properties: iii) date of deployment, iv) depth, v) department or organization that owns or deployed the reef (e.g. government department, aquaculture operation, recreational fishing club). Physical properties: vi) dimensions of entire reef area (i.e. length x width x height) and also including the space between modules, vii) weight and displacement volume of the modules (so not incorporating space between modules), viii) construction materials, ix) receiving substrate (e.g. silt, clay, mud, sand, rock, coral, macrophytes). Objectives and monitoring: x) type of objectives (i.e. biological/ecological, social and/or economic, xi) specified objectives (ideally quantitative; see Becker et al., 2018), xii) target species, xiii) indicators that objective has been met, xiv) monitoring method, frequency and duration and xv) self-assessed success against objective (ideally quantitative, but could be qualitative and use the five-level Likert scale). The collation of such data would allow comparisons to be made that could enhance our understanding and evaluation of these structures and help improve the efficiency and success of artificial reefs deployments in the future.

41

Fig. 2.9. Pictoral description of a single module artificial reef constructed using the criteria outlined in the current study. Figure taken from Becker et al. (2017).

Fig. 2.10. Pictoral description of a mixed module artificial reef constructed using the criteria outlined in the current study. Figures kindly provided by Subcon.

42 Chapter 3: Comparisons between the fish faunas of the Mandurah artificial reef and adjacent unstructured habitat

3.0. Abstract Fish assemblages associated with the Mandurah artificial reef and a nearby natural habitat of sand and small rocky outcrops were surveyed using Baited Remote Underwater Video systems between February 2019 and July 2020. Data from 136 one hour-long videos were analyzed using PERMANOVA to investigate differences in the number of species, total MaxN and Simpson’s diversity index and the abundance of three key recreationally targeted species between the two sites over the duration of the study. Differences in the composition of habitat and feeding guilds and the faunal community composition were also investigated between sites and months using a range of multivariate statistical approaches. A total of 69 species were recorded across the two sites, including 55 teleosts, five sharks, five rays and one marine mammal. Significant differences were found for most variables, with those between site accounting for almost all of the observed variation. A significantly greater mean number of species and individuals were found on the artificial reef (~ 12 and ~ 138, respectively) than the control site (~ 8 and ~ 63) and this was also reflected in the abundances of the three key recreational species. The artificial reef contained more pelagic species and individuals likely due to its design facilitating an upward flow diversion and the effect this had on the availability of plankton as food. Significant differences in fish composition were also driven largely by site variation, due to higher relative abundances of Neatypus obliquus, Chrysophrys auratus and Platycephalus speculator on the artificial reef and Parequula melbournensis at the control site. Habitat and food preferences were identified as likely stimuli for the observed differences between sites. While this study does not address the attraction vs production debate, it provides evidence that the Mandurah artificial reef is being utilized by recreationally important fish species.

43 3.1. Introduction Artificial reefs are used throughout the world for a range of purposes (Baine, 2001), with one of the most common being to improve fish abundances, including for fisheries (Chapter 2; Becker at al., 2017). Recreational fishing is a popular practice in industrialised countries, such as those in Europe and North America, where the time available for leisure activities and financial resources are greater (Arlinghaus et al., 2015). Participation in Australia exceeds that of most countries and is driven by cultural characteristics (Department of Agriculture, Water and the Environment, 2003). In an effort to provide safe, easily assessable fishing spots, greater catches and an enhanced fishing experience, many artificial reefs are specifically designed for recreational fishing (Chapter 2; Florisson et al., 2018a). Recreational fisher participation in the state of Western Australia is twice the national average at almost 30% (Department of Primary Industries and Regional Development, 2018) and many of the most recent artificial reef deployments and planned developments are designed to facilitate recreational fishing (see review in Chapter 1).

Differences in the fish community present on various types of natural marine habitats have long been recognised, particularly involving comparisons between bare sand and those that are complex such as macrophytes and rocky reefs (Guidetti, 2000; Gratwicke and Speight, 2005). Identifying the potential differences and ascertaining the reasons why is a topic of particular importance for artificial reef research (Carr and Hixon, 1997), as understanding the drivers can be used to evaluate and enhance the design of artificial reefs. Numerous studies have compared the faunal composition of artificial reefs to those on a range of natural habitats including rocky reef (Folpp et al., 2013; Mills et al., 2017), coral (Carr and Hixon, 1997) and/or flat substrates, i.e. sand and mud (Fabi and Fiorentini, 1994; Wu et al., 2019). Folpp et al. (2013) found that concrete reef balls deployed in three estuaries in south-eastern Australia all harboured a different fish assemblage to nearby natural rocky reef and bare sand. Species richness and abundances were greater on the artificial reefs, which were attributed to the reef harbouring its own endemic fauna, together with supporting populations of some

44 species found on the adjacent habitats (Folpp et al., 2013). Similarly, in the Gulf of Mexico, Streich et al. (2017) found that fish assemblages on a bare, soft substrate were greatly enhanced by the addition of 470 concrete pyramid modules, particularly for those species of importance to fishing. Whilst differences between fish assemblages on artificial reefs and natural habitat have been documented, some studies have found that this is not the case (Cresson et al., 2014). A study in China found that the fish and macroinvertebrate composition recorded on a mixed design artificial reef, constructed from concrete, stone and disused vessels and unconsolidated muddy sediment were similar (Wu et al., 2019). This was however attributed to the sinking of artificial reef components into the mud, rending them ineffectual, as a similar reef on coarse sediment harboured the same species as natural rocky reef. In addition to spatial changes, temporal changes have been shown to influence faunal composition in coastal marine environments (Hyndes et al., 1999; Wilber et al., 2003). This phenomenon has also been observed on artificial reefs (Hastings et al., 1976;

Valle et al., 2007). For example, Wu et al. (2019) documented pronounced temporal changes in the community on the artificial reef located in temperate waters of the Yellow Sea. These authors recorded an average of ~ 2 species and ~ 5 individuals in each trap deployment during winter months, compared to ~ 7 species and ~ 60 individuals in late summer/autumn. These changes were due to occurrence of transient warmer- water species during times of higher water (Wu et al., 2019). Similarly, but at a coarser level, Florisson et al. (2018b) found that the abundance of fish found on two concrete module artificial reefs in an embayment in south-western Australia was nearly three times greater in summer than in winter and there was an associated difference in faunal composition. Walker (2016) analysed the same data but on a monthly level, also finding a greater number of species and fish and, in particular, the abundances of two recreationally important taxa during summer months. Temporal and seasonal changes in fish assemblages on a habitat can be driven by many factors, commonly oceanographic processes influencing larval supply (Milicich, 1994; Huret et al., 2010) and seasonal migration (Hyndes et al., 1999; Cappo et al., 2000).

45 The south-west artificial reef trial in Geographe Bay (Western Australia) aimed to promote recreational fishing and therefore harbour recreationally targeted fish species (Chapter 1; Florisson et al., 2018b). Following its success, an identical artificial reef comprising 30 large 3 x 3 m concrete modules was deployed in Mandurah in 2016 ~ 100 km to the north of the Geographe Bay reefs to achieve the same aims . Almost three years after installation, the current study used a similar sampling method as Florisson et al. (2018b), i.e. a citizen science approach to deploying Baited Remote Underwater Video (BRUV) systems to sample the fish fauna (which includes associated taxa such as cephalopods and marine mammals) of the Mandurah artificial reef in each month between February 2019 and July 2020. This study specifically aimed to determine whether the fish faunal characteristics, i.e. the number of species, abundance, diversity, composition of residency, habitat and feeding guilds and community, differed between the Mandurah artificial reef and a nearby natural (control) site and over time. It was hypothesised that the Mandurah artificial reef would support a more speciose and abundant fish fauna that the nearby control site. Moreover, as in temperate environments in south-western Australian and elsewhere, that more species and individuals would be recorded during the summer than winter months. This study, which incorporates the first comparison between an artificial reef designed to enhance recreational fishing and a control site in Western Australia, will provide valuable insights into the fish communities utilising artificial reefs in temperate south-western Australia. Moreover, critical evaluation of the results will aid the design and site selection for the future artificial reefs being deployed in this region.

3.2. Materials and methods 3.2.1. Study region and artificial reef Mandurah is a coastal city (~ 100,000 residents) located ~ 65 km south of Perth, in south-western Australia, stretching over ~ 30 km of coastline (Fig. 3.1). The bathymetry of the nearshore marine environment is relatively flat and shallow with two major inshore reef lines running parallel to the land, one at ~ 9 km from shore and 10 m

46 deep (Five Fathom Bank) and the other at ~ 15 km and 30 m. Substrate in this region comprise a combination of sand and limestone reef with limited areas of seagrass (Fig. 3.2), the two main reef lines having a large degree of vertical relief (~ 3 - 5 m), also providing habitat for some corals. The nearby Peel-Harvey Estuary covers an area of ~ 130 km2 and is the largest inland water body in the southern half of Western Australia (Valesini et al., 2019). The estuary drains a catchment of 9,400 km2 and discharges that water and associated nutrients through a natural and an artificial entrance channel (Potter et al., 2016). The region experiences Mediterranean climate, with hot, dry summers and cool, wet winters (Hallett et al., 2017,). The majority of rainfall occurs during winter, the annual magnitude of which has decreased by > 200 mm since the 1970s (Valesini et al., 2019). Local wind processes have a large influence on the ocean hydrodynamics, with southerly and south-westerly winds stimulating swells from a south-west direction (Pattiaratchi et al., 1997). Any upwelling from strong southerly winds during summer is suppressed by the Leeuwin current (Twomey et al., 2007; Waite et al., 2007). The strength of this current is most prominent in autumn and winter, when austral winds are weakest (Waite et al., 2007). Coastal waters in this region are oligotrophic, due to the lack of nutrients that run-off from the ancient landscape and the Leeuwin current bringing nutrient-poor water from the tropics, southwards (Molony et al., 2011).

Fig. 3.1. Satellite image of Mandurah, Western Australia ◼, showing the location of the artificial reef  and the control site , also indicating the location of five boat ramps that are close to the artificial reef site . Satellite image provided by Google and CNES/Astrium. 47 Local fishing effort is predominantly directed from recreational fishers, although a small number of commercial rock lobster and octopus fishers operate in the area. While recreational fishing effort was considered moderate (Sumner and Williamson,

1999), fishing participation in Western Australia is amongst the highest in the world at ~ 30% and Mandurah is one of the fastest growing cities in Australia, which will increase the number of recreational fishers utilizing the marine environment (Australian Bureau of Statistics, 2018). A fish survey over limestone reef habitat off Rottnest island (~ 60 km north of Mandurah), found large numbers of the pomacentrid Chromis westaustralis, the pempherid Pempheris klunzingeri and the kyphosid Kyphosus cornelii (Holmes et al., 2013). Additionally, Goldsworthy et al., (2020) identified the labrids Notolabrus parilus and Coris auricularis and pomacentrid Parma mccullochi as typifying the local fish fauna. The Mandurah artificial reef is located between 10 and 12 km from local boat ramps in the City of Mandurah. The reef is constructed from 30 steel-reinforced concrete ‘Fish Box’ modules, each of which is 3 m3 and weighs 10 tonnes (Fig. 1.1a).

Modules are placed in six clusters of five units, deployed over a four-hectare area. The hollow cube design with curved cross braces was designed to promote localised ‘upwelling’ (Recfishwest, 2020). As upwelling is typically regarded as a large-scale phenomenon, we will now refer to this effect in our study as ‘flow diversion’. The modules and design of the Mandurah artificial reef are identical to those deployed off Bunbury and Dunsborough in Geographe Bay ~ 100 km to the south (Florisson et al., 2018b). The reef was deployed in a depth of 25 m in 2016, following consultation with local fishers and the collection of sediment samples to ensure the modules did not sink into the substrate. The reefs were deployed in an effort to increase the abundance of recreationally important fish species, such as the sparid Chrysophorus auratus, and the carangids Pseudocaranx spp. and Seriola hippos, and thus improve recreational fishing opportunities.

3.2.2. Sampling regime This study aimed to follow that of Florisson et al. (2018b) using a citizen science approach (i.e. Reef Vision). This involved using print and social media to recruit avid 48 recreational fishers who lived in close proximity to the artificial reef and fished in the region regularly. A suite of eight independent fishers, together with members of the Mandurah Offshore Fishing and Sailing Club, were provided with a BRUV (see later) and trained at a workshop. Fishers were asked to provide one video from one of the clusters of the Mandurah artificial reef and one from a nearby natural ‘control’ site (each with six hypothetical clusters) located ~ 2 km away and in the same depth in each month (Fig. 3.2). Training occurred in February 2019 with sampling scheduled to occur monthly until January 2021. A suite of the same BRUVs were kept at a local tackle shop and Murdoch University to allow researchers to collect additional samples if an insufficient number of videos (i.e. four per month from each site) were provided by recreational fishers.

Fig. 3.2. Photographs extracted from BRUV videos from the (a-c) Mandurah artificial reef and nearby (d-f) control site. Note the high density of fish and epibiont growth on the modules in (a) and (b) and the lack of habitat surrounding the reef when the BRUV faced away from the modules. Photographs on the control site (d-f) showcase the sandy habitat interspersed with small rocky outcrops with an associated macrophyte community that characterised this area.

49 The BRUV frame was constructed from a marine grade 316 stainless steel frame that was 34 cm in height and 31 cm x 31 cm in length and width, respectively. Two 1 kg lead were mounted to each side of the base to ensure the unit was negatively buoyant and sank quickly. A wide-angle action camera (GoPro Hero 2018; GoPro, Inc. California, USA) in a waterproof housing rated to 40 m was mounted on the underside of the frame and recorded footage in 1080p (1920 × 1080 pixels) with a frame rate of 60 s-1. A bait arm with a length of 500 mm from the BRUV central point and a plastic mesh bait bag (180 mm x 100 mm) was mounted above the camera. Prior to deployment, 500 g of Australian Sardine Sardinops sagax was placed in the bait bag. This fish is regarded as the most effective bait for BRUVs in Western Australia (Florisson et al., 2018b). It should be noted that length observations, including estimations, were made for fish species of interest, however, this could not be measured quantitatively with the mono BRUV units employed.

3.2.3. Data extraction Over the course of the 18 months of this study (February 2019 to July 2020) a total of 174 BRUV videos were collected from the Mandurah artificial reef (89) and control site (85), exceeding the minimum needed (four per from each site in each month) and comprising > 3 TB of data. Unfortunately, due to hard-drive failure, data from both sites in August 2019 and the control site in November 2019 were lost and thus not analysed (12 out of an expected 144 lost). Prior to analysis, for those site and month combinations where excess videos were available, those replicates in which the camera faced into the sediment or towards the surface of the water were excluded, as were videos that did not capture the reef modules and those in which the water and/or light clarity precluded the accurate identification of fish. For each video, the MaxN, i.e. the maximum number of individuals of a particular species seen in any one video frame (Whitmarsh et al., 2017), was recorded for each five-minute interval starting from the time immediately after the BRUV touched the substrate until 60 minutes later. Taxa were identified to the lowest possible taxonomic

50 level, which was typically species. Each species was assigned to a level in a suite of guilds, i.e. residency, habitat, and feeding guilds, and given a recreational fishing status. Residency guilds followed those used by Walker (2016) on the Bunbury and

Dunsborough artificial reefs where; Transient: species never present in more than two consecutive months (Costello and Myers, 1996), Resident level three: species that occurred in two or more consecutive months (Talbot et al., 1978), Resident level two: species present in at least half of the months (Relini et al., 1994), and Resident level one: species occurring in at least 87.5% of the months (Costello and Myers, 1996).

Each taxon was also assigned to a habitat guild, assigned on the basis of the part of the water column that a species primarily inhabits, their position relative to the artificial reef modules and their behaviour (Wartenberg and Booth 2015). Benthic: species that primarily had contact with the reef surface, or occupied its structure, Epibenthic: species that associated with the reef, but rarely made contact, and Pelagic: species that tended to swim above the reef. Note that this study followed Walker (2016) in assigning true benthic species, such as rays, to the epibenthic guild because of their limited contact with the reef structure.

The feeding guilds used followed Elliot et al. (2007). Namely; Detritivore: species which feed on decaying organic matter and associated organisms, Herbivore: species that consume plant material, including phytoplankton, Omnivore: species that feed on both plant and animal material, Zooplanktivore: species that predate primarily on small crustaceans in the water column, Zoobenthivore: species that feed on animals that live in, on, or immediately above the substratum, Piscivore: species predating predominantly on fish, and Opportunist: those species whose feeding behaviour and food preferences change depending on food availability.

Finally, species were designated as either targeted or non-targeted by recreational fishers in this region and, if so, were assigned a quantitative score ranging from 1 (low value) to 5 (iconic). This subjective classification considered fish edibility, sports fishing value and local fishing culture and was based on L. Ramm’s 15 years of experience as a recreational fisher, including seven years working in tackle stores in

51 Mandurah. Although the data was not collected, observations were made on the proportion of males and females for those fish species that exhibit sequential hermaphroditism.

3.2.4. Statistical analyses All statistical analyses were performed using the PRIMER v7 multivariate statistics software package (Clarke and Gorley, 2015) with the PERMANOVA+ add-on (Anderson et al., 2008).

3.2.4.1. Preliminary analyses As there has been conjecture on the length of time a BRUV needs to be deployed to provide a representative sample of the fish fauna (Whitmarsh et a., 2017), preliminary analyses were undertaken to determine whether the 60-minute soak time used was sufficient. To this end, the MaxN of each species in each five-minute interval of each of the 132 videos analysed from the Mandurah artificial reef and nearby control site were subjected to DIVERSE (Clarke and Gorley, 2015) to calculate the number of species, total MaxN (i.e. the sum of the MaxN values for individual species) and Simpson’s Diversity Index. The resultant values were used to produce a rarefaction curve with associated 95% confidence limits for each of the variables for the artificial reef and control site. Changes in the fauna community were examined using a range of multivariate techniques. Firstly, the MaxN values of each species in each five-minute interval of each sample were dispersion-weighted, by dividing the counts for each species by their mean index of dispersion, i.e. the average of the variance to mean ratio in replicate videos (Clarke et al., 2006). This ensures all species have equivalent variability by down- weighting the abundances of heavily-schooling species, such as the carangids Trachurus novaezelandiae and Pseudocaranx dentex, whose numbers are erratic among samples relative to those species which return more consistent values, e.g. the aracanid Anoplocapros amygdaloides and labrid Coris auricularis. These data were then square- root transformed to balance the contribution of relatively abundant species, compared to those with lower MaxN values (Clarke et al., 2014a). 52 The transformed data for each five-minute time interval were then averaged across the samples for each site and used to construct a Bray-Curtis resemblance matrix. This was subjected to hierarchical agglomerative clustering (CLUSTER; Clarke et al.,

2014a) to determine the time intervals that were ≥ 95 % similar in terms of their species composition. The same matrix was used to construct a non-metric Multi-Dimensional Scaling (nMDS) ordination plot (Clarke, 1993), which provides a visual representation of the changes in fish faunal composition over time for both sites. Finally, the dispersion-weighted and square-root transformed MaxN data for each time interval at each site was used to construct a shade-plot (Clarke et al., 2014b). The shade plot is a visualization of the data matrix, where a white space for a species demonstrates that it was not recorded, while the depth and colour of shading, ranging from grey shades through to black, represents increasing values for the abundance of that species in that time interval. The averaged samples (x axis of the plot) are ordered from lowest to highest time intervals for each of the two sites. Species (y axis of the plot) are ordered to optimise the seriation statistic ρ by non-parametrically correlating their resemblances to the distance structure of a linear sequence and constrained by a cluster dendrogram (Clarke et al., 2014a).

3.2.4.2. Diversity measures and abundance of key species and functional guilds Prior to subjecting the data for six univariate variables (i.e. number of species, total MaxN and Simpson’s Diversity Index and the MaxN abundances of C. auratus, Pseudocaranx dentex and Seriola hippos) to Permutational Multivariate Analysis of

Variance (PERMANOVA; Anderson et al., 2008), each variable was tested to ascertain if a transformation was required to meet the test assumptions of homogeneity of variance and normality. This was achieved by plotting the loge mean against the loge standard deviation of every group of samples (i.e. each month for each site) to determine the linear slope of the relationship and comparing it to the criteria in Clarke et al. (2014a).

This analysis indicated that total MaxN required loge(X+1) transformation. The data for each of the six variables were used to construct a Euclidean distance matrix, which were, in turn, subjected to a two-way PERMANOVA to determine if the

53 values for that variable differed significantly between Site (2 levels; Mandurah artificial Reef and control site) and Date (17 levels; each month between February 2019 and July 2020). In these, and all subsequent tests, the null hypothesis of no significant difference among a priori groups was rejected if the significance level (P) was ≤ 0.05 and the relative influence of each term in the model was quantified using the percentage contribution of the mean squares of that term to the total mean squares. Given the large number of variables, to reduce the number of subsequent plots only those terms that were significant and had a percentage mean squares of ≥ 30% were visualised. The same two-way PERMANOVA design and approach was employed to investigate whether the square-root transformed percentage contribution of species and individuals to each of the habitat guilds varied with Site or Date. The any trends identified were examined using stacked bar graphs of mean values. This same process was repeated for feeding guilds.

3.2.4.3. Faunal composition The data for the MaxN abundance of each species in each replicate sample from each site in each month were dispersion-weighted and square-root transformed. These pre-treated data were used to construct a Bray-Curtis resemblance matrix and subjected to the same two-way PERMANOVA design used above. To visually examine trends in the main effects, the above Bray-Curtis matrix was separately subjected to the bootstrap averages routine (Clarke and Gorley 2015). The averages of repeated bootstrap samples (bootstrapped averages) for level of the main effect were used to construct an mMDS ordination plot. Superimposed on the plot was (i) a point representing the group average (i.e. the average of the bootstrapped averages) and (ii) the associated, smoothed and marginally bias-corrected 95% bootstrap region, in which 95% of the bootstrapped averages fall. Note that due to the relatively large number of levels in the Date factor (17), the two-dimensional plot did not provide an accurate representation (stress value

> 0.3) and thus a three-dimensional plot was employed, which is calculated without the 95% bootstrap regions. Patterns in the Date × Site interaction were displayed visually using a centroid nMDS plot. This was produced using a distances among centroids

54 matrix, which creates averages in the ‘Bray-Curtis space’ from the four replicate samples representing each site in each of the 17 months (Lek et al., 2011). Shade plots were constructed from the pre-treated data to illustrate the trends exhibited by species with respect to Site and the Date × Site interaction. Note that as 69 species were recorded throughout the study, many of which only occurred in a few samples, each shade plot was restricted to the 35 ‘most important’ as determined by Whittaker’s Index of Association (Clarke and Gorley 2015).

3.3. Results 3.3.1. Preliminary analysis Accumulation curves for the mean number of species, total MaxN and Simpson’s Index reached an asymptote before 60 minutes on both the artificial reef and control site (Fig. 3.1). The mean number of species on the artificial reef reached an asymptote faster than the control site (cf. Figs 3.1a,b). For example, in comparison to the maximum value (i.e. that after 60 minutes), after five minutes, 59% of the mean number of species had been detected on the artificial reef, whereas at the control site this value was only 28%. Furthermore, 90% of the maximum value was detected after 35 minutes on the artificial reef, but after 50 minutes at the control site. A similar trend was present for total MaxN, with values for the artificial reef plateauing after just 20 minutes, where they reached 82% of the maximum value, compared to only 22% at the control site (Fig. 3.1c). The accumulation curve for total MaxN at the control site increased markedly between 50 and 55 minutes (i.e. 44 vs 98 fish, respectively), which was due to large schools of fish recorded in two of the 64 videos (Fig. 3.1d). In contrast to the other variables, values for Simpson’s Index changed very little over time, ranging from 0.69 to 0.74 and 0.72 to 0.78 on the artificial reef and control site, respectively (Figs 3.1f,g). The composition of fish fauna at both sites increased in similarity as the length of the videos increased (Fig. 3.2). The samples after 5 and 10 minutes were the most distinct from the longer time intervals, with similarities of 80 (artificial reef) and 60% (control site). Samples from the artificial reef obtained a similarity of 95% after 40 minutes, with those from the control site reaching the same state after 45 minutes

55 (Fig. 3.2a). Although the faunal community at both sites decreased in similarity over time, those of the control sites were slightly more dispersed, indicating they were more variable (Fig. 3.2.b). The total MaxN of most species recorded, particularly those that were abundant e.g. Neatypus obliquus and Parequula melbournensis, reached a stable value after 20 minutes (Fig. 3.3). While there were some species whose abundance took longer to stabilize, e.g. the batoids Myliobatis australis and Dasyatis brevicaudata, this typically occurred by 45 minutes and these species were generally observed in only low numbers. These results suggest that, in the case of the univariate metrics of abundance and diversity and faunal composition, the values had stabilized well before the end of the footage and thus the 60 minutes of footage was considered sufficient to provide a representative sample of the fish community. Note this time is longer than that used by Florisson et al. (2018b) on the artificial reefs in Bunbury and Dunsborough ~ 100 km to the south.

56

Artificial reef Control site (a) (b)

14 14 s

s 12 12

e

e

i

i c

c 10 10

e

e

p

p

s

s

8 8

f

f

o

o

r

r 6 6

e

e b

b 4 4

m

m

u

u N N 2 2

0 0

0 5 0 5 0 5 0 5 0 5 0

5

0 5 0 5 0 5 0 5 0 5 0

5

1 1 2 2 3 3 4 4 5 5 6

1 1 2 2 3 3 4 4 5 5 6 Time (mins) Time (mins)

(c) (d) 200 200

150 150

N N

x x

a a

M M

100 100

l l

a a

t t

o o T 50 T 50

0 0

0 5 0 5 0 5 0 5 0 5 0 0 5 0 5 0 5 0 5 0 5 0

5 5

1 1 2 2 3 3 4 4 5 5 6 1 1 2 2 3 3 4 4 5 5 6 Time (mins) Time (mins)

(e) (f)

0.9 0.9

x x

e e

d d n

0.8 n 0.8

I I

s s

' '

n n

o o

s s p

0.7 p 0.7

m m

i i

S S

0.6 0.6

0 5 0 5 0 5 0 5 0 5 0 0 5 0 5 0 5 0 5 0 5 0

5 5

1 1 2 2 3 3 4 4 5 5 6 1 1 2 2 3 3 4 4 5 5 6 Time (mins) Time (mins)

Fig. 3.1. Accumulation curves for the mean (a,b) number of species, (c,d) total MaxN and (e,f) Simpson’s Diversity Index for the Mandurah artificial reef (a, c, e) and control site (b, d, f), respectively. Error bars represent ±95% confidence limits.

57 (a) 20 Site AR CS 30

40

50

y

t

i

r

a l

i 60

m

i S

70

80

90

100

0 5 0 5 0 5 0 5 0 5 0 0 5 0 5 0 5 0 5 0 5 0

5 5

1 1 2 2 3 3 4 4 5 5 6 1 1 2 2 3 3 4 4 5 5 6 Samples (b) 5 2D Stress: 0.01 Site 10 15 2025 AR 303540 45505560 CS Site AR CS

5 15 20 10 25 30 35 4045505560

Fig. 3.2. Cluster dendogram (a) and nMDS ordination plot (b), derived from a Bray-Curtis resemblance matrix, constructed from the dispersion-weighted and square-root transformed and averaged MaxN abundances of each species in five-minute intervals, captured from BRUV footage on the Mandurah artificial reef (AR) and the control site (CS). The dashed horizontal line in (a) indicates a Bray-Curtis similarity of 95%.

58

5 0 5 0 5 0 5 0 5 0 0 5 0 5 0 5 0 5 0 5 0

2 0

1 1 2 2 3 3 4 4 5 5 6 5 1 1 2 2 3 3 4 4 5 5 6 5 Site AR Ophthalmolepis lineolata Pentapodus vitta CS Neatypus obliquus Chrysophrys auratus Platycephalus speculator Coris auricularis Acanthaluteres vittiger 1 Parequula melbournensis Upeneichthys vlamingii Notolabrus parilus Pseudocaranx dentex Austrolabrus maculatus Eupetrichthys angustipes Chelmonops curiosus Seriola hippos Pempheris klunzingeri Caesioscorpis theagenes Rhabdosargus sarba 0 Trachurus novaezelandiae Dasyatis brevicaudata Anoplocapros lenticularis Eubalichthys mosaicus Arripis georgianus Sillago ciliata Meuschenia flavolineata Seriola lalandi Enoplosus armatus Meuschenia freycineti Sphyraena novaehollandiae Anoplocapros amygdaloides Seriola dumerili Zeus faber Octopus tetricus Chaetodon assarius Myliobatis australis Aracana aurita Sillaginodes punctata Scomber australasicus Goniistius gibbosus Sepioteuthis australis Aptychotrema vincentiana Pomatomus saltatrix Neosebastes pandus Clupeidae spp. Mustelus antarcticus Scobinichthys granulatus Sepia apama Heterodontus portusjacksoni Trygonoptera ovalis Trygonorrhina fasciata Gymnothorax woodwardi Omegophora armilla Charcharhinus brevipinna Siphonognathus beddomei Gobiesocidae sp. Labridae sp. Chromis klunzingeri Parupeneus chrysopleuron Ostorhinchus victoriae Thunnus maccoyii Pseudolabrus biserialis Aulohalaelurus labiosus Arripis truttaceus Meuschenia venusta Tursiops truncatus Chaetodermis penicilligerus Orectolobus maculatus

Fig. 3.3. Shade plot constructed from the dispersion-weighted, square-root transformed and averaged MaxN abundances of each species recorded from consecutive five-minute intervals of BRUV footage from the Mandurah artificial reef (AR) and a control site (CS). Darker shading indicates a higher abundance of a species.

3.3.2. Primary analysis 3.3.2.1. Description of the fauna A total of 69 taxa were recorded on the 132 videos from the Mandurah artificial reef and nearby control site (Table 3.1). Among these, 55 were teleosts, whilst the remainder consisted of five sharks, five rays, three cephalopods and a marine mammal (Tursiops truncatus). Six species contributed more than 5% to the overall total MaxN i.e. Clupeidae spp. (26.2%), the carangid Trachurus novaezelandiae (15.4%), the labrid

59 Coris auricularis (11.5%), the serranid Caesioscorpis theagenes (8.0%), the pempherid Pempheris klunzingeri (6.2%) and the microcanthid Neatypus obliquus (6.0%). Notable contributions were also made by the sparid C. auratus and carangid P. dentex (4.8 and

3.9%, respectively), which are highly sought-after recreational species and are target species for the artificial reef. (Table 3.1). While the total number of species found on the artificial reef (49), was less than the 56 recorded at the nearby control site, over twice as many fish were recorded on the reef habitat (i.e. 9,447 vs 4,044; Table 3.1). Each of the eight species listed above and the sphyraemid Sphyraena novaehollandiae represented ~ 5% or more of the total fish fauna. In contrast, only four species Clupeidae spp., T. novaezelandiae, C. auricularis and P. klunzingeri contributed ~ 5% or more of the total fish fauna at the control site. Among the 69 species recorded, 36 were recorded at both sites (albeit sometimes in markedly different abundances), while 13 and 20 species were found at the artificial reef and control site alone, respectively (Table 3.1). Twenty-four of the species that occurred at both sites were more abundant on the artificial reef, with the reverse being true for

11 species and one (Sillaginodes punctatus) was recorded in equal amounts at both sites. Some of the pelagic schooling species, i.e. C. theagenes, S. novaehollandiae and the carangid Seriola hippos were all recorded in far greater numbers (93 - 153 times more abundant) on the artificial reef (Table 3.1). Less marked but still pronounced differences were also observed for the recreationally important C. auratus, P. dentex and platycephalid Platycephalus speculator whose abundances were 8 to 12 times more abundant on the artificial reef. Considering recreationally targeted species only, 15 (nine teleosts, three cephalopods and three sharks) were found at the artificial reef, compared to 12 species (nine teleosts, two cephalopods and one shark) at the control site (Table 3.1.). While the latter does include the iconic scombrid Thunnus maccoyii, only a single individual was recorded. The percentage contribution of recreationally targeted species with a score of three or greater (considered of particular importance) was 10.8% overall, but substantially greater on the artificial reef than the control site, i.e. 13.8% vs 3.9%.

60 Table 3.1. Number of individuals observed (N) mean MaxN abundance (X), standard deviation (SD), rank by abundance (R) and percentage contribution (%) of each of the 69 species recorded on the Mandurah Artificial Reef and nearby control site using BRUVs between February and July 2020. The presence of a species in only one of the two sites is denoted by a ‘Y’ and the numbers > 1 in the ×AR column (green shading) represent the magnitude of the increased abundance of that species on the artificial reef. The total number of individuals, MaxN and species are also provided. Species shaded in grey contributed ≥ 5% of the individuals at a site. Habitat guilds (HG) P = pelagic; E = Epibenthic and Benthic = B. Feeding guild (FG) Omnivore = OV; Zooplanktivore = ZP; Zoobenthivore = ZB; Piscivore = PV and Opportunist = O. Species targeted by recreational fishers (Rec) and denoted with a Y are assigned a score from 1 to 5 based on their prestige.

Overall Artificial reef Control site Comparison Species HG FG Rec Rec # N X SD R % N X SD R % N X SD R % AR CS x AR Clupeidae spp. P ZP N 0 3535 26.78 216.79 1 26.20 1500 22.06 181.90 2 15.88 2035 31.80 249.97 1 50.32 0.74 Trachurus novaezelandiae P ZP Y 2 2078 15.74 60.25 2 15.40 1877 27.60 82.13 1 19.87 201 3.14 7.87 4 4.97 9.34 Coris auricularis E ZP Y 1 1556 11.79 11.95 3 11.53 1053 15.49 11.24 4 11.15 503 7.86 11.50 2 12.44 2.09 Caesioscorpis theagenes P ZP N 0 1081 8.19 28.24 4 8.01 1074 15.79 37.92 3 11.37 7 0.11 0.67 26 0.17 153.43 Pempheris klunzingeri B ZP N 0 838 6.35 15.36 5 6.21 475 6.99 13.17 8 5.03 363 5.67 17.47 3 8.98 1.31 Neatypus obliquus E ZP N 0 812 6.15 6.68 6 6.02 735 10.81 6.11 5 7.78 77 1.20 2.18 8 1.90 9.55 Sphyraena novaehollandiae P O Y 2 659 4.99 37.33 7 4.88 652 9.59 51.78 6 6.90 7 0.11 0.44 26 0.17 93.14 Chrysophrys auratus E ZB Y 5 530 4.02 4.84 8 3.93 490 7.21 4.74 7 5.19 40 0.63 1.45 14 0.99 12.25 Pseudocaranx dentex E O Y 4 513 3.89 6.83 9 3.80 458 6.74 8.42 9 4.85 55 0.86 1.91 12 1.36 8.33 Acanthaluteres vittiger B OV Y 2 271 2.05 2.35 10 2.01 179 2.63 2.60 10 1.89 92 1.44 1.88 6 2.27 1.95 Scomber australasicus P ZP Y 2 186 1.41 11.39 11 1.38 177 2.60 15.81 11 1.87 9 0.14 0.79 21 0.22 19.67 Parequula melbournensis E ZP N 0 137 1.04 1.11 12 1.02 30 0.44 0.90 21 0.32 107 1.67 0.94 5 2.65 0.28 Upeneichthys vlamingii E ZB N 0 133 1.01 1.30 13 0.99 84 1.24 1.32 14 0.89 49 0.77 1.24 13 1.21 1.71 Platycephalus speculator E PV Y 3 131 0.99 1.14 14 0.97 120 1.76 1.07 12 1.27 11 0.17 0.42 20 0.27 10.91 Seriola hippos P PV Y 5 102 0.77 2.45 15 0.76 101 1.49 3.26 13 1.07 1 0.02 0.13 39 0.02 101.00 Ophthalmolepis lineolata E ZB N 0 91 0.69 1.24 16 0.67 91 1.42 1.46 7 2.25 Y Notolabrus parilus B ZB Y 1 77 0.58 0.74 17 0.57 14 0.21 0.44 24 0.15 63 0.98 0.79 10 1.56 0.22 Pentapodus vitta E ZB N 0 72 0.55 1.02 18 0.53 2 0.03 0.17 42 0.02 70 1.09 1.24 9 1.73 0.03 Austrolabrus maculatus B ZB N 0 66 0.50 0.80 19 0.49 37 0.54 0.82 19 0.39 29 0.45 0.78 15 0.72 1.28 Siphonognathus beddomei E OV N 0 62 0.47 4.00 20 0.46 62 0.97 5.72 11 1.53 Y Rhabdosargus sarba E ZB Y 2 52 0.39 1.18 21 0.39 52 0.76 1.56 15 0.55 Y Arripis georgianus P PV Y 3 45 0.34 1.47 22 0.33 44 0.65 2.01 16 0.47 1 0.02 0.13 39 0.02 44.00 Pomatomus saltatrix P PV Y 2 41 0.31 3.48 23 0.30 40 0.59 4.85 17 0.42 1 0.02 0.13 39 0.02 40.00 Sepioteuthis australis E ZB Y 5 38 0.29 0.88 24 0.28 14 0.21 0.91 24 0.15 24 0.38 0.85 17 0.59 0.58 Chelmonops curiosus B ZB N 0 38 0.29 0.67 24 0.28 38 0.56 0.85 18 0.40 Y Sillago ciliata E ZB Y 4 37 0.28 0.96 26 0.27 31 0.46 1.26 20 0.33 6 0.09 0.39 29 0.15 5.17 Eupetrichthys angustipes B ZB N 0 30 0.23 0.50 27 0.22 5 0.07 0.31 34 0.05 25 0.39 0.61 16 0.62 0.20 Dasyatis brevicaudata E O N 0 28 0.21 0.45 28 0.21 24 0.35 0.54 22 0.25 4 0.06 0.24 31 0.10 6.00 Eubalichthys mosaicus B OV N 0 22 0.17 0.41 29 0.16 20 0.29 0.52 23 0.21 2 0.03 0.18 37 0.05 10.00 Myliobatis australis E ZB Y 1 21 0.16 0.39 30 0.16 12 0.18 0.42 29 0.13 9 0.14 0.35 21 0.22 1.33 Scobinichthys granulatus B OV Y 2 20 0.15 0.40 31 0.15 2 0.03 0.17 42 0.02 18 0.28 0.52 18 0.45 0.11 Sillaginodes punctatus E ZB Y 5 18 0.14 0.39 32 0.13 9 0.13 0.42 31 0.10 9 0.14 0.35 21 0.22 1.00 Enoplosus armatus E ZB N 0 16 0.12 0.39 33 0.12 13 0.19 0.47 28 0.14 3 0.05 0.28 34 0.07 4.33

61 Table 3.1. Continued.

Overall Artificial reef Control site Comparison Species HG FG Rec Rec # N X SD R % N X SD R % N X SD R % AR CS x AR Anoplocapros lenticularis B ZB N 0 15 0.11 0.34 34 0.11 14 0.21 0.44 24 0.15 1 0.02 0.13 39 0.02 14.00 Seriola dumerili P PV Y 4 14 0.11 0.75 35 0.10 14 0.21 1.04 24 0.15 Y Gymnothorax woodwardi B ZB N 0 12 0.09 0.31 36 0.09 12 0.19 0.43 19 0.30 Y Trygonorrhina dumerilii E ZB N 0 11 0.08 0.30 37 0.08 2 0.03 0.17 42 0.02 9 0.14 0.39 21 0.22 0.22 Seriola lalandi P O Y 5 11 0.08 0.43 37 0.08 11 0.16 0.59 30 0.12 Y Aracana aurita B ZB N 0 10 0.08 0.27 39 0.07 8 0.12 0.32 32 0.08 2 0.03 0.18 37 0.05 4.00 Chromis klunzingeri E ZP N 0 8 0.06 0.61 40 0.06 8 0.13 0.88 25 0.20 Y Trygonoptera ovalis E ZB N 0 7 0.05 0.22 41 0.05 7 0.11 0.31 26 0.17 Y Neosebastes pandus B ZB Y 1 7 0.05 0.22 41 0.05 3 0.04 0.21 37 0.03 4 0.06 0.24 31 0.10 0.75 Sepia apama E ZB Y 3 6 0.05 0.21 43 0.04 1 < 0.01 0.12 47 0.01 5 0.08 0.27 30 0.12 0.20 Meuschenia flavolineata B OV Y 2 6 0.05 0.24 43 0.04 6 0.09 0.33 33 0.06 Y Mustelus antarcticus E ZB Y 5 5 0.04 0.19 45 0.04 2 0.03 0.17 42 0.02 3 0.05 0.21 34 0.07 0.67 Omegophora armilla E ZB N 0 4 0.03 0.17 46 0.03 4 0.06 0.24 31 0.10 Y Zeus faber E PV Y 4 4 0.03 0.17 46 0.03 4 0.06 0.24 35 0.04 Y Anoplocapros amygdaloides B ZB N 0 4 0.03 0.17 46 0.03 4 0.06 0.24 35 0.04 Y Carcharhinus brevipinna P PV Y 4 3 0.02 0.15 49 0.02 3 0.05 0.21 34 0.07 Y Heterodontus portusjacksoni E O N 0 3 0.02 0.15 49 0.02 2 0.03 0.17 42 0.02 1 0.02 0.13 39 0.02 2.00 Goniistius gibbosus E ZB N 0 3 0.02 0.15 49 0.02 3 0.04 0.21 37 0.03 Y Octopus aff. O.tetricus B ZB Y 2 3 0.02 0.15 49 0.02 3 0.04 0.21 37 0.03 Y Meuschenia freycineti B OV Y 2 3 0.02 0.15 49 0.02 3 0.04 0.21 37 0.03 Y Chaetodon assarius B ZB N 0 3 0.02 0.15 49 0.02 3 0.04 0.21 37 0.03 Y Arripis truttaceus P ZP Y 2 1 < 0.01 0.09 55 0.01 1 0.02 0.13 39 0.02 Y Gobiesocidae sp. B ZB N 0 1 < 0.01 0.09 55 0.01 1 0.02 0.13 39 0.02 Y Ostorhinchus victoriae E ZB N 0 1 < 0.01 0.09 55 0.01 1 0.02 0.13 39 0.02 Y Parupeneus chrysopleuron E ZB N 0 1 < 0.01 0.09 55 0.01 1 0.02 0.13 39 0.02 Y Pseudolabrus biserialis B ZB N 0 1 < 0.01 0.09 55 0.01 1 0.02 0.13 39 0.02 Y Thunnus maccoyii P PV Y 5 1 < 0.01 0.09 55 0.01 1 0.02 0.13 39 0.02 Y Meuschenia venusta B OV Y 2 1 < 0.01 0.09 55 0.01 1 0.02 0.13 39 0.02 Y Tursiops truncatus P PV N 0 1 < 0.01 0.09 55 0.01 1 0.02 0.13 39 0.02 Y Chaetodermis penicilligerus B ZB N 0 1 < 0.01 0.09 55 0.01 1 0.02 0.13 39 0.02 Y Aulohalaelurus labiosus B ZB N 0 1 < 0.01 0.09 55 0.01 1 0.02 0.13 39 0.02 Y Labridae sp. E ZB N 0 1 < 0.01 0.09 55 0.01 1 0.02 0.13 39 0.02 Y Aptychotrema vincentiana E ZB Y 1 1 < 0.01 0.09 55 0.01 1 < 0.01 0.12 47 0.01 Y Orectolobus maculatus E ZB Y 3 1 < 0.01 0.09 55 0.01 1 < 0.01 0.12 47 0.01 Y Siphamia cephalotes E ZP N 0 1 < 0.01 0.09 55 0.01 1 0.02 0.13 39 0.02 Y Odax acroptilus E ZB N 0 1 < 0.01 0.09 55 0.01 1 0.02 0.13 39 0.02 Y Total number of individuals 13,491 9,447 4,044 Total mean MaxN 102.11 138.88 63.19 Total number of species 69 49 56

62 Across both sites, eight species were identified in more than 50% of the videos, of which five were present in over 60%; i.e. C. auricularis (78.8%), the monacanthid Acanthaluteres vittiger (75.8%), N. obiquus (68.2%), the mullid Upeneichthys vlamingii (62.9%) and C. auratus

(60.6%; Table 3.2). Only six of the ten most abundant species were amongst the ten that were most frequently recorded, with two others contributing < 1% to the overall abundance

(Parequula melbournensis and U. vlamingii; cf. Tables 3.1, 3.2). On the artificial reef, six species of teleost were found in more than 85% of the videos, with only one, the gerreid

P. melbournensis, being recorded more frequently at the control site. Thirty-five species were recorded more frequently on the artificial reef (Table. 3.2). Among these, the most pronounced was N. obliquus, which was identified in almost every video on the artificial reef (97.1%), but present in only 37.5% of those from the control site. Each of the three target species had a much higher affinity for the artificial reef, i.e. C. auratus (92.7% vs 26.6%), P. dentex (85.3% vs 21.9%) and S. hippos (36.8% vs 1.6%). This trend was also observed in P. speculator. Thirty-three species were observed more often at the control site, including P. melbournensis, the labrids Notolabrus parilus and Ophthalmolepis lineolata and the nemiperid Pentapodus vitta, but typically these species are of little recreational value.

63 Table 3.2. Number of observations (Obs) and associated frequency of occurrence (%F) for each of the 69 species recorded on the Mandurah artificial reef and control site using BRUVs between February 2019 and July 2020. The total number of samples is also provided. Each species is assigned to a Residency guild (Res); Transient (T) or residency levels 1, 2 or 3. Species with a %F ≥ 50% are shaded in grey.

Overall Artificial reef Control site Species Res Rec Rec # x Obs %F x Obs %F x Obs %F Coris auricularis 1 Y 1 104 78.79 61 89.71 43 67.19 Acanthaluteres vittiger 1 Y 2 100 75.76 60 88.24 40 62.50 Neatypus obliquus 1 N 0 90 68.18 66 97.06 24 37.50 Upeneichthys vlamingii 1 N 0 83 62.88 49 72.06 34 53.13 Chrysophrys auratus 1 Y 5 80 60.61 63 92.65 17 26.56 Parequula melbournensis 3 N 0 74 56.06 17 25.00 57 89.06 Pseudocaranx dentex 1 Y 4 72 54.55 58 85.29 14 21.88 Platycephalus speculator 1 Y 3 71 53.79 61 89.71 10 15.63 Pempheris klunzingeri 1 N 0 60 45.45 43 63.24 17 26.56 Notolabrus parilus 2 Y 1 57 43.18 13 19.12 44 68.75 Trachurus novaezelandiae 1 Y 2 52 39.39 35 51.47 17 26.56 Austrolabrus maculatus 2 N 0 44 33.33 25 36.76 19 29.69 Pentapodus vitta T N 0 39 29.55 2 2.94 37 57.81 Ophthalmolepis lineolata T N 0 38 28.79 38 59.38 Seriola hippos 2 Y 5 26 19.70 25 36.76 1 1.56 Dasyatis brevicaudata 2 N 0 26 19.70 22 32.35 4 6.25 Chelmonops curiosus 2 N 0 26 19.70 26 38.24 Eupetrichthys angustipes 3 N 0 25 18.94 4 5.88 21 32.81 Caesioscorpis theagenes 2 N 0 24 18.18 22 32.35 2 3.13 Sepioteuthis australis 3 Y 5 23 17.42 7 10.29 16 25.00 Eubalichthys mosaicus 2 N 0 20 15.15 18 26.47 2 3.13 Myliobatis australis 3 Y 1 20 15.15 11 16.18 9 14.06 Rhabdosargus sarba 2 Y 2 19 14.39 19 27.94 Sphyraena novaehollandiae 2 Y 2 18 13.64 13 19.12 5 7.81 Arripis georgianus 2 Y 3 18 13.64 17 25.00 1 1.56 Sillago ciliata 3 Y 4 18 13.64 14 20.59 4 6.25 Scobinichthys granulatus T Y 2 18 13.64 2 2.94 16 25.00 Sillaginodes punctatus 3 Y 5 16 12.12 7 10.29 9 14.06 Anoplocapros lenticularis 2 N 0 14 10.61 13 19.12 1 1.56 Scomber australasicus 2 Y 2 13 9.85 11 16.18 2 3.13 Enoplosus armatus 3 N 0 13 9.85 11 16.18 2 3.13 Gymnothorax woodwardi T N 0 11 8.33 11 17.19 Aracana aurita 3 N 0 10 7.58 8 11.76 2 3.13 Trygonorrhina dumerilii T N 0 10 7.58 2 2.94 8 12.50 Trygonoptera ovalis T N 0 7 5.30 7 10.94 Neosebastes pandus T Y 1 7 5.30 3 4.41 4 6.25 Sepia apama T Y 3 6 4.55 1 1.47 5 7.81 Seriola lalandi 3 Y 5 6 4.55 6 8.82 Meuschenia flavolineata 3 Y 2 5 3.79 5 7.35 Mustelus antarcticus T Y 5 5 3.79 2 2.94 3 4.69 Seriola dumerili 3 Y 4 4 3.03 4 5.88 Zeus faber 3 Y 4 4 3.03 4 5.88 Anoplocapros amygdaloides 3 N 0 4 3.03 4 5.88 Omegophora armilla T N 0 4 3.03 4 6.25 Chaetodon assarius T N 0 4 3.03 4 5.88 Clupeidae spp. T N 0 3 2.27 1 1.47 2 3.13 Goniistius gibbosus T N 0 3 2.27 3 4.41 Octopus aff. O.tetricus T Y 2 3 2.27 3 4.41 Heterodontus portusjacksoni T N 0 3 2.27 2 2.94 1 1.56 Meuschenia freycineti 3 Y 2 3 2.27 3 4.41 Charcarhinus brevipinna T Y 4 3 2.27 3 4.69 Siphonognathus beddomei T N 0 2 1.52 2 3.13 Pomatomus saltatrix T Y 2 2 1.52 1 1.47 1 1.56

64 Table 3.2. Continued.

Overall Artificial reef Control site Species Res Rec Rec # x Obs %F x Obs %F x Obs %F Chromis klunzingeri T N 0 2 1.52 2 3.13 Arripis truttaceus T Y 2 1 0.76 1 1.56 Gobiesocidae sp. T N 0 1 0.76 1 1.56 Ostorhinchus victoriae T N 0 1 0.76 1 1.56 Parupeneus chrysopleuron T N 0 1 0.76 1 1.56 Aptychotrema vincentiana T Y 1 1 0.76 1 1.47 Pseudolabrus biserialis T N 0 1 0.76 1 1.56 Thunnus maccoyii T Y 5 1 0.76 1 1.56 Meuschenia venusta T Y 2 1 0.76 1 1.56 Tursiops truncatus T N 0 1 0.76 1 1.56 Orectolobus maculatus T Y 3 1 0.76 1 1.47 Chaetodermis penicilligerus T N 0 1 0.76 1 1.56 Aulohalaelurus labiosus T N 0 1 0.76 1 1.56 Labridae sp. T N 0 1 0.76 1 1.56 Siphamia cephalotes T N 0 1 0.76 1 1.56 Odax acroptilus T N 0 1 0.76 1 1.56 Number of samples 132 68 64

3.3.2.2. Diversity measures and abundance of key species The number of species was shown by two-way PERMANOVA to differ between Date and Site and the interaction between these two main effects was also significant (Table 3.3a). The main effect of Site explained the vast majority of the variance (87%), with Date and the two-way interaction effect only accounting for 7% and 4%, respectively. The mean number of species found at the artificial reef (12) was substantially greater than the 8 at the control site (Figs 3.4a,b). The relatively minor Date effect was driven by the fact that in most sampling occasions 10 - 13 species were recorded, except during June 2019 and May and June 2020, when this declined to a mean of ~ 7 (data not shown). The decline in species richness in June 2019 was more pronounced at the artificial reef than control site, thereby accounting for the significant Date × Site interaction (data not shown). Similarly, total MaxN differed significantly with all terms in the model, with most of the variability explained by Site (81.6%; Table 3.3b). The mean total MaxN value for the artificial reef was more than double the corresponding value for the control site (138 vs 63; Fig. 3.4b). Once again, temporal changes were relatively minor (%MS = 5.5), with differences due to peaks in abundance in March 2019 and 2020 (~ 300), low values between April and July of both years (20 - 70) and intermediate values (90 - 140) in the remaining months (data not shown). The interaction was caused by peaks in abundance

65 in March 2019 and 2020 occurring on different sites, i.e. the artificial reef and control site, respectively (data not shown).

Table 3.3. Degrees of freedom (df), mean squares (MS), percentage contribution of mean squares (%MS), pseudo-F values (pF) and P values (P) from two-way PERMANOVA tests on (a) number of species, (b) total MaxN, (c) Simpson’s diversity index and the MaxN abundances of three recreationally important species, (d) Chrysophorus auratus, (e) Seriola hippos and (f) Pseudocaranx dentex recorded on the Mandurah artificial reef and control site between February 2019 and July 2020. Significant differences (P < 0.05) are bolded and those considered to have had a large contribution to the total variance of each test (> 30%) are shaded in grey.

(a) Number of species (b) Total MaxN Factors df MS %MS pF P MS %MS pF P Date 16 34 7.21 4.093 0.001 4 5.52 4.935 0.001 Site 1 406 87.12 49.424 0.001 59 91.26 81.640 0.001 Date × Site 15 18 3.91 2.216 0.011 1 2.11 1.888 0.044 Residual 99 8 1.76 1 1.12

(c) Simpson's Diversity Index (d) Chrysophorus auratus MaxN Factors df MS %MS pF P MS %MS pF P Date 16 0.089 44.58 2.807 0.002 26 1.78 2.676 0.004 Site 1 0.007 3.47 0.218 0.659 1398 96.30 145.060 0.001 Date × Site 15 0.072 36.07 2.271 0.004 18 1.26 1.892 0.032 Residual 99 0.032 15.88 10 0.66

(e) Seriola hippos MaxN (f) Pseudocaranx dentex MaxN Factors df MS %MS pF P MS %MS pF P Date 16 13 12.20 4.011 0.001 44 3.40 1.177 0.274 Site 1 78 73.31 24.095 0.001 1170 90.68 31.403 0.001 Date × Site 15 12 11.44 3.760 0.001 39 3.03 1.050 0.440 Residual 99 3 3.04 37 2.89

Simpson’s diversity index values did not differ significantly with Site but did vary with Date (%MS = 44.6%) and the Date × Site interaction (%MS = 36.1; Table 3.3c). Index values ranged from 0.70 to 0.87 in all sampling occasions except March 2019 (0.55) and May 2020 (0.45; Fig. 3.4c). The temporal pattern of values at both sites was similar except in May 2020, with those at the control site markedly lower than the artificial reef (i.e. 0.16 and 0.74, respectively; cf. Fig. 3.4d,e).

66 (a) (b) 14 200

12

s e

i 150

c 10

e

N

p

x

s a

8

f

M

o 100

l

r

6 a

t

e

o

b T

m 4

u 50 N 2

0 0

S S

R R

A C A C Site Site (c)

1.6 x

e

d

n

I

y 1.2

t

i

s

r

e v

i 0.8

D

s

' n

o 0.4

s

p

m i

S 0

9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0

1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2

l l

t

r r r r

y v c y

b n p n b n

c

u u

a p a p

a o e a

e u e a e u

J J

O

A A

J J J

M M

F F

N D

S

M M Date (d) (e)

1.5 1.5

x x

e e

d d

n n

I I

1.2 1.2

y y

t t

i i

s s r

r 0.9 0.9

e e

v v

i i

D D

s

s 0.6 0.6

' '

n n

o o s

s 0.3 0.3

p p

m m

i i S

S 0 0

9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0

1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2

l l l l

t t

r r r r r r r r

y v c y y c y

b n p n b n b n p n b n

c c

u u u u

a p a p a p a p

a o e a a e a

e u e a e u e u e a e u

J J J J

O O

A A A A

J J J J J J

M M M M

F F F F

N D D

S S

M M M M Date Date Fig. 3.4. Mean (a) number of species and (b) total MaxN recorded on the Mandurah artificial reef (AR) and control site (CS) between February 2019 and July 2020. Mean Simpson’s diversity index in each date (c) combined, (e) artificial reef only and (f) control site only. Error bars represent ± 95% confidence intervals.

The abundance of C. auratus significantly differed with all terms in the model, with Site accounting for almost all the variation at 96.3% (Table 3.3di). Far more C. auratus were observed on the artificial reef than control site, i.e. 7.2 vs 0.6 (Fig. 3.5a). Although significant, differences among sampling occasions for both sites combined showed no consistent trend, with the interaction likely being caused by the almost consistent presence of C. auratus at the artificial reef site (15 of 16 sampling occasions), compared to being recorded in 10 of the 16 sampling occasions at the control sites and only with a Max of > 1 in April 2020 (data not shown).

67 As with C. auratus, the MaxN abundance of S. hippos differed significantly with Date, Site and the Date × Site interaction (Table 3.3e). On the basis of %MS, Site accounted for the majority of the variance (73.3%), followed by Date (12.2%) and their interaction (11.4%). Site differences were driven by the almost absence of S. hippos on the control site, whilst on average 1.5 individuals were recorded on the artificial reef per video (Fig. 3.5b). Over time, MaxN abundances were low (< 1 fish) between February and November 2019 and May and July 2020, with a greater number of fish recorded in the remaining months and particularly February 2020 (5; data not shown). Seriola hippos was only recorded once on the control site (February 2020), but in ten months on the artificial reef. The abundance of P.dentex only differed with Site, (Table 3.3f), with a far larger mean number observed on the artificial reef than control site (i.e. 6.7 and 0.9, respectively; Fig. 3.5c).

(a) (b)

10 2.5

s u

t 8 2.0

a

s

r

o

u

p

a

6 p 1.5

i

s

h

y

r

a

l

h o

p 4 1.0

i

o

r

s

e

y

S r

h 2 0.5 C

0 0

S S

R R

A C A C Site Site (c)

10

x e

t 8

n

e

d

x 6

n

a

r a

c 4

o

d

u e

s 2 P

0

S

R

A C Site Fig. 3.5. Mean total MaxN of three recreationally important species, (a) Chrysophorus auratus, (b) Seriola hippos and (c) Pseudocaranx dentex recorded at the Mandurah artificial reef (AR) and control site (CS) between February 2019 and July 2020. Error bars represent ± 95% confidence intervals.

68 3.3.2.3. Contribution of taxa and individuals to feeding and residency guilds Almost half of the 69 species recorded across both sites during the 17 months of the study were not present in more than two consecutive months, classifying them as transient (Fig. 3.6). Furthermore, only 14 (20%) of the species had a residency level of 3, meaning that they were identified in two or more consecutive months and 12 (17%) were level 2 residents (i.e. found in ≥ 50% of the videos). Species that had a high affinity towards their habitat (either artificial reef or control site), classified as those found in ≥ 87.5% of videos, were assigned to resident level 1 and included nine species (13%).

40 35 30 25 20 15

Number Number species of 10 5 0 Transient Resident 3 Resident 2 Resident 1 Residency guild

Fig. 3.6. Number of taxa in each residency guild found on the Mandurah artificial reef and the control site between February 2019 and July 2020.

Two-way PERMANOVA tests detected significant differences in the percentage composition of habitat and feeding guilds among Date, Site and their interaction using both the number of species and number of individuals (Table 3.4). In all cases Site explained most of the variability (60 to 85% of the total mean squares), with the remaining terms in the model never accounting for more than 19% and in some cases were as low as 7%. Differences in the contribution of the various habitat guilds by the number of species across sites was due mainly to a higher percentage of pelagic species on the artificial reef than control site (i.e. 14 vs 6%) and correspondingly slightly lower contributions of epibenthic (58 vs 62%) and benthic species (27 vs 31%; Fig. 3.7a). A

69 similar pattern was detected when based on the number of individuals with a far greater proportion of pelagic species on the artificial reef (30 vs 12%) and lower contribution of benthic species (13 vs 25%; Fig. 3.7b).

Based on feeding guilds, the artificial reef contained a more even distribution of species about such guilds (i.e. each > 10% of the total; Fig. 3.7c). In contrast, the control site harboured very low numbers of piscivorous (2.9%) and opportunist (3.7%) species compared to the artificial reef (12.7 and 11.5%, respectively). Instead, almost half of the species found on the control site were zoobenthivores (49.9%), compared to 34.9% on the artificial reef (Fig. 3.7c). When based on abundance, piscivores and opportunists were more prevalent on the artificial reef (5.0 and 10.1%, respectively) than the control site (1.4 and 3.3%, respectively; Fig. 3.7d). The control site did, however, support a higher proportion of zoobenthivores than the artificial reef (29.5 vs 18.3%) and also contained twice the contribution of omnivores (8.4 vs 4.0%).

Table 3.4. Degrees of freedom (df), mean squares (MS), percentage contribution of mean squares (% MS), pseudo-F values (pF) and P values (P) from two-way PERMANOVA tests on the percentage contribution of habitat guilds by (a) number of species and (b) individuals and feeding guilds by (c) number of species and (d) individuals, recorded on the Mandurah artificial reef and control site between February 2019 and July 2020. Significant differences (P < 0.05) are bolded and those considered to have had a large contribution to the total variance of each test (> 30%) are shaded in grey.

(a) Habitat guilds by number of species (b) Habitat guilds by number of individuals Factors df MS %MS pF P MS %MS pF P Date 16 1014 18.52 4.390 0.001 1578 15.14 3.774 0.001 Site 1 3295 60.19 14.265 0.001 6791 65.16 16.242 0.001 Date × Site 15 934 17.07 4.045 0.001 1634 15.68 3.908 0.001 Residual 99 231 4.22 418 4.01

(c) Feeding guilds by number of species (d) Feeding guilds by number of individuals Factors df MS %MS pF P MS %MS pF P Date 16 1126 6.71 3.023 0.001 1376 11.60 3.783 0.001 Site 1 14164 84.50 38.043 0.001 8977 75.71 24.683 0.001 Date × Site 15 1101 6.57 2.957 0.001 1141 9.62 3.136 0.001 Residual 99 372 2.22 364 3.07

70 (a) (b)

(c) (d)

Fig. 3.7. Percentage contribution of fish to (a) habitat guilds by number of species and (b) number of individuals, (c) feeding guilds by number of species and (d) number of individuals, recorded on the Mandurah artificial reef and control site between February 2019 and July 2020. Habitat guilds: P = pelagic; E = Epibenthic and Benthic = B. Feeding guild: Piscivore = PV; Omnivore = OV; Opportunist = O; Zoobenthivore = ZB and Zooplanktivore = ZP.

3.3.2.4 Faunal composition Two-way PERMANOVA detected a significant difference in the composition of the faunal communities between Date, Site and the Date × Site interaction (Table 3.5). Over 90% of the variability was explained by Site, with Date and the Date × Site interaction contributing only 3.9% and 3.2%, respectively. Thus, subsequent focus is mainly placed on the investigating differences between Site. Given the significant two- way interaction, temporal trends are briefly explored for each site separately.

Table 3.5. Degrees of freedom (df), mean squares (MS), percentage contribution of mean squares (%MS), pseudo-F values (pF) and P values (P) from a two-way PERMANOVA test on the faunal composition recorded on the Mandurah artificial reef and control site in each month between February 2019 and July 2020. Significant differences (P < 0.05) are bolded and those considered to have had a large contribution to the total variance of each test (> 30%) are shaded in grey.

Factors df MS %MS pF P Date 16 3462 3.87 2.676 0.001 Site 1 81826 91.51 63.235 0.001 Date × Site 15 2835 3.17 2.191 0.001 Residual 99 1294

71 The significant and influential Site effect is clearly illustrated on the mMDS plot where the bootstrapped averages and 95% bootstrapped regions for each site are widely separated and entirely discrete (Fig. 3.8a). Differences between the site are mainly due to larger abundances of species such as N. obliquus, C. auratus, P. dentex and P. speculator on the artificial reef, of which the latter three are all recreationally important (Fig. 3.9). Other targeted species were also present in relatively moderate numbers on the artificial reef, but had essentially limited or no presence on the control site, e.g. the sparid Rhabdosargus sarba, S. novaehollandiae, S. hippos and the arripid Arripis georginaus. There was a suite of species that were more abundant on the control site, including P. melbournensis, N. parilus, P. vitta and O. lineolata, but they are typically not targeted by recreational fishers (Fig. 3.9). Several of the most abundant species, i.e. U. vlamingii, C. auricularis and A. vittiger were recorded in relatively equal abundances at both sites. Although hard to visualise due to the large number of months where data were collected and thus coded for season, there was a seasonal shift on composition. On the three-dimensional bootstrapped mMDS plot, winter (blue symbols), spring (green) and summer (red) exhibit relatively tight grouping albeit with some overlap, while those for autumn (yellow) are more widely dispersed (Fig. 3.8b). This overlap even when ‘simplified’ to seasons illustrates why the Date term contributed little to the percentage mean squares (Table 3.5). A centroid nMDS plot was constructed to better visualise the interaction term. As expected, points representing the two sites are clearly separated, with those for the artificial reef located near the top and away from those of the control site near the bottom (Fig. 3.8c). For each site there is a cluster of samples and then a group of ‘outliers’ for June and July 2019 and May, June and July 2020, which is more pronounced for the control site and thus helps contribute to the significant Date × Site interaction. Pairwise PERMANOVA tests on the faunal compositions between each pair sampling occasions yielded 73 differences (53% of all pairwise comparisons) on the artificial reef and 45 differences (38%) at the control site (Appendix 3). At both sites, most of the comparisons involved winter (June, July) and the autumn month of April,

72 due to the reduced abundance or absence of common species at both sites, specifically U. vlamingii and A. maculatus, both which were found in relatively high abundances in other months (Fig. 3.10). Differences in April 2019 at the artificial reef and in April 2020 at the control site were largely driven by the reduced abundances of T. novaezelandiae and C. curiosis and of U. vlamingii and E. angustipes, respectively, which were present in majority of other months. On the artificial reef, where more temporal differences were detected, additional pairwise differences were detected in March (67% of pairwise tests across both years), January (75% in 2020) and October (67% in 2019; Appendix 3a). The low abundance of C. auratus in March 2019 and its absence in that month in 2020 was the main reason for this difference (Fig. 3.10). The fauna in October 2019 and January 2020 contained a relatively low species richness compared to adjacent months. At the control site, September 2019 was different from 60% of the other months due to the absence of common species, e.g. A. maculatus, N. obliquus and C. auratus (Fig. 3.10). Key recreational fishing species such as C. auratus and S. hippos were more abundant on the artificial reef in autumn and summer months, whilst P. dentex was abundant in all months except those in winter of both years (Fig. 3.10).

73 (a) 2D Stress: 0.05 Site AR 20 CS av:AR

av:CS

2 S

D 0 M

-20

-20 0 20 MDS1 (b)

2D Stress: 0.08 Site (c) Oct19 AR Jul19 Feb20 Dec19Jan20 CS Jun19 May19Nov19 Apr19Sep19 Site Feb19Mar19 Mar20 Apr20 AR CS Jun20 May20 Jul20

Apr20

Jul20 Mar19 Jun20 Jul19Mar20Jan20 Jun19 Feb20 Dec19 Apr19 Sep19Oct19 May19

May20 Feb19

Fig. 3.8. (a) Two and (b) three-dimensional mMDS plots constructed from bootstrap averages of the dispersion-weighted and square-root transformed MaxN abundances of each species in each (a) Site and (b) Date. Group averages (larger symbols) and 95% region estimates (shaded areas on 2D plot only) fitted to the bootstrap averages are provided. (c) nMDS plot derived from a distance-among-centroid matrix constructed from a Bray-Curtis resemblance matrix of dispersion-weighted and square-root transformed MaxN abundances of each species in each site on each sampling occasion.

74 S

3 R Rec

A C Scomber australasicus N Sphyraena novaehollandiae Y Enoplosus armatus 2 Anoplocapros lenticularis Arripis georgianus Sillago ciliata Caesioscorpis theagenes Rhabdosargus sarba 1 Eubalichthys mosaicus Seriola hippos Chelmonops curiosus Dasyatis brevicaudata 0 Trachurus novaezelandiae Myliobatis australis Pempheris klunzingeri Austrolabrus maculatus Pseudocaranx dentex Platycephalus speculator Chrysophrys auratus Upeneichthys vlamingii Coris auricularis Acanthaluteres vittiger Neatypus obliquus Parequula melbournensis Notolabrus parilus Sillaginodes punctata Sepioteuthis australis Eupetrichthys angustipes Pentapodus vitta Ophthalmolepis lineolata Scobinichthys granulatus Trygonorrhina fasciata Gymnothorax woodwardi Trygonoptera ovalis Sepia apama

Fig. 3.9. Shade plot constructed from the dispersion-weighted, square-root transformed MaxN abundance of each species found on the Mandurah artificial reef (AR) and the control site (CS). Darker shading indicates a higher abundance of a species and those that are targeted by recreational fishers are denoted by a blue diamond.

75

Fig. 3.10. Shade plot constructed from the dispersion-weighted, square-root transformed MaxN abundance of each species recorded on the Mandurah artificial reef and the control site in each month where sampling occurred. Darker shading indicates a higher abundance of a species and those that are targeted by recreational fishers are denoted by a blue diamond.

76 3.4. Discussion

This study used BRUVs to determine the fish and associated fauna (e.g. cephalopods and marine mammals) present on the Mandurah artificial reef between February 2019 and July 2020. As previous studies on the fish fauna of artificial reefs designed to enhance recreational fishing have focused solely on the artificial habitat (Walker, 2016; Florisson et al., 2018b), sampling was also undertaken on nearby natural habitat (i.e. control site). Although not directly addressing the attraction vs production debate, the results showed that the characteristics of the fish fauna on the two sites differed greatly and helped to estimate the success of the artificial reef in meeting its pre-deployment goals of increasing the abundance of C. auratus, Pseudocaranx spp. and S. hippos and providing improved recreational fishing opportunities.

3.4.1. Number of species and total MaxN

3.4.1.1. Site differences

Although the number of species and total MaxN differed between sites and over time, the former factor explained most of the variability in both cases. Greater numbers of species and individuals were recorded on artificial reefs than over rocky reef and sand in three estuaries in south-eastern Australia (Folpp et al., 2013), and between an artificial reef and natural coral reef in the (Rilov and Benayahu, 2002). A similar result was also found by Streich et al. (2013), with species richness being greater at the artificial reef than a sandy control site. Given that the Mandurah artificial reef and control site are only ~ 2 km away, located equidistant from the natural reef line (Five- fathom Bank) and in a similar depth, the differences are almost certainly generated by the provision of habitat and feeding preferences of different fish species rather than geographical differences. It should be noted that some small patch reefs (bombies) do exist in the area, which could influence the extent of connectivity for mobile benthic species. One of the key attributes provided by the Fish Box modules on the Mandurah artificial reef is the 3 m vertical relief. This has been recognized as having a positive

77 influence on the richness and abundance of fish on natural (Molles, 1978) and artificial reefs (Strelcheck et al., 2005) due to a potential increase in the provision of food and shelter, compared to flat substrate. Rilov and Benayahu (2002) found that larval recruitment of fish was up to three times higher on a vertical than horizontal artificial reef. If this phenomenon did occur on the Mandurah artificial reef, then it may have promoted a greater recruitment of fish over the three years prior to the initiation of sampling. Although not experimentally investigated in this study as mono and not stereo- BRUVs were employed, it was observed that the average length of some species did increase over the 17 months of sampling (L. Ramm, personal observation). Scuba dives on the artificial reef in January 2019, immediately prior to the commencement of sampling in February of that year, confirmed that the average length of C. auricularis, A. vittiger, C. auratus and P. dentex was well below the average in the area on established natural reef systems. Additionally, only females of the protogynous hermaphrodites C. auricularis and A. vittiger were recorded during particularly the first six months of sampling. Around November 2019, some individuals of these species began the change from female to male and by February 2020 a relatively normal proportion of male and female fish existed (L. Ramm, unpublished data). These observations are pertinent to the production vs attraction debate and, if substantiated, could indicate that production is more dominant than attraction. However, as such analyses cannot be performed using the data in this thesis, this notion is speculative. Twice the number and abundance of pelagic species were recorded on the artificial reef than the control site. This could be due to the design of the artificial reef modules, which use cross-braces through the center of the modules to promote flow diversion, aimed to increase plankton above the reef. This provides a food source for large schools of zooplanktivores, e.g. Clupeidae spp. and T. novaezelandiaei at a low energy cost. These were the two most abundant species and occupied a position above the modules where zooplankton and phytoplankton plumes are formed, which, in turn, attracted piscivorous fish such as Seriola hippos. It is

78 noteworthy that 101 individuals of this species were recorded on the artificial reefs compared to only one at the control site.

3.4.1.1. Site differences in abundance of recreationally important species Three key recreationally important species (C. auratus, S. hippos and P. dentex) were identified as targets for the Bunbury and Dunsborough artificial reefs of identical design (Florisson et al., 2018b). These species were also mentioned by Recfishwest (2020) as targets for the Mandurah artificial reef and thus, a statistical comparison was undertaken to evaluate the performance of the reef against this objective. The abundances of each species were markedly greater on the artificial reef than the control site. For example, the mean number of C. auratus was seven times higher on the artificial reef and occurred in 93% of videos compared to only 27% at the control site, strongly indicating that this species prefers structured habitat. Visual analysis and supported by recreational fishing indicated that the C. auratus recorded were juveniles, (10 and 25cm in length; L. Ramm, personal observation). In New Zealand, the densities of juvenile

C. auratus less than two years old was related to the availability of food (Kingett and Choat, 1981). Similarly, Parsons et al. (2015) found that post-settlement of individuals below six cm fork length on artificial seagrass units occurred in intermediate water velocities. The moderate water velocity increased copepod availability without also increasing energic costs to the fish (i.e. as would be the case in higher velocities). Although the current study cannot confirm the size of C. auratus on the artificial reef, it appears clear that the availably of food, particularly zooplankton, has a major influence on habitat selection. Upward flow diversion generated by the artificial reef may therefore attract post-settlement C. auratus and potentially provide habitat for successive life stages. Several studies also demonstrate that C. auratus has a preference for complex habitat, but that the type of habitat was not influential (Terres et al., 2015;

Parsons et al., 2016). This would suggest that an artificial reef can provide adequate habitat for juvenile C. auratus. Interestingly, Ross et al. (2008) found that 1-2 year old pink snapper (mean fork length = 13.9 cm) preferred to occupy sand flats adjacent to rocky reef habitat. Thus, the Mandurah artificial reef may be ideal as it is located on an

79 expansive sandflat. This is supported by the fact that densities of C. auratus on the Mandurah artificial reef were far greater than on identical reefs in Bunbury and Dunsborough (Chapter 5).

The abundance of S. hippos was also significantly greater on the artificial reef. These large opportunistic piscivores (Rowland, 2009) are commonly found over rocky reef and adjacent sandy areas (Hutson et al., 2007), of which the artificial reef provides the only such high-profile habitat in the vicinity. Vertical relief has been identified as having a positive correlation with abundance of the congeneric Seriola dumerili (Kellison and Sedberry, 1998) which was found in small numbers only at the artificial reef site. This may explain the higher abundance of S. hippos on the Mandurah artificial reef. Individuals and/or schools of this species can move large distances during their life (Gillanders et al., 2001; Hutson et al., 2007; Rowland, 2009), typically travelling to spawn in summer before returning to a resident location during winter (Rowland, 2009). As individuals on the artificial reef were estimated to be ~ 30 to 50 cm long, they may move less than those of a larger size as was found in South Australia by Gillanders et al. (2001).

This lack of movement would help explain the relatively uninfluential effect of time on the abundance of this species. Over 90% of the variance in the spatial and temporal abundance of P. dentex was explained by Site, which is related to its association with reefs (Thresher and Gunn, 1986) and artificial structures (Cermak, 2002). Caragnids appear to utilize the habitat provided by artificial reefs and were recorded in intermediate-high numbers on the Dunsborough and Bunbury artificial reefs (Florisson et al., 2018b). As sexual maturation in P. dentex does not occur until a length of 30 cm (Afonso et al., 2009), the tendency for individuals to reside on the Mandurah artificial reef (i.e. they were present in 16 of the 17 months and in 85% of videos, thus explaining the lack of a temporal shift in abundance), may be due to their relatively small size (L. Ramm, personal observation).

80 3.4.2. Differences among habitat and feeding guilds Significant differences were detected amongst the contribution of species and individuals to habitat guilds with, once again, site, being the main driver. The control site harbored less than half the number of pelagic species and individuals. This is likely due to the absence of the flow diversion effect provided by the design of the cross-braces in the artificial reef modules, which enhances concentrations of plankton. It is thus relevant that the three most abundant species recorded on the artificial reef are pelagic zooplanktivores. This is consistent with a study by Brodeur and Pearcy (1992) who found increased numbers of zooplanktivores in years of strong upwelling. While benthic species contributed to the fauna in equal proportions, the proportion of such individuals was greater on the control site. This was driven by higher counts of N. parilus and E. angustipes, which, due to their camouflage, prefer the more macrophyte dominated control site. The proportion of species and individuals to feeding guilds was also influenced strongly by Site, with the artificial reef having a more even distribution of the five feeding guilds than the control site. Higher proportions of omnivores on the artificial reef are mostly influenced by species such as A. vittiger and E. mosaicus, whilst differences in piscivores are due to the presence of P. speculator and S. hippos, the latter of which was found in 90% of videos on the reef. The occurrence of these abundant species on the artificial reef could be related to the shelter and food that the artificial reef provides. While studies have shown the A. vittiger has a preference for seagrass (Hyndes et al., 2003; Harvey et al., 2013) and macroalgal habitats (Curley et al., 2002), which were more common at the control site, this species has been recorded over reef where adjacent seagrass and macrophyte were not present (Harvey et al., 2012). Thus, it appears that their habitat preference is complex and likely influenced by the surrounding seascape. The increased proportion of piscivores on the artificial reef (e.g. S. hippos) could be due to the relatively large abundances of zooplanktivorous pelagic species such as

T. novaezelandiae, Clupeidae spp. and C. theagenes that would be predated upon.

81 3.4.3. Faunal composition It appears that the presence of an artificial reef over bare substrate in the marine environment can shape the faunal characteristics of such habitat to a large extent.

Following previous trends, differences in the faunal composition was also predominantly driven by site. Consistently high abundance of N. obliquus, C. auratus and P. speculator on the artificial reef differentiated this site for the control which contained P. melbournensis. Brooker et al. (2020) found that N. obliquus were positively correlated with reef cover and invertebrate density. Although not directly measured in the current study, the artificial reef does provide significant cover in comparison to the control site, that this small species could utilise to avoid predators. As previously discussed, the greater abundances of C. auratus on the artificial reef were likely influenced by increased food availability and the provision of structural complexity. Parequula melbournensis is known to prefer sand and seagrass habitats (Platell et al., 1997) which were the dominant substrate types at the control site, with, in contrast, little macrophyte cover present on or near to the artificial reef.

Species that were found exclusively on the artificial reef and in relatively high abundances included R. sarba and C. curioisis. Folpp et al. (2013) also recorded large numbers of R. sarba on two artificial reefs constructed from Reef Balls compared to an adjacent sandy site in three south-east Australian estuaries. This could indicate that vertical relief has an influence on the habitat preference of this sparid. While C. curiosus has not been the subject of much research, Walker (2016) found that it was twice as abundant on the Dunsborough than the Mandurah artificial reef (see also Chapter 5), which could be due to the presence of nearby natural reef in Geographe Bay. Exclusive species found on the control site included the labrids O. lineolata and S. beddomei. The former species is usually associated with rocky habitat (Russell et al., 2010; Peregrin, 2017) and its absence on the artificial reef indicates its food and/or habitat preferences may not be satisfied. This finding is interesting, as the morphologically similar and confamilial C. auricularis was consistently abundant at both sites. As these species are known to partition their food resources (Lek et al., 2011), prey availability may be the driver for the spatial differences in abundance. On the other

82 hand, S. beddomei is closely associated with seagrass habitat (MacArthur and Hyndes, 2001), which is absent at the artificial reef. Although explaining < 4% of the variability in the faunal data, seasonal differences are observed, with, in general, more species found in summer compared to winter. Visibility was far less in winter due to frequent storm events and larger swells, a limitation of using BRUVs to sample fish assemblages (Whitmarsh et al., 2017). It is recognised that various environmental factors are also likely to drive this temporal variability, however, these were not investigated in this study.

3.4.4. Study limitations and future directions

3.4.4.1. Limitations of the BRUV sampling method As with all monitoring methods, there is a degree of bias when using BRUVs to survey fish faunas. Work by Florisson (2015) suggested, albeit with limited data, that the types of fish recorded using BRUVs can vary depending on the direction of the field of view in relation to the artificial reef. In this study, videos were collected in which extent of the artificial reef in shot varied; from filling the entire field of view, when the BRUV was dropped in or directly in front of the reef modules, to none when the BRUVs faced directly away from the modules. This would have had negligible influence on active, epibenthic species like C. auratus and P. dentex, but would influence the ability to record benthic fish that were closely associated with the reefs structure. Although not studied empirically, it was observed that BRUVs dropped closer to the artificial reef recorded a higher species richness, particularly of benthic species (L. Ramm, personal observation). As a visual sampling methodology, the quality of data produced from BRUVs, like underwater visual census, diver operated video and remotely operated video, is heavily influenced by water clarity. There was a large variation in the extent of the visibility over the sampling period, being better during summer and worst in winter and following storms and heavy . Although unable to be documented quantitatively, visibility in May 2020 was particularly poor, barely allowing the bait basket to be seen (~ 50 cm away from the camera). This would prevent the recording of pelagic schooling species and cautious benthic species that do not come within close proximity to the BRUV.

83 Moreover, poor visibility will reduce the ability to correctly identify taxa (Kingsford and Battershill, 2000). BRUVs, through the provision of bait, are also known to be more selective towards carnivorous and omnivorous species (Watson et al., 2010). Despite these limitations, BRUVs are among the most commonly used methods for surveying the fish faunas of artificial reefs (Chapter 2) and considered a viable method for surveying the fish assemblages of marine habitats, where visibility is generally good, and for identifying the presence of recreationally targeted fish species, which often have a strong attraction towards bait.

3.4.4.2. Future directions This study has highlighted some environmental and methodological factors that could potentially be investigated in future studies on artificial reefs to help produce more representative data. This includes rigorous analysis on whether the presence of artificial reef modules in the field of view of BRUV videos influences characteristics of the fauna. Given the issues with poor visibility at times in the current study, there would be value in incorporating a standardized visibility measure, as suggested by Whitmarsh et al. (2017), particularly as most studies fail to consider this factor. More specifically, being able to compare the fish fauna of the Mandurah, Dunsborough and Bunbury artificial reefs, which are comprised of 3 m high modules, to more high-profile natural reefs would also be beneficial, to better understand the effectiveness of these structures. It is recommended that the large and ‘hollow’ Fish Box modules in Mandurah artificial reef are augmented with smaller modules like those used in the Esperance artificial reef, e.g. Apollo, Abitat and Reef Dome modules (Fig. 2.10) and has been undertaken at Bunbury. Due to the position of the Mandurah artificial reef on an expansive sand flat, fish appear to spend large amounts of energy maintaining their position in the water column during surge and swell, which is common during winter (L. Ramm, personal observation). It is also unusual for a natural reef system to contain multiple hollow structures with large spaces of bare sand between, as provided by the artificial modules. Infilling of the artificial reef area with some of the aforementioned

84 modules may provide a more stable layer of water close to the substrate, reducing energy costs during high swell activity, providing additional habitat for existing species and potentially encouraging others to colonize onto structurally different modules.

Augmentation of the reef would also provide opportunity to undertake a similar study to determine the influence of these smaller modules.

85 Chapter 4: General conclusion

This study aimed to, firstly, conduct a systematic literature review on the characteristics of deployed artificial reefs around the world and determine how their attributes varied spatially and temporally. Secondly, to describe the fish fauna present on the Mandurah artificial reef and nearby natural habitat of sand and small rocky outcrops over a 17 month period between February 2019 and July 2020.

4.1. Global trends in artificial reef construction and deployment Data from 1,074 unique artificial reefs and 71 countries was successful in providing insights into the various characteristics of artificial reef design, purpose and monitoring programs. The surge in artificial reef deployments after 1965 is driven by our need to replenish fish stocks and rehabilitate degraded marine habitat, in which can be mitigated by artificial reefs, as well as improve recreational fishing. Majority of artificial reefs resided in waters 10 – 30 m deep, indicating that most reefs are desired to be easily accessible for a range of purposes including the use by recreational and commercial fishers and for scientific studies. The use of concrete as a reef material steadily increased, while those of steel and rubber decreased. This coincides with the transition from materials of opportunity to purpose-built reefs as knowledge progresses and affluent countries with access to funds aim to, for example, improve recreational fishing experiences whilst preserving the natural environment. Future artificial reef research should consider the inclusion of the reef characteristics mentioned in Chapter 2 to benefit the progression of this field of science.

4.2. Fish fauna of the Mandurah artificial reef BRUVs were used to describe the fish and associated faunal assemblages on the Mandurah artificial reef and an adjacent control site between February 2019 and July

2020. The higher number of species and individuals on the artificial reef was determined by differences amongst sites, with temporal differences being relatively uninfluential. Three recreationally targeted fish species C. auratus, S. hippos, and P. dentex were also found to have a higher abundance on the artificial reef than the control site, likely due 86 to a combination of habitat and feeding preferences. The upwards flow diversion effect of the artificial reef modules has been successful, with larger proportions of zooplanktivores and piscivores on the artificial reef compared to the control site.

Significant differences in the fish composition was also due to site variation, driven by high proportionate abundances of N. obliquus, C. auratus and P. speculator on the artificial reef and P. melbournensis on the control site.

4.3. Comparisons of the fish faunas on the Mandurah artificial reef to reefs of the same design in Geographe Bay Given that the Mandurah artificial reef is constructed from 30 3 m3 Fish Box modules arranged into six clusters of five modules and is thus identical those of the Bunbury and Dunsborough artificial reefs, there is an opportunity to compare the fish assemblages from the two reefs in Geographe Bay (Florisson et al., 2018b) to the Mandurah artificial reef and control site (Chapter 3). All data were collected using BRUVs and although the frames were slightly different sizes, the cameras were similar (both

GoPros) and used the same resolution and frame rate and the same type and quantity of bait (cf. Section 3.2 and Florisson et al., 2018b). Note, however, that while all BRUVs were deployed for at least one hour, only data from the first 45 minutes was extracted for the Geographe Bay reefs. However, as the MaxN counts of each species were recorded every five minutes at the Mandurah sites, MaxN counts after 45 minutes were able to be calculated to ensure the data between all four sites were comparable. The aim of this comparison was to undertake some preliminary analyses to statistically compare the characteristics of the fish fauna (i.e. number of species, total MaxN, Simpson’s diversity index and faunal composition), between the three artificial reefs and the control site taking into account the potentially confounding impact of any temporal variability. Note that, to simplify the analyses and their interpretation, data for each of the months was assigned to an austral season. The combined data for each site in each season was subjected to the data pre-treatment, PERMANOVA, AMOSIN, boot- strapped and centroid MDS ordination and shade plot routines described in Chapter 3. The only difference was the design for the analyses were two-way and involved i.e. Site

87 (four levels; Bunbury, Dunsborough artificial reefs and Mandurah control site) and Season (four levels; summer, autumn, winter and spring), with both factors being fixed. The number of species was found to differ significantly between sites and seasons (P = 0.001) but not the interaction between these factors (Table 4.1a). Most of the variability (58%) was explained by differences between sites with greater mean values recorded on the Dunsborough and Mandurah reefs (12) than the Bunbury reef and the control site (8; Fig. 4.1a). A further 33% of the variability was due to seasonal changes with a greater number of species present over the summer months than during winter (12 vs 8; Fig. 4.1b). The lack of a significant Site x Season interactions suggests that the trends among sites were maintained in all seasons. Although Site, Season and their interaction all significantly influenced total MaxN, 73% of this variation was explained by Site (Table 4.1b). Mean total MaxN values at the Mandurah reef (164) were double that recorded at Dunsborough (79) and more than five times greater than Bunbury and the Mandurah control site (30 vs 33, respectively; Fig. 4.1c). Lower MaxN values were recorded in winter than the other seasons (Fig. 4.1d). Values for Simpson’s (diversity) index did not differ with any main effect or their interaction, with values ranging between 0.7 and 0.8 (Fig. 4.1.e,f).

Table 4.1. Mean squares (MS), percentage contribution of mean squares to the total mean squares (%MS), pseudo-F ratios (F) and significance levels (P) from two-way PERMANOVA tests on the (a) number of species, (b) total MaxN, (c) Simpson’s (Diversity) Index and (d) community composition of the fish fauna of three artificial reefs and control site in each season. df = degrees of freedom. Significant results (P = < 0.05) are highlighted in bold and those with a %MS ≥ 30 shaded in grey.

(a) Number of species (b) Total MaxN Factors df MS %MS pF P MS %MS pF P Site 3 128.37 58.24 10.931 0.001 17.39 73.38 28.674 0.001 Season 3 73.44 33.32 6.253 0.001 4.41 18.61 7.273 0.001 Site x Season 9 6.85 3.11 0.584 0.823 1.29 5.45 2.132 0.034 Residual 142 11.74 5.33 0.61 2.56

(c) Simpson's Index (d) Community composition Factors df MS %MS pF P MS %MS pF P Site 3 0.03 28.70 1.465 0.242 53745 82.04 28.041 0.001 Season 3 0.04 35.78 1.826 0.157 5874 8.97 3.065 0.001 Site x Season 9 0.02 15.92 0.813 0.601 3978 6.07 2.075 0.001 Residual 142 0.02 19.60 1917 2.93

88

(a) (b) 14 14

12 12

s

s

e

i

e i c 10

c 10

e

e

p

p s

8

s f

8

f

o

o

r 6 r

e 6

e

b b

m 4 m

u 4

u N

2 N 2 0

0

n n n n

u u a o

u p

A

W

B

D C

M S S Site Season (c) (d) 250 250

200 200

N

N x

150 x 150

a

a

M

M

l

l

a

a t

100 t 100

o

o

T T 50 50

0 0

n n n n

u p

A

W

u u a o

S S

B

D C M Season Site (e) (f) 0.85 0.85

0.80

x 0.80

x

e

e

d

d

n

I

n

0.75 I

0.75

s

'

s

'

n

n

o o

s 0.70 s

p 0.70

p

m

i

m

i S

0.65 S 0.65

0.60

0.60

n n n n

u u a o

u p

A

W

B

D C

M S S Site Season Fig. 4.1. Mean number of species, total MaxN and Simpson’s (diversity) index among sites and seasons. Error bars represent ± 95% confidence limits. Note the interaction plot for total MaxN was excluded as it only contributed 5% to the total mean squares and, although not significantly different, the graphs for Simpson’s index are included to provide an indication of the range of values and their associated variability. Bun = Bunbury; Dun = Dunsbrough; Man = Mandurah; Con = control site.

Faunal composition was shown by PERMANOVA to differ between sites and seasons and a significant interaction between the two main effects was also detected (Table 4.1d). Site explained the majority of the variability (82%), whereas season and the interaction only accounted for 9 and 6%, respectively. This is consistent with the global

R statistic values one-way ANOSIM tests, where the value for site (0.756) was much larger than that for season (0.068). Such results indicate that differences in composition were predominantly influenced by site rather than season. Two dimensional nMDS plots derived from boot-strapping support this suggestion, with the points representing the

89 four different sites each tightly clustered into a discrete group with no overlap (Fig. 4.2a), whereas those for seasons are more dispersed and some overlap between autumn and both summer and spring occurs (Fig. 4.2b). The reason for the relatively minor Site x Season is shown on the centroid nMDS plot, where the arrangement of seasons within each site is not consistent (Fig. 4.2c)

Among sites, each was found by one-way ANOSIM to be significant to each other (P = 0.001), with among the largest differences being between the three artificial reefs and the control site (R = 0.680 - 0.826). This was due to a greater range of species being on the reefs and species such as Neatypus obliquus and P. dentex being recorded in large numbers on the complex structures (Fig. 4.3). Among the artificial reefs Mandurah was the most distinct (R = 0.771 and 0.789), due to larger numbers of the recreationally important C. auratus and Platycephalus speculator and lower abundances of Anoplocapros amygdaloides (Fig. 4.3). The reefs with the smallest difference in faunal composition were Bunbury and Dunsborough (R = 0.221), which contained, on average, a similar suite of species, only in greater numbers on the latter reef. The negligible differences in composition were mainly due to winter being different from both summer and autumn (P = 0.002; R = 0.251 and 0.130, respectively) and usually reflected a reduction in the abundance of common species such as Coris auricularis. The preliminary univariate and multivariate analyses conducted here demonstrate the major influence of Site on the fish fauna. Seeing as all three artificial reefs are constructed using the same components and design, the results highlight influence that site selection has on the species that are able to inhabit the artificial reefs. If the goal of future reef deployments is to improve the abundances of key recreational species, such as C. auratus, then understanding the drivers behind the above trends would be valuable and help relate the biological performance on the reef to oceanographic processes, connectivity to adjacent habitats and ultimately guide site selection. Without investigation, latitudinal differences are a possible factor driving this difference. This would suggest that the closer similarity between Dunsborough and Bunbury reefs is potentially explained by the distance between them (~ 50 km), compared to that of Dunsborough and Mandurah (~ 125 km). 90

(a) 40 2D Stress: 0.1 Site Bunbury Dunsborough Mandurah Control 20 av:Bunbury av:Dunsborough av:Mandurah

av:Control

2 S

D 0 M

-20

-40 -20 0 20 40 60 MDS1 (b)

20 2D Stress: 0.17 Season Su A W Sp av:A av:Su av:W av:Sp

2 0

S

D M

-20

-20 0 20 MDS1 (c)

W 2D Stress: 0.05 Site Su Bunbury ASp Dunsborough Sp A W Su Mandurah Control Site Bunbury Dunsborough Mandurah Control

Su

A Sp W W Su A

Sp

Fig. 4.2. mMDS ordination plots constructed from bootstrap averages for (a) Site and (b) Season calculated from a Bray-Curtis resemblance matrix of the transformed MaxN values of each species in each replicate sample. Group averages (black symbols) and approximate 95% region estimates fitted to the bootstrap averages (shaded areas) are provided. (c) Centroid nMDS ordination plots constructed from a distance among centroids matrix of the above Bray- Curtis resemblance matrix. Arrows on (c) denote the order of the seasons at each site.

91

u u p p u u p

5 p Site

W S A S S W A S S A S W S A S W Dunsborough 4 Sphyraena novaehollandiae Bunbury Gymnothorax woodwardi Sepia apama Mandurah Pentapodus vitta 3 Ophthalmolepis lineolata Control Sillago ciliata Scobinichthys granulatus Rec 2 Eupetrichthys angustipes Eubalichthys mosaicus N Arripis georgianus Y 1 Caesioscorpis theagenes Acanthaluteres vittiger Pempheris klunzingeri 0 Chrysophrys auratus Platycephalus speculator Upeneichthys vlamingii Notolabrus parilus Coris auricularis Pseudocaranx dentex Neatypus obliquus Austrolabrus maculatus Parequula melbournensis Chelmonops curiosus Myliobatis australis Anoplocapros amygdaloides Seriola hippos Dasyatis brevicaudata Anoplocapros lenticularis Trygonorrhina fasciata Aracana aurita Monocanthidae spp. Goniistius gibbosus Parapercis haackei Pentaceropsis recurvirostris Parapercis ramsayi Diodon nicthemerus Aptychotrema vincentiana Mustelus antarcticus Trygonoptera ovalis Chaetodon assarius Fig. 4.3. Shade plot of the transformed MaxN values of each of the 40 most important species recorded at each of the four sites in each season. Shading intensity is proportional to abundance. Rec = recreationally targets species; N = No and Y = Yes.

4.4. Future directions The systematic literature review (Chapter 2) has highlighted that a standardized approach should be taken when describing the characteristics of an artificial reef and any associated monitoring program. It is hoped proponents of artificial reefs will follow the guidelines provided, as this will aid comparisons among reefs and could potentially lead to a database that describe artificial reefs around the world. Future studies using BRUVs should be undertaken, particularly if programs like Reef Vision are to continue with the future deployment of artificial reefs in Western Australia. These should investigate the effect of field of view and visibility on the number and types of fish recorded on artificial reefs in south-western Australia and thus identify and potentially enable any bias to be accounted for. Furthermore, the augmentation of the Mandurah artificial reef with smaller, low-profile reefs, as occurred in on the similarly designed Bunbury artificial reef, would most likely have a positive effect on the number of species and individuals and thus be beneficial. If this was to occur, this would allow a second study to be conducted on the artificial reef and opportunity to compare those fish fauna data with that obtained during this study.

92 References

Abelson, A. & Shlesinger, Y. 2002. Comparison of the development of coral and fish communities on rock-aggregated artificial reefs in Eilat, Red Sea. ICES Journal of Marine Science, 59, S122-S126. Afonso, P., Fontes, J., Holland, K. N. & Santos, R. S. 2009. Multi-scale patterns of habitat use in a highly mobile reef fish, the white trevally Pseudocaranx dentex, and their implications for marine reserve design. Marine Ecology Progress Series, 381, 273-286. Ajemian, M. J., Wetz, J. J., Shipley-Lozano, B., Shively, J. D. & 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, e0126354. Alabama Department Of Conservation And Natural Resources. 2017. Standard operating protocol for offshore reef construction [Online]. Montgomery, Alabama. Available: https://www.outdooralabama.com/artificial-reefs/construction-protocol [Accessed 15/05/2020]. Allgeier, J. E., Yeager, L. A. & Layman, C. A. 2013. Consumers regulate nutrient limitation regimes and primary production in seagrass ecosystems. Ecology, 94, 521-529. Anderson, M., Gorley, R. N., & Clarke, K. R. 2008. PERMANOVA + for PRIMER. Guide to software and statistical methods, Plymouth, UK, PRIMER-E. Arlinghaus, R., Abbott, J. K., Fenichel, E. P., Carpenter, S. R., Hunt, L. M., Alós, J., Klefoth, T., Cooke, S. J., Hilborn, R. & Jensen, O. P. 2019. Opinion: Governing the recreational dimension of global fisheries. Proceedings of the National Academy of Sciences, 116, 5209-5213. Arlinghaus, R., Tillner, R. & Bork, M. 2015. Explaining participation rates in recreational fishing across industrialised countries. Fisheries Management and Ecology, 22, 45-55. Bailey-Brock, J. H. 1989. Fouling community development on an artificial reef in Hawaiian waters. Bulletin of Marine Science, 44, 580-591. Baine, M. 2001. Artificial reefs: A review of their design, application, management and performance. Ocean and Coastal Management, 44, 241-259. Bateman, T. 2015. Artificial Reefs: Types, applications, trends in deployment and the development of a cost-effective method for monitoring their fish faunas. Honours thesis, Murdoch University. Beck, M. W., Brumbaugh, R. D., Airoldi, L., Carranza, A., Coen, L. D., Crawford, C., Defeo, O., Edgar, G. J., Hancock, B., Kay, M. C., Lenihan, H. S., Luckenbach, M. W., Toropova, C. L., Zhang, G. & Guo, X. 2011. Oyster reefs at risk and recommendations for conservation, restoration, and managament. BioScience, 61, 107-116. Beck, M. W., Heck, K. L., Able, K. W., Childers, D. L., Eggleston, D. B., Gillanders, B. M., Halpern, B., Hays, C. G., Hoshino, K., Minello, T. J., Orth, R. J., Sheridan, P. F. & Weinstein, M. P. 2001. The identification, conservation, and management of estuarine and marine nurseries for fish and invertebrates: A better understanding of the habitats that serve as nurseries for marine species and the factors that create site-specific variability in nursery quality will improve conservation and management of these areas. BioScience, 51, 633- 641. Becker, A., Smith, J. A., Taylor, M. D., Mcleod, J. & Lowry, M. B. 2019. Distribution of pelagic and epi-benthic fish around a multi-module artificial reef-field: Close module spacing supports a connected assemblage. Fisheries Research, 209, 75-85. Becker, A., Taylor, M. D., Folpp, H. & Lowry, M. B. 2018. Managing the development of artificial reef systems: The need for quantitative goals. Fish and Fisheries, 19, 740-752. Becker, A., Taylor, M. D. & Lowry, M. B. 2017. Monitoring of reef associated and pelagic fish communities on Australia’s first purpose built offshore artificial reef. ICES Journal of Marine Science, 74, 277-285. Beets, J. & Hixon, M. A. 1994. Distribution, persistence, and growth of groupers (Pisces: Serranidae) on artificial and natural patch reefs in the Virgin Islands. Bulletin of Marine Science, 55, 470-483.

93 Bell, J. D., Leber, K. M., Blankenship, H. L., Loneragan, N. R. & Masuda, R. 2008. A new era for restocking, stock enhancement and sea ranching of coastal fisheries resources. Reviews in Fisheries Science, 16, 1-9. Blakeway, D., Byers, M., Stoddart, J. & Rossendell, J. 2013. Coral colonisation of an artificial reef in a turbid nearshore environment, Dampier Harbour, Western Australia. PloS One, 8, e75281. Bohnsack, J. A. & Sutherland, D. L. 1985. Artificial reef research: A review with recommendations for future priorities. Bulletin of Marine Science, 37, 11-39. Bohnsack J. A. 1989. Are high densities of fishes at artificial reefs the result of habitat limitation or behavioral preferences? Bulletin of Marine Science, 44, 631-645. Bombace, G. 1989. Artificial reefs in the Mediterranean Sea. Bulletin of Marine Science, 44, 1023- 1032. Bornmann, L. & Mutz, R. 2015. Growth rates of modern science: A bibliometric analysis based on the number of publications and cited references. Journal of the Association for Information Science and Technology, 66, 2215-2222. Bortone, S. A. 2015. CARAH (International Conference on Artificial Reefs and Related Aquatic Habitats): An historical perspective of accomplishments. Journal of Applied Ichthyology, 31, 3-14. Bortone, S. A., Turpin, R. K., Cody, R. C., Bundrick, C. M. & Hill, R. L. 1997. Factors associated with artificial-reef fish assemblages. Gulf of Mexico Science, 15, 17-34. Branden, K. L., Pollard, D. A. & Reimers, H. A. 1994. A review of recent artificial reef developments in Australia. Bulletin of Marine Science, 55, 982-994. Brandini, F. 2014. Marine biodiversity and sustainability of fishing resources in Brazil: A case study of the coast of Paraná state. Regional Environmental Change, 14, 2127-2137. Brickhill, M., Lee, S. & Connolly, R. 2005. Fishes associated with artificial reefs: Attributing changes to attraction or production using novel approaches. Journal of Fish Biology, 67, 53-71. Broadley, A. D., Tweedley, J. R. & Loneragan, N. R. 2017. Estimating biological parameters for penaeid restocking in an Australian temperate estuary. Fisheries Research, 186, 488-501. Brodeur, R. D. & Pearcy, W. G. 1992. Effects of environmental variability on trophic interactions and food web structure in a pelagic upwelling ecosystem. Marine Ecology Progress Series, 84, 101-119. Brooker, M. A., De Lestang, S., Fairclough, D. V., Mclean, D., Slawinski, D., Pember, M. B. & Langlois, T. J. 2020. Environmental and anthropogenic factors affect fish abundance: Relationships revealed by automated cameras deployed by fishers. Frontiers in Marine Science, 7, 1-12. Bull, A. S. & Love, M. S. 2019. Worldwide oil and gas platform decommissioning: A review of practices and reefing options. Ocean & Coastal Management, 168, 274-306. Burt, J. A., Feary, D. A., Cavalcante, G., Bauman, A. G. & Usseglio, P. 2013. Urban breakwaters as reef fish habitat in the Persian Gulf. Marine Pollution Bulletin, 72, 342-350. Calver, M. C., Goldman, B., Hutchings, P. A. & Kingsford, R. T. 2017. Why discrepancies in searching the conservation biology literature matter. Biological Conservation, 213, 19-26. Campbell, D. & Murphy, J. 2005. The 2000-01 national recreational fishing survey: Economic report. Canberra, Australia. Cappo, M., Walters, C. J. & Lenanton, R. C. 2000. Estimation of rates of migration, exploitation and survival using tag recovery data for western Australian “salmon” (Arripis truttaceus: Arripidae: Percoidei). Fisheries Research, 44, 207-217. Carr, M. H. & Hixon, M. A. 1997. Artificial reefs: The importance of comparisons with natural reefs. Fisheries, 22, 28-33. Cermak, M. J. 2002. Caranx latus (Carangidae) chooses dock pilings to attack silverside schools: A tactic to interfere with stereotyped escape behavior of prey? The Biological Bulletin, 203, 241-243.

94 Cisneros-Montemayor, A. M. & Sumaila, U. R. 2010. A global estimate of benefits from ecosystem-based marine recreation: Potential impacts and implications for management. Journal of Bioeconomics, 12, 245-268. Clarke, D. G., Fredette, T. J. & Imsand, D. 1988. Creation of offshore topographic features with dredged material. Information Exchange Bulletin, D-88-5, 1-5. Claisse, J. T., Pondella, D. J., Love, M., Zahn, L. A., Williams, C. M., Williams, J. P. & 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, 15462-15467. Clarke, K. R. (1993). Non-parametric multivariate analyses of changes in community structure. Australian Journal of Ecology, 18, 117–143. Clarke, K. R., Chapman, M., Somerfield, P. & Needham, H. 2006. Dispersion-based weighting of species counts in assemblage analyses. Marine Ecology Progress Series, 320, 11-27. Clarke, K. R., Gorley, R. N., Somerfield, P. J., & Warwick, R. M. (2014a). Change in marine communities: an approach to statistical analysis and interpretation. Primer-E Ltd. Clarke, K. R, & Gorley, R. N. 2015. Primer: User manual/tutorial. Primer-E Ltd., Plymouth, 93. Clarke, K. R., Tweedley, J. R. & Valesini, F. J. 2014b. Simple shade plots aid better long-term choices of data pre-treatment in multivariate assemblage studies. Journal of the Marine Biological Association of the United Kingdom, 94, 1-16. Clark, S. & Edwards, A. 1999. An evaluation of artificial reef structures as tools for marine habitat rehabilitation in the Maldives. Aquatic Conservation: Marine and Freshwater Ecosystems, 9, 5-21. Coll, J., Abad, R., Álvarez, E., Deudero, S., Mas, R., Riera, F. & Moreno, I. 2009. State of fish populations and influence on the trammel net fishery at three Balearic Island (western Mediterranean) artificial reefs a decade after their deployment. Bulletin of Marine Science, 85, 77-100. Collins, K., Jensen, A. & Albert, S. 1995. A review of waste tyre utilisation in the marine environment. Chemistry and Ecology, 10, 205-216. Collins, K., Jensen, A., Mallinson, J., Roenelle, V. & Smith, I. 2002. Environmental impact assessment of a scrap tyre artificial reef. ICES Journal of Marine Science, 59, S243-S249. Cooke, S. J. & Cowx, I. G. 2004. The role of recreational fishing in global fish crises. BioScience, 54, 857-859. Cresson, P., Ruitton, S., Ourgaud, M. & Harmelin-Vivien, M. 2014. Contrasting perception of fish trophic level from stomach content and stable isotope analyses: A Mediterranean artificial reef experience. Journal of Experimental Marine Biology and Ecology, 452, 54-62. Crisp, J., Loneragan, N., Tweedley, J., D’souza, F. & Poh, B. 2018. Environmental factors influencing the reproduction of an estuarine Penaeid population and implications for management. Fisheries Management and Ecology, 25, 203-219. Curley, B. G., Kingsford, M. J. & Gillanders, B. M. 2002. Spatial and habitat-related patterns of temperate reef fish assemblages: Implications for the design of marine protected areas. Marine and Freshwater Research, 53, 1197-1210. Dauterive, L. Rigs-To-Reefs policy, progress, and perspective. In: Halloc, P. & French, L., eds. Proceedings of the AAUS 20th Symposium Proceedings 2000. Diving for Science in the Twenty-First Century, 2000 Nahant, Massachusetts. American Academy of Underwater Sciences, 64-66. d'Cruz, T., Creech, S. & Fernandez, J. 1994. Comparison of catch rates and species composition from artificial and natural reefs in Kerala, India. Bulletin of Marine Science, 55, 1029-1037. Deegan, L. A., Johnson, D. S., Warren, R. S., Peterson, B. J., Fleeger, J. W., Fagherazzi, S. & Wollheim, W. M. 2012. Coastal eutrophication as a driver of salt marsh loss. Nature, 490, 388-392. Delos Reyes, M. & Martens, R. 1996. Low-cost artificial reef program in the Philippines: An evaluation in the management of a tropical coastal ecosystem. Oceanographic Literature Review, 2, 1-6.

95 Department Of Agriculture, Fisheries and Forestry, Government of Australia. 2003. National recreational and indigenous fishing survey. In: Henry, G. W. & Lyle, J. M. (eds.). Canberra, Australia. Department Of Agriculture, Water And The Environment, Government Of Australia. 2020b. Sea dumping [Online]. Australian Government. Available: https://www.environment.gov.au/marine/marine-pollution/sea-dumping [Accessed 08/10/2020]. Department Of Agriculture, Water and the Environment, Government of Australia. 2020c. Recreational fishing [Online]. Australian Government. Available: https://www.agriculture.gov.au/abares/research-topics/fisheries/fisheries-and- aquaculture-statistics/recreational-fishing#australian-capital-territory [Accessed 15/05/2020]. Department Of Agriculture, Water and the Environment, Government of Australia. 2020a. Permit application forms and fees [Online]. Australian Government Available: http://www.environment.gov.au/topics/marine/marine-pollution/sea-dumping/forms- and-application-fees [Accessed 15/05/2020]. Department Of Primary Industries And Regional Development, Government of Australia. 2018. Annual Report. Perth, Australia. Department Of Primary Industries And Regional Development, Government of Australia. 2019. Tender released to design and build northern metro artificial reef [Online]. Available: http://www.fish.wa.gov.au/About-Us/News/Pages/_archive/Tender-released-to-design- and-build-northern-metro-artificial-reef.aspx [Accessed 14/05/2020]. Dickson, L. & Mcculloch, A. 1996. Shell, the Brent Spar and Greenpeace: A doomed tryst? Environmental Politics, 1, 122-129. Dowling, R. K. & Nichol, J. 2001. The HMAS Swan artificial diving reef. Annals of Tourism Research, 28, 226-229. Dupont, J. M. 2008. Artificial reefs as restoration tools: A case study on the west Florida shelf. Coastal Management, 36, 495-507. Elliott, M., Whitfield, A. K., Potter, I. C., Blaber, S. J., Cyrus, D. P., Nordlie, F. G. & Harrison, T. D. 2007. The guild approach to categorizing estuarine fish assemblages: A global review. Fish and Fisheries, 8, 241-268. Evans, P. & Ranasinghe, R. Artificial surfing reefs: A new paradigm in coastal protection? Coasts & Ports 2001: Proceedings of the 15th Australasian Coastal and Ocean Engineering Conference, the 8th Australasian Port and Harbour Conference, 2001 Barton, Australia. Institution of Engineers, Australia, 128-133. Fabi, G. & 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, 538-558. Fabi, G., Spagnolo, A., Bellan-Santini, D., Charbonnel, E., Çiçek, B. A., García, J. J. G., Jensen, A. C., Kallianiotis, A. & Santos, M. N. D. 2011. Overview on artificial reefs in Europe. Brazilian Journal of Oceanography, 59, 155-166. FAO 2012. Investing in agriculture for a better future. Rome, Italy: FAO. FAO 2018. The State of World Fisheries and Aquaculture 2018 - Meeting the sustainable development goals. Rome, Italy. Feigenbaum, D., Bushing, M., Woodward, J. & Friedlander, A. 1989. Artificial reefs in Chesapeake Bay and nearby coastal waters. Bulletin of Marine Science, 44, 734-742. Ferse, S. Multivariate responses of the coral reef fish community to artificial structures and coral transplants. Proceedings of the 11th International Coral Reef Symposium. Ft. Lauderdale, Florida, 2008. 1225-1229. Fitzhardinge, R. & Bailey-Brock, J. 1989. Colonization of artificial reef materials by corals and other sessile organisms. Bulletin of Marine Science, 44, 567-579. Florisson, J. H. 2015. Can recreational fishers provide an effective means of monitoring artificial reefs? Honours thesis, Murdoch University.

96 Florisson, J. H., Rowland, A. J., Harvey, E. S., Allen, M. B., Watts, S. L. & Saunders, B. J. 2020. King Reef: An Australian first in repurposing oil and gas infrastructure to benefit regional communities. The APPEA Journal, 60, 435-439. Florisson, J. H., Rowland, A. J., Matthews, A. C., Tweedley, J. R. & Campbell, L. L. 2018a. The application, needs, costs and benefits of habitat enhancement structures in Western Australia and cost effective monitoring methods. Report to the Fisheries Research and Development Corporation Perth, Australia: Recfishwest. Florisson, J. H., Tweedley, J. R., Walker, T. H. & Chaplin, J. A. 2018b. Reef vision: A citizen science program for monitoring the fish faunas of artificial reefs. Fisheries Research, 206, 296-308. Folpp, H., Lowry, M., Gregson, M. & Suthers, I. M. 2013. Fish assemblages on estuarine artificial reefs: Natural rocky-reef mimics or discrete assemblages? PLoS One, 8, e63505. Freire, K. M. F., Belhabib, D., Espedido, J. C., Hood, L., Kleisner, K. M., Lam, V. W., Machado, M. L., Mendonça, J. T., Meeuwig, J. J. & Moro, P. S. 2020. Estimating global catches of marine recreational fisheries. Frontiers in Marine Science, 7, 1-18. Giansante, C., Fatigati, M., Ciarrocchi, F., Milillo, G. S., Onori, L. & Ferri, N. 2010. Monitoring of ichthyic fauna in artificial reefs along the Adriatic coast of the Abruzzi region of Italy. Veterinaria Italiana, 46, 365-374. Gillanders, B. M., Ferrell, D. J. & Andrew, N. L. 2001. Estimates of movement and life-history parameters of yellowtail kingfish (Seriola lalandi): How useful are data from a cooperative tagging programme? Marine and Freshwater Research, 52, 179-192. Gratwicke, B. & Speight, M. R. 2005. Effects of habitat complexity on Caribbean marine fish assemblages. Marine Ecology Progress Series, 292, 301-310. Greenwell, C. N., Loneragan, N. R., Admiraal, R., Tweedley, J. R. & Wall, M. 2019a. Octopus as predators of abalone on a sea ranch. Fisheries Management and Ecology, 26, 108-118. Greenwell, C. N., Loneragan, N. R., Tweedley, J. R. & Wall, M. 2019b. Diet and trophic role of octopus on an abalone sea ranch. Fisheries Management and Ecology, 26, 638-649. Guidetti, P. 2000. Differences among fish assemblages associated with nearshore Posidonia oceanica seagrass beds, rocky-algal reefs and unvegetated sand habitats in the Adriatic Sea. Estuarine, Coastal and Shelf Science, 50, 515-529. Hallett, C. S., Hobday, A. J., Tweedley, J. R., Thompson, P. A., Mcmahon, K. & Valesini, F. J. 2018. Observed and predicted impacts of climate change on the estuaries of south-western Australia, a Mediterranean climate region. Regional Environmental Change, 18, 1357- 1373. Halpern, B. S., Frazier, M., Potapenko, J., Casey, K. S., Koenig, K., Longo, C., Lowndes, J. S., Rockwood, R. C., Selig, E. R. & Selkoe, K. A. 2015. Spatial and temporal changes in cumulative human impacts on the world’s ocean. Nature Communications, 6, 1-7. Halpern, B. S., Walbridge, S., Selkoe, K. A., Kappel, C. V., Micheli, F., D'agrosa, C., Bruno, J. F., Casey, K. S., Ebert, C. & Fox, H. E. 2008. A global map of human impact on marine ecosystems. Science, 319, 948-952. Harmelin, J. G. 2000. Mediterranean marine protected areas: Some prominent traits and promising trends. Environmental Conservation, 27, 104-105. Harris, H. E., Patterson Iii, W. F., Ahrens, R. N. & Allen, M. S. 2019. Detection and removal efficiency of invasive lionfish in the northern Gulf of Mexico. Fisheries Research, 213, 22- 32. Harris, P. T. 2012. Seafloor geomorphology - Coast, shelf, and abyss. In: Harris, P. T. & Baker, E. K. (eds.) Seafloor Geomorphology as Benthic Habitat. London, United Kingdom: Elsevier. Harvey, E. S., Butler, J., Mclean, D. & Shand, J. 2012. Contrasting habitat use of diurnal and nocturnal fish assemblages in temperate Western Australia. Journal of Experimental Marine Biology and Ecology, 426, 78-86. Harvey, E. S., Cappo, M., Kendrick, G. A. & Mclean, D. L. 2013. Coastal fish assemblages reflect geological and oceanographic gradients within an Australian zootone. PLoS One, 8, e80955.

97 Hastings, R., Ogren, L. & Mabry, M. 1976. Observations on the fish fauna associated with offshore platforms in the northeastern Gulf of Mexico. Fishery Bulletin. United States Fish and Wildlife Service, 74. Hegde, A. V. 2010. Coastal erosion and mitigation methods - Global state of art. Indian Journal of Geo-Marine Sciences, 39, 521-530. Hill, J. & Wilkinson, C. 2004. Methods for ecological monitoring of coral reefs - A resource for managers. In: SCIENCE, A. I. O. M. (ed.) Australian Institute of Marine Science, Townsville. Townsville, Australia: Australian Institute of Marine Science. Hughes, R. M. 2015. Recreational fisheries in the USA: Economics, management strategies, and ecological threats. Fisheries Science, 81, 1-9. Huret, M., Petitgas, P. & Woillez, M. 2010. Dispersal kernels and their drivers captured with a hydrodynamic model and spatial indices: A case study on anchovy (Engraulis encrasicolus) early life stages in the Bay of Biscay. Progress in Oceanography, 87, 6-17. Hutson, K., Smith, B., Godfrey, R., Whittington, I., Chambers, C., Ernst, I. & Gillanders, B. 2007. A tagging study on yellowtail kingfish (Seriola lalandi) and Samson fish (S. hippos) in south Australian waters. Transactions of the Royal Society of South Australia, 131, 128-134. Hyndes, G., Platell, M., Potter, I. & Lenanton, R. 1999. Does the composition of the demersal fish assemblages in temperate coastal waters change with depth and undergo consistent seasonal changes? Marine Biology, 134, 335-352. Hyndes, G. A., Kendrick, A., Macarthur, L. D. & Stewart, E. 2003. Differences in the species-and size-composition of fish assemblages in three distinct seagrass habitats with differing plant and meadow structure. Marine Biology, 142, 1195-1206. Iannibelli, M. & Musmarra, D. 2008. Effects of anti‐trawling artificial reefs on fish assemblages: The case of Salerno Bay (Mediterranean Sea). Italian Journal of Zoology, 75, 385-394. Ito, Y. 2011. Artificial reef function in fishing grounds off Japan. In: Bortone, S. A., Brandini, F. P., Fabi, G. & Otake, S. (eds.) Artificial Reefs in Fisheries Management. Boca Raton, Florida: CRC Press. Jan, R.-Q., Liu, Y.-H., Chen, C.-Y., Wang, M.-C., Song, G.-S., Lin, H.-C. & Shao, K.-T. 2003. Effects of pile size of artificial reefs on the standing stocks of fishes. Fisheries Research, 63, 327- 337. Jensen, A. 2002. Artificial reefs of Europe: Perspective and future. ICES Journal of Marine Science, 59, S3-S13. Jørgensen, D. 2012. OSPAR’s exclusion of rigs-to-reefs in the North Sea. Ocean & Coastal Management, 58, 57-61. Kaiser, M. J. 2006. The Louisiana artificial reef program. Marine Policy, 30, 605-623. Kaiser, M. J. & Pulsipher, A. G. 2005. Rigs-to-reef programs in the Gulf of Mexico. Ocean Development & International Law, 36, 119-134. Keller, K., Smith, J. A., Lowry, M. B., Taylor, M. D. & Suthers, I. M. 2017. Multispecies presence and connectivity around a designed artificial reef. Marine and Freshwater Research, 68, 1489-1500. Kellison, T. G. & Sedberry, G. R. 1998. The effects of artificial reef vertical profile and hole diameter on fishes off South Carolina. Bulletin of Marine Science, 62, 763-780. Kerr, S. 1992. Artificial reefs in Australia. Their construction, location and function. Bureau of Rural Resources. Kim, C. G., Lee, J. W. & Park, J. S. 1994. Artificial reef designs for Korean coastal waters. Bulletin of Marine Science, 55, 858-866. Kingett, P. & Choat, J. 1981. Analysis of density and distribution patterns in Chrysophrys auratus (Pisces: Sparidae) within a reef environment: An experimental approach. Marine Ecology Progress Series, 5, 283-290. Kingsford, M. & Battershill, C. 2000. Studying temperate marine environments: A handbook for ecologists, Boca Raton, Florida, CRC Press. Kitada, S. 2018. Economic, ecological and genetic impacts of marine stock enhancement and sea ranching: A systematic review. Fish and Fisheries, 19, 511-532.

98 Langhamer, O. 2012. Artificial reef effect in relation to offshore renewable energy conversion: State of the art. The Scientific World Journal, 2012, 1-8. Langlois, T., Williams, J., Monk, J., Bouchet, P., Currey, L., Goetze, J., Harasti, D., Huveneers, C., Malcolm, H. & Whitmarsh, S. 2018. Marine sampling field manual for benthic stereo BRUVS (Baited Remote Underwater Videos). In: Przelsawski, R. & Foster, S. (eds.) Field Manuals for Marine Sampling to Monitor Australian Waters, Version 1. Canberra, Australia. Layman C. A., Allgeier J. E. 2020. An ecosystem ecology perspective on artificial reef production. Journal of Applied Ecology. 57, 2139-2148. Lechanteur, Y. & Griffiths, C. 2001. Fish associated with artificial reefs in , South Africa: A preliminary survey. African Zoology, 36, 87-93. Lee, M. O., Otake, S. & Kim, J. K. 2018. Transition of artificial reefs (ARs) research and its prospects. Ocean & Coastal Management, 154, 55-65. Leeworthy, V. R., Maher, T. & Stone, E. A. 2006. Can artificial reefs alter user pressure on adjacent natural reefs? Bulletin of Marine Science, 78, 29-37. Leitao, F., Santos, M., Erzini, K. & Monteiro, C. 2009. Diplodus spp. assemblages on artificial reefs: Importance for near shore fisheries. Fisheries Management and Ecology, 16, 88-99. Lek, E., Fairclough, D., Platell, M., Clarke, K., Tweedley, J. & Potter, I. 2011. To what extent are the dietary compositions of three abundant, co‐occurring labrid species different and related to latitude, habitat, body size and season? Journal of Fish Biology, 78, 1913-1943. Lewis, P. 2015. Further investigation into critical habitat for juvenile dhufish (Glaucosoma hebraicum), artificial habitats and the potential to monitor annual juvenile recruitment. North Beach, Australia. Lima, J. S., Zalmon, I. R. & Love, M. 2019. Overview and trends of ecological and socioeconomic research on artificial reefs. Marine Environmental Research, 145, 81-96. London Convention And Protocol/UNEP 2009. London convention and protocol/UNEP guidelines for the placement of artificial reefs. London, United Kingdom. Macarthur, L. & Hyndes, G. 2001. Differential use of seagrass assemblages by a suite of odacid species. Estuarine, Coastal and Shelf Science, 52, 79-90. Macleod, I., Morrison, P., Richards, V. & West, N. 2004. Corrosion monitoring and the environmental impact of decommissioned naval vessels as artificial reefs. 70, 592-297. Mcleod, P. & Lindner, R. 2018. Economic dimension of recreational fishing in Western Australia. Perth, Australia: Recfishwest, Perth, Western Australia. Milicich, M. 1994. Dynamic coupling of reef fish replenishment and oceanographic processes. Marine Ecology Progress Series, 110, 135-144. Mills, K. A., Hamer, P. A. & Quinn, G. P. 2017. Artificial reefs create distinct fish assemblages. Marine Ecology Progress Series, 585, 155-173. Moher, D., Liberati, A., Tetzlaff, J. & Altman, D. G. 2010. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. Int J Surg, 8, 336-341. Molles, M. C. 1978. Fish species diversity on model and natural reef patches: Experimental insular biogeography. Ecological Monographs, 48, 289-305. Molony, B., Newman, S., Joll, L., Lenanton, R. & Wise, B. 2011. Are Western Australian waters the least productive waters for finfish across two oceans? A review with a focus on finfish resources in the Kimberley region and North Coast Bioregion. Journal of the Royal Society of Western Australia, 94, 323. Morley, D. M., Sherman, R. L., Jordan, L. K., Banks, K., Quinn, T. P. & Spieler, R. E. 2008. Environmental enhancement gone awry: characterization of an artificial reef constructed from waste vehicle tires. Environmental Problems in Coastal Regions. Southampton, United Kingdom: WIT Press. Muñoz-Pérez, J. J. 2008. Artificial reefs to improve and protect fishing grounds. Recent Patents on Engineering, 2, 80-86. Murphy, H. M. & Jenkins, G. P. 2010. Observational methods used in marine spatial monitoring of fishes and associated habitats: A review. Marine and Freshwater Research, 61, 236-252.

99 Myers, R. A. & Worm, B. 2003. Rapid worldwide depletion of predatory fish communities. Nature, 423, 280-283. Obregón, C., Hughes, M., Loneragan, N. R., Poulton, S. J. & Tweedley, J. R. 2020. A two-phase approach to elicit and measure beliefs on management strategies: Fishers supportive and aware of trade-offs associated with stock enhancement. Ambio, 49, 640-649. Okano, T., Takeda, M., Nakagawa, Y., Hirata, K., Mitsuhashi, K., Kawaguchi, S. & Ito, J. 2011. Artificial reefs to induce upwelling to increase fishery resources. Artificial Reefs in Fisheries Management. Boca Raton, Florida: CRC Press. Okoli, C., Schabram, K. 2010. A guide to conducting a systematic literature review of information systems research. Sprouts: Working Papers on Information Systems. 10 (26). http://sprouts.aisnet.org/10-26. Osenberg, C. W., St. Mary, C. M., Wilson, J. A. & Lindberg, W. J. 2002. A quantitative framework to evaluate the attraction - production controversy. ICES Journal of Marine Science, 59, S214-S221. Parsons, D., Buckthought, D., Middleton, C. & Mackay, G. 2016. Relative abundance of snapper (Chrysophrys auratus) across habitats within an estuarine system. New Zealand Journal of Marine and Freshwater Research, 50, 358-370. Pattiaratchi, C. & Gould, B. H. J. 1997. Impact of sea-breeze activity on nearshore and foreshore processes in. Continental Shelf Research, 17, 1539-1560. Pattiaratchi, C. Performance of an artificial surfing reef: Cable Station Western Australia. Proceedings of the Sixth International Conference on Coastal and Port Engineering in Developing Countries (Colombo, Sri Lanka), 18p, 2003 Colombo, Sri Lanka. 16. Pauly, D. 2018. A vision for marine fisheries in a global blue economy. Marine Policy, 87, 371- 374. Pauly, D., Christensen, V., Guénette, S., Pitcher, T. J., Sumaila, U. R., Walters, C. J., Watson, R. & Zeller, D. 2002. Towards sustainability in world fisheries. Nature, 418, 689-695. Pautasso, M. 2012. Publication growth in biological sub-fields: Patterns, predictability and sustainability. Sustainability, 4, 3234-3247. Pautasso, M. 2014. The jump in network ecology research between 1990 and 1991 is a Web of Science artefact. Ecological Modelling, 286, 11-12. Paxton, A. B., Shertzer, K, W. Bacheler, N. M., Kellison, G. T., Riley, K. L., Taylor, J.C. 2020. Meta- analysis reveals artificial reefs can be effective tools for fish community enhancement but are not one-size-fits-all. Frontiers in Marine Science, 7. https://doi.org/10.3389/fmars.2020.00282. Peregrin, L. S. 2017. Study of fish assemblages in Ecklonia radiata dominated rock reefs in subtropical eastern Australia and dietary analyses of three species of generalist predators of the family Labridae. Master of science, Southern Cross University. Pickering, H. & 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, 39-59. Platell, M. E., Sarre, G. A. & Potter, I. C. 1997. The diets of two co-occurring marine teleosts, Parequula melbournensis and Pseudocaranx wrighti, and their relationships to body size and mouth morphology, and the season and location of capture. Environmental Biology of Fishes, 49, 361-376. Polidoro, B. A., Carpenter, K. E., Collins, L., Duke, N. C., Ellison, A. M., Ellison, J. C., Farnsworth, E. J., Fernando, E. S., Kathiresan, K. & Koedam, N. E. 2010. The loss of species: Mangrove extinction risk and geographic areas of global concern. PLoS One, 5, e10095. Polovina, J. J. & Sakai, I. 1989. Impacts of artificial reefs on fishery production in Shimamaki, Japan. Bulletin of Marine Science, 44, 997-1003. Pomeroy, R. S. 2012. Managing overcapacity in small-scale fisheries in Southeast Asia. Marine Policy, 36, 520-527. Potter, I. C., Veale, L., Tweedley, J. R. & Clarke, K. R. 2016. Decadal changes in the ichthyofauna of a eutrophic estuary following a remedial engineering modification and subsequent environmental shifts. Estuarine, Coastal and Shelf Science, 181, 345-363.

100 Qin, M., Yue, C. & Du, Y. 2020. Evolution of China's marine ranching policy based on the perspective of policy tools. Marine Policy, 117, e103941. Recfishwest. 2020. Artificial Reefs [Online]. Available: https://recfishwest.org.au/our- services/artificial-reefs/ [Accessed 09/08/2020]. Reubens, J. T., Degraer, S. & Vincx, M. 2014. The ecology of benthopelagic fishes at offshore wind farms: A synthesis of 4 years of research. Hydrobiologia, 727, 121-136. Riggio, S., Badalamenti, F. & D’anna, G. 2000. Artificial reefs in Sicily: An overview. In: Jensen, A. C. (ed.) Artificial reefs in European seas. Southampton, United Kingdom: Kluwer Academic Publishers. Rilov, G. & Benayahu, Y. 2000. Fish assemblage on natural versus vertical artificial reefs: The rehabilitation perspective. Marine Biology, 136, 931-942. Rilov, G. & Benayahu, Y. 2002. Rehabilitation of coral reef-fish communities: The importance of artificial-reef relief to recruitment rates. Bulletin of Marine Science, 70, 185-197. Ross, P., Thrush, S., Montgomery, J., Walker, J. & Parsons, D. 2008. Habitat complexity and predation risk determine juvenile snapper (Pagrus auratus) and goatfish (Upeneichthys lineatus) behaviour and distribution. Marine and Freshwater Research, 58, 1144-1151. Rowland, A. J. 2009. The biology of samson fish Seriola hippos with emphasis on the sportfishery in Western Australia. PhD, Murdoch University. Russell, B., Fairclough, D. & Pollard, D. 2010. Ophthalmolepis lineolata. In: IUCN 2011. IUCN Red List of Threatened Species. Version 2011.2. Ryan, K., Hall, N., Lai, E., Smallwood, C., Tate, A., Taylor, S. & Wise, B. 2019. Statewide survey of boat-based recreational fishing in Western Australia 2017/18. Fisheries Research Report. Perth, Australia. Ryan, K., Hall, N., Lai, E., Smallwood, C., Taylor, S. & Wise, B. 2015. State-wide survey of boat- based recreational fishing in Western Australia 2013/14. Perth, Australia: Fisheries Research Division. Sánchez-Jerez, P. & Ramos-Esplá, A. 2000. Changes in fish assemblages associated with the deployment of an antitrawling reef in seagrass meadows. Transactions of the American Fisheries Society, 129, 1150-1159. Sanders, M. Artificial reefs in Australia. International Conference on Artificial Reefs, 1974 Houston, Texas. 84-90. Schaffer, V. 2011. Valuing artificial reefs: The case of the ex-HMAS Brisbane conservation park. PhD, University of the Sunshine Coast. Schroepfer, R. L. & 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, 93-101. Seaman Jr, W. 2002. Unifying trends and opportunities in global artificial reef research, including evaluation. ICES Journal of Marine Science, 59, S14-S16. Seaman, W. & Lindberg, W. J. 2009. Artificial Reefs. In: Steele, J. H., Thorpe, S. A. & Turekian, K. K. (eds.) Encyclopedia of ocean sciences. 2nd ed.: Academic press. Shani, A., Polak, O. & Shashar, N. 2012. Artificial reefs and mass marine ecotourism. Tourism Geographies, 14, 361-382. Sherman, R. L., Gilliam, D. S. & Spieler, R. E. 2002. Artificial reef design: Void space, complexity, and attractants. ICES Journal of Marine Science, 59, S196-S200. Sherman, R. L. & Spieler, R. E. 2006. Tires: Unstable materials for artificial reef construction. In: Brebbia, C. A. (ed.) Transactions on Ecology and the Environment. Southampton, United Kingdom: WIT Press. Si, Y. 2012. Legal regulation of artificial reef. Ocean University of China Silva, R., Mendoza, E., Mariño-Tapia, I., Martínez, M. L. & Escalante, E. 2016. An artificial reef improves coastal protection and provides a base for coral recovery. Journal of Coastal Research, 75, 467-471.

101 Simon, T., Pinheiro, H. T. & Joyeux, J.-C. 2011. Target fishes on artificial reefs: Evidences of impacts over nearby natural environments. Science of the Total Environment, 409, 4579- 4584. Smith, J. A., Cornwell, W. K., Lowry, M. B. & Suthers, I. M. 2017. Modelling the distribution of fish around an artificial reef. Marine and Freshwater Research, 68, 1955-1964. Smith, J. A., Lowry, M. B. & Suthers, I. M. 2015. Fish attraction to artificial reefs not always harmful: A simulation study. Ecology and Evolution, 5, 4590-4602. Steinback, S. R., Gentner, B. & Castle, J. The economic importance of marine angler expenditures in the United States. Eleventh Biennial Conference of the International Institute of Fisheries Economics and Trade, 2004 Seattle, Washington. US Department of Commerce, National Oceanic and Atmospheric Administration. Steneck, R. S. & Pauly, D. 2019. Fishing through the Anthropocene. Current Biology, 29, 987-992. Stephan, C. D., Dansby, B. G., Osburn, H. R., Matlock, G. C., Riechers, R. K. & Rayburn, R. 1990. Texas artificial reef fishery management plan, Austin, Texas, Texas Parks & Wildlife. Stobutzki, I. C., Silvestre, G. T. & Garces, L. R. 2006. Key issues in coastal fisheries in South and Southeast Asia, outcomes of a regional initiative. Fisheries Research, 78, 109-118. Stolk, P., Markwell, K. & Jenkins, J. M. 2007. Artificial reefs as recreational scuba diving resources: A critical review of research. Journal of Sustainable Tourism, 15, 331-350. Stone, R., Coston, L., Hoss, D. & Cross, F. 1975. Experiments on some possible effects of tire reefs on pinfish (Lagodon rhomboides) and black sea bass (Centropristis striata). Marine Fisheries Review, 37, 18-20. Streich, M. K., Ajemian, M. J., Wetz, J. J., Shively, J. D., Shipley, J. B. & Stunz, G. W. 2017. Effects of a new artificial reef complex on red snapper and the associated fish community: An evaluation using a before-after control-impact approach. Marine and Coastal Fisheries, 9, 404-418. Strelcheck, A., Cowan, J. & Shah, A. 2005. Influence of reef location on artificial reef fish assemblages in the northcentral Gulf of Mexico. Bulletin of Marine Science, 77, 425-440. Sutton, S. G. & Bushnell, S. L. 2007. Socio-economic aspects of artificial reefs: Considerations for the Marine Park. Ocean & Coastal Management, 50, 829-846. Syc, T. S. & Szedlmayer, S. T. 2012. A comparison of size and age of red snapper (Lutjanus campechanus) with the age of artificial reefs in the northern Gulf of Mexico. Fishery Bulletin, 110, 458-469. Szedlmayer, S. T. & Shipp, R. L. 1994. Movement and growth of red snapper, Lutjanus campechanus, from an artificial reef area in the northeastern Gulf of Mexico. Bulletin of Marine Science, 55, 887-896. Taylor, M. D., Chick, R. C., Lorenzen, K., Agnalt, A.-L., Leber, K. M., Blankenship, H. L., Vander Haegen, G. & Loneragan, N. R. 2017. Fisheries enhancement and restoration in a changing world. Fisheries Research, 186, 407-412. Techera, E. J. & Chandler, J. 2015. Offshore installations, decommissioning and artificial reefs: Do current legal frameworks best serve the marine environment? Marine Policy, 59, 53- 60. Terres, M. A., Lawrence, E., Hosack, G. R., Haywood, M. D. E. & Babcock, R. C. 2015. Assessing habitat use by snapper (Chrysophrys auratus) from baited underwater video data in a coastal marine park. PLoS One, 10, e0136799. Thierry, J.-M. 1988. Artificial reefs in Japan - a general outline. Aquacultural Engineering, 7, 321- 348. Thresher, R. E. & Gunn, J. S. 1986. Comparative analysis of visual census techniques for highly mobile, reef-associated piscivores (Carangidae). Environmental Biology of Fishes, 17, 93- 116. Tweedley, J. R., Warwick, R. M. & Potter, I. C. 2016. The contrasting ecology of temperate macrotidal and microtidal estuaries. In: Hughes, R. N., Hughes, D. J., Smith, I. P. & Dale, A. C. (eds.) Oceanography and marine biology: An annual review. 1st ed. Boca Raton, Florida: CRC Press.

102 Twomey, L., Waite, A., Pez, V. & Pattiaratchi, C. 2007. Variability in nitrogen uptake and fixation in the oligotrophic waters off the south west coast of Australia. Deep Sea Research Part II: Topical Studies in Oceanography, 54, 925-942. Valesini, F., Hallett, C., Hipsey, M., Kilminster, K., Huang, P. & Hennig, K. 2019. Peel- Harvey Estuary, Western Australia. Coasts and Estuaries. Elsevier. Valle, C., Bayle-sempere, J. T., Dempster, T., Sanchez-Jerez, P. & Gimenez-Casalduero, F. 2007. Temporal variability of wild fish assemblages associated with a sea-cage fish farm in the south-western Mediterranean Sea. Estuarine, Coastal and Shelf Science, 72, 299-307. Waite, A. M., Thompson, P. A., Pesant, S., Feng, M., Beckley, L., Domingues, C., Gaughan, D., Hanson, C., Holl, C. & Koslow, T. 2007. The Leeuwin Current and its eddies: An introductory overview. Deep Sea Research Part II: Topical Studies in Oceanography, 54, 789-796. Walker, T. H. E. 2016. Characteristics of the fish faunas of artificial reefs in Geographe Bay determined from video footage collected by recreational fishers. Honours thesis, Murdoch University. Walter, R. C., Buffler, R. T., Bruggemann, J. H., Guillaume, M. M., Berhe, S. M., Negassi, B., Libsekal, Y., Cheng, H., Edwards, R. L. & Von Cosel, R. 2000. Early human occupation of the Red Sea coast of Eritrea during the last interglacial. Nature, 405, 65-69. Watson, D. L., Harvey, E. S., Fitzpatrick, B. M., Langlois, T. J. & Shedrawi, G. 2010. Assessing reef fish assemblage structure: How do different stereo-video techniques compare? Marine Biology, 157, 1237-1250. Waycott, M., Duarte, C. M., Carruthers, T. J., Orth, R. J., Dennison, W. C., Olyarnik, S., Calladine, A., Fourqurean, J. W., Heck, K. L. & Hughes, A. R. 2009. Accelerating loss of seagrasses across the globe threatens coastal ecosystems. Proceedings of the National Academy of Sciences, 106, 12377-12381. Whitmarsh, S. K., Fairweather, P. G. & Huveneers, C. 2017. What is big BRUVver up to? Methods and uses of baited underwater video. Reviews in Fish Biology and Fisheries, 27, 53-73. Wilber, D., Clarke, D., Burlas, M., Ruben, H. & Will, R. 2003. Spatial and temporal variability in surf zone fish assemblages on the coast of northern New Jersey. Estuarine, Coastal and Shelf Science, 56, 291-304. Willis, T. J., Millar, R. B. & Babcock, R. C. 2000. Detection of spatial variability in relative density of fishes: Comparison of visual census, angling, and baited underwater video. Marine Ecology Progress Series, 198, 249-260. Wu, Z., Tweedley, J. R., Loneragan, N. R. & Zhang, X. 2019. Artificial reefs can mimic natural habitats for fish and macroinvertebrates in temperate coastal waters of the Yellow Sea. Ecological Engineering, 139, 105579. Wu, Z., Zhang, X., Dromard, C. R., Tweedley, J. R. & Loneragan, N. R. 2018. Partitioning of food resources among three sympatric scorpionfish (Scorpaeniformes) in coastal waters of the northern Yellow Sea. Hydrobiologia, 826, 331-351. Zhou, X., Zhao, X., Zhang, S. & Lin, J. 2019. Marine ranching construction and management in East China Sea: Programs for sustainable fishery and aquaculture. Water, 11, 1237.

103 Appendices

Appendix 1: Copy of Ramm et al. presentation given at the 10th Florida State University- Mote International Symposium on Fisheries Ecology and 6th International Symposium on Stock Enhancement and Sea Ranching in Sarasota, Florida (United States of America) in November 2019.

104

105

106

107

108

109

110 Appendix 2: Copy of Ramm et al. poster given at the 10th Florida State University-Mote International Symposium on Fisheries Ecology and 6th International Symposium on Stock Enhancement and Sea Ranching in Sarasota, Florida (United States of America) in November 2019.

111 Appendix 3: T-values from a pairwise PERMANOVA tests on the Date × Site interaction recorded on the (a) Mandurah artificial reef and (b) control site in each month between February 2019 and July 2020. Interactions that returned a significant difference between dates (P < 0.05) are shaded in grey. Data were unavailable for August 2019 on the artificial reef and August and November 2019 on the control site.

2019 2020 (a) Month Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Mar -0.20 Apr 0.21 0.27 May 0.06 0.07 -0.10 Jun 0.32 0.42 0.32 0.15 Jul 0.43 0.68 0.47 -0.00 0.22 2019 Aug Sep 0.19 0.27 0.14 0.13 0.30 0.51 Oct 0.17 0.48 0.45 0.18 0.31 0.60 0.05 Nov 0.08 0.31 0.49 0.06 0.17 0.53 0.00 -0.00 Dec 0.15 0.21 0.64 0.10 0.30 0.73 0.05 0.33 0.00 Jan 0.17 0.35 0.51 0.21 0.44 0.77 0.53 0.53 0.55 0.43 Feb 0.17 0.25 0.8 0.44 0.44 0.79 0.56 0.30 0.44 0.25 0.52 Mar 0.25 0.35 0.81 0.75 0.69 0.98 0.82 0.97 0.99 0.90 0.90 0.80 2020 Apr 0.45 0.24 0.31 0.06 0.19 0.37 0.37 0.65 0.54 0.34 0.5 0.12 0.64 May -0.00 0.07 0.38 0.40 0.28 0.83 0.21 0.69 0.68 0.27 0.63 0.39 0.44 0.02 Jun 0.43 0.47 0.56 0.28 0.06 0.08 0.47 0.56 0.46 0.57 0.63 0.59 0.65 0.18 0.41 Jul 0.59 0.75 0.55 0.37 0.31 0.39 0.47 0.81 0.67 0.77 0.93 0.77 0.89 0.00 0.23 0.31

112 Appendix 3: Continued.

2019 2020 (b) Month Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Mar -0.00 Apr 0.12 0.06 May -0.10 0.16 0.41 Jun 0.21 0.08 -0.10 0.35 Jul 0.34 0.15 0.23 0.39 0.24 2019 Aug Sep 0.23 0.06 0.06 0.53 0.17 -0.10 Oct -0.10 0.03 0.24 -0.00 0.20 0.17 0.34 Nov Dec -0.00 0.00 0.23 -0.10 0.15 0.18 0.40 0.15 Jan 0.02 0.13 0.38 0.01 0.30 0.16 0.41 0.18 0.02 Feb 0.10 0.15 0.42 0.14 0.29 0.07 0.40 0.14 -0.10 -0.10 Mar 0.05 0.13 0.37 0.03 0.23 -0.00 0.24 0.00 -0.20 -0.20 -0.30 2020 Apr 0.67 0.49 0.37 0.84 -0.00 0.69 0.68 0.69 0.53 0.59 0.63 0.69 May 0.67 0.63 0.37 0.83 0.12 0.63 0.46 0.76 0.69 0.84 0.78 0.67 0.39 Jun 0.41 0.37 0.22 0.58 -0.00 0.19 0.32 0.23 0.45 0.48 0.50 0.44 0.27 0.32 Jul 0.57 0.49 0.26 0.99 0.10 0.65 0.49 0.41 0.88 0.76 0.71 0.78 0.26 0.40 0.00

113

114