Population Demographics of Sevengill Sharks (Notorynchus Cepedianus) in Paterson Inlet, Stewart Island

Population demographics of sevengill sharks (Notorynchus cepedianus) in Paterson Inlet, Stewart Island.

A thesis submitted for the partial fulfilment of the degree of Master of Science

At the University of Otago, Dunedin, New Zealand

Robert Clarence Lewis 2019 Abstract

The broadnose sevengill shark (Notorynchus cepedianus) is an abundant large marine predator found around coastal New Zealand. Data on the ecology of the species are severely limited and anthropogenic impacts are unquantified. To partially address this, a study of the demographics of a sevengill shark population was conducted in Sawdust Bay, a shallow coastal embayment within Paterson Inlet, Stewart Island/Rakiura. Underwater video of the sharks was captured for photo-identification (photo-ID) capture-recapture and photogrammetry. Sevengill sharks were detected on all but one of 133 deployments, resulting in 349 separate encounters. Extracted images were subjected to strict image quality criteria resulting in usable images of 149 individuals that were included in the analysis. Capture-recapture analysis of the photo-ID data was conducted using a Huggins Robust Design. Abundance varied seasonally from 34 (95% CI = 21 - 55) during winter 2016, to 94 (95% CI = 44 - 199) during spring 2017. These matched seasonal trends seen in other studies of areas with high sevengill shark abundance around Tasmania. Temporary emigration rate was greatest before autumn and winter (0.721 and 0.752 respectively) and lower before spring and summer (0.331 and 0.409 respectively). Apparent annual survival rate was estimated as 0.51, but with low precision (95% CI = 0.16 - 0.86). The observed sex ratio was heavily biased, with females making up 95% of the population in winter 2016 and 73% in summer 2017. Using stereo-photogrammetric methods, 78 of the 149 individuals were measured using the image analysis software Vidsync 1.1661. Total lengths ranged from 149.0 to 276.5 cm (mean = 211.4, SE = 2.97). There was no apparent difference among seasons in the size classes that used the study area. All measured males (n = 15) were found to be at sizes consistent with maturity (>150 cm), while females (n = 75) consisted of a mix of juvenile (60%) and mature (40%) females. Not all individuals measured over the duration of the study appeared to grow, probably because sevengill sharks are slow growing when mature, and potential growth may have been swamped by measurement error. This study presents the first data on abundance, apparent annual survival, sex and size structure of broadnose sevengill sharks in New Zealand. The methods used here will be continued in a wider study with the aim of quantifying demographics and ecological role of the species.

i Acknowledgements I would firstly like to express my sincere gratitude and appreciation to my supervisors, Dr Will Rayment and Professor Steve Dawson of the Marine Science Department at the University of Otago. Will was immensely supportive over the course of my study and his guidance continues to steer me toward being a better writer and researcher. Likewise, Steve’s door was always open, and I could not have gotten this far without his advice and wisdom. Thank you both for your patience, dedication and encouragement over the journey that was my masters. I would like to thank Dr Pascal Sirguey and Dr Aurelien Vivancos who were involved in advising me when I was planning my stereo-photogrammetry methodology. I would also like to thank Clinton Duffy of the Department of Conservation and Malcolm Francis of the National Institute of Water and Atmospheric Research for their advice on many things shark related. I would like to express my gratefulness to Jordan Housiaux, who originally started the study, and got me involved. To my lab group, and those who assisted me in the field, I give you my utmost thanks for giving your time and enthusiasm to the project: Dave Johnston, Tom Brough, Marta Guerra, Tamlyn Somerford, Steph Bennington, Lindsay Wickman, Jesu Valdes, Will Carome, Toby Dickson, Eva Leunissen, Trudi Webster, Stina Kolodzey, Simone Jarret, Janice Huang, Charles Ellen, and Manon Knapen. Thank you to Rosa Edwards and Michael Heldsinger for continuing the research and officially making us “Shark Lab”. And thanks to those who helped organise and support each of my expeditions Steve Little, Bill Dickson, Evan Kenton, Julie-Anne Parsons and Chris Fitzpatrick. I would like to give many thanks to the community of Stewart Island, who welcomed me and showed genuine interest in my research. In particular Southern Seafoods, who supply bait in my times of need; Rakiura of the Ulva Island ferry, and Chris of Rakiura Charters who kept an eye out for me while we were on the water. And a big thank you to Ken McAnergney who very kindly let me use his moorings. Finally, I must express my very profound gratitude to my parents, David and Liza; my sister Isedora; and my brothers Marcus and Jared for their unconditional support and love throughout my long journey. And a very special thanks to my loving partner Jerusha who provided me with unfailing support and continuous encouragement throughout my study and writing this thesis. This accomplishment would not have been possible without everyone involved, and I sincerely thank you all.

-Robert Lewis

ii Table of contents

Abstract………………………………………………………………………………………. i Acknowledgements…………………………………………………………………...... …….ii List of figures………………………………………………………………………………… iv List of tables…………………………………………………………………………………..v Chapter one - General introduction …………………………………………………………..1 1.1 Difficulties of elasmobranch research………………………………………...1 1.2 The role of predators in marine ecosystems ………………………………….1 1.3 Methods of obtaining demographic information……………………………...2 1.4 Broadnose sevengill sharks…………………………………………………...4 1.4.1 Taxonomy………………………………………………………………4 . 1.4.2 Distribution and biogeography………………………………………. 5 1.4.3 Diet and feeding……………………………………………………… 6 1.4.4 Life history……………………………………………………………7 1.4.5 Threats………………………………………………………………...8 1.4.6 Broadnose sevengills in New Zealand……………………………….. 8 1.5 Motivation for research and objectives………………………………………..9 1.6 Objectives……………………………………………………………………...10 1.7 Thesis outline……………………………………………………………….....10

Chapter two – Estimating abundance and survival of sevengill sharks in Paterson Inlet, Stewart Island, using photo-ID capture recapture………………….…………………………..11 2.1 Introduction……………………………………………………………………11 2.2 Aims…………………………………………………………………………...14 2.3 Methods………………………………………………………………………..14 2.3.1 Study site………………………………………………………………14 2.3.2. Equipment …………………………………………………………… 16 2.3.3 Field methods………………………………………………………….16 2.3.4 Inlet wide sampling……………………………………………………18 2.3.5 Individual identification……………………………………………….18 2.3.6 Image grading………………………………………………………….20 2.3.7 Photo catalogue and encounter history………………………………...21 2.3.8 Pollock’s Robust Design and model parameters………………………22 2.3.9 Robust Design assumptions…………………………………………...25 2.4 Results…………………………………………………………………………27 2.4.1 Sighting rates, site fidelity and conditioning…………………………..27 2.4.2 Image quality results…………………………………………………...31 2.4.3 Capture-recapture analysis…………………………………………….32 2.5 Discussion……………………………………………………………………...36 2.5.1 Abundances and seasonality…………………………………………...36 2.5.2 Apparent annual survival rate………………………………………….38 2.5.3 Residency and site fidelity……………………………………………..38 2.5.4 Limitations……………………………………………………………..40 2.5.5 Conclusion……………………………………………………………..41

iii Chapter three – Describing size and sex structure of sevengill sharks in Sawdust Bay, Stewart Island, using stereo-photogrammetry……………………………………………….…42. 3.1 Introduction……………………………………………………………………42 3.2 Aims…………………………………………………………………………....44 3.3 Methods………………………………………………………………………..45 3.3.1 Stereo-photogrammetry………………………………………………...45 3.3.2 Calibration process……………………………………………………..46 3.3.3 Measuring sevengill sharks and calculating total length……………….48 3.3.4 Growth………………………………………………………………….52 3.3.5 Determining sex………………………………………………………...52 3.3.6 Determining life stage………………………………………………….55 3.3.7 Statistical methods……………………………………………………...55 3.4 Results………………………………………………………………………….55 3.4.1 Sex structure……………………………………………………………55 3.4.2 Accuracy and precision of the stereo-photogrammetry system….…….56 3.4.3 Estimates of lengths of sevengill sharks……………………………….57 3.4.4 Growth……………………………………………………………….. 62 3.5 Discussion……………………………………………………………………. 64 3.5.1 Accuracy and precision………………………………………………. 64 3.5.2 Size structure…………………………………………………………. 65 3.5.3 Sex structure…………………………………………………………..66 3.5.4 Implications of size and sex structures………………………………. 67 3.5.5 Growth……………………………………………………………….. 67 3.5.6 Limitations…………………………………………………………… 68 3.5.7 Future research…………………………………………………………69 3.5.8 Conclusions…………………………………………………………….70

Chapter 4 – General Discussion………………………………………………………………..71 4.1 Ecological importance…………………………………………………………71 4.2 Research and management of shark populations in New Zealand……………72 4.3 Main findings………………………………………………………………… 73 4.4 Implications of the results……………………………………………………. 73 4.5 Wider context………………………………………………………………… 74 4.6 Critique of baited remote underwater video…………………………………..74 4.7 Future research…………………………………………………………………76 4.8 Conclusion……………………………………………………………………..77

Appendix A. All PDL and TL measurements of sevengill sharks in Sawdust Bay, Paterson Inlet. Sex and 95% confidence intervals are included.……………………………….78 References cited………………………………………………………………………………..81

iv List of figures

Figure 1.1. Lateral view of a broadnose sevengill shark, Notorynchus cepedianus……….5 Figure 1.2. Map of the known distribution of broadnose sevengill sharks…………………6 Figure 1.3. Schematic diagram of the developmental stages of a sevengill shark…………..7 Figure 2.1. Map of the study area…………………………………………………………..15 Figure 2.2. BRUV design……………………………………………………………………16 Figure 2.3. University of Otago’s research vessel ‘Moki’…………………………………..17 Figure 2.4. Examples of images used for identification…………………………………….19 Figure 2.5. Image grading scale……………………………………………………………21 Figure 2.6. Diagram of the Robust Design sampling structure………………………………23 Figure 2.7. The mean number of individual sevengill sharks encountered per Deployment in each primary period……………………………………………28 Figure 2.8. Frequency distribution for encounter frequency of individual sevengill sharks……………………………………………………………………………29 Figure 2.9. Discovery curve of individual sevengill sharks………………………………..30 Figure 2.10 Frequency distribution for proportion of primary periods in which.…………..31 individual broadnose sevengill sharks………………………………………….35 Figure 2.11. Time-series of model averaged estimates of abundance………………..……..44 Figure 3.1. Diagram of stereo-photogrammetry design…………………………………….46 Figure 3.2. Stereo-camera calibration frame for use in image analysis…………………….47 Figure 3.3. Example of the VidSync calibration process.…………………………………..48 Figure 3.4. Diagram showing the measurements of total length and pre-caudal length……49 Figure 3.5. A 2.61m female sevengill shark that washed up in Hoopers Inlet, Dunedin…..50 Figure 3.6. An example of the light patch of skin that indicates the start of the dorsal fin……………………………………………………………………………….51 Figure 3.7. Examples of measuring sevengill sharks in Vidsync…………………………..50 Figure 3.8. An image that displays the dark band that runs along the spine of sevengill sharks………………………………………………………………..52 Figure 3.9. Captured images of female and male sevengill sharks…………………………54 Figure 3.10. Seasonal sex ratio of sevengill sharks in Sawdust Bay…………………………56 Figure 3.11. Comparison of measured TL of sevengill sharks using stereo-photogrammetry against the calculated TL using the PDL conversion formula……………………………………………………………...57 Figure 3.12. Total length distribution of sevengill sharks in Sawdust Bay……………………59 Figure 3.13. Seasonal total length distribution of female sevengill sharks…………………..60 Figure 3.14. Average total length for male and female sevengill sharks………………………61 Figure 3.15. Initial and final PDL measurements of eight sevengill sharks over a period of >300 days…………………………………………………………….62 Figure 3.16. Size distribution of sevengill sharks…………………………………………….63 Figure 3.17. A smaller broadnose sevengill pushes a larger individual out of the way……..68

v List of tables

Table 2.1. Image quality grades…………………………………………………………….20 Table 2.2. Time span and total video hours for all sampled seasons………………………27 Table 2.3. Regression analysis of individuals encountered over consecutive deployment each season…………………………………………………………29 Table 2.4. Proportion of image grades for all encounters with sevengill sharks by season. 32 Table 2.5 Ranking of Huggins Robust Design models to estimate survival probability (j), capture probability (p) and temporary emigration probability (γ) for sevengill sharks in Sawdust Bay…………………………….33 Table 2.6. Model averaged parameter estimates, standard errors (SE.) and 95% confidence intervals for apparent annual survival probability (j), capture probability (p) and temporary emigration probability (γ) for sevengill sharks…35 Table 2.7. Model averaged abundance estimates of sevengill sharks……………………… Table 3.1. The number of individuals measured and the percentage they made up in each season separated by sex…………………………………………………….58 Table 3.2. Measurements of eight sevengill sharks with a minimum 300 days at liberty….63

vi Chapter 1 - General introduction

1.1 The challenges of elasmobranch research Characterising and understanding the biology and ecology of mobile marine species can be challenging. Obtaining sufficient data to estimate important population parameters such as abundance and fecundity is often not possible without large investment of time and resources (Santana-Garcon et al., 2014). Sharks exemplify this problem. Many species undergo large migrations, inhabit the deep sea, and/or are highly pelagic. Dedicated studies are usually expensive, or reliant on data derived from fisheries which may not be accurate (Murphy & Jenkins, 2010). Overfishing and habitat degradation have greatly altered the abundances of marine species, in particular large predators such as sharks (Baum & Worm, 2009; Dulvy et al., 2014). To make sound management decisions it is essential to have a better understanding of how anthropogenic pressures are changing abundances of predatory species within food webs (Myers et al, 2007). The broadnose sevengill shark (Notorynchus cepedianus) is a common predator found in temperate coastal habitats around the world (Compagno, 1984). Although there is some evidence for a wide distribution around New Zealand (Anderson et al., 1998), the species has been poorly studied, and is described as “data deficient” by the IUCN (Compagno, 2005; IUCN, 2009). The aim of this study is to obtain the first systematic demographic data on a population of broadnose sevengill sharks in New Zealand using underwater stereo-video methods. In doing so, I attempt to address the lack of knowledge on sevengill sharks in New Zealand and improve our understanding of this species.

1.2 The role of predators in marine ecosystems Predators often play an important role in maintaining ecosystem health by limiting the population growth potential of their prey, and/or their grazing potential. This is known as top- down control (Estes, 1996; Schmitz et al., 1997). By limiting the growth or predation potential of meso-trophic species, low trophic level species can be saved from unsustainable predation pressures (Estes et al., 2011; Wallach et al., 2015). In some cases, removal of high-trophic level predators has resulted in a trophic cascade, where meso-trophic predators over-exploit low- trophic level species leading to knock-on effects two or three trophic levels down (Ducklow, 2009; Grubbs et al., 2016). The importance of predator species varies among ecosystems depending, in part, on the complexity of the food web (Strong, 1992; Estes, 1996). Marine

1 ecosystems can be particularly complex owing to linkages between different ocean zones (e.g. abundance of pelagic species affecting the health of benthic species; Sieben et al., 2011), and high levels of redundancy (multiple predators feeding on the same prey species; Ruttenberg et al., 2011; Ritchie et al., 2012; Grubbs et al., 2016). Sharks have long been promoted as being responsible for structuring marine food webs, both directly by regulating prey abundances through predation (consumptive effects), and indirectly by modifying prey behaviour, physiology or morphology (non-consumtive effects; Heithaus et al., 2008; Jermacz & Kobak, 2017). It is no surprise then, that the loss of sharks as apex predators from marine ecosystems can have a range of impacts. For example, the health of reef structures around tropical islands has been shown to decline (Sandin et al., 2008; Ruppert et al., 2013; Rasher et al., 2017), seagrass beds have degraded from increased grazing (Heithaus et al., 2012), meso-predator abundances have increased (Shepherd & Myers, 2005; Kitchell et al., 2006; Barley et al., 2017a), and teleost reef fish assemblages have substantially changed (Friedlander & DeMartini, 2002; Barley et al., 2017b). The impacts of removing sharks from marine ecosystems are rarely straightforward however, and the full range of effects is often hard to predict (Sieben et al., 2011). For instance, large sharks (>2.5m) often inhabit the same trophic guild as large teleost fish, dolphins, toothed whales, seals and other sharks (Frisch et al., 2016). The effects on the ecosystem of removing sharks may then be masked by the presence of these other species (Grubbs et al., 2016). Before we can tease apart which impacts are attributable to changes in shark populations, we need basic information on the abundance, distribution, demographics, and foraging ecology of a shark species (Dudgeon et al., 2015; Siqueira et al., 2017).

1.3 Methods of obtaining demographic information In the past, the logistical difficulty of conducting research on large marine predators has profoundly limited the ability to obtain accurate demographic data (Ferretti et al., 2010). Traditional methods were dependent on data from either commercial or recreational fisheries, or from independent research which caught and measured them (Santana-Garcon et al., 2014). However, recent advances in technology have started to make this research more feasible (Barnett et al., 2010; Jacoby et al., 2012). Physical capture, via fishing, has allowed collection of a variety of data including relative abundances, size and sex structure, skin, and blood samples. Furthermore, live-captured specimens can be tagged and released, in an attempt to gain insights into movement ecology

2 and survival rates (Santana-Garcon et al., 2014). These methods do come with drawbacks. Firstly, the data have intrinsic biases such as size-selectivity or catchability arising from the fishing gear that was employed (e.g. trawls, longlines, set nets; McAuley et al., 2007). Secondly, the animals have to be caught and handled, resulting in some level of physical and physiological stress (Skomal, 2007). Elasmobranchs show a variety of responses and susceptibility to physiological stress, ranging from no obvious effects, to behavioural modification (Hammerschlag et al., 2011), to mortality, some of which may happen post- release (Skomal, 2007: Mandelman & Skomal, 2012; Marshall et al., 2012). Post-release mortality is very difficult to quantify (Skomal 2007; Afonso & Hazin, 2014; Escalle et al., 2016), so the effects of studies which employ direct capture methods are largely unknown. Tag and release is a capture-recapture methodology which can also be used to estimate abundance, survival and emigration. It has been employed on sharks since the 1950s (Kohler & Turner, 2001) in tandem with fisheries based methods. Because it involves capture, tag and release has the same drawbacks listed above, especially that post-release survival is usually unknown. Tag and release methods are used on elasmobranchs not only in professional research (e.g. Templeton, 1954; Kohler, 2002; Holmes et al., 2014), but also in citizen science, such as the National Oceanic and Atmospheric Administration’s Cooperative Shark Tagging Program (NOAA NMFS CSTP). There is a very real possibility that recreational anglers, believing their actions to be responsible and helpful for understanding elasmobranch populations, may in fact be having a negative impact, especially on larger, vulnerable species such as thresher sharks, Alopias vulpinus (Heberer et al., 2010). Lastly, physical capture of large sharks undoubtedly involves some risk to the fishers. Natural markings on elasmobranchs can often be used to identify individuals (e.g. great white sharks, Carcharodon carcharias, Ryklief et al., 2014), meaning individual identification does not always have to rely on physical tagging. In these cases, obtaining a photograph of the identifying marking is equivalent to an individual being ‘tagged’. The use of photography or videography to study sharks has become more prevalent in recent years as cameras have become smaller, more powerful, and cheaper (Marshall & Pierce, 2012). Photo and video based studies are non-invasive, can be cheaper than using fishing methods, and can produce information on relative abundances, sex, and size structure (Shortis & Harvey, 1998; Santana- Garcon et al., 2014). The approach is especially suitable for large species which would be difficult to handle, such as basking sharks, Cetorhinus maximus (Gore, 2016) and whale sharks, Rhinchodon typus (Meekan et al., 2006). Capture-recapture analysis methods are often used in conjunction with photo-identification (photo-ID) studies to produce estimates of abundance

3 and survival rates (Auger-Méthé & Whitehead, 2007). Photo-ID data can also be used to investigate migration, site fidelity and habitat use by elasmobranch species (e.g. Ryklief, 2014).

1.4 Broadnose sevengill sharks The broadnose sevengill shark is a large coastal predator found around the world, including in New Zealand waters. Although, there has been little published research on the ecological role they play in their respective ecosystems (Barnett et al., 2012), Cortes (1999) ranked them as more influential in shaping ecosystems than great white sharks, owing to their varied diet. As with other sharks that spend time in coastal areas (e.g. tiger shark, Galeocerdo cuvier, blue shark, Prionace glauca, great white shark; Baum & Worm, 2009), they have been shown to be important in maintaining species diversity, stability, and health of marine ecosystems (Barnett et al., 2012). Sevengill sharks have individually distinctive pigmentation patterns on their dorsal surfaces, so individuals can be easily distinguished via photography or videography (Housiaux et al, 2018). This makes photo-ID capture-recapture methods especially well-suited for studies of sevengill shark abundance, demographics and distribution (Housiaux et al, 2018).

1.4.1 Taxonomy Broadnose sevengills are placed within the order Hexanchiformes, which contains only two families; Chlamydoselachidae or frilled sharks, and Hexanchidae, or cow sharks. The Hexanchidae is a globally distributed family that contains four extant species: the sharpnose sevengill (Heptranchias perlo), bluntnose sixgill (Hexanchus griseus), big eye sixgill (Hexanchus vitulus), and the broadnose sevengill shark (Notorynchus cepedianus) (Fig. 1.1). Hereafter, ‘sevengill shark’ refers only to the broadnose sevengill shark. As a part of the Hexanchidae family, sevengills are modern sharks that retain ancient features compared to other shark families, such as extra pairs of gills (seven in total, compared to the typical five), a single dorsal fin and an anal fin (Barnett et al., 2012). Their dorsal surface is dark grey, spotted with white and black pigment, and they have a white underbelly. The top lobe of the caudal fin is greatly elongated while the bottom lobe is reduced (Compagno, 1984). Sevengill sharks have a broad blunt snout and a sub-terminal mouth (Young, 1962; Compagno, 1984; Compagno et al., 2005; Fig. 1.1).

Figure 1.1. Lateral view of a broadnose sevengill shark, Notorynchus cepedianus (adapted from Compagno, 1984).

1.4.2 Distribution and biogeography Broadnose sevengill sharks usually inhabit temperate, turbid coastal waters (Williams et al., 2011). They can be found in bays, estuaries and inlets from 1m-200m deep (Barnett et al., 2012), but are typically found in depths less than 50m (Compagno, 1984; Compagno et al., 2005; Abrantes & Barnett, 2011). In the Northern Hemisphere, sevengill sharks are found off the west coast of North America and in the waters around Japan and China. In the Southern Hemisphere they are found in the coastal waters of South America, Namibia, South Africa, South Eastern Australia, Tasmania, and New Zealand (Compagno, 1984; Compagno et al., 2005; Fig. 1.2). Although there have been no systematic surveys in New Zealand, data from bottom trawl fisheries suggest they have a wide distribution (Anderson et al., 1998). They also occur in relatively high densities year-round in Paterson Inlet, Stewart Island (Housiaux et al., 2018).

Figure 1.2. Map of the known distribution of broadnose sevengill sharks, Notorynchus cepedianus (adapted from Compagno, 1984; 2005).

Sevengill sharks exhibit seasonal movement patterns, with high abundances found in shallow coastal embayments during spring and summer (Barnett 2012). During winter they remain present at low abundances or leave for deeper waters (Lucifora et al., 2005; Barnett et al., 2010; 2011; Williams et al., 2012). Other marine species that exhibit seasonal migrations, for example turtles and whales, do so for reproductive reasons, or to access seasonal foraging areas (Stevick et al., 2011; Narazaki et al., 2015). The reasons for the observed seasonal migrations by sevengill sharks are not fully understood, but it has been suggested that pregnant females move offshore during winter to give birth to pups in deeper water, while the remainder of the population remains in coastal habitats (Barnett et al., 2011).

1.4.3 Diet and feeding Sevengill sharks are generalist feeders and are able to forage on a wide variety of prey through opportunistic scavenging as well as active hunting (Ebert, 1991a). Stomach contents of dissected sharks have included cephalopods, teleost fish, skates, rays, sharks, pinnipeds, and cetaceans (Ebert, 1991b). Populations from Namibia, South Africa, South America and California show dietary differences depending on what prey are available in each respective area (Ebert, 1991b). Other sharks, such as dogfish (Squalidae), rig (Triakidae) and catsharks (Scyliorhinidae) make up a considerable proportion of their prey (up to 50%), regardless of geographical region (Ebert, 1991a, b; Barnett et al., 2012).

6 An ontogenetic shift in diet occurs, with juvenile individuals feeding on small bony fish, cephalopods and small sharks (<1m), while adults feed on larger reef fishes, sharks, rays and marine mammals (Ebert, 1991a; Lucifora et al., 2005; Abrantes & Barnett, 2011). Cannibalism is displayed by sevengill sharks which target smaller or distressed conspecifics (e.g. wounded individuals or those caught on a line; Ebert, 1991b; 1996; Barnett et al., 2012). Ebert (1991a) described sevengill sharks as opportunistic feeders that exploit locally abundant prey species, in contrast to more selective elasmobranchs such as great white sharks which focus on a preferred prey (Ryklief et al., 2014). Sevengill sharks may forage in loose aggregations when hunting larger prey such as seals or dolphins, which a single individual could not handle on its own (Ebert, 1991b; Motta & Wilga, 2001; Compagno et al., 2005). For example, aggregations of sharks predating cape fur seals (Arctocephalus pusillus) have been observed in St. Helena Bay, South Africa, and in Lüderitz Lagoon, Namibia (Ebert, 1991a). No direct observations of predation on dolphins have been reported, however, the rate at which cetacean remains are found in stomach contents seems to be too high to be accounted for by scavenging alone (Cawthorn, 1988; Ebert, 1991a; b).

1.4.4 Life history Life history studies of sevengill sharks have taken place in California (Ebert, 1986), South Africa (Ebert, 1996), and Argentina (Lucifora et al., 2005).

Figure 1.3. Schematic diagram of the developmental stages of a sevengill shark (adapted from Ebert, 1986; 1996; Dykhuizen & Mollet, 1992).

Gestation is thought to last around 12 months, followed by a one-year break to regain condition (Ebert, 1989). This results in a mature female population that in the breeding season contains equal numbers of gravid and non-gravid females (Ebert, 1996). Reproduction is ovoviviparous and females are recorded pupping between 55 and 100 pups at a time (Ebert, 1996; 2001), which is considered high for an elasmobranch (Ebert, 1996). Pups are born at 35 - 55 cm total length (Ebert, 1986; Compagno et al, 2005). Size at maturity for male sevengill sharks is reported to be between 130 - 180 cm, at an age of 4-5 years; females are reported to be mature at between 220 - 250 cm at 11 - 21 years of age (Ebert, 1989, Dykhuizen & Mollet, 1992) (Fig. 1.3). Longevity is estimated to be around 30 - 50 years based on growth models using captive specimens (Compagno et al, 2005; Braccini et al., 2010; Barnett et al., 2012).

1.4.5 Threats From the late 1970s, sevengill sharks were taken in recreational and commercial fisheries along the California coast, with peak landings occurring in 1981, before a sharp decline when the fishery crashed in 1986 (Fowler et al., 2005; Barnett 2012). Similarly, commercial harvest in the Lüderitz lagoon, Namibia, crashed just nine months after starting in 1990 (Fowler et al., 2005). Sevengill sharks are also regularly caught as bycatch throughout their distribution (e.g. Ebert, 1986, 1996; Fowler et al., 2005). Bycatch in fisheries is the principal threat to most large shark species such as tiger shark, mako shark (Isurus oxyrinchus), and blue shark (Vas, 1990; Myers et al., 2007; Bustamante & Bennett, 2013; Levesque, 2013). Recreational fishing pressure on sevengill sharks is unquantified, but given the overlap in distribution there are undoubtedly many recreational catches each year (Barnett et al., 2012).

1.4.6 Broadnose sevengill sharks in New Zealand Until recently, knowledge of sevengill sharks in New Zealand waters was based mainly on reported commercial catch data (e.g. Ford et al., 2018) and application of findings from populations in other countries (e.g. Ebert, 1986; 1996; Lucifora et al, 2005; Barnett et al, 2012). Housiaux (2016) conducted the first dedicated study of sevengill sharks in New Zealand, providing data on occurrence, distribution, and genetics for populations at two locations in the South Island. However, the methods employed meant the study resulted in only limited information on abundance, sex ratios and size structure. Therefore, much work remains to be completed on the key demographic parameters for sevengill sharks in New Zealand.

8 It is unclear to what extent sevengill sharks are affected by anthropogenic impacts in New Zealand. There is no targeted fishery, and the species is not within the New Zealand quota management system (QMS), but bycatch occurs in longline and trawl fisheries (Ford et al., 2018). Bycaught sevengill sharks have at times been sold by commercial fishers, but on a very small scale (Steve Little, pers. comm.). As the species is present in other parts of the world and has relatively high fecundity for an elasmobranch, broadnose sevengill sharks are regarded as being at low risk from commercial impacts in New Zealand (Ford et al., 2018). Ministry of Primary Industry (MPI) reports indicate that catch data on sevengill sharks in New Zealand are of very poor quality, that there are no abundance indices available, and bycatch reporting of this species is likely to be unreliable (Ford et al., 2018). Annual reported landings in New Zealand between 2008 - 2016 were approximately 10t per fishing year (Ford et al., 2015; Clinton Duffy, pers. comm.). Given the lack of incentive to report bycatch, and recent evidence of high levels of unreported bycatch in New Zealand fisheries (Simmons et al., 2016), reported catch figures are likely to be a significant underestimate of true catch levels. Additionally, there are no recreational catch data available. Even if robust data for commercial and recreational catch existed for New Zealand however, we lack the biological knowledge to put such catch data into perspective.

1.5 Motivation for research and objectives The broadnose sevengill shark is a widely distributed species present in many coastal habitats around the world. In New Zealand, sevengill sharks may play an important role in shaping coastal ecosystems, especially considering the decline of other large shark species (e.g., great white sharks) in recent decades (Malcolm et al., 2001; Barnett et al., 2012; Hillary et al., 2018). There are, however, large gaps in our knowledge; population abundances, size structure, sex structure and growth rates are essentially unknown. Sensible and effective management requires understanding of these basic parameters. The present study builds on the information obtained in the first dedicated study of sevengill sharks in New Zealand (Housiaux , 2016; Housiaux et al., 2018). She attracted sevengill sharks to the surface of the water for boat-based photography, but had difficulties obtaining consistent photo-ID data, and suggested that encounter rates may have been significantly negatively biased by the methodology used. In the current study, I used baited remote underwater stereo- video in an attempt to improve the encounter and recapture rates. Using these methods, it was

9 possible to estimate of population parameters of sevengill sharks via mark-recapture analysis methods, and measure sharks using stereo-photogrammetry.

1.6 Objectives The objectives of this study include: 1. To develop a baited-remote underwater video system tailored to capturing images of broadnose sevengill sharks for photo-ID capture-recapture analyses.

2. To estimate the seasonal abundances and survival rate of the broadnose sevengill shark population in Paterson Inlet, Stewart Island/Rakiura.

3. To characterise the size and sex structure of the broadnose sevengill shark population in Paterson Inlet, via photogrammetry using a baited stereo video system.

1.7 Thesis outline In Chapter two I estimate seasonal abundances and survival rate of sevengill sharks in Paterson Inlet using Robust Design capture-recapture analysis (Pollock, 1990) of photo-ID data gained via baited underwater video. This chapter explores the seasonal changes in broadnose sevengill shark numbers using Sawdust Bay within Paterson Inlet from June 2015 to November 2016. These are the first abundance estimates produced within New Zealand. Chapter three focusses on the population structure of sevengill sharks in Paterson Inlet using stereo-photogrammetry (e.g. Harvey & Shortis, 1995). Stereo-photogrammetry allows for the total lengths of the sharks to be measured, thus characterising size structure of the population. Size structure relates to age and maturity, which when combined with sex structure, indicates what role this particular sub-population plays in the metapopulation. Chapter four provides a general discussion of the findings of this study, outlines how they could be applied, and provides recommendations for future research.

10 Chapter 2 – Estimating abundance and survival of sevengill sharks in Paterson Inlet, Stewart Island, using photo-ID capture-recapture. 2.1 Introduction The majority of elasmobranch species around the world lack basic demographic data (Dulvy et al., 2008). New Zealand elasmobranchs are no different; the most recent chondrichthyan risk assessment stated that of 50 assessed taxa only 18 had sufficient data available to make a robust assessment of their status (Ford et al., 2018). Numerous studies emphasise the importance of sharks to ecosystem and food web health (e.g. Stevens et al., 2000; Ferretti et al., 2010; Heithaus et al., 2012), yet basic population characteristics such as abundance, distribution, size structure, and reproductive output are unknown for most species (e.g. Ford et al., 2018). This information is vital for understanding ecological role, but also crucial for assessing conservation status (Lettink & Armstrong, 2003). The collapse of the spiny dogfish (Squalus acanthias) population in European waters for example, was caused in part by the limited understanding of spiny dogfish migrations outside of established protected areas (Fordham, 2006). Porbeagle sharks (Lamna nasus) have faced unregulated fishing pressure in the northeastern Atlantic since the 1930s resulting in serious declines (Fowler et al., 2004). Despite this, appropriate management was not put into place as there was a lack of sound data for the species (Fowler et al., 2004). Basking sharks (Cetorhinus maximus) around New Zealand have become rare due to continued bycatch (Francis & Smith, 2010). Protection for the species had been planned, but the lack of basic demographic information prevented effective management (Francis & Smith, 2010). One approach to generate these demographic data is through the use of capture-recapture methods (Lincoln, 1930; Pollock et al., 1990). These methods require that individuals are individually identifiable and can be encountered on more than a single occasion (Seber, 1973; Auger-Méthé & Whitehead, 2007). The utility of capture-recapture studies is dependent on the species, capture method, and environmental situation (Lettink & Armstrong, 2003). Under the right circumstances, capture-recapture studies can generate reliable estimates of a population’s demographic status, and may contribute to understanding the home-range and site-fidelity of individuals within that population (Kohler & Turner, 2001; Lettink & Armstrong, 2003). To render individuals identifiable, it is normally required to uniquely ‘mark’ them for later re- identification. Over a wide range of terrestrial and marine species, this can be achieved using

11 bands, brands, tags or ID clips (e.g. Lincoln, 1930; Carver et al., 1999; Baker et al., 2006; Bruce & Bradford, 2013). Such marking or tagging can be invasive (e.g. clipping a toe, implanting a tag), and usually requires handling (e.g. Holmes et al., 2014; Vandenabeele et al., 2014). Depending on the species, capture, handling, and tagging can lead to high levels of stress which may modify behaviour, such as leaving the study area or avoiding future capture (Mali et al., 2012; Marshall & Pierce, 2012; Santana-Garcon et al., 2014). Marking may also impact the animals’ fitness, reducing the ability to move, feed, reproduce or survive (Dicken et al., 2006; Chivers et al., 2016). This is not only inherently undesirable in terms of the animals’ welfare, but it may lead to biased estimates of demographic parameters from that population (Seber, 1986). Confounding stress or behavioural issues related to tagging can be avoided if individuals are naturally distinguishable (e.g. via pigmentation patterns, or physical features). Photo-ID capture-recapture studies use captured images of natural identifying features, ‘marking’ the individual without the need for invasive techniques (Würsig & Würsig, 1977; Harvey & Shortis, 1995; Harvey et al., 2003; Arzoumanian et al., 2005). Along with minimising handling issues, photo-ID further avoids some of the problems traditionally associated with physical tags, as natural marks are less likely be removed, lost, or fouled (although they may change with time; Marshall & Pierce, 2012). Photo-ID capture-recapture studies have been successfully conducted on a variety of vertebrate taxa including salamanders (Bailey et al., 2004), zebra (Grange et al., 2015), cheetah (Marnewick et al., 2008), dolphins (Gormley et al., 2005; James et al., 2015), whales (Bradford et al., 2006; Auger-Méthé & Whitehead, 2007), and sharks (Porcher, 2005; Speed et al., 2008; Ryklief et al., 2014; Gore et al., 2016). For photo-ID to be appropriate, individuals need to be reliably distinguished and re-identified over time (Stevick et al., 2001). Inconsistency among different analysts can cause bias when matching images for photo-ID studies, so steps should be taken to minimise this (Holmberg et al., 2009; Marshall & Pierce, 2012). For example, implementing a clear set of image quality criteria minimises the risk of making errors in the matching process (Urian et al., 2015). Furthermore, standardising a specific area of the animal to focus on for identification (e.g. pigmentation on the dorsal fin only; Ryklief et al., 2014) can reduce subjectivity and misidentification. If these requirements are met, photo-ID capture-recapture is a tested and robust method for obtaining demographic information on many wildlife populations (Urian et al., 2015). Photo-ID of elasmobranchs has become more widely used in recent years owing to increased accessibility to appropriate camera technology (Marshall & Pierce, 2012). Photo-ID capture-

12 recapture, using body pigmentation and/or dorsal fin notches to identify individuals, has been successfully used with whale sharks (Rhinchodon typus, Speed et al., 2008; Holmberg et al., 2009), great white sharks (Carcharadon carcharias, Chapple et al., 2011, Ryklief et al., 2014), basking sharks (Gore et al., 2016), raggedtooth sharks (Carcharias taurus, Van Tienhoven et al., 2007), and sicklefin lemon sharks (Negaprion acutidens, Buray et al., 2009). Capturing photo-ID images of sharks usually requires the photographer to either be diving with an underwater camera (e.g. Gunn, 1999), or working from a boat and attracting individuals to the surface using bait (e.g. Ryklief et al., 2014). These are not ideal methods as there is an inherent risk in diving with some species of shark, and not all species can be attracted to the water surface easily. The broadnose sevengill shark (Notorynchus cepedianus) for example (hereafter sevengill shark), is a poorly studied species that fits these criteria. They are large enough to pose some risk to divers (Compagno, 1984), and although they can be attracted to the surface with bait (Housiaux et al., 2018), they are predominantly demersal foragers (Barnett et al, 2012). They do, however, possess markings that vary between individuals and which are suitable for photo-ID capture-recapture (Housiaux, 2016). As discussed in section 1.3, what little we know about sevengill sharks in New Zealand comes from one MSc thesis and an associated paper (Housiaux, 2016; Housiaux et al., 2018), and unpublished data from incidental catch in fisheries for other species. Housiaux (2016) compared photo-ID capture-recapture with physical tagging for obtaining demographic data. Dorsal spot markings which allowed for individual identification were shown to be stable over at least five years on captive sevengill sharks (Housiaux, 2016), and so provided suitable marks for capture-recapture methods. Sevengill sharks were attracted to the water surface with bait, and photo-ID images were obtained before attempting to implant an ID tag in the dorsal surface of the free-swimming shark. Due to tag loss and biofouling of the tags, photo-ID of natural marks was concluded to be the more promising method. However, the sample size in the photo- ID study was relatively small, most likely because the demersal sharks were reluctant to follow the bait to the surface. This limited the effectiveness of the method and prevented the calculation of an abundance estimate. To date, no abundance estimates have been produced for sevengill sharks in New Zealand. Baited remote underwater video (BRUV) provides an alternative to surface photography (Santana-Garcon et al. 2014) for photo-ID. BRUV has become widely used for conducting underwater scientific surveys (e.g. Cappo et al., 2003; Ruppert et al., 2013), due to its non- invasive nature, cost effectiveness, and lower subjectivity compared to counts by divers (Langlois et al., 2006; Lowry et al., 2012; Santana-Garcon et al., 2014; Letessier et al., 2015).

13 BRUV is easily deployed near the seabed, which should increase the encounter rate of sevengill sharks. The resulting photo-ID data can then be used in capture-recapture analyses to estimate abundances and survival rates, and examine site fidelity (e.g. Bailey et al., 2004; Siqueira et al., 2017). Gaining robust estimates of these demographics is imperative to improve our understanding of the ecology of sevengill sharks and facilitate conservation and management of the species, both in New Zealand and abroad.

2.2 Aims To date there have been no estimates of demographic rates for sevengill sharks anywhere in New Zealand. The aim of this chapter is to produce the first estimates of abundance and survival rate by employing photo- ID capture-recapture analyses on a population of sevengill sharks in Paterson Inlet, Stewart Island.

2.3 Methods

2.3.1 Study site The primary study site was Sawdust Bay, located at the entrance to the North Arm of Paterson Inlet, Stewart Island (46.9012°S, 168.0127°E) (Fig. 2.1). Sawdust Bay has a maximum water depth of 15m, and the seabed consists of sand and mud. Sawdust Bay was chosen because the previous research by Housiaux et al. (2018) and local anecdotal information indicated that sevengill sharks could be reliably encountered there.

Figure 2.1. Map of the study area located in Paterson Inlet, Stewart Island. The primary sample site in Sawdust Bay, the secondary sample sites and the Ulva Island marine reserve are indicated. Insets show the location of Paterson Inlet on Stewart Island, within New Zealand.

15 2.3.2. Equipment Video images were obtained using a downward-facing baited underwater video (BRUV) system equipped with two Go-Pro Hero 4 Silver cameras mounted in waterproof housings in a stainless-steel frame (Fig. 2.2). The cameras were aligned so the lens axes converged at an angle of 20 degrees and were positioned 1.2m above a 1m long pole marked with 10 cm graduations. A bait box was attached in the middle of the length pole. Two Qudos 300 lumen LED lights were used to illuminate the field of view (Fig. 2.2).

Figure 2.2. BRUV design. Go-Pro cameras are positioned 40 cm apart looking down onto the length pole and bait box 120 cm away. Dashed lines indicate approximate field of view of cameras.

2.3.3 Field Methods Data were collected between 19/06/2016 and 6/11/2017 during six field expeditions ranging from two weeks to a month in duration. There was one field expedition per season. Seasons were classified as winter (June - August), spring (September - November), summer (December - February), and autumn (March – May). Trips to the study site were conducted twice a day, in the morning and afternoon if weather allowed. A single deployment was classified as when the BRUV was submerged and successfully captured images for one hour. Once the first deployment of a trip was completed, the BRUV was retrieved, the batteries were changed, and

16 a second deployment completed if weather conditions allowed. Deployments were taken across a range of tidal states. All deployments were conducted in Beaufort sea-state <4 and swell height <1m from a 4.3m Kiwikraft vessel (Fig. 2.3).

Upon reaching the study site, the vessel was anchored and chumming commenced. Chum consisted of a mixture of ground fish frames and seawater. The primary species used as chum was blue cod (Parapercis colias), but as this was not always available, barracouta (Thyrsites atun), salmon (Oncorhynchus tshawytscha) were used as a temporary replacement until more blue cod could be sourced. Housiaux (2018) showed that bait type was not a significant predictor in the number of sevengill sharks that were attracted to bait. The bait box connected to the BRUV was filled with the same bait used in the chumming to lure the sharks into view of the cameras when deployed. The nature of the bait box and the type of bait used ensured that sharks were not able to consume any of the bait. Recording on the Go-Pros was initiated, and a calibration process conducted (see Chapter 3 methods). Once this was complete the BRUV was deployed to within two meters of the seabed. Chumming continued steadily throughout the deployment to ensure a continuous slick was produced. Date, time, location (GPS coordinates), weather conditions and bait type were recorded as accompanying data.

Figure 2.3. University of Otago’s 4.3m research vessel ‘Moki’ from which the data were collected.

17 2.3.4 Inlet wide sampling In addition to the data collected in Sawdust Bay, deployments were also undertaken between 26/10/17 and 06/11/17 at three secondary sites; Trumpeter Point, Papatiki Bay and Prices Inlet (see Fig 2.1). This was to compare the density of sharks at different locations within Paterson Inlet. These sites were selected using local anecdotal information on where sevengill sharks had also been previously observed. Four deployments (two in the morning and two in the afternoon) were conducted at each secondary site over a span of two weeks. The sampling methods at these sites were the same as those undertaken at the primary site, Sawdust Bay.

2.3.5 Individual identification Video footage was examined using Avidemux 2.7.1 which allowed for frame by frame viewing. During examination of the footage, each time a sevengill shark entered the field of view, the video was paused, and the video time recorded. The dorsal region of the shark from the rostrum to the trailing edge of the pectoral fins was used for identification, as it provided the easiest area to compare spot patterns (Fig. 2.4). The best still frame was extracted from each encounter, and hereafter referred to as an “image” for photo-ID purposes. If a video frame indicated the shark’s sex, that frame’s video time was recorded, and an additional image was extracted. Sharks were sexed based on the presence or absence of claspers (elongated modifications of the pelvic fins that are calcified on male sharks; Ebert, 1989). These are easily observed in sevengill sharks owing to their undulating swimming motion.

Figure 2.4. Examples of images used for identification. The animals are perpendicular to the camera so that the white pigmentation on the dorsal surface (from the trailing edge of the pectoral fins to the tip of the rostrum) is clear.

19 2.3.6 Image grading Images from each encounter were graded before being used for comparisons. A quality score (1 - 4) was allocated based on image sharpness and the animal’s position in the frame (Table 2.1, Fig. 2.5). Ideally the shots were in sharp focus, with the spot markings clearly visible, and the shark was perpendicular to the camera so that the markings on both left and right of the dorsal surface could be identified. The best image of each individual was retained. Only images of grade 1 or 2 were used in comparing to other images.

Table 2.1. Image quality grades dependant on sharpness, clarity, and shark position in the frame. Grade Description 1 The image is in sharp focus, markings are clearly distinguishable, and the animal is positioned perpendicularly to the cameras.

2 Lighting and/or water clarity mean that the dorsal markings are clearly distinguishable but not completely sharp. The animal is perpendicular to the cameras.

3 The image is in sharp focus, and markings are distinguishable, but the animal is not positioned well in the frame.

4 The image is not sharp, and the markings are not distinguishable.

Figure 2.5. Grades of images ranging from 1 - 4 depending on clarity, focus and body position (refer to Table 2.1).

2.3.7 Photo catalogue and encounter history For each BRUV deployment, the best image of each individual (if it passed the image quality criteria) was compared to the existing catalogue. If the individual already existed in the catalogue, its sighting history was updated. If not, it was given a unique alphanumeric code and added to the catalogue. Sharks were coded in the order they were discovered (i.e. the first shark identified was S001, the second was S002, etc.). As sharks were identified and added to the catalogue their relevant data were recorded into an encounter database. Relevant data included the ID code (e.g. S007), date an individual was seen (encounter history), and sex (if known).

21 2.3.8 Pollock’s Robust Design and model parameters Photo-ID capture histories were analysed to obtain estimates of abundance and survival rate for sevengill sharks using Sawdust Bay. Survival rate can only be estimated using “open” capture-recapture models, i.e. those that allow additions and deletions from the population between sampling periods (Pollock, 1990; Kendall et al., 1997). In contrast, a “closed” model is one which assumes that no additions or deletions occur between sampling periods, and hence the population size remains constant (Pollock, 1990; Kendall et al., 1997). Most open capture- recapture models assume that all individuals alive in the population are available to be captured during each sampling session (Kendall et al., 1997). As the site fidelity and migration behaviour of sevengill sharks is poorly understood, this assumption may not be reasonable. Therefore, it may be more appropriate to use a model which allows temporary emigration from the study area. One such model is the Robust Design (Pollock 1990). The Robust Design is a combination of open and closed capture-recapture models, utilising strengths from both. The sampling structure of the Robust Design is nested, consisting of primary sampling periods each containing several secondary sample periods (Fig. 2.6). The Robust Design requires a minimum of three primary samples, each with at least two secondary sample periods. There is no upper limit on the number of primary and secondary samples. During each primary period, the target population is assumed to be closed, with no deaths, births, immigration, or emigration (Kendall et al., 1997). Abundance estimates can therefore be generated for each primary sample period using closed capture models (Huggins, 1989; Pollock, 1990). These models require multiple capture sessions, which is why primary periods are divided into multiple secondary periods. Between primary periods the population is assumed to be demographically and geographically open, allowing births and deaths, with individuals moving in and out of the study area. This means estimates of apparent survival (j), and emigration rates (g) can be made using open population models (e.g. Cormack-Jolly-Seber; Pollock, 1990; Kendall et al., 1997). This nested design allows for flexibility in sampling strategy, as well as accommodating fieldwork which often occurs in short sessions (e.g. one or two weeks within a season). Shorter secondary sample periods are also more likely to meet the assumption of a closed population.

Figure 2.6. Diagram of the Robust Design sampling structure (adapted from Pollock, 1982). Each primary sample is broken up into multiple secondary samples. The population is assumed to be demographically closed within primary samples, but open between them.

To employ the Robust Design in the current study, data were organised seasonally from winter 2016 to spring 2017 resulting in six primary periods consisting of up to 25 deployments per period. Each primary period was divided into two secondary capture periods of approximately equal duration. To ensure that the secondary periods were independent, the last deployment of the first secondary period and the first deployment of the second period could not occur on the same day. Encounter histories were transcribed into binary presence or absence (1 or 0) strings for each individual shark. For example, if an individual was seen in winter 2016 during both secondary capture periods but was then absent for the other five seasons of the study, it would have an encounter history of “110000000000”. These capture histories were saved as text files which were analysed using the Huggins Robust Design framework in Program MARK (v. 6.2, White & Burnham, 1999). The Huggins Robust Design closed model derives abundance from capture probabilities, removing abundance as a parameter from the remaining model calculations (Huggins, 1989; Siqueira et al., 2017). The Huggins model has been widely used in photo-ID capture-recapture studies of marine megafauna (e.g. Bradford et al., 2006; Holmberg et al., 2009; Santostasi et al., 2016; Gimenez, 2016; Siqueira et al., 2017).

A suite of models was fitted to the data using the parameters apparent annual survival rate (j), temporary emigration rate (g), and capture/recapture probabilities (p & c). Apparent survival rate was kept constant (j(.)) in all models as the data were too sparse to investigate time-

23 varying survival. It should be noted that apparent survival is calculated as a function of mortality and permanent emigration, (i.e. the models cannot distinguish between true mortality and permanent emigration). In program MARK, temporary emigration is denoted by g” which is the probability of emigrating from the study area between primary periods, or g’ which is the probability of staying outside the study area once emigrated. All models which allowed temporary emigration employed random temporary emigration, i.e. where g’=g”. This means that the probability of an individual being present in the study area in a particular primary period is independent of whether or not it was present in the previous primary period.

The rate of random emigration was kept constant between all primary periods (g’=g”(.)) or allowed to take a different value for each season (g’=g”(Season)). No movement models (g’=g”(0)), in which individuals were forced to be present during each primary period, were also constructed to investigate the applicability of models which did not allow temporary emigration. Capture and recapture probabilities were kept equal (p=c) for reasons of parsimony, and because there was no reason to expect a behavioural effect of capture. In other words, it was assumed to be unlikely that a shark which has been videoed once was more or less likely to be videoed again due to trap shy or trap happy responses (i.e. being positively or negatively reinforced to return or stay away from a trap; Hammond, 1986; Kendall & Nichols 1995; Kendall, 1999; Bailey et al., 2004; Carter et al., 2012; Urian et al., 2015; Huggins et al., 2018). The probability of capturing an individual was kept constant (p=c(.)), allowed to vary among primary periods (p=c(T)), or allowed to vary among primary and secondary periods (p=c(Tt)).

The various combinations of the parameters described above yielded nine possible models in the model set. The models were ranked using Akaike’s Information Criterion (AICc) scores, with the model having the lowest score deemed the best model (Akaike, 1974; Symonds & Moussalli, 2011). To account for uncertainty in model selection, model averaging was employed using the Akaike weight of each model (Burnham & Anderson, 1998). The top three models were averaged for generating estimates of abundance, survival rate and capture probability. Log-normal confidence intervals were generated for the model averaged abundances (for calculations see Buckland et al., 1993, p. 118).

2.3.9 Robust Design assumptions According to Pollock (1990) the assumptions of the Robust Design are a combination of closed and Jolly-Seber methods (Kendall, 1990): - Marks cannot be lost. - Marks must be correctly read. - The population is closed to deletions and additions during primary periods. - Survival rate is the same for all individuals in the population, regardless of availability for capture. - Capture probability is the same for all individuals in the population.

To maximise confidence in the results, violations of assumptions were minimised where possible. In this study, mark loss was not a concern as natural marks with long-term stability were used for identification (as evidenced by Housiaux, 2016). As there was no physical handling of the sharks or tag attachment there was unlikely to be an impact from the study that would change survivability between sampling occasions. Image quality criteria were imposed on the images to minimise the possibility of misidentification (see section 2.3.6). Individuals with less distinct markings would have a higher possibility of not being recognised if they were re-sighted later on in low quality images. This creates a source of heterogeneity in capture probability. Applying image quality criteria reduces this heterogeneity in capture probability by ensuring that even animals with less distinct spot patterns would be individually recognisable (e.g. James et al., 2015; Gore et al., 2016). The use of bait may add a source of heterogeneity in capture probability. While the Robust Design is relatively robust to heterogeneity of capture probability (Kendall et al., 1997), if the sharks become trap happy, this could still bias the results (Seber, 1986; Urian et al., 2015). Bait could provide a positive association with the BRUV causing them to be more willing to approach it in future (Johnson & Kock, 2006). As stated in the methods, however, the sharks did not actually get to eat the bait, meaning there was no reward, which should act to minimise trap happy effects. Conversely, long term exposure to chum with no reward may stop individuals from responding to it all together (i.e. trap-shyness; Hammond, 1986; Laroche et al., 2007). The BRUV itself is unlikely to cause any trap shyness as there is no harm or observable stress caused to individuals who inspect it. Nonetheless, to investigate whether

25 conditioning was occurring, the mean number of unique individual sharks seen per hour of deployment in each season was calculated. A linear model was fitted to investigate whether there was a trend in number of encounters per hour of video over successive sampling days during each season. A significant increasing or decreasing trend would indicate a conditioning effect. Finally, the assumption of geographic closure is one that is rarely met in studies of mobile species (Lettink & Armstrong, 2003) such as sharks, as they can freely move to and from the study area. There are three situations that can occur resulting in deletions and additions without significantly biasing the resulting abundance estimates (Kendall, 1999): - If movement in and out of the study area is random. - If the entire population is present at the first secondary capture session within a period, but some individuals emigrate before the last session, and detection histories are pooled for all sessions. - If sampling begins before all animals in the population have arrived but they are all present in the last session, and all sessions are pooled.

In the present study the chance of additions or deletions to the population was minimised by having short time spans for each secondary period, effectively shrinking the window for emigration, immigration, deaths and births to occur. Furthermore, there is no reason to expect that movements in and out of the study area, should they occur, were anything other than random.

26 2.4 Results A total of 133 deployments were conducted on 47 separate sampling days between 19/6/16 and 6/10/17. Although the intention was to obtain equal effort in each season, due to variable weather conditions there was subtle variation in the duration of video recorded per sampling trip (Table 2.2).

Table 2.2. Time span and total video hours for all sampled seasons in Sawdust Bay, Stewart Island. Primary Time span Total video (hours) period (days)

Winter 2016 9 23 Spring 2016 9 22 Summer 2017 11 19 Autumn 2017 20 20 Winter 2017 9 25 Spring 2017 12 24

2.4.1 Sighting rates, site fidelity and conditioning Sevengill sharks were encountered on all but one deployment. The number of individuals encountered per deployment ranged from 0 - 15 (mean = 4.95, SE = 0.45). There was a seasonal pattern in number of individuals encountered per deployment, with the maximum in summer 2017 (mean = 8.50, SE = 1.29) and the minimum in winter 2017 (mean = 2.69, SE = 0.43; Fig. 2.7).

27 12

2 Mean no. of individuals deployment per

0 Winter 16 Spring 16 Summer 17 Autumn 17 Winter 17 Spring 17 Season

Figure 2.7. The mean number of individual sevengill sharks encountered per deployment in each primary period (±standard error) from winter 2016 to spring 2017 in Sawdust Bay, Stewart Island.

There was no evidence of a consistent positive or negative conditioned response to the presence of the BRUV. A regression analysis of the number of individuals encountered over consecutive deployments in each primary period yielded only one significant relationship, in spring 2016 (Table 2.3). Overall there was little evidence for a consistent positive or negative trend for encountering individuals over consecutive deployments which would suggest conditioning.

28 Table 2.3. Regression analysis of individual sevengill sharks encountered over consecutive deployments each season. The direction of the trend (positive or negative), R2, F and p values are reported. N.B: * indicates a significant result at alpha = 0.05.

2 Season Direction R F-value p-value

(+ or -) value (degrees of freedom)

Winter 2016 + 0.230 3.583 1,12 0.083

Spring 2016 + 0.383 6.203 1,10 0.032*

Summer 2017 - 0.038 0.320 1,8 0.587

Autumn 2017 - 0.221 2.554 1,9 0.144

Winter 2017 + 0.003 0.031 1,11 0.86

Spring 2017 - 0.082 0.897 1,10 0.366

During the 132 deployments in which sevengill sharks were detected, there was a total of 349 encounters with sharks, comprising 149 different individuals. Of these different individuals, 80 were only encountered once, 54 individuals were encountered between two and four times, and 15 individuals were encountered more than four times (mean = 2.3, SE = 0.19, max = 15; Fig 2.8).

Frequency 30

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Number of encounters

Figure 2.8. Frequency distribution for encounter frequency of individual sevengill sharks (n = 149), during 133 BRUV deployments from winter 2016 to spring 2017 in Sawdust Bay, Stewart Island.

29 The majority of individual sharks were encountered in only one primary period (Fig. 2.9). However, 17 individuals were encountered in three or more primary periods. No individuals were seen in all primary periods.

100

Frequency 40

0 0.16 0.33 0.5 0.66 0.83 1 Proportion of primary periods

Figure 2.9 Frequency distribution for proportion of primary periods in which individual sevengill sharks (n =149) were encountered during 133 BRUV deployments from winter 2016 to spring 2017 in Sawdust Bay, Stewart Island.

A discovery curve of individuals over the entire study shows that new individuals were continuously being identified in the study area, and hence immigration and/or births were occurring (Fig. 2.10).

30 160

140

120

100

60 Known individuals Known 40

0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 Deployment days

Figure 2.10. Discovery curve of individual sevengill sharks across all sampling days (n = 47) in Sawdust Bay, Stewart Island.

During the 12 deployments at secondary sites (see Fig. 2.1) only one sevengill shark was encountered; at Trumpeter Point. During this time sevengill sharks were frequently encountered in Sawdust Bay (mean = 3.25 sharks per deployment, SE = 0.82). While not quantified, other fish identified from BRUV footage included rig (Mustelus lenticulatus), blue cod (Parapercis colias), spottie (Notolabrus celidotus), and rough skate (Dipturus nasutus).

2.4.2 Image quality results Encounters with sevengill sharks on average resulted in grade 1 images 86% of the time, and grade 2 images 10% of the time. Grades 3 and 4 images were each captured 2% of the time but were not used for the capture-recapture analysis. There was no evidence for a seasonal pattern in image quality (table 2.4). The proportion of rejected images varied only slightly among seasons, ranging from 0% in autumn 2017 to 5% in winter 2017. The 149 images in the photo catalogue (one for each known individual) comprised of 128 grade 1 images (85%), and 21 grade 2 images (15%).

31 Table 2.4. Proportion of image grades for all encounters with sevengill sharks by season. Total Season encounters Grade 1 % Grade 2 % Grade 3 % Grade 4 % Winter 2016 57 84 12 4 0 Spring 2016 68 87 10 0 3 Summer 2017 91 76 21 2 1 Autumn 2017 41 93 7 0 0 Winter 2017 41 93 2 5 0 Spring 2017 51 86 10 2 2

2.4.3 Capture-recapture analysis The model with the lowest AICc score, and therefore the top-ranking model of those trialled, featured constant survival rate, random temporary emigration varying by season, and constant capture probability (j(.)g’=g”(Season)p=c(.); Table 2.5). This model had an Akaike weight of over 50%. The only difference in the second ranking model (Akaike weight = 27%) was that it allowed capture probabilities to vary between primary periods (p=c(T)). The third ranked model was the only other model that had an Akaike weight of more than 5%. The major difference between this model and the top two was that the rate of temporary emigration did not vary between primary periods (g’=g”(.)). The models that allowed for different capture probabilities between secondary periods (p=c (Tt)) had little support, each accounting for less than 0.04% of model weights. The best no movement model (j(.)g’=g”(0)p=c(T)) was ranked 6th with a weight of 1%, showing there was little support for models which did not allow temporary emigration.

32 Table 2.5. Ranking of Huggins Robust Design models to estimate survival probability (j), capture probability (p) and temporary emigration probability (γ) for sevengill sharks in Sawdust Bay, Stewart Island. Notation: (.) constant; (T) capture probability rate varies between primary periods; (Tt) capture probability varies between capture sessions within primary periods; (p=c) no recapture effect; (g’=g”) random emigration model; (Season) different rate of emigration between each season; (g’=g”(0)) no movement model. No. AIC Rank Model AICc dAICc Parameters Weight 1 j(.)g’=g”(Season)p=c(.) 6 754.800 0.000 0.508 2 j(.)g’=g”(Season)p=c(T) 11 756.032 1.239 0.274 3 j(.)g’=g”(.)p=c(T) 15 757.433 2.632 0.136 4 j(.)g’=g”(Season)p=c(Tt) 17 759.575 5.129 0.039 5 j(.)g’=g”(.)p=c(Tt) 14 760.770 5.969 0.026 6 j(.)g’=g”(0)p=c(T) 7 761.998 7.197 0.013 7 j(.)g’=g”(0)p=c(Tt) 13 765.578 10.777 0.002 8 j(.)g’=g”(.)p=c(.) 3 767.903 13.103 0.001 9 j(.)g’=g”(0)p=c(.) 2 786.665 31.864 0.000

In order to make the most robust parameter estimates, model averaging was employed on the top three models in the set (Table 2.6). Apparent annual survival rate (j) was estimated at 0.51, however, the precision of this estimate was low as evidenced by the wide confidence intervals (95% CI = 0.16 - 0.86). The probability of being absent from the study area (g) ranged from 0.33 to 0.75 depending on the season. There was a low probability of an individual being absent in either spring or summer (33% - 41%). In autumn and winter however, there was a high chance of absence (72% - 75%), indicating increased rates of emigration prior to these seasons. Capture probability across the whole study ranged from 0.26 - 0.40 (Table 2.6).

33 Table 2.6. Model averaged parameter estimates, standard errors (SE) and 95% confidence intervals for apparent annual survival probability (j), capture probability (p) and temporary emigration probability (γ) for sevengill sharks in Sawdust Bay, Stewart Island, for the top three ranked models according to AICc. N.B: Temporary emigration probabilities (g) are the probability of an individual being absent from the named season.

95% confidence intervals Parameter Estimate SE Lower Upper j 0.512 0.221 0.157 0.856 g (Spring) 0.331 0.228 0.061 0.789 g (Summer) 0.409 0.158 0.161 0.714 g (Autumn) 0.721 0.159 0.354 0.924 g (Winter) 0.752 0.157 0.369 0.940 p (Winter 2016) 0.399 0.123 0.196 0.644 p (Spring 2016) 0.339 0.065 0.225 0.475 p (Summer 2017) 0.334 0.061 0.227 0.461 p (Autumn 2017) 0.266 0.098 0.119 0.493 p (Winter 2017) 0.282 0.098 0.132 0.504 p (Spring 2017) 0.256 0.085 0.126 0.452

The model-averaged estimates of abundance revealed a seasonal signal of lower abundance in autumn and winter, compared with higher abundances in spring and summer (Table 2.7, Fig. 2.11). The minimum estimate was 34 (SE = 8.5; 95% CI = 21 - 55) in winter 2016, while the maximum was 94 (SE = 37.4; 95% CI = 44 - 199) in spring 2017. There were fewer recaptures during the secondary periods in 2017 compared to those in 2016. This resulted in lower precision of the abundance estimates, as shown by the wider confidence intervals.

34 Table 2.7. Model averaged abundance estimates of sevengill sharks in Sawdust Bay, Stewart Island with standard errors (SE) and log-normal 95% confidence intervals for all primary periods. Log-normal 95% Abundance confidence Primary period SE estimate intervals Lower Upper Winter 2016 34 8.5 21 55 Spring 2016 82 14.4 58 115 Summer 2017 92 15.9 66 129 Autumn 2017 55 30.0 20 148 Winter 2017 55 30.7 20 153 Spring 2017 94 37.4 44 199

200 180 160 140 120 normal CI)95% - 100 log - 80 60 40 20 Abundance (+/ 0 Winter 2016 Spring 2016 Summer 2017 Autumn 2017 Winter 2017 Spring 2017 Season

Figure 2.11. Time-series of model-averaged estimates of abundance (+/- log-normal 95% confidence intervals) for sevengill sharks in Sawdust Bay, Stewart Island.

35 2.5 Discussion In this chapter I used photo-ID data to generate the first estimates of demographic parameters for sevengill sharks in New Zealand. BRUVs were successfully employed to capture images of sharks in their demersal habitat. The natural pattern of spots on their dorsal surface was found to be an appropriate way of distinguishing among individuals and identifying them over time. Estimates of abundance and apparent annual survival rate were generated using Huggins Robust Design models.

2.5.1 Abundances and seasonality This study presents the first abundance estimates for broadnose sevengill sharks anywhere in New Zealand. I found that Sawdust Bay in Paterson Inlet was used by between 30 and 100 sevengill sharks per field season. A study in Norfolk Bay, Tasmania, revealed much higher abundances of sevengill sharks (n = 1000 - 1200; Dudgeon et al., 2015). Although the area of Paterson Inlet sampled in the current study is smaller than Norfolk Bay, both studies indicate that sevengill sharks aggregate in large numbers. Sawdust Bay was used by sevengill sharks throughout the year, although abundances fluctuated seasonally, being highest in summer and lowest in winter. Understanding the extent of the study area that these estimates apply to is challenging, as a bait plume was used to attract sharks to the BRUV. The area a bait plume can cover is dependent on the local conditions (e.g. tide or wind; Heagney et al., 2007; Taylor et al., 2013). At a minimum, the current study represents sampling the north-western section of Paterson Inlet. Given the abundances found, and that sevengill sharks have been shown to exert substantial predation pressure on prey in areas they inhabit (Cortes, 1999; Barnett et al., 2010; Barnett et al., 2011), they are likely to have a considerable impact on the Paterson Inlet ecosystem and food-web. A seasonal pattern of emigration was apparent, with high rates of temporary emigration prior to the cooler months of the year (autumn and winter) compared with spring and summer, when emigration was lower. Reinforcing this, “no movement” models (which allowed no temporary emigration) were ranked low via AICc (Table 2.5). There were some differences between the 2016 and 2017 spring seasons. Spring 2017 had half as many encounters per unit effort compared with the year before and fewer recaptures. The reasons for these differences are unknown but regardless, the abundances estimated by program MARK for the two spring seasons were similar and support the trend of lower emigration rates prior to spring. A seasonal pattern of temporary emigration helps explain the seasonal pattern of abundance and matches observations by Housiaux et al. (2018) from their study in Sawdust Bay. On a

36 wider scale it fits seasonal movement patterns described from studies of sevengill sharks in Tasmania (Barnett et al., 2011; Stehfest et al., 2014), the Western Cape of South Africa (Ebert, 1989; Ebert, 1996), and Washington State in North America (Williams et al., 2011) which all found a consistent seasonal pattern of movement away from coastal aggregation sites during winter. These patterns have varied explanations in the literature. While water temperature appears to be a key driving force behind seasonal movements of sevengill sharks (Housiaux et al., 2018), it is likely that it influences them indirectly by impacting prey abundances (e.g. rig shark; Francis & Mace, 1980; King, 1984; Ebert, 1986; Barnett et al., 2011). Intraspecific resource partitioning has previously been hypothesised to play a role in seasonal movements of sevengill sharks, with populations becoming more dispersed in cooler months when prey abundances are low (Barnett et al, 2011). Studies in other locations have shown that sevengill sharks are capable of foraging on a wide range of prey species (see section 1.2). Major prey categories include teleost fish, chondrichthyans and pinnipeds (Ebert, 1991; Barnett et al., 2010a,b). While analysing prey abundance was not a part of this study, the footage at least confirmed blue cod, rough skate and rig shark as being commonly present in Sawdust Bay, all of which could represent potential prey species for sevengill sharks. Previous literature has highlighted the significance of chondrichthyan prey in sevengill shark diets (Ebert, 1986; 1991a; 2002). Sevengill sharks in Tasmania have been shown to target gummy sharks (Mustelus antarcticus) which are closely related to rig shark (Mustelus lenticulatus; Barnett et al 2010a,b). Rig shark abundances also show an increase in shallow bays and estuaries in spring and summer, for the purposes of parturition (King, 1984; Massey & Francis, 1989). This pattern overlaps with increased abundances in sevengill sharks. Rig shark then, seem likely to be under predation pressure by sevengill sharks in the study area. Aside from prey, the reproductive cycle of sevengill sharks likely plays a role in seasonal movements (see section 1.2). Reproduction affects the seasonal movements of sevengill sharks in California (Ebert, 1986; 1989), Washington (Williams et al., 2012) and Patagonia (Lucifora et al., 2005). Seasonal use of embayments in these areas was likely for the purpose of mating and pupping. Neonates and gravid females were reported in these study areas during spring and summer seasons (Ebert, 1989; Lucifora et al., 2005). The BRUV technique used in this study did not allow confirmation of whether pregnant or gravid females were present. Other techniques, such as ultrasound scans (e.g. Sulikowski et al., 2016), would be required to further investigate the influence of reproductive status on abundance and distribution. This would

37 however require capturing and handling sharks which could cause spontaneous abortion of the pups (Adams et al., 2018).

2.5.2 Apparent annual survival rate The model-averaged rate of apparent annual survival (j) over the time of the study was estimated at 0.51. This estimate is low for a species reported to live >30 years of age (Fowler et al., 2005), and other large sharks such as great white sharks (40 - 73yrs; Hamady et al, 2014), porbeagle sharks (65yrs; Francis et al., 2007) and sandbar sharks (21yrs; Joung et al, 2004) have similarly long life spans. It could be that survival really is this low due to high rates of predation. For example, around Stewart Island there is a semi-resident population of white sharks (Francis et al. 2015), a species which is a known predator of sevengill sharks (Reyes & Garcia-Borboroglu, 2004). However, apparent survival is inherently confounded. It is calculated as a function of mortality and permanent emigration and is further confounded with capture probability (Cooch & White, 2006). The combination of imperfect capture probability and temporary emigration over the relatively short duration of the present study are more likely causes of why the apparent survival estimate is low. It should be noted that the uncertainty around the estimate of survival rate was very high (95% CI = 0.16 - 0.86). The true value is likely to lie in the upper part of this range. With an increased study duration, the precision of the estimate should improve (Kendall, 1999). Likewise increasing the size of the study area could also solve this. Dudgeon et al. (2015) for example, had previously reported much higher survival rates (j = 0.75 - 0.90) for sevengill sharks around Tasmania and the east coast of Australia. In their case however, acoustic telemetry was used to disentangle emigration and mortality (effectively removing the issue of study area). Even so, the estimate of apparent survival in the present study was the first generated for sevengill sharks anywhere in New Zealand, and acts as a useful starting point for future studies.

2.5.3 Residency and site fidelity Of the 149 individual sharks, 104 were seen in only one primary sampling period indicating either low rates of individual residency, or large home ranges. This reinforces the previous conclusion that there is immigration and emigration occurring in and out of the study area. The shape of the discovery curve, which did not plateau, further supports the idea that new individuals were moving into Sawdust Bay over the duration of the study. Sevengill sharks are thought to use certain areas for specific purposes (e.g. breeding, pupping, foraging) with

38 individuals moving between sites as required (Ebert, 1986; Larson et al., 2015). Sawdust Bay might provide a seasonal foraging area in this case. There is, however, the possibility that some individuals were present in the study area over consecutive primary periods but did not approach the BRUV, or if they did, suitable photo-ID images were not obtained. Alternatively, home ranges of sevengill sharks may be considerably larger than the study area and a particular individual may only spend a subset of their time in Sawdust Bay while remaining resident in Paterson Inlet. Ebert (1996) reported that 11 sevengill sharks tagged in Namibia moved less than 54 km from their tagging sites during up to 1615 days at liberty. While 54 km is not a large distance to cover for a large mobile marine predator over those time spans, the distance from the western shore of Paterson Inlet to the entrance is only approximately 15 km. It is possible that a sevengill shark would make use of multiple areas of preferable habitat by moving among them within Paterson Inlet. Future sampling that encompasses more of the inlet may elucidate residency patterns of individual sharks. Some evidence of annual site fidelity was observed, with 13 individuals encountered in winter or spring in both 2016 and 2017, but not encountered in summer or autumn (i.e. it appears they had emigrated out of the study area and returned). Due to the study period lasting only a year and a half, it was not possible to confirm site fidelity in consecutive summers. Previous studies with longer sampling periods (e.g. Barnett, 2011; Williams et al., 2011; 2012) have found that sevengill sharks display seasonal migrations, but exhibit long term site fidelity by returning from year to year (Larson et al., 2015). It is therefore plausible that individual sharks return to Sawdust Bay over several years. Continued long term sampling is required to increase confidence in the observed site fidelity of sevengill sharks to Sawdust Bay. Although no individuals were encountered in every season sampled, 17 sharks were encountered in three of the six seasons, providing some evidence for residence by at least a portion of the population. Sampling outside Sawdust Bay indicated that sevengill sharks were not uniformly distributed throughout Paterson Inlet and that fine-scale habitat selection appears to be taking place. This has previously been suggested for populations in Tasmania with specific individuals remaining in particular areas of the inlet they inhabited (Barnett et al., 2012). In this study however, inlet- wide sampling was limited to 12 deployments over two weeks in spring 2017. More intensive sampling would be required to help resolve the preferred habitat in Paterson Inlet.

39 2.5.4 Limitations When interpreting the results of this study it is important to consider limitations and potential biases. One limitation which should be acknowledged in this study is that the range of the bait plume probably varied among deployments, thus affecting the area over which sevengill sharks were sampled. I attempted to minimise this by standardising the bait, and the duration of each deployment. However, depending on conditions (e.g. wind speed or tidal state), the distance a bait plume can travel may differ by orders of magnitude (Heagney et al., 2007; Taylor et al., 2013). An alternative would be to use un-baited remote videos, which would eliminate this problem but presumably significantly reduce encounter rate. The use of un-baited remote underwater videos has been shown to decrease the precision of count data, and baited methods are recommended (Bernard & Götz, 2012). The other potential disadvantage of a baited sampling method is positive or negative conditioning of the sharks. Results showed little evidence to support positive reinforcement as trends in sightings per deployment were inconsistent over the study period. While there has been no targeted research on sevengill sharks becoming conditioned, studies on other large sharks (>2m) which used similar baiting methods indicated that trap happy responses are unlikely. For example, great white sharks continued to approach unobtainable bait but spent less time around it depending on the number of previous encounters (Johnson & Kock, 2006; Laroche et al., 2007). Furthermore, Laroche et al. (2007) suggested that the use of chum alone (without feeding) is not enough of a reward to significantly alter the behaviour of great whites. Tiger sharks also appeared to show no behavioural change in response to eco-tourism that uses chum (Hammerschlag et al., 2012). These studies used eco-tourism situations, which offer bait at a far higher frequency than during the current study. Trap shyness is also a potential bias, especially in shark species where larger individuals can perturb smaller ones (Jacoby et al., 2012). For example, great whites and tiger sharks sometimes appear to be uncomfortable in the company of conspecifics, particularly larger ones (Dudley et al., 2000). This does not appear to be true of broadnose sevengill sharks however (Ebert, 1991). Despite records of cannibalism, aggregations of sevengill sharks of mixed sizes have been observed in South Africa, South America, and Tasmania (Ebert, 1996; Lucifora et al., 2005; Barnett et al., 2010). There were several occurrences in this study of multiple individuals of varying sizes in the same frame (see Chapter 3 Discussion). This suggests that avoidance of the BRUV due to larger individuals being present was not a consistent problem. Finally, while the variation in the duration of video among primary periods was subtle, the time

40 span of the autumn 2017 primary period was double that of the other primary periods (due to weather conditions). A longer time period allows a greater opportunity for sharks to be sampled, especially if they are moving in and out of the study area. Therefore, the autumn 2017 abundance estimate is potentially biased high compared to the other seasonal estimates.

2.5.5 Conclusion This study presents the first estimates of abundance and survival rates for sevengill sharks in New Zealand. The appropriateness of the Robust Design to estimate population parameters was confirmed as shown by the lack of support for “no movement” models and relatively high capture probabilities. The use of natural marks on sevengill sharks proved effective for photo- ID and will continue to be utilised for future sampling. The results of this study mirror seasonal patterns observed in Tasmania (e.g. Barnett, 2010; Dudgeon et al., 2015) and North America (Ebert, 1989). Sevengill sharks exhibited a seasonal pattern of emigration with high abundances in summer and fewer sharks in winter. The drivers of this are most likely linked to intraspecific resource partitioning, reproduction or both. Encounters per deployment varied through time but gave a clear indication of relative abundances for Sawdust Bay, reinforcing the seasonal capture-recapture estimates. Lastly there was some evidence of annual site fidelity. This study provides the starting point for a long-term data set characterising parameters of sevengill shark populations in southern New Zealand, which in future will allow for monitoring of population stability and health.

41 Chapter 3 – Describing size and sex structure of the sevengill shark population in Sawdust Bay, Stewart Island, using stereo-photogrammetry.

3.1 Introduction A thorough understanding of a species’ population demographics is required for sound conservation and management (Lande, 1988; Harry, 2017). In addition to knowing the abundance of a population, basic information on age and sex structure is also important, and can drastically change how a population should be managed (Natanson et al., 2018). Age and sex parameters can be used to estimate reproductive potential and therefore current and future population growth, and extinction risk (Lebreton et al., 1992). The age and sex structure of a population is also crucial to understanding life history and how the population interacts with its environment (Lucifora et al., 2005; Natanson et al., 2018). Over different life stages an animal may substantially change its diet, trophic level, and home range, all of which are important considerations in conservation management (Simpfendorfer et al., 2011). Sharks are notoriously difficult to study. When physically captured (e.g. by long line or trawl; Santana-Garcon et al., 2014) sex can be directly confirmed, and biometric information recorded. However, being caught and restrained can have negative effects on the shark (Skomal & Mandelman, 2012). While sexing sharks and other elasmobranchs is relatively easy thanks to the presence of claspers, aging them is more challenging because they lack the calcified structures (e.g. otoliths) used to age teleost fish (Goldman, 2005; Cailliet, 2015). Aging techniques using other structures such as vertebrae, spines or caudal thorns (e.g. Soldat, 1982; Cailliet, 1990; Carlson & Goldman, 2006; Taylor et al., 2013) have been developed, but their reliability varies among species and is influenced by environmental factors (e.g. food availability, Mohan et al., 2004; Cailliet, 2015; Natanson et al., 2018). Additionally, the relationships between growth of these structures and actual age have been validated in relatively few species of sharks (Goldman, 2005; Cailliet, 2015). Many elasmobranch species appear to lack structures that can be used for quantifying age (Braccini et al., 2010; Cailliet, 2015). Broadnose sevengill sharks (and the family Hexanchidae) for example, have poorly calcified vertebrae (Ebert, 1986), and no spines that can be used for estimating age (Braccini et al., 2010). A method using growth bands in the

42 neural arches of bluntnose sixgill sharks (Hexanchus griseus) has been developed, but the banding pattern was only observed on the outer portion of the arches and was inconsistent along the length of the neural arch (McFarlane et al. 2002). Further research is required to demonstrate the reliability of this technique on bluntnose sixgill sharks, and investigate whether it can be applied to other members of the Hexanchid family (Braccini et al., 2010). When age data are unavailable or difficult to obtain, size is frequently used as a proxy (e.g. sandbar shark, Carcharhinus plumbeus, Casey & Natanson, 1992; and Atlantic tope shark, Galeorhinus galeus, Dureuil & Worm, 2015). This requires substantial data on size and growth of captive or recaptured individuals, and accuracy can vary widely due to individual variability in growth (Dureuil & Worm, 2015). Despite not being able to estimate age from size directly, it is at least possible to relate the total length (TL) of sevengill sharks to different life-history stages (Ebert, 1989; Dykhuizen & Mollet, 1992; Braccini et al., 2010). Using previous literature, it is possible to place individual sevengill sharks into functional cohorts (e.g. neonate, juvenile, adult) if TL is known. In Chapter 2, I described obtaining photo-ID capture-recapture data for estimation of population size. Applying photogrammetric methods to the same footage provides length measurements that can be used to classify life stages of individual sharks without the need for physical capture. Photogrammetry is the process of deriving measurement information about objects in photos or videos. Stereo-photogrammetry is the use of two cameras recording images with overlapping fields of view (FOV) to capture the same object at slightly different angles. The separation of the cameras and degree of image overlap allow three dimensional measurement of objects captured in both fields of view (Fig. 3.1; Klimley & Brown, 1983). Stereo-photogrammetry was first applied to sharks by Klimley & Brown (1983) using film based methods with scalloped hammerhead sharks (Sphyrna lewini). The transition from film to digital photography has dramatically improved its ease of use, principally because the physical photogrammetric model usually required in analysis of images on film has been replaced by a computer-based digital model (e.g. Growcott et al., 2012). Harvey & Shortis (1995) were among the first to apply digital stereo-photogrammetry in an underwater setting and emphasised the need for a strong calibration methodology to ensure accuracy and precision of extracted measurements. In recent years photogrammetric methods have become relatively common in marine research, in large part due to the low cost and relatively high quality of imagery from GoPro and other action cameras (Marshall & Pierce, 2012; Raoult et al., 2016).

43 Figure 3.1. Diagram of stereo-photogrammetry design with overlapping fields of view (FOV). When an object enters the overlapping zone, accurate measurements can be calculated.

For obtaining measurement data, photogrammetry has many advantages over fishery or diver based methods, including being non-invasive, cost-effective and producing a permanent data record that can be reanalysed later (Marshall & Pierce, 2012; Letessier et al., 2013). Stereo- photographs can capture multiple individuals simultaneously, and record natural interspecific and intraspecific behaviours which may not be seen otherwise (Letessier et al., 2013; Santana- Garcon et al., 2014). There is far less subjectivity in photogrammetry compared with size estimates made by divers and reduced behavioural modification of animals that might be wary of a diver (although a camera rig may still have some effect; Marshall & Pierce, 2012). Lastly, still-frames can be extracted from videos for individual identification (see Chapter 2).

3.2 Aims To date there is very little known about demographic structure of sevengill shark populations in New Zealand. Therefore, the overall aim of this chapter is to make the first attempt at describing size and sex structure of a population of sevengill sharks. To accomplish this, I developed a stereo-photogrammetry system capable of measuring sharks and identifying their sex. As photogrammetry methods are novel for this species, I tested the practicality of the system, by investigating the accuracy and precision of measurements, and the reliability of the hardware and methodology. Then, using stereo-photogrammetric methods, I described the

44 seasonal sex and size structure of the sevengill shark population in Sawdust Bay, Paterson Inlet, and investigated whether changes occurred over time. With an understanding of the seasonal sex and size structure, I identified potential reasons as to why sevengill sharks congregate in Sawdust Bay. Lastly, I explored whether the stereo-photogrammetry system was capable of measuring growth in sevengill sharks using measurements of the same individuals over subsequent seasons.

3.3 Methods The study site was Sawdust Bay, located at the entrance to the North Arm of Paterson Inlet, Stewart Island (46.9012°S, 168.0127°E). Data were collected between 19th June 2016 and 6th November 2017, during six field expeditions ranging from two weeks to a month in duration. For detailed information on the study site, equipment and field methods used please refer to the chapter 2 methods section.

3.3.1 Stereo-photogrammetry The camera system used to collect footage was as described in Chapter 2. A flashing LED light was filmed for use as a “clapper” for synchronising videos for analysis. Vidsync (www.vidsync.org) is freeware developed for Macintosh operating systems specifically for scientific stereo-photogrammetry analysis (Neuswanger et al., 2016). The software relies on a calibration process in order to calculate differences between the cameras’ fields of view (Neuswanger et al., 2016). Therefore, in addition to the methods described in chapter 2, the stereo-photogrammetry required filming a calibration frame to quantify and correct for image distortion. A custom-built calibration frame was produced, consisting of two 150 cm x 50 cm grid panels (10 cm x 10 cm grid) separated by 40 cm depth and held in place with PVC tube (Fig. 3.2). Different coloured markers were used to identify the front and rear grid intersections during analysis.

Figure 3.2. Stereo-camera calibration frame for use in image analysis. Two 150 cm x 50 cm grids (10 cm x 10 cm grid) are separated by 40 cm PVC pillars. Grid intersections are colour marked to distinguish the front grid from the rear (yellow on the front grid, and red on the rear grid).

Prior to deployment of the BRUV system in the field, the calibration frame was deployed over the side of the vessel to 1.5 m water depth. Cameras and lights were turned on and recording initiated. The calibration frame was filmed underwater for one minute, at a distance of 1.2 m. The camera rig was raised and lowered to ensure both cameras filmed the whole frame. The BRUV was then deployed to within 2 m of the seafloor and secured to the vessel. This procedure was carried out for each deployment.

3.3.2 Calibration process Before analysis, the two videos from the left and right cameras had to be synchronised to match frames. This was accomplished using the LED ‘clapper’ filmed at the start of all videos. Once synchronised, two calibration processes were required before measurements of a shark could be taken. The first was distortion correction of the camera lenses, using plumb-lines generated along the straight edges of the calibration frame seen in both left and right video files. In this case I used the lines of the front plane of the calibration grid and created plumb-lines by selecting points along the frame (an automatic process also exists in Vidsync, but it had trouble identifying the edges of my frame). These points would automatically connect into straight lines and calculate distortion correction.

46 The second calibration process established the scale of the images, using the front and rear planes of the calibration frame. Vidsync generates a co-ordinate system of x, y, and z (x = horizontal component of grid, y = vertical component of grid, z = depth) values of every intersection on the grid (Fig. 3.3). I selected each intersection of the calibration frame in a pre- determined order. As an example, the bottom left corner of the grid is listed as 0,0. The next intersection to the right is 0, 100 (to identify that it is 100mm from the origin), the next is 0, 200 and so on. I selected each intersection from left to right, starting on the bottom row and working upwards. This was repeated for the rear grid of the same frame. The whole process was then repeated for the other camera. Selection of the intersections was aided by a zoomed view of the frame and could be finely adjusted.

Figure 3.3. Example of the VidSync calibration process for the left camera. Blue circles represent an x, y, and z co-ordinate on the grid (mm) on the front plane of the frame. Orange circles represent the coordinates for the rear plane. An identical image for the right-side camera was also generated and linked to the coordinates of the left.

Once this was all completed, Vidsync then links all four sets of co-ordinates (left camera: front and rear, right camera: front and rear) and automatically calculates distance and angle differences between left and right cameras. As long as the cameras do not change position within the rig, the calibration will remain valid. Calibration was conducted at the beginning of each new video, or whenever the cameras had been removed from their housings.

47 Once the calibration process was completed, replicated test measurements of a standard 80 cm length of the calibration frame were taken and used to check calibration accuracy. This was conducted at the beginning of each video that was used for measuring sharks.

3.3.3 Measuring sevengill sharks and calculating total length Whilst reviewing footage for the purposes of measuring sevengill sharks it was found that due to the length of the top lobe of the caudal fin, it was rarely possible to capture the total length (TL) of a shark in frame. However, the pre-dorsal length (PDL; the distance from the tip of the rostrum to the base of the dorsal fin; Fig. 3.4), was frequently measurable.

Figure 3.4. Diagram showing the measurements of total length (TL) and pre-dorsal length (PDL) on a sevengill shark.

Dykhuizen et al. (1992) found a very close relationship between the PDL and TL in sevengill sharks from California. They measured 25 sharks with TL between 65 cm and 188 cm and found that it could be accurately predicted from the PDL (r2 = 0.996), according to the equation (where TL and PDL are measured in cm): (�� + 11.7) �� = 0.576

48 As there is limited published biometric data available on sevengill sharks, there were no other suitable measurements that could be used to confidently calculate TL. When TL was able to be measured in VidSync, the PDL was also measured and tested using the relationship described by Dykhuizen et al. (1992). The formula was also tested on a specimen washed up in Hoopers Inlet, Dunedin in 2016 (Fig. 3.5).

Figure 3.5. Female sevengill shark (TL = 261 cm) that washed up in Hoopers Inlet, Dunedin, during January 2017.

In the process of identifying individuals for the capture-recapture analysis described in Chapter 2, all instances of when sharks presented opportunities for being measured were recorded for easy re-location in VidSync. In order to measure the PDL of the sharks, it was necessary to identify and select the tip of the rostrum and the origin (anterior point) of the dorsal fin in the two synchronised images from each camera. I attempted to be as accurate as possible selecting the same points, although VidSync did present a guideline and a zoomed view to assist with this. Sevengill sharks have a light patch of skin at the anterior origin of their dorsal fin, which aided in accurately selecting the pre-dorsal point each time (Fig. 3.6).

Figure 3.6. An example of the light patch of skin that indicates the start of the dorsal fin (the origin) which was used during selecting the start or end of a measurement in Vidsync.

Due to the sinusoidal nature of shark swimming, measuring from the tip of the rostrum to the pre-dorsal point in a straight line (Fig. 3.7a) was not always possible. Vidsync can however, measure along the curve of the body in multiple segments (Fig. 3.7b). Sevengill sharks have a dark band that runs along the centre of the dorsal surface which can be used as a guide (Fig. 3.8). Increasing the number of segments leads to increased possibility of user error in selecting the same start and end points in the images from each camera. To minimise this, no more than four segments were used to measure PDL. If the position of the shark required more than this to match the curvature of its body, that frame was not used.

B. Figure 3.7. Examples of measuring sevengill sharks in Vidsync. A is an example of where a straight-line measurement was appropriate, while B displays how the program can cope with curves in the body of the shark by measuring in segments.

Figure 3.8. An image that displays the dark band that runs along the spine of sevengill sharks which was used as a guide when selecting measurement segments in Vidsync.

Five replicate PDL measurements were taken for each individual shark and averaged to increase accuracy and obtain an estimate of precision (Santana-Garcon et al., 2014). The five replicates could not always be obtained from the same encounter. In that case, the additional replicates were obtained from other encounters with the same shark, but always within the same season to minimise variation due to growth.

3.3.4 Growth To investigate the adequacy of the stereo-photogrammetry system for measuring growth of sevengill sharks over time, individuals that were recaptured over periods of more than 300 days were measured on multiple occasions.

3.3.5 Determining sex Sexing sevengill sharks relied on sighting the pelvic fins and looking for the presence of claspers (Fig. 3.9). In order to do this reliably it was necessary to watch the area around the pelvic fins closely as sharks swam through the field of view. Owing to the swimming motion

52 of sevengill sharks, the pelvic fins of females and claspers of mature males were often visible. There is potentially a risk that the claspers of young or juvenile male sevengill sharks would not have developed sufficiently to be obvious, and therefore these individuals could be misidentified as females. Ebert (1986) reported that males can be mature as small as 130 cm TL, and still juveniles as large as 150 cm TL (n = 85, size range from 44 – 248 cm). Clasper length in individuals larger than 150 cm was reported to increase abruptly and become obvious (Ebert, 1986). In the present study, I had to be confident that juvenile males were not misidentified as females. Therefore, if a shark was less than 160 cm TL, it was only classified as female if a full ventral view of the pelvic fins was obtained, or as a male if claspers were observed. Otherwise the shark was classified as “unknown sex”.

B, Figure 3.9. Captured images of female (A) and male (B) sevengill sharks with the absence or presence of claspers highlighted.

54 3.3.6 Determining life stage Sevengill shark neonates are between 40 cm and 80 cm in size (Ebert 1986; 1996). Male sexual maturity is considered to be at ~140 cm while female sexual maturity is at ~220 cm (Ebert 1986, 1996; Barnett et al., 2010). All sharks less than 80 cm TL were therefore classified as neonates. Males between 80 cm and 140 cm were classified as juvenile, and >140 cm as mature. Females between 80 cm and 220 cm were classified as juvenile, and >220 cm as mature.

3.3.7 Statistical methods The size distribution of sevengill sharks in Sawdust Bay was visualised using a histogram of estimated TL. To explore seasonal size distribution, histograms of all females in each season were plotted. Since males are generally smaller, and as they were comparatively fewer, they were omitted to prevent skewing the size distributions. Growth data were plotted to compare estimated TL before and after a minimum of 300 days at liberty. A log-likelihood ratio test was conducted to investigate whether sex ratio was significantly different between seasons. Statistical differences were considered significant at an alpha of 0.05.

3.4 Results

3.4.1 Sex structure Images that allowed for confident determination of sex were obtained for all but one individual (n=148). There were four sharks under 160 cm, three of which were confirmed to be male due to the presence of claspers. The fourth shark did not display a clear view of its pelvic fins and hence was the only individual that could not be confidently sexed in this study. The sample was skewed towards females, with males comprising only 17% of the observed population. Sex ratio varied significantly among seasons (G = 20.43, p = 0.001, DF = 5; Fig. 3.10). There was a noticeable increase in male presence during the summer and autumn periods, peaking in summer at 27% of the sampled individuals.

55 Female Male Unknown 100

90 80

50 40

proportion of population (%) of population proportion 20

0 Winter16 (21) Spring16 (46) Summer17 (59) Autumn17 (36) Winter17 (27) Spring17 (39) Season

Figure 3.10. Seasonal sex ratios of sevengill sharks in Sawdust Bay between winter 2016 and spring 2017. Sample size is indicated in brackets following each season.

3.4.2 Accuracy and precision of the stereo-photogrammetry system Repeated measurements (n = 30) of a reference length (80 cm) on the calibration frame from different videos (n = 15) showed high consistency and accuracy. There was a slight positive bias when measuring the 80 cm reference length (mean = 80.5 cm, range = 80.1 - 80.9 cm). The mean error was therefore +0.63%, and the maximum error was +1.13%. The coefficient of variation (CV = SD/mean) of repeated measurements of the reference length was 0.30%. Opportunities for comparison of the measured TL to the estimated TL (using the relationship with PDL) were rare (n = 4), but suggested a strong relationship (r2 = 0.997, Fig 3.11). A single instance of a washed up sevengill shark allowed direct measurement and comparison of the PDL to TL conversion, which was accurate to within 1cm (physically measured TL = 2.60 m, estimated TL = 2.61 m).

56 240 230 220 210 200 190 180 170 160 150 Calculated total length (cm) length total Calculated 140 140 160 180 200 220 240 Measured total length (cm)

Figure 3.11. Comparison of measured TL of sevengill sharks using stereo-photogrammetry against the calculated TL using the PDL conversion formula (n = 4). The dotted line shows a linear regression with R2 = 0.997.

3.4.3 Estimates of lengths of sevengill sharks Of the 149 individual sevengill sharks that were identified, 87 could be measured, representing between 47% and 90% of the sharks seen per season. The remaining individuals were not positioned sufficiently well for accurate measurement. Five replicate measurements of PDL were made each time a shark was measured. Averaged PDLs of individual sharks ranged from 72.1 - 147.5 cm, resulting in estimated TLs of 149 – 276 cm (mean TL= 211.6 cm; Table. 3.1). Measurements of individual sharks showed a mean CV (SD/mean) of 2.1% (range 0.27 - 10.80%).

57 Table 3.1. The number of individual sevengill sharks measured per season in Sawdust Bay, Stewart Island. Notation: PDL = pre-dorsal length; TL = extrapolated total length. Season Sex Individuals Mean PDL SE Minimum Maximum measured (TL) PDL (TL) PDL (TL) cm cm cm Winter 2016 Female 18 111.2 (213) 4.1 92.6 (181) 131.1 (248) Winter 2016 Male 1 118.7 (226) - 118.7 (226) 118.7 (226) Spring 2016 Female 29 112.8 (216) 3.6 93.4 (182) 138.3 (260) Spring 2016 Male 2 92.1 (180) 4.5 89.3 (175) 94.3 (184) Summer 2017 Female 33 107.3 (213) 4.0 72.1 (149) 140.6 (264) Summer 2017 Male 12 95.4 (186) 5.8 77.0 (154) 118.2 (225) Autumn 2017 Female 8 104.9 (209) 9.8 75.0 (151) 127.5 (242) Autumn 2017 Male 4 97.9 (190) 15.7 74.5 (150) 118.2 (225) Winter 2017 Female 12 116.1 (222) 8.2 94.7 (185) 147.5 (276) Winter 2017 Male 0 - - - - Spring 2017 Female 16 112.9 (216) 5.5 80.4 (160) 131.1 (248) Spring 2017 Male 0 - - - - Overall - 23.8 106.9 (206) 6.8 72.1 (149) 147.5 (276)

58 The distribution of individual TLs revealed that size classes between 180 cm and 240 cm predominated in Sawdust Bay, with the modal size class being 210 - 219 cm (Fig 3.12).

18 16 14 12 10 8

Frequency 6 4 2 0

140-149 150-159 160-169 170-179 180-189 190-199 200-209 210-219 220-229 230-239 240-249 250-259 260-269 270-279 Total length (cm)

Figure 3.12. Size distribution of sevengill sharks (n = 87) measured in Sawdust Bay, Stewart Island, between winter 2016 and spring 2017.

Size distribution of female sevengill sharks showed little variation between seasons with the 200 - 229 cm size bin having the highest frequency in all seasons (Fig.3.13). Males were excluded from seasonal comparisons of size distributions as there were relatively few of them. Males were on average smaller than females (mean female TL = 217 cm; mean male TL = 186 cm; Fig. 3.14).

Figure 3.13. Comparison of total length distributions of female sevengill sharks during winter 2016 (A), spring 2016 (B), summer 2017 (C), autumn 2017 (D), Winter 2017 (E), and spring 2017 (F) in Sawdust Bay, Stewart Island.

60 230

220

210

200

190 Total length (cm) length Total

180

170 Male Female Sex

Figure 3.14. Average total length for male (n = 15) and female (n = 72) sevengill sharks estimated from photogrammetric measurements in Sawdust Bay (±95% CI).

All males that were measured were above 140 cm TL, placing them all into the mature life stage. In comparison, females were a mixture of juvenile (60%) and mature (40%) individuals. No measured individuals fell into the neonate life stage (Fig. 3.15).

61 18 Males Females = Female sexual 16 maturity 14 = Male sexual maturity 12 10 8 Frequency 6 4 2 0

140-149 150-159 160-169 170-179 180-189 190-199 200-209 210-219 220-229 230-239 240-249 250-259 260-269 270-279 Size range (cm) Figure 3.15. Size distribution of sevengill sharks (n = 87) in Sawdust Bay, Stewart Island between winter 2016 and spring 2017, partitioned by sex. The red and blue dotted lines represent size at female and male sexual maturity respectively (Ebert 1986, 1996; Barnett et al., 2010).

3.4.4 Growth Eight individuals (all females) were photographically recaptured with at least 300 days between first and last sightings, and with images that allowed measurement. There was no consistent evidence for growth over time; some individuals showed small positive increments, while some showed small negative increments (mean difference in TL = 1.89 cm, SE = 1.18; Table 3.2, Fig. 3.16). The smallest individual did show the best evidence for growth however, with an estimated growth rate of 8.14 cm per year.

62 Table 3.2. Total length measurements of eight female sevengill sharks with a minimum of 300 days at liberty between the initial and final encounter in Sawdust Bay, Stewart Island. Shark Days Initial encounter Final encounter Difference Growth/yr ID at (cm) (cm) liberty Average 95% CV% Average 95% CV% TL (cm) CI TL (cm) CI 001 483 214.5 2.91 3.06 214.7 2.67 3.04 0.26 0.15 013 373 205.5 1.50 1.80 204.3 1.59 1.92 -1.15 -1.15 018 373 213.5 1.76 2.02 214.1 2.04 2.33 0.56 0.56 022 370 230.9 4.56 4.80 237.2 1.66 1.69 6.32 6.32 027 367 233.2 1.73 1.80 236.4 1.10 1.13 3.23 3.23 028 505 220.8 2.70 2.98 220.5 2.33 2.58 -0.28 -0.20 060 348 211.4 2.05 2.38 210.2 1.05 1.22 -1.18 -1.24 062 345 189.1 1.46 1.92 196.4 4.07 5.12 7.33 8.14

245 Shark ID S001 235 S013

S018 225 S022

S027 215 S028 PDL (cm) PDL S060 205 S062

195

185 0 100 200 300 400 500 Days at liberrty

Fig 3.16. Initial and final TL of eight female sevengill sharks over a period of >300 days at liberty.

63 3.5 Discussion This study provides the first description of the size and sex structure of a population of sevengill sharks in New Zealand. This was achieved using a downward facing stereo-photogrammetry system which allowed me to measure, with high accuracy and precision, 58% of the individuals identified over the course of the 18-month study. All individuals but one could be sexed. The population was shown to consist of individuals between 149 cm and 276 cm, with the modal size class being 2.1 m - 2.2 m. Sawdust Bay was dominated by female sharks, with males making up only a small proportion of the sampled population. The proportion of males increased during summer and autumn (January and April/May respectively). The males that were encountered were generally much smaller than the females. Attempting to measure growth of individuals over a period of >300 days resulted in mean differences (ca.1 cm) that were too small to be reliably resolved given the uncertainties in the measurements of the individual sharks.

3.5.1 Accuracy and precision Before any measurements of sevengill sharks could take place the accuracy and precision of the stereo-photogrammetric methodology had to be tested. After calibration, the stereo-video system produced accurate, repeatable measurements of objects of known length (80 cm target, CV = 0.30%, mean error = + 0.49 cm). Mean accuracy (99.4%) was comparable or better than most photogrammetric systems used to measure wildlife (Harvey et al., 2001; 2003; Growcott et al, 2013; Durban et al 2015; Dawson et al., 2017) which shows that the approaches used in calibration and distortion correction were successful. The greater variation seen in the five replicate measurements of each shark (mean CV = 2.1%) was therefore likely the result of variation in shark body curvature during measurement, and human error in precisely locating the measurement points. Deciding to measure from the tip of the rostrum to the base of the dorsal fin (PDL) removed the need to contend with the long upper lobe of the caudal fin and increased the total number of sharks available to be measured. The conversion equation from PDL resulted in accurate TL estimations (mean difference between measured and calculated TL = 1.43 cm, range = 0.4 – 3.1 cm). In all, the system was highly capable of measuring sevengill sharks for the purpose of placing them into size bins, and by extension, a life history stage which informs us of the ecological role the Sawdust Bay population may play.

3.5.2 Size structure Sevengill sharks in Sawdust Bay ranged from 149cm to 276 cm in total length. Males were significantly smaller than females, which has been observed in other areas populated by sevengill sharks (e.g. California, South Africa, Tasmania; Ebert 1986, Barnett 2010). Total length ranged between 150 - 220 cm for males in Sawdust Bay (n = 15). Female sharks had a wider range of sizes between 149 cm and 276 cm (n = 72), but note that only one had a TL < 180 cm. These results fit the general pattern of shark morphometry. Sharks are usually sexually dimorphic, with males regularly being smaller than females as they do not require the girth to carry pups (Compagno, 1984; Carlson & Parsons, 1997; Lucifora et al., 2005). The observed TL size ranges in Sawdust Bay showed that measured males were probably sexually mature (>140cm; Ebert, 1986), while there was a mix of mature (40%) and juvenile (60%) females (female sexual maturity occurs at ~220 cm TL; Ebert 1986). This does assume however, that the New Zealand population of sevengill sharks mature at the same approximate TLs as populations studied previously in California and Tasmania (Ebert, 1986; 1996; Barnett et al., 2010). The distribution of TLs of sevengill sharks in Sawdust Bay was similar to that of populations in the Derwent Estuary and Norfolk Bay in Tasmania, which had average sizes of 216 cm and 208 cm respectively (Barnett et al., 2010; compared with 211 cm in Sawdust Bay). Unfortunately, these previously published size distributions were not separated by males and females, so it is not possible to compare if males inhabit the smaller size classes as they do in Sawdust Bay. While a large proportion of female sevengill sharks encountered in Sawdust Bay were mature, it is not known if any were pregnant. The reproductive period of sevengill sharks is thought to have a two-year cycle, with a year spent in pregnancy, followed by a year spent foraging to regain condition (Ebert, 1986; 1989; 1996). In the South African population of sevengill sharks, Ebert (1996) found females with fresh wounds, as well as females with healed wounds. Since biting is an integral part of courtship in many sharks Ebert (1996) suggested that females with fresh wounds had probably recently mated, while those with scars were individuals in the recovery year. The wounds and scars are produced when a male shark bites a female on the pectoral fin or gill area and are usually evident from dorsal or lateral views (Ritter & Amin, 2019). Therefore, if scarred female sharks are present in Sawdust Bay, they may well be in

65 their recovery year. During this study there were no females with fresh mating scars, although few had scars that may have been from past mating events. Studies in California (Ebert, 1986) and northern Patagonia (Lucifora et al., 2005) found neonates (individuals <80 cm TL) on a seasonal basis, which indicated these sites function as pupping or nursery areas. The lack of neonates in the present study suggests that Sawdust Bay does not serve these functions. The large litter size that sevengill sharks produce (60 - 104 pups; Ebert, 1986) implies that if pupping occurs in Paterson Inlet, at least some neonates would have been detected in Sawdust Bay.

3.5.3 Sex structure The majority of sharks sampled in Sawdust Bay during winter and spring (June and October/November respectively) were female. During these seasons’ males made up less than 3% of all encountered sharks. Summer and autumn seasons saw an increase in the proportion of males, up to 25% of the encountered population. Female biased sex ratios in sevengill shark populations have similarly been reported around Patagonia and Tasmania (Lucifora et al., 2005; Barnett et al., 2010). An overall female:male ratio of 2.7:1 was reported in Tasmania, with the months of July, August and September consisting of entirely female catches (Barnett et al., 2010). Around Patagonia the ratio of females to males was as high as 5:1 during March (Lucifora et al., 2005). Sexual segregation is often seen in other species of shark (e.g. great white sharks, Carcharadon carcharias, Robbins, 2007; Kock et al., 2013; scalloped hammerhead sharks, Sphyrna lewini, Klimley, 1987; nurse sharks, Ginglymostoma cirratum, Pratt & Carrier, 2001; small-spotted catsharks, Scyliorhinus canicular; Sims et al., 2001; and leopard sharks, Triakis semifasciata, Hight & Lowe, 2007). The drivers of sexual segregation are often unclear and probably vary among species, but may be due to reproductive, environmental, behavioural or habitat selection factors (Klimley, 1987; Sims et al., 2001; Robbins, 2007). Use of satellite tags on sevengills sharks in Tasmania revealed that males exhibit a northern seasonal migration during winter (Barnett et al., 2011). It has been suggested that this is so that male sevengills make use of seperate resources and reduce intraspecific competition with larger females which allows them to maintain reproductive fitness (Barnett et al., 2010). Additionally, Williams et al. (2011) found a male dominated population off Washington State, which suggests that males and females may both remain close to the coast, but segregate to different areas to reduce

66 competition. Future research could examine whether New Zealand sevengill sharks show a similar northerly migration during winter.

3.5.4 Implications of size and sex structures The combination of results from the size and sex ratios implies that Sawdust Bay is not primarily a breeding area for sevengill sharks, although it may occur opportunistically. Awruch et al. (2014) similarly reported a low proportion of males to females, and few mature females with mating scars in Tasmania. They also concluded that mating was not the primary reason for sevengill sharks to congregate in their study area. As stated above, the lack of neonates also indicates that the study site does not function as a pupping or nursery site. Removing these options then, Sawdust Bay most likely offers a foraging area for mature or near mature female sevengill sharks. Sevengill sharks may use Sawdust Bay between mating years for foraging. When females are ready, they may move to specific breeding areas before returning to Sawdust Bay to rest after pupping. Long-term monitoring of this population could establish whether females have a pattern of departure and return based on reproductive events. An increased presence of males appears to be the main driver behind the lower average TL of the population observed during summer and autumn. The reason for this increase is not clear, but it could be related to either increased prey or opportunistic mating. Further sampling in summer and autumn seasons would be required to further understand if the average TL continues to be smaller during these seasons, or if this was a function of small sample size in the current study.

3.5.5 Growth Attempting to monitor the growth of individuals proved to be challenging. When growth had previously been studied in sevengill sharks in captivity, it was shown that while neonates exhibited high growth rates over short time periods (38 cm/year), mature individuals exhibited relatively small amounts of growth per year, if any (0 - 6.4 cm/year) (Dykhuizen & Mollet, 1992; Ebert, 1996). An issue with attempting to track growth over time in this study is that real growth could be swamped by measurement error. Some of the individuals that were measured showed small reductions in size, which can only be attributed to measurement errors. The smallest individual that was remeasured did show the most growth (7.33 cm in 345 days), as would be expected. However, the two largest individuals had the second and third highest growth rates, which does not make sense given that larger sharks should have slower growth.

67 Ebert (1996) similarly reported large variance in growth rates for 12 sevengill sharks that were at liberty for between 34 and 396 days. An individual that was at liberty for 359 days grew 8.9 cm, while another which was at liberty for 396 days grew only 3.6 cm. Unfortunately, there were no sizes reported for these individuals to which the growth rates could be related. Dykhuizen & Mollet (1992) measured no difference in length of two adult sevengill sharks (one male and one female, both ca.225 cm TL) over a four-year period. In summary, tracking the growth of sevengill sharks with stereo-photogrammetry proved difficult due to the slow growth rates of large sharks, and the relatively short duration of the study. The opportunity to remeasure the smaller individuals (140 -180 cm) may have provided better evidence for growth in this population. Unfortunately, all the individuals in these size classes were either only seen once or did not present an opportunity to be remeasured when later identified.

3.5.6 Limitations There are two reasons why biased estimates of the size distribution of sevengill sharks in Sawdust Bay may have been obtained. Firstly, using a BRUV means that individuals must be attracted to bait in order to capture footage. As previously mentioned, smaller sharks can become intimidated by the presence of larger sharks (Jacoby et al., 2012), and may not approach. In this study small individuals were often seen approaching the bait while a larger individual was still present (e.g. Fig. 3.17). This could be due to sevengill sharks residing in loose aggregations and hunting together when required (Ebert, 1991b). Secondly, as sevengill sharks go through an ontogenetic shift in diet (Ebert, 2002), there is a possibility that the bait used (blue cod) was not highly attractive to a particular size class.

Figure 3.17. A small sevengill shark pushes a larger individual out of the way to inspect the bait-box.

68 The use of a regression relationship to convert from PDL to TL could lead to bias in length estimates, especially as the relationship was based on only 25 individuals. The relationship had a very high R2 value however, and while opportunities to measure TL using the stereo- photogrammetry system were rare, the few data available suggest that the extrapolation was accurate. Further ground truthing would certainly be welcome and would increase the confidence in estimated TLs, especially for sharks outside the size range encountered in this study. Another limitation in estimating the TL was variance in multiple measurements of the same individual. Steps were taken to minimise variation by implementing the criterion that the shark’s body needed to be straight enough between the rostrum and the origin of the dorsal fin such that no more than four measurement segments were required to cope with curvature. Additionally, the use of features such as the light patch of skin at the origin of the dorsal fin, as well as the dark band that runs along the dorsal surface, aided in selecting the same measurement points consistently. Finally, a limitation in obtaining TL estimates was the set-up of the stereo-photogrammetry rig and its field of view. Due to working in a low visibility environment, the bait box had to be fastened sufficiently close to the cameras for the sharks’ identifying markings to be clear. This limited the field of view and resulted in only 58% of individuals being measurable with five replicates. Frustratingly there were often occasions when a measurement opportunity was possible in the view of one camera, but not on the other. A wider field of view would increase the number of measurable events, as well as the opportunities for PDL to TL comparisons (although potentially with a decrease in accuracy). Future studies may choose to employ a camera with a wider field of view (at the risk of greater distortion), or sample in an environment with better visibility so that the bait box can be positioned further from the camera. In the wider New Zealand context, this study has the limitation that it is from a single location within Paterson Inlet, Stewart Island. Other sites around Stewart Island may have different size or sex structures, and so applicability outside of Paterson Inlet is unknown.

3.5.7 Future research The data presented here describe the size and sex structure of the Sawdust Bay sevengill shark population over a time span of 18 months. Long-term seasonal monitoring of the population would enable me to better understand the observed variability. Wider sampling of Paterson Inlet would provide a clearer picture of how sevengill sharks segregate their population over a larger scale. Increasing the number of monitoring sites around New Zealand is equally

69 important to investigate long distance movements, demographic differences among areas, and identify pupping or nursery areas for this species. Further life history research could be conducted (e.g. the width of a sevengill shark as an indication of gravidness) or biometric correlations calculated (e.g. the relationship between head width and TL) on the currently available data.

3.5.8 Conclusions This study presents the first attempt at describing the population demographics of sevengill sharks in New Zealand. The use of BRUVs and stereo-photogrammetry to non-invasively describe the sex and size structure of a population of sevengill sharks was innovative, successful and provided useful results. The data showed that the Sawdust Bay sevengill shark population is female-biased and consists of a mix of immature and mature individuals. The low proportion of males and complete lack of neonates suggests that Sawdust Bay functions largely as a foraging ground for female sevengill sharks, as opposed to a mating, pupping or nursery area (although mating may occur opportunistically). The abundance of males does increase during the breeding seasons and opportunistic breeding attempts may occur, although this study showed little evidence of active breeding (such as fresh breeding scars; e.g. Awruch et al., 2014). Determining growth using stereo-photogrammetry was not successful due to measurement uncertainty and the sharks’ very slow growth rate. A much longer study would be required to better capture this. Overall, this study has increased our knowledge of the population demographics of this little studied species. This information can be used in future for researchers to aid the understanding of the ecology of this species, and for managers to produce more robust policy.

70 Chapter 4 – General Discussion 4.1 Ecological importance Broadnose sevengill sharks occupy a high trophic level and are capable of exploiting multiple prey sources (Ebert, 1986; 1992; Barnett, 2010). They occur all around New Zealand (Anderson et al., 1998; Housiaux et al., 2018) and are abundant in some areas (Cox & Francis, 1997, Ford et al., 2018). Considering this, sevengill sharks may play an important part in shaping New Zealand’s coastal food webs (see also Barnett, 2010; Williams, 2011). To better understand their importance, basic demographic information such as seasonal abundances, movements and estimates of survival rates are required (Dulvy et al., 2004; Siqueira et al., 2017). As sevengill sharks have shown ontogenetic shifts in diet ( Ebert, 2002), the role they play in ecosystems may change as they grow. Therefore, understanding the age and life stage structure of sevengill sharks around New Zealand is also important. The aim of this thesis was to describe the demography and life-stage structure of a population of sevengill sharks in southern New Zealand, as an initial step towards addressing our lack of understanding of their ecological role and importance in New Zealand coastal ecosystems. Studies of abundance and demographics are crucial for understanding life history patterns and establishing monitoring baselines (Dulvy et al., 2004). A well-known demonstration of this is the recent controversy over the study by Myers et al. (2007), which reported declining abundances of five large coastal shark species as an indication of a trophic cascade. The decline of large sharks was thought to be the driver for an increase in cownose rays, which in turn were thought to be responsible for declining local shellfish stocks (Myers et al., 2007). In contrast, Grubbs et al. (2016) argued that the life history traits (e.g. age at maturity and low potential for population growth) of cownose rays did not match the reported increase in abundances reported by Meyers et al. (2007), and the data used to reflect the whole population of large sharks was geographically limited. Precisely where the truth lies is uncertain. A similar case is currently being seen in South Africa, where a study of great white shark abundance at False Bay and Seal Island has shown a declining trend since 2014 (Hammerschlag et al., 2019). Meanwhile, there has been a sharp increase in abundance of sevengill sharks around Seal Island (Hammerschlag et al., 2019). These studies exemplify the complexity of marine ecosystems and show why sound demographic and life history data are required to tease apart interactions between species.

71 4.2 Research and management of shark populations in New Zealand There is currently a paucity of dedicated research on elasmobranchs in New Zealand. A National Plan Of Action (NPOA; Ministry of Primary Industry, 2013) concerning the management and conservation of sharks was most recently re-evaluated in early 2014. One of the outcomes was a qualitative risk assessment of all known shark species around New Zealand, made by a panel of experts using data sourced primarily from fisheries (Ford et al., 2015). The panel considered that of the 112 species known to occur in New Zealand waters, 81 had enough data to reach a consensus opinion. Of these 81 however, only 14 species were deemed to have sound data, with the majority (including sevengill sharks) considered to have data of poor quality. An updated risk assessment released in 2018 focused on fewer species (50 instead of 81) and again reported that only eight of these were thought to have sound data. These assessments emphasise the dearth of shark research occurring in New Zealand. Most species lack sufficient data to make robust decisions for conservation or management, leaving only qualitative methods such as surveys of expert opinion. Nevertheless, the NPOA currently stands as one of the best measures by which to direct elasmobranch research, with the risk assessment highlighting the species thought to be at most risk, and those about which we know least. Without robust data it is difficult to construct effective conservation and management plans for any species. This is especially true for sharks, as they often have complex life history traits and migratory patterns (Ferretti et al., 2010). Mismanagement of shark fisheries has led to large scale declines via direct targeting and incidental bycatch (Baum et al., 2003; Robbins et al., 2006; Dulvy et al., 2017) and habitat degradation/loss have impacted home ranges, nursery/pupping sites, and juvenile survival (Chin et al., 2016; Rosa et al., 2016). While the effects of reducing any particular shark species is hard to quantify and the true extent of the effects could be wide ranging (Ferretti et al., 2010), there is little doubt that large predators have a fundamental influence on the ecosystem they inhabit (Heithaus et al., 2008). This thesis provides a starting point for understanding the role that sevengill sharks play in local New Zealand ecosystems. With increased research on New Zealand elasmobranchs, we can better ensure the future health of our marine coastal systems.

72 4.3 Main findings The abundance of sevengill sharks in Sawdust Bay exhibited seasonal fluctuations, with highest numbers present in spring and summer. The high emigration rates out of the study area prior to autumn and winter, coupled with lower estimates of abundance during these seasons, suggest a seasonal movement pattern. The proportion of males in the population was much lower than that of females, especially so in winter and spring. The size structure of the population placed females into either immature or mature stages, while all males were of mature size (based on size at maturity data from overseas populations; Ebert, 1986; 1996, Lucifora et al., 2005; Barnett, 2010).

4.4 Implications of the results It has been reported that mating by sevengill sharks occurs in summer (e.g. Ebert, 1986; 1996). Given that mature females make up a large proportion of the population, the shift in sex ratio during summer implies some reproductive purpose. However, as no fresh mating scars were detected in the BRUV footage, there was little physical evidence of mating (e.g. Ebert, 1986; Awruch et al., 2014) observed during my study. Foraging therefore, seems to be the primary motivation for the population of sevengill sharks using Sawdust Bay. Rig (Mustelus lenticulatus) were often seen in the video footage and are highly likely to be preyed upon by juvenile and mature sevengill sharks (Ebert, 1991; Francis, 2013). Rig are viviparous, and increase in abundance in shallow bays and estuaries during spring and summer for parturition (King, 1984; Massey & Francis, 1989). This influx of prey could be a driver of seasonality in sevengill shark abundance in Sawdust Bay. It could be the case that female sevengill sharks increase in abundance in accordance with prey levels during spring and summer, which in turn attracts males for mating purposes. My data suggest that many of the female sharks leave Sawdust Bay in autumn, potentially in response to declining prey availability. Low abundances of females were recorded over winter, and no males. The same pattern was observed in a study from Tasmania (Barnett et al., 2011). In that study, male sevengill sharks travelled large distances along the Australian coast. It seems likely that males emigrate out of Paterson Inlet on a seasonal basis to different sites around New Zealand for alternative foraging and breeding opportunities.

73 4.5 Wider Context As previously stated, there are not sufficient data on sevengill shark abundances and survival in New Zealand to make confident assessments of their population status. According to commercial catch data, sevengill sharks are ubiquitous around New Zealand and caught on >300 days per year (Ford et al., 2018). In combination with the abundance data from Sawdust Bay, this suggests that the species is reasonably abundant in southern New Zealand waters, at least in comparison to other large coastal sharks (e.g. great whites which are estimated at between 280 - 650 adults for the total New Zealand population; Hillary et al., 2018). As previously noted, Norfolk Bay in Tasmania also has large abundances of sevengill sharks (n = 1246) in an area of approximately 180km2 (Dudgeon et al., 2015). Norfolk Bay has many physical similarities to Sawdust Bay. Both are shallow, with low-visibility water, and both contain commercially fished chondrichthyan species (gummy shark in Norfolk Bay; Barnett et al., 2010; and rig shark in Sawdust Bay). This combination of characteristics may explain why Sawdust Bay is an important year-round habitat for sevengill sharks. Sampling outside Sawdust Bay (albeit with limited survey effort) yielded low densities of sevengill sharks, which suggests that distribution is patchy, even at the scale of Paterson Inlet. Aggregations of sevengill sharks off Washington State were found to be similarly seasonal and patchy, within similar spatial scales (Williams et al., 2012).

4.6 Critique of baited remote underwater video Housiaux (2016) was the first to explore the population of sevengill sharks in Sawdust Bay, using methods that baited sharks to the surface for above-water photographic identification. The current study improved on these methods by employing BRUVs, which increased the number of encounters with sharks and hence allowed more opportunity for photo-ID. The results and conclusions of my study confirm the previous observations of seasonality, and provide further evidence that foraging is the most likely reason for the presence of sevengill sharks in Sawdust Bay. The study by Housiaux (2016) laid the foundation for the current capture-recapture study by showing that ID markings on sevengill sharks are stable, illustrating that photo-ID is possible with sevengill sharks. Furthermore, Housiaux (2016) identified important areas within Paterson Inlet to sample and outlined useful recommendations for future research. The BRUVs in this study performed admirably for capturing photo-ID images of sevengill sharks in Sawdust Bay. They also provided a non-invasive way to assess the size and sex of

74 the individuals encountered. One of the great advantages of BRUV data is that the record is permanent and can be re-analysed in future. Other biometrics could be calculated (e.g. head width to TL relationship) which could make estimating TL easier (as the head was almost always captured in frame). Future studies could also make use of the video data for other research (e.g. abundance and seasonality of other species, such as rig shark) without added cost or field effort. The limitations of the BRUV system should be considered when planning future research. The standard Go-Pro housings have a depth rating limited to 30m, battery life is relatively short when filming on high-quality (needed for photo-ID of spot patterns), and image quality is dependent on underwater visibility. The first two limitations can be addressed via high pressure housings and high-capacity batteries. Filming underwater is conditional on the visibility, which was managed in this study by optimising the distance from the cameras to the bait box. This was a compromise between field of view and capturing sufficient detail to identify individuals in water visibility that was sometimes limited to 1.5m. In areas that have clearer water (e.g. Fiordland) the cameras could be set further away from the bait box resulting in a wider field of view, and therefore a better chance of the animal being captured by both cameras. Lastly, the effectiveness of BRUVs is limited to species that are attracted to bait, and if used for photo-ID the target species must have some form of identifiable feature. In contrast, methods in which fish are physically captured allow for sample collection for genetics or stable isotope analysis, and much easier discrimination among species with subtle morphological differences (Marshall & Pierce, 2012; Santana-Garcon et al., 2014). Physical capture allows for a more direct measurement process, but this may not be more accurate than photogrammetry of swimming sharks (Klimley & Brown, 1983). While BRUVs have limitations, they are especially well suited to studying elasmobranch populations. My system used readily available, inexpensive “action cameras” inside a simple stainless-steel frame. The initial set up cost was low in comparison to purpose-built underwater video equipment. If a shark was to interfere with the equipment the damage would have been easily and inexpensively repaired. Given that this BRUV system is re-useable, it can continue to perform underwater surveys for future research without requiring new or additional equipment (compared to acoustic, spaghetti and band tags which are usually single use). The BRUV system implemented in this study serves as a reliable design for future research on sevengill sharks or other demersal species.

75 4.7 Future research The present study made important steps towards what will hopefully become a long-term study of sevengill sharks using Paterson Inlet. Continued seasonal monitoring of Sawdust Bay would increase the value of the current dataset. A longer data set would allow better estimates of survival rate and more robust quantification of seasonal trends (in both abundance and sex ratios; Kendall, 1999). Sampling could also be expanded to incorporate other sites within Paterson Inlet. This would allow insight into fine-scale distribution and habitat preferences, and potentially indicate effects of protection due to the Ulva Island marine reserve. Additionally, any population segregation within the inlet (by sex or size as seen in other populations; e.g. Barnett et al., 2010; Williams et al., 2011) could be investigated. Beyond Paterson Inlet, sampling new sites around the mainland is required to gain a better understanding of sevengill shark distribution. Fiordland, Otago Harbour, Marlborough Sounds and Akaroa Harbour all have regular reports of sevengill sharks and could have different population structures and ecological roles compared with the Sawdust Bay population. Photo- ID studies in new areas may also encounter individuals seen in Sawdust Bay, which would reveal coastal movement patterns. If photo-ID capture recapture research on sevengill shark populations around New Zealand is continued, management of the photo-ID catalogue would need to be streamlined. As the photo catalogue grows, matching individuals to the catalogue becomes more time consuming and the risk of misidentification increases. Use of purpose-built identification software is recommended. The Interactive Individual Identification System (I3S; http://www.reijns.com/i3s/) software would be appropriate but requires a strict data entry protocol and continued maintenance to perform at its best. This is because markings on animals are manually selected and therefore the process contains some subjectivity. Automated matching software is another option (e.g., https://www.wildbook.org/), and has been applied to sevengill sharks for the Ocean Sanctuaries Project (https://sevengill.oceansanctuaries.org/). This option could provide the best long-term solution to managing the ID catalogue but would require additional development. The scope of the studies on sevengill sharks could be expanded to gain information on fishing pressure, the state of preferred habitat, genetic structure, dietary studies, ontogenetic shifts, life history traits and nursery/pupping sites around New Zealand. These would all add valuable information for assessing the ecological role, conservation status, and management needs of sevengill sharks in New Zealand.

76 4.8 Conclusion This study achieved the first systematic investigation of the abundance and demographic structure of sevengill sharks in New Zealand, while making use of minimally invasive survey techniques. It is hoped the results will form the beginning of a long-term data set which can contribute to future research, risk assessments, conservation and management.

77 Appendix A. All PDL and TL measurements of sevengill sharks in Sawdust Bay, Paterson Inlet. Sex and 95% confidence intervals are included. Shark ID Sex Mean Lower Upper Estimated PDL 95%CI 95%CI TL 001 F 111.8 108.8 114.8 214.5 002 F 104.9 102.6 107.1 202.4 006 F 110.5 107.1 113.9 212.1 007 M 94.5 94.1 95.0 184.4 010 F 97.2 94.2 100.2 189.1 012 M 118.7 116.7 120.7 226.3 013 F 106.6 105.0 108.3 205.5 014 F 118.8 107.6 130.1 226.6 015 F 116.5 112.8 120.3 222.6 016 F 98.5 96.5 100.4 191.3 017 F 111.2 108.0 114.5 213.4 018 F 111.3 109.3 113.3 213.5 019 F 131.1 128.5 133.8 248.0 022 F 121.3 116.2 126.4 230.9 023 F 104.6 102.8 106.4 201.9 024 F 117.6 117.3 117.9 224.5 025 F 96.2 92.0 100.4 187.3 026 F 92.6 90.5 94.7 181.0 027 F 122.6 120.7 124.6 233.2 028 F 115.5 112.4 118.5 220.8 029 F 104.7 103.6 105.9 202.2 030 F 102.9 101.0 104.7 198.9 033 F 118.3 116.8 119.9 225.7 035 F 115.1 113.6 116.6 220.1 036 M 89.3 87.2 91.3 175.3 039 F 110.0 109.4 110.5 211.2 042 F 100.1 98.8 101.5 194.2 044 F 123.7 119.3 128.1 235.0 045 F 130.2 129.3 131.1 246.3

78 Shark ID Sex Mean Lower Upper Estimated PDL 95%CI 95%CI TL 046 F 102.2 101.8 102.7 197.8 049 F 94.7 92.1 97.2 184.7 050 F 103.2 101.1 105.3 199.4 051 F 134.4 132.8 136.0 253.6 052 F 111.8 109.9 113.6 214.3 053 F 93.4 90.5 96.2 182.4 054 F 109.7 107.8 111.7 210.8 055 F 114.4 112.7 116.1 219.0 056 F 138.3 134.3 142.2 260.4 057 F 124.8 122.8 126.7 236.9 059 F 113.9 113.3 114.4 218.0 060 F 111.0 109.2 112.7 212.9 061 F 110.4 109.0 111.7 211.9 062 F 97.2 95.6 98.9 189.1 066 F 103.6 102.5 104.6 200.1 067 M 84.8 83.5 86.1 167.6 068 M 101.3 97.6 104.9 196.1 069 F 74.1 73.2 75.0 149.0 070 F 114.8 112.1 117.5 219.7 071 F 103.5 101.8 105.3 200.1 072 M 93.7 91.6 95.8 183.0 073 F 134.3 132.6 136.0 253.5 074 M 99.3 97.3 101.2 192.6 075 M 100.0 98.9 101.2 194.0 077 M 106.9 106.2 107.6 205.9 078 F 135.9 133.7 138.1 256.3 079 F 113.0 111.9 114.1 216.5 080 M 76.8 74.6 78.9 153.6 081 F 112.4 110.8 113.9 215.4 082 F 120.8 118.2 123.4 230.0 083 F 102.2 101.8 102.7 197.8

79 Shark ID Sex Mean Lower Upper Estimated PDL 95%CI 95%CI TL 088 F 122.7 120.8 124.7 233.4 089 M 118.2 116.1 120.3 225.5 091 F 108.7 106.9 110.4 209.0 092 F 127.5 126.9 128.1 241.7 093 F 140.6 138.4 142.8 264.4 094 F 137.0 135.9 138.2 258.2 095 M 93.3 92.3 94.4 182.3 096 M 80.4 79.0 81.7 159.8 097 F 105.9 104.6 107.2 204.2 098 F 102.3 102.0 102.6 198.0 099 M 98.2 95.8 100.5 190.7 101 F 96.9 95.0 98.8 188.6 102 F 93.0 92.8 93.2 181.7 103 F 103.5 102.3 104.7 200.0 104 F 98.3 96.5 100.1 190.9 105 F 125.2 123.1 127.3 237.7 106 F 122.0 120.1 124.0 232.2 107 F 127.6 126.1 129.2 241.9 117 M 75.0 72.3 77.7 150.5 119 F 94.7 91.3 98.1 184.7 129 F 139.5 138.1 140.9 262.4 130 F 98.8 98.0 99.5 191.8 132 F 116.8 116.0 117.7 223.1 133 F 113.3 112.7 114.0 217.1 134 F 147.5 144.3 150.8 276.5 141 F 133.6 131.4 135.9 252.3 143 F 110.5 108.9 112.1 212.1

80 References cited

Abrantes, K. G., Barnett, A. (2011). Intrapopulation variations in diet and habitat use in a marine apex predator, the broadnose sevengill shark Notorynchus cepedianus. Marine Ecology Progress Series, 442, 133–148. Adams, K. R., Fetterplace, L. C., Davis, A. R., Taylor, M. D., Knott, N. A. (2018). Sharks, rays and abortion: The prevalence of capture-induced parturition in elasmobranchs. Biological Conservation, 217, 11–27. Afonso, A. S., Hazin, F. H. V. (2014). Post-release survival and behavior and exposure to fisheries in juvenile tiger sharks, Galeocerdo cuvier, from the South Atlantic. Journal of Experimental Marine Biology and Ecology, 454, 55–62. Akaike, H. (1974). A new look at the statistical model identification. IEEE Transactions on Automatic Control, 19(6), 716–723. Anderson, O. F., Bagley, N. W., Hurst, R. J., Francis, M. P., Clark, M. R., McMillan, P. J. (1998). Atlas of New Zealand fish and squid distributions from research bottom trawls. NIWA Technical Report, 42. Arzoumanian, Z., Holmberg, J., Norman, B. (2005). An astronomical pattern-matching algorithm for computer-aided identification of whale sharks Rhincodon typus. Journal of Applied Ecology, 42(6), 999–1011. Auger-Méthé, M., Whitehead, H. (2007). The use of natural markings in studies of long- finned pilot whales (Globicephala melas). Marine Mammal Science, 23(1), 77–93. Awruch, C. A., Jones, S. M., Asorey, M. G., & Barnett, A. (2014). Non-lethal assessment of the reproductive status of broadnose sevengill sharks (Notorynchus cepedianus) to determine the significance of habitat use in coastal areas. Conservation Physiology, 2(1), 1–14. Bailey, L. L., Simons, T. R., Pollock, K. H. (2004). Estimating detection probability parameters for Plethodon Salamanders Using the Robust Capture-Recapture Design. The Journal of Wildlife Management, 68(1), 1-13. Baker, J. D., Harting, A. L., Johanos, T. C. (2006). Use of discovery curves to assess abundance of Hawaiian monk seals. Marine Mammal Science, 22(4), 847–861. Barley, S. C., Meekan, M. G., Meeuwig, J. J. (2017a). Species diversity, abundance, biomass, size and trophic structure of fish on coral reefs in relation to shark abundance. Marine Ecology Progress Series, 565, 163-179. Barley, S. C., Meekan, M. G., Meeuwig, J. J. (2017b). Diet and condition of mesopredators on coral reefs in relation to shark abundance. PLoS ONE, 12(4). Barnett, A., Abrantes, K. G., Stevens, J. D., Bruce, B. D., Semmens, J. M. (2010a). Fine- Scale Movements of the Broadnose Sevengill Shark and Its Main Prey, the Gummy Shark. PLoS ONE, 5(12). Barnett, A., Abrantes, K. G., Stevens, J. D., Semmens, J. M. (2011). Site fidelity and sex- specific migration in a mobile apex predator: Implications for conservation and ecosystem dynamics. Animal Behaviour, 81(5), 1039–1048.

81 Barnett, A., Braccini, J. M., Awruch, C. A., Ebert, D. A. (2012). An overview on the role of Hexanchiformes in marine ecosystems: Biology, ecology and conservation status of a primitive order of modern sharks. Journal of Fish Biology, 80(5), 966–990. Barnett, A., Stevens, J. D., Frusher, S. D., Semmens, J. M. (2010b). Seasonal occurrence and population structure of the broadnose sevengill shark Notorynchus cepedianus in coastal habitats of south-east Tasmania. Journal of Fish Biology, 77(7), 1688–1701. Baum, J. K., Baum, J. K., Myers, R. A., Kehler, D. G., Worm, B., Harley, S. J., Doherty, P. A. (2003). Collapse and conservation of shark populations in the Northwest Atlantic. Science, 299(17), 389–392. Baum, J. K., Worm, B. (2009). Cascading top-down effects of changing oceanic predator abundances. Journal of Animal Ecology, 78(4), 699–714. Bernard, A. T. F., Götz, A. (2012). Bait increases the precision in count data from remote underwater video for most subtidal reef fish in the warm-temperate Agulhas bioregion. Marine Ecology Progress Series, 471, 235–252. Braccini, J. M., Troynikov, V. S., Walker, T. I., Mollet, H. F., Ebert, D. A., Barnett, A., Kirby, N. (2010). Incorporating heterogeneity into growth analyses of wild and captive broadnose sevengill sharks Notorynchus cepedianus. Aquatic Biology, 9(2), 131–138. Bradford, A. L., Wade, P. R., Weller, D. W., Burdin, A. M., Ivashchenko, Y. V., Tsidulko, G. A., VanBlaricom, G. R., Brownell Jr, R. L. (2006). Survival estimates of western gray whales Eschrichtius robustus incorporating individual heterogeneity and temporary emigration. Marine Ecology Progress Series, 315, 293–307. Bruce, B. D., Bradford, R. W. (2013). The effects of shark cage-diving operations on the behaviour and movements of white sharks, Carcharodon carcharias, at the Neptune Islands, South Australia. Marine Biology, 160(4), 889–907. Buray, N., Mourier, J., Planes, S., Clua, E. (2009). Underwater photo-identification of sicklefin lemon sharks, Negaprion acutidens, at Moorea (French Polynesia). Cybium, 33(1), 21–27. Burnham, K. P., Anderson, D. R. (1998). Model Selection and Inference - A Practical Information - Theoretic Approach. New York, Springer-Verlag. Bustamante, C., Bennett, M. B. (2013). Insights into the reproductive biology and fisheries of two commercially exploited species, shortfin mako (Isurus oxyrinchus) and blue shark (Prionace glauca), in the south-east Pacific Ocean. Fisheries Research, 143, 174–183. Cailliet, G. M. (1990). Elasmobranch age determination and verification an updated review. Elasmobranchs as living resources: advances in the biology, ecology, systematics, and the status of the fisheries, NOAA Tech. Rep. NMFS 90, 157-165. Cailliet, G. M. (2015). Perspectives on elasmobranch life-history studies : a focus on age validation and relevance to fishery management. Journal of Fish Biology, 87, 1271– 1292. Campana, S. E., Natanson, L. J., Myklevoll, S. (2002). Bomb dating and age determination of large pelagic sharks. Canadian Journal of Fisheries and Aquatic Sciences, 59, 450-455. Cappo, M., Harvey, E., Malcolm, H., Speare, P. (2003). Aquatic Protected Areas - What Works Best and How Do We Know? : Proceedings of the World Congress on Protected Areas. Cairns, Australia, August 2002. Australian Society for Fish Biology, 455-464.

82 Carlson, J. K., Goldman, K. J. (2006). Age and growth of chondrichthyan fishes: New methods, techniques and analysis. Environmental Biology of Fishes, 77(3), 209-421. Carlson, J. K., Parsons, G. R. (1997). Age and growth of the bonnethead shark, Sphyrna tiburo, from northwest Florida, with comments on clinal variation. Environmental Biology of Fishes, 50(3), 331–341. Carter, A. J., Heinsohn, R., Goldizen, A. W., Biro, P. A. (2012). Boldness, trappability and sampling bias in wild lizards. Animal Behaviour, 83(4), 1051–1058. Carver, L., Burger, V. A., Brennan, L. A. (1999). Passive integrated transponders and patagial tag markers for northern bobwhite chicks. The Journal of Wildlife Management, 63(1), 162-166. Casey, J. G., Natanson, L. J. (1992). Revised Estimates of Age and Growth of the Sandbar Shark (Carcharhinus plumbeus) from the Western North Atlantic. Canadian Journal of Fisheries and Aquatic Sciences, 49(7), 1474–1477. Cawthorn, M. W. (1988). Recent observations of hectors dolphin, Cephalorhynchus hectori, in New Zealand. Report for the International Whaling Commission, 9, 303-314. Chapple, T. K., Jorgensen, S. J., Anderson, S. D., Kanive, P. E., Klimley, A. P., Botsford, L. W., Block, B. A. (2011). A first estimate of white shark, Carcharodon carcharias, abundance off Central California. Biology Letters, 7(4), 581–583. Chin, A., Heupel, M. R., Simpfendorfer, C. A., Tobin, A.J. (2016). Population organisation in reef sharks: new variations in coastal habitat use by mobile marine predators. Marine Ecology Progress Series, 544,197-211. Chivers, L. S., Hatch, S. A., Elliott, K. H. (2016). Accelerometry reveals an impact of short- term tagging on seabird activity budgets. The Condor, 118(1), 159–168. Compagno, L.J.V. (1984). Sharks of the World. An annotated and illustrated catalogue of the shark species known to date. Volume 1. Hexaniformes, Squaliformes, Squatiniformes and Pristiophoriformes. FAO, Rome. Compagno, L.J.V. (2005). Notorynchus cepedianus. The IUCN Red List of Threatened Species 2005: e.T39324A10200310. Cooch, E.G., White, G. (2002). Progam MARK: a gentle introduction. Available from http://www.phidot.org/software/mark/docs/book/

Cortes, E. (1999). Standardized diet compositions and trophic levels of sharks. ICES Journal of Marine Science, 56, 707–717. Cox, G., Francis, M. (1997). Sharks and rays of New Zealand. Christchurch, New Zealand, Canterbury University Press. Dicken, M. L., Booth, A. J., Smale, M. J. (2006). Preliminary observations of tag shedding, tag reporting, tag wounds, and tag biofouling for raggedtooth sharks (Carcharias taurus) tagged off the east coast of South Africa. ICES Journal of Marine Science, 63(9), 1640– 1648. Ducklow, H. W., Doney, S. C., Steinberg, D. K. (2009). Contributions of Long-Term Observations to Marine Ecology and Biogeochemistry. Annual Review of Marine Science, 1, 279-302.

83 Dudgeon, C. L., Pollock, K. H., Braccini, J. M., Semmens, J. M., Barnett, A. (2015). Integrating acoustic telemetry into mark–recapture models to improve the precision of apparent survival and abundance estimates. Oecologia, 178(3), 761–772. Dudley, S. F. J., Anderson-Reade, M. D., Thompson, G. S., & McMullen, P. B. (2000). Concurrent scavenging off a whale carcass by great white sharks, Carcharodon carcharias, and tiger sharks, Galeocerdo cuvier. Fishery Bulletin, 98, 646–649. Dulvy, N. K., Baum, J. K., Clarke, S., Compagno, L. J. V., Cortés, E., Domingo, A., Fordham, S., Fowler, S., Francis, M. P., Gibson, C., Martínez, J., Musick, J. A., Soldo, Alen., Stevens, J. D. Valenti, S. (2008). You can swim but you can't hide: The global status and conservation of oceanic pelagic sharks and rays. Aquatic Conservation: Marine and Freshwater Ecosystems, 18(5), 459-482. Dulvy, N. K., Ellis, J. R., Grant, A., Jennings, S., Reynolds, J. D., Goodwin, N. B. (2004). Methods of assessing extinction risk in marine fishes. Fish and Fisheries, 5(3), 255–276. Dulvy, N. K., Fowler, S. L., Musick, J. A., Cavanagh, R. D., Kyne, M., Harrison, L. R., Carlson, J. K., Davidson, L. N. K., Fordham, S. V., Francis, M. P., Pollock, C. M., Simpfendorfer, C. A., Burgess, G. H., Carpenter, K. E., Compagno, L. J., Ebert, D. A., Gibson, C., Heupel, M. R., Livingstone, S. R., Sanciangco, J. C., Stevens, J. D., Valenti, S., White, W. T. (2014). Extinction risk and conservation of the world’s sharks and rays. ELife, 3, 1–35. Dulvy, N. K., Simpfendorfer, C. A., Davidson, L. N. K., Fordham, S. V., Bräutigam, A., Sant, G., Welch, D. J. (2017). Challenges and Priorities in Shark and Ray Conservation. Current Biology, 27(11), 565–572. Dureuil, M., Worm, B. (2015). Estimating growth from tagging data: An application to northeast Atlantic tope shark Galeorhinus galeus. Journal of Fish Biology, 87(6), 1389–1410. Dykhuizen, G. Van, Mollet, H. F. (1992). Growth, age estimation and feeding of captive sevengill sharks, notorynchus cepedianus, at the Monterey Bay aquarium. Marine and Freshwater Research, 43(1), 297–318. Ebert, D. A. (1986). Aspects of the life history of California's two cowshark species, Notorynchus cepedianus and Hexanchus griseus. (Masters thesis). San Jose State University, California. Ebert, D. A. (1989). Life history of the sevengill shark, Notorynchus cepedianus peron, in two northern California bays. California Fish and Game, 75(2), 102–112. Ebert, D. A. (1991a). Diet of the sevengill shark Notorynchus cepedianus in the temperate coastal waters of southern Africa. South African Journal of Marine Science, 11(1), 565– 572. Ebert, D. A. (1991b). Observations on the predatory behaviour of the Sevengill Shark Notorynchus Cepedianus. South African Journal of Marine Science, 11(1), 455–465. Ebert, D. A. (1996). Biology of the sevengill shark Notorynchus cepedianus (Peron, 1807) in the temperate coastal waters of southern Africa. South African Journal of Marine Science, 17(1), 93–103. Ebert, D. A. (2001). Cow Sharks. California’s Living Marine Resources: A Status Report, 470–471. Ebert, D. A. (2002). Ontogenetic changes in the diet of the sevengill shark. Marine and

84 Freshwater Research, 53, 517–523.

Escalle, L., Murua, H., Amande, J. M., Arregui, I., Chavance, P., Delgado de Molina, A., Gaertner, D., Fraile, I., Filmalter, J. D., Santiago, J., Forget, F., Arrizabalaga, H., Dagorn, L., Merigot, B. (2016). Post-capture survival of whale sharks encircled in tuna purse-seine nets: tagging and safe release methods. Aquatic Conservation: Marine and Freshwater Ecosystems, 26(4), 782–789. Estes, J. A., Terborgh, J., Brashares, J. S., Power, M. E., Berger, J., Bond, W. J. Carpenter, S. R., Essington, T. E., Holt, R. D., Jackson, J. B. C., Marquis, R. J., Oksainen, L., Oksanen, T., Paine, R. T., Pikitch, E. K., Ripple, W. J., Sandin, S. A., Scheffer, M., Schoener, T. W., Shurin, J. B., Sinclair, A. R. E., Soule, M. E., Virtanen, R., Wardle, D. A. (2011). Trophic downgrading of planet Earth. Science, 333, 301–306. Estes, J. A. (1996). Predators and ecosystem management. Wildlife Society Bulletin 24(3), 390–396. Ferretti, F., Osio, G. C., Jenkins, C. J., Rosenberg, A. A., Lotze, H. K. (2013). Long-term change in a meso-predator community in response to prolonged and heterogeneous human impact. Scientific Reports, 3, 1–11. Ferretti, F., Worm, B., Britten, G. L., Heithaus, M. R., Lotze, H. K. (2010). Patterns and ecosystem consequences of shark declines in the ocean. Ecology Letters, 13(8), 1055– 1071. Ford, R. B., Galand, A., Clark, M. R., Crozier, P., Duffy, C. A. J., Dunn, M., Francis, M. P., Wells, R. (2015). Qualitative ( Level 1 ) risk assessment of the impact of commercial fishing on New Zealand chondrichthyans. New Zealand Aquatic Environment and Biodiversity Report No. 157. Ford, R. B., Francis, M. P., Holland, L., Clark, M. R., Duffy, C. A. J., Dunn, M., Jones, E., Wells, R. (2018). Qualitative ( Level 1 ) risk assessment of the impact of commercial fishing on New Zealand chondrichthyans: an update for 2017. New Zealand Aquatic Environment and Biodiversity Report No. 201. Fordham, S. V. (2006). Shark alert: Revealing Europe’s impact on shark populations. The Shark Alliance. Accessed via https://oceana.org/reports/shark-alert-revealing- europes-impact-shark- population.

Fowler, S. L., Cavanagh, R. D., Camhi, M., Burgess, G. H., Cailliet, G. M., Fordham, S. V., Simpfendorfer, C. A., Musick, J. A. (2005). Sharks, rays and chimaeras: the status of the chondrichthyan fishes. Status Survey. IUCN/ SSC Shark Specialist Group. IUCN, Gland, Switzerland and Cambridge, UK. Fowler, S., Raymakers, C., Grimm, U. (2004). Trade in and Conservation of two Shark Species , Porbeagle (Lamna nasus) and Spiny Dogfish (Squalus acanthias). Bonn, Germany, Bundesamt für Naturschutz.

Francis, M.P., Campana, S.E., Jones, C.M. (2007). Age under-estimation in New Zealand porbeagle sharks (Lamna nasus): is there an upper limit to ages that can be determined from shark vertebrae? Marine and Freshwater Research 58: 10-23.

Francis, M. P. (2013). Temporal and Spatial Patterns of Habitat Use by Juveniles of a Small

85 Coastal Shark (Mustelus lenticulatus) in an Estuarine Nursery. PLoS ONE, 8(2).

Francis, M. P., Duffy, C., Lyon, W. (2015). Spatial and temporal habitat use by white sharks (Carcharodon carcharias) at an aggregation site in southern New Zealand. Marine and Freshwater Research, 66, 900–918. Francis, M. P., Mace, J. T. (1980). Reproductive biology of Mustelus lenticulatus from Kaikoura and Nelson. New Zealand Journal of Marine and Freshwater Research, 14(3), 303–311. Francis, M. P., Smith, M. H. (2010). Basking shark (Cetorhinus Maximus) bycatch in New Zealand fisheries, 1994-95 to 2007-08. New Zealand Aquatic Environment and Biodiversity Report No. 49. Friedlander, A. M., DeMartini, E. E. (2002). Contrasts in density, size, and biomass of reef fishes between the northwestern and the main Hawaiian islands: The effects of fishing down apex predators. Marine Ecology Progress Series, 230, 253–264. Frisch, A. J., Ireland, M., Rizzari, J. R., Lönnstedt, O. M., Magnenat, K. A., Mirbach, C. E., Hobbs, J. P. A. (2016). Reassessing the trophic role of reef sharks as apex predators on coral reefs. Coral Reefs, 35(2), 459–472. Goldman, K. (2005). Chapter 6. Age and growth of elasmobranch fishes. Management Techniques for Elasmobranch. 97–132. FAO fisheries technical paper 474. Gore, M. A., Frey, P. H., Ormond, R. F., Allan, H., Gilkes, G. (2016). Use of photo- identification and mark-recapture methodology to assess basking shark (Cetorhinus maximus) populations. PLoS ONE, 11(3). Gormley, A. M., Dawson, S. M., Dawson, S. M., Slooten, E., Brager, S. (2005). Capture- recapture estimates of hector’s dolphin abundance at Banks Peninsula, New Zealand. Marine Mammal Science, 21(2), 204–216. Grange, S., Barnier, F., Duncan, P., Gaillard, J. M., Valeix, M., Ncube, H., Periquet, S., Fritz, H. (2015). Demography of plains zebras (Equus quagga) under heavy predation. Population Ecology, 57(1), 201–214. Growcott, A., Sirguey, P., Dawson, S. (2012). Development and Assessment of a Digital Stereo Photogrammetric System to Measure Cetaceans at Sea. Photogrammetric Engineering and Remote Sensing. 78. 237-246. Grubbs, R. D., Carlson, J. K., Romine, J. G., Curtis, T. H., McElroy, W. D., McCandless, C. T., Cotton, C. F., Musick, J. A. (2016). Critical assessment and ramifications of a purported marine trophic cascade. Scientific Reports, 6(1), 20970. Gunn, J. S., Stevens, J. D., Davis, T. L. O., Norman, B. M. (1999). Observations on the short-term movements and behaviour of whale sharks (Rhincodon typus) at Ningaloo Reef, Western Australia. Marine Biology, 135(3), 553–559. Hamady, L. L., Natanson, L. J., Skomal, G.B., Thorrold, S. R. (2014). Vertebral bomb radiocarbon suggests extreme longevity in white sharks. PLoS ONE, 9(1): e84006. Hammerschlag, N., Gallagher, A. J., Lazarre, D. M. (2011). A review of shark satellite tagging studies. Journal of Experimental Marine Biology and Ecology, 398(1–2), 1–8. Hammerschlag, N., Gallagher, A. J., Wester, J., Luo, J., Ault, J. S. (2012). Don’t bite the

86 hand that feeds: Assessing ecological impacts of provisioning ecotourism on an apex marine predator. Functional Ecology, 26(3), 567–576. Hammerschlag, N., Williams, L., Fallows, M., Fallows, C. (2019). Disappearance of white sharks leads to the novel emergence of an allopatric apex predator, the sevengill shark. Scientific Reports, 9(1), 1908. Harry, A. V. (2017). Evidence for systemic age underestimation in shark and ray ageing studies. Fish and Fisheries, 19(2), 185–200. Harvey, E., Cappo, M., Shortis, M., Robson, S., Buchanan, J., Speare, P. (2003). The accuracy and precision of underwater measurements of length and maximum body depth of southern bluefin tuna (Thunnus maccoyii) with a stereo-video camera system. Fisheries Research, 63(3), 315–326. Harvey, E., Shortis, M. (1995). A system for stereo-video measurement of sub-tidal organisms. Marine Technical Society Journal, 29. Harvey, E., Fletcher, D., Shortis, M. (2001). A comparison of the precision and accuracy of estimates of reef-fish lengths determined visually by divers with estimates produced by a stereo-video system. Fishery Bulletin, 99, 63–71. Heagney, E. C., Lynch, T. P., Babcock, R. C., Suthers, I. M. (2007). Pelagic fish assemblages assessed using mid-water baited video: Standardising fish counts using bait plume size. Marine Ecology Progress Series, 350, 255–266. Heberer, C., Aalbers, S. A., Bernal, D., Kohin, S., DiFiore, B., Sepulveda, C. A. (2010). Insights into catch-and-release survivorship and stress-induced blood biochemistry of common thresher sharks (Alopias vulpinus) captured in the southern California recreational fishery. Fisheries Research, 106(3), 495–500. Heithaus, M. R., Frid, A., Wirsing, A. J., Worm, B. (2008). Predicting ecological consequences of marine top predator declines. Trends in Ecology and Evolution, 23(4). Heithaus, M. R., Wirsing, A. J., Dill, L. M. (2012). The ecological importance of intact top- predator populations: A synthesis of 15 years of research in a seagrass ecosystem. Marine and Freshwater Research, 63(11). Hight, B. V., Lowe, C. G. (2007). Elevated body temperatures of adult female leopard sharks, Triakis semifasciata, while aggregating in shallow nearshore embayments: Evidence for behavioral thermoregulation? Journal of Experimental Marine Biology and Ecology, 352(1), 114–128. Hillary, R. M., Bravington, M. V., Patterson, T. A., Grewe, P., Bradford, R., Feutry, P., Gunasekera, R., Peddemors, V., Werry, J., Francis, M. P., Duffy, C. A. J., Bruce, B. D. (2018). Genetic relatedness reveals total population size of white sharks in eastern Australia and New Zealand. Scientific Reports, 8(1), 1–9. Holmberg, J., Norman, B., Arzoumanian, Z. (2009). Estimating population size, structure, and residency time for whale sharks Rhincodon typus through collaborative photo- identification. Endangered Species Research, 7(1), 39–53. Holmes, B. J., Pepperell, J. G., Griffiths, S. P., Jaine, F. R. A., Tibbetts, I. R., Bennett, M. B. (2014). Tiger shark (Galeocerdo cuvier) movement patterns and habitat use determined by satellite tagging in eastern Australian waters. Marine Biology, 161, (11). Housiaux, J. A. (2016). Exploring the murky world of the sevengill shark , Notorynchus

87 cepedianus , in southern New Zealand (Unpublished masters thesis). University of Otago, New Zealand. Housiaux, J. A., Hepburn, C. D., Rayment, W. J.(2018). Seasonal variation in occurrence of the sevengill shark , Notorynchus cepedianus , in two inshore habitats of southern New Zealand. New Zealand Journal of Zoology, 46(1), 48-60. Huggins, A. R. M. (1989). On the statistical analysis of capture experiments. Biometrika, 76(1), 133–140. Huggins, R., Stoklosa, J., Roach, C., Yip, P. (2018). Estimating the size of an open population using sparse capture-recapture data. Biometrics, 74(1), 280–288. IUCN. 2009. IUCN Red List of Threatened Species (ver. 2009.2). Available at: www.iucnredlist.org. (Accessed: 3 November 2009). Jacoby, D. M. P., Croft, D. P., Sims, D. W. (2012). Social behaviour in sharks and rays: Analysis, patterns and implications for conservation. Fish and Fisheries, 13(4), 399– 417. James, B., Bester, M., Penry, G., Gennari, E., Elwen, S. (2015). Abundance and degree of residency of humpback dolphins Sousa plumbea in Mossel Bay, South Africa. African Journal of Marine Science, 37(3), 383–394. Jermacz Ł, Kobak J (2017) Keep calm and don’t stop growing: Non-consumptive effects of a sympatric predator on two invasive Ponto-Caspian gammarids Dikerogammarus villosus and Pontogammarus robustoides. PLoS ONE, 12(8). Johnson, R., & Kock, A. (2006). South Africa’s White Shark cage-diving industry - is there cause for concern? Finding a balance: White shark conservation and recreational safety in the inshore waters of Cape Town, South Africa; proceedings of a specialist workshop, 40–59. Joung, S & Liao, Yih-Yia & Chen, Che-Tsung. (2004). Age and growth of sandbar shark, Carcharhinus plumbeus, in northeastern Taiwan waters. Fisheries Research, 70, 83-96. Neuswanger, J. R., Wipfli, M. S., Rosenberger, A. E., Hughes, N. F. (2016). Measuring fish and their physical habitats: versatile 2D and 3D video techniques with user-friendly software. Canadian Journal of Fisheries and Aquatic Sciences., 73(12),1861-1873. Kendall, W. L. (1999). Robustness of closed capture-recapture methods to violations of the closure assumption. Ecology, 80(8), 2517–2525. Kendall, W. L., Nichols, J. D. (1995). On the use of secondary capture-recapture samples to estimate temporary emigration and breeding proportions. Journal of Applied Statistics, 22(5–6), 751–762. Kendall, W. L., Nichols, J. D., Hines, J. E. (1997). Estimating temporary emigration using capture-recapture data with Pollock’s robust design. Ecology, 78(2), 563–578. King, K. J. (1984). Changes in condition of mature female rig (Mustelus lenticulatus) from Golden Bay in relation to seasonal inshore migrations. New Zealand Journal of Marine and Freshwater Research, 18(1), 21–27. King, K. J., Clark, M. R. (1984). The food of rig (Mustelus lenticulatus) and the relationship of feeding to reproduction and condition in Golden Bay. New Zealand Journal of Marine and Freshwater Research, 18(1), 29-42.

Kiszka, J. J., Aubail, A., Hussey, N. E., Heithaus, M. R., Caurant, F., Bustamante, P. (2015). Plasticity of trophic interactions among sharks from the oceanic south-western Indian Ocean revealed by stable isotope and mercury analyses. Deep-Sea Research Part I: Oceanographic Research Papers, 96, 49–58. Kitchell, J. F., Essington, T. E., Boggs, C. H., Schindler, D. E., Walters, C. J. (2002). The role of sharks and longline fisheries in a pelagic ecosystem of the Central Pacific. Ecosystems, 5(2), 202–216. Kitchell, J. F., Martell, S. J. D., Walters, C. J., Jensen, O. P., Kaplan, I. C., Watters, J., Essignton, T. E., Boggs, C. H. (2006). Billfishes in an ecosystem context. Bulletin of Marine Science, 79(3), 669–682. Klimley, A. P. (1987). The determinants of sexual segregation in the scalloped hammerhead shark, Sphyrna lewini. Environmental Biology of Fishes, 18(1), 27–40. Klimley, A. P., Brown, S. T. (1983). Stereophotography for the field biologist: measurement of lengths and three-dimensional positions of free-swimming sharks. Marine Biology, 74(2), 175–185. Kock, A., O’Riain, M. J., Mauff, K., Meÿer, M., Kotze, D., Griffiths, C. (2013). Residency, habitat use and sexual segregation of white sharks, Carcharodon carcharias in False Bay, South Africa. PLoS ONE, 8(1). Kohler, N. E., Turner, P. A. (2001). Shark tagging: A review of conventional methods and studies. Environmental Biology of Fishes, 60, 191–223. Kohler, N. E., Turner, P. A, Hoey, J. J., Natanson, L. J., Briggs, R. (2002). Tag and recapture data for three pelagic shark species: blue shark (Prionace glauca), shortfin mako (Isurus oxyrinchus) and porbeagle (Lamna nasus) in the North Atlantic Ocean. International Commission for the Conservation of Atlantic Tunas, Collective Volume of Scientific Papers SCRS/2001/64, 54(4), 1231–1260. Lande, R. (1988). Genetics and demography in biological conservation. Science, 241(4872), 1455–1460. Langlois, T., Chabanet, P., Pelletier, D. (2006). Baited underwater video for assessing reef fish populations. Secretariat of The South Pacific Community Fisheries Newsletter, 118, 53-56. Laroche, R. K., Kock, A. A., Dill, L. M., Oosthuizen, W. H. (2007). Effects of provisioning ecotourism activity on the behaviour of white sharks Carcharodon carcharias. Marine Ecology Progress Series, 338, 199–209. Larson, S., Farrer, D., Lowry, D., Ebert, D. A. (2015). Preliminary observations of population genetics and relatedness of the broadnose sevengill shark, Notorynchus cepedianus, in Two Northeast Pacific Estuaries. PLoS ONE, 10(5). Lebreton, A. J., Burnham, K. P., Clobert, J., David, R. (1992). Modelling survival and testing biological hypotheses using marked animals : A unified approach with case. Ecological Society of America, 62(1), 67–118.

89 Letessier, T. B., Juhel, J. B., Vigliola, L., Meeuwig, J. J. (2015). Low-cost small action cameras in stereo generates accurate underwater measurements of fish. Journal of Experimental Marine Biology and Ecology, 466, 120–126. Letessier, T. B., Meeuwig, J. J., Gollock, M., Groves, L., Bouchet, P. J., Chapuis, L., Vianna, G. M. S., kemp, K., Koldewey, H. J. (2013). Assessing pelagic fish populations: The application of demersal video techniques to the mid-water environment. Methods in Oceanography, 8, 41–55. Lettink, M. & Armstrong, D.P. (2003) An introduction to using mark–recapture analysis for monitoring threatened species. New Zealand Department of Conservation Technical Series, 28A, 5–32. Levesque, J.C. (2013). Commercial fishery catch characteristics and population assessment of the shortfin mako shark (Isurus oxyrinchus) in the Western North Atlantic Ocean. Fisheries and Aquaculture Journal, 77, 1-18. Lincoln, F. C. (1930). Calculating waterfowl abundance on the basis of banding returns. U.S. Department of Agriculture, (Circular No. 118), 1–4. Lowry, M., Folpp, H., Gregson, M., Suthers, I. (2012). Comparison of baited remote underwater video (BRUV) and underwater visual census (UVC) for assessment of artificial reefs in estuaries. Journal of Experimental Marine Biology and Ecology, 416– 417, 243–253. Lucifora, L. O., Menni, R. C., Escalante, A. H. (2005). Reproduction, abundance and feeding habits of the broadnose sevengill shark Notorynchus cepedianus in north Patagonia, Argentina. Marine Ecology Progress Series, 289, 237–244. Malcolm, H., Bruce, B. D., Stevens, J. D. (2001). A review of the biology and status of white sharks in Australian waters. Report to Environment Australia, Marine Species Protection Program, CSIRO Research, Hobart, 81. Mali, I., Brown, D. J., Jones, M. C. (2018). Switching bait as a method to improve freshwater turtle capture and recapture success with hoop nets. Southeastern Naturalist, 11(2), 311-318. Mandelman, J. W., Skomal, G. B. (2012). The physiological stress response in elasmobranch fishes: Selected papers from a symposium at the 26th annual meeting of the American Elasmobranch Society. Comparative Biochemistry and Physiology - A Molecular and Integrative Physiology, 162(2), 71–72. Marnewick, K., Funston, P. J., Karanth, K. U. (2008). Evaluating camera trapping as a method for estimating cheetah abundance in ranching areas. South African Journal of Wildlife Research, 38(1), 59–65. Marshall, A. D., Pierce, S. J. (2012). The use and abuse of photographic identification in sharks and rays. Journal of Fish Biology, 80(5), 1361–1379. Marshall, H., Field, L., Afiadata, A., Sepulveda, C., Skomal, G., Bernal, D. (2012). Hematological indicators of stress in longline-captured sharks. Comparative Biochemistry and Physiology - A Molecular and Integrative Physiology, 162(2), 121– 129.

90 Massey, B. R., Francis, M. P. (1989). Commercial catch composition and reproductive biology of rig (mustelus lenticulatus) from Pegasus Bay, Canterbury, New Zealand. New Zealand Journal of Marine and Freshwater Research, 23(1), 113–120. Matich, P., Ault, J. S., Boucek, R. E., Bryan, D. R., Gastrich, K. R., Harvey, C. L., Heithaus, M. R., Kiszka, J. J., Paz, V., Rehage, J. S., Rosenblatt, A. E. (2017). Ecological niche partitioning within a large predator guild in a nutrient-limited estuary. Limnology and Oceanography, 62(3), 934–953. McAuley, R. B., Simpfendorfer, C. A., Wright, I. W. (2007). Gillnet mesh selectivity of the sandbar shark (Carcharhinus plumbeus): Implications for fisheries management. ICES Journal of Marine Science, 64(9), 1702–1709. McCauley, D. J., Papastamatiou, Y. P., Young, H. S. (2010). An observation of mating in free-ranging blacktip reef sharks, Carcbarbinus melanopterus. Pacific Science, 64(2), 349–352. McFarlane, G. A., King, J. R., Saunders, M. W. (2002). Preliminary study on the use of neural arches in the age determination of bluntnose sixgill sharks (Hexanchus griseus). Fishery Bulletin, 100(4), 861–864. Meekan, M. G., Bradshaw, C. J. A., Press, M., McLean, C., Richards, A., Quasnichka, S., Taylor, J. G. (2006). Population size and structure of whale sharks Rhincodon typus at Ningaloo Reef, Western Australia. Marine Ecology Progress Series, 319, 275–285. Motta, P. J., Wilga, C. D. (2001). Advances in the study of feeding mechanisms, mechanics and behaviours of sharks. Environmental Biology of Fishes, 60(3), 131–156. 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(2), 236-252. Myers, R. A., Baum, J. K., Shepherd, T, D., Powers, S. P., Peterson, C. H. (2007). Cascading effects of the loss of apex predatory sharks from a coastal ocean. Science, 315, 1846- 1850. Narazaki, T., Sato, K., Miyazaki, N. (2015). Summer migration to temperate foraging habitats and active winter diving of juvenile loggerhead turtles Caretta caretta in the western North Pacific. Marine Biology, 162(6), 1251–1263. Natanson, L. J., Skomal, G. B., Hoffmann, S. L., Porter, M. E., Goldman, K. J., Serra, D. (2018). Age and growth of sharks: Do vertebral band pairs record age? Marine and Freshwater Research, 69(9), 1440–1452. Pollock, K. H. (1982). A capture-recapture design robust to unequal probability of capture. The Journal of Wildlife Management, 46(3), 752–757. Pollock, K. H. (1986). Estimating the size of wildlife populations using capture techniques. The Fascination of Statistics, (8), 253–282. Pollock, K. H., Nichols, J. D., Brownie, C., Hines, J. (1990). Statistical inference for capture- recapture experiments. Wildlife Monographs, 107, 3-97. Porcher, I. F. (2005). On the gestation period of the blackfin reef shark, Carcharhinus melanopterus, in waters off Moorea, French Polynesia. Marine Biology, 146(6), 1207– 1211.

91 Pratt, H. L., Carrier, J. C. (2001). A review of elasmobranch reproductive behaviour with a case study on the nurse shark, Ginglymostoma cirratum. Environmental Biology of Fishes, 60, 157–188. Raoult, V., David, P. A., Dupont, S. F., Mathewson, C. P., O’Neill, S. J., Powell, N. N., Williamson, J. E. (2016). GoProsTM as an underwater photogrammetry tool for citizen science. PeerJ, 4, e1960. Rasher D. B., Hoey, A. S., Hay, M. E. (2017). Cascading predator effects in a Fijian coral reef ecosystem. Science Reports, 7(1), 15684-15684. Reyes, L. M., Garcia-Borboroglu, P. (2004). Killer Whale (Orcinus orca) Predation on Sharks in Patagonia, Argentina: A First Report. Aquatic Mammals, 30(3), 376-379. Ritchie, E. G., Elmhagen, B., Glen, A. S., Letnic, M., Ludwig, G., McDonald, R. A. (2012). Ecosystem restoration with teeth: What role for predators? Trends in Ecology and Evolution, 27(5), 265–271. Ritter, E. K., Amin, R. W. (2019). Mating scars among sharks: evidence of coercive mating? Acta Ethologica, 22(1), 9-16. Robbins, R. L. (2007). Environmental variables affecting the sexual segregation of great white sharks Carcharodon carcharias at the Neptune Islands South Australia. Journal of Fish Biology, 70(5), 1350–1364. Robbins, W. D., Hisano, M., Connolly, S. R., Choat, J. H. (2006). Ongoing Collapse of Coral-Reef Shark Populations. Current Biology, 16(23), 2314–2319. Rosa, R., Baptista, M., Lopes, V. M., Pegado, M. R., Paula, J. R., Trübenbach, K., Miguel Leal, C., Calado, R., Repolho, T. (2014). Early-life exposure to climate change impairs tropical shark survival. Proceedings. Biological sciences, 281(1793), 20141738. Ruppert, J. L. W., Travers, M. J., Smith, L. L., Fortin, M. J., Meekan, M. G. (2013). Caught in the middle: Combined impacts of shark removal and coral loss on the fish communities of coral reefs. PLoS ONE, 8(9), 1–9. Ruttenberg, B. I., Hamilton, S. L., Walsh, S. M., Donovan, M. K., Friedlander, A., DeMartini, E., Sala, E., Sandin, S. A. (2011). Predator-induced demographic shifts in coral reef fish assemblages. PLoS ONE, 6(6). Ryklief, R. (2012). Population dynamics of the white shark, Carcharodon carcharias, at Mossel Bay South Africa (unpublished masters thesis). Nelson Mandela Metropolitan University, South Africa. Ryklief, R., Pistorius, P., Johnson, R. (2014). Spatial and seasonal patterns in sighting rate and life-history composition of the white shark Carcharodon carcharias at Mossel Bay, South Africa. African Journal of Marine Science, 1–5. Sandin, S. A., Smith, J. E., DeMartini, E. E., Dinsdale, E. A., Donner, S. D., Friedlander, A. M., Konotchick, T., Malay, M., Maragos, J. E., Obura, D., Pantos, O., Paulay, G., Richie, M., Rohwer, F., Schroeder, R. E., Walsh, S., Jackson, J. B. C., Knowlton, N., Sala, E. (2008). Baselines and degradation of coral reefs in the Northern Line Islands. PLoS ONE, 3(2). Santana-Garcon, J., Braccini, M., Langlois, T. J., Newman, S. J., Mcauley, R. B., Harvey, E. S. (2014). Calibration of pelagic stereo-BRUVs and scientific longline surveys for sampling sharks. Methods in Ecology and Evolution, 5(8), 824–833.

92 Santostasi, N. L., Bonizzoni, S., Bearzi, G., Eddy, L., Gimenez, O. (2016). A Robust Design capture-recapture analysis of abundance, survival and temporary emigration of three odontocete species in the gulf of Corinth , Greece, 38, 1–21. Schmitz, O. J., Beckerman, A. P., Brien, K. M. O. (1997). Behaviourally mediated trophic cascades: Effects of predation risk on food web interactions. Ecology, 78(5), 1388–1399. Seber, G. A. F. (1973). The estimation of animal abundance and related parameters. Griffin, London, U.K. Seber, G. A. F. (1986). A Review of Estimating Animal Abundance. Biometrics, 42(2), 267– 292. Shepherd, T. D., Myers, R. A. (2005). Direct and indirect fishery effects on small coastal elasmobranchs in the northern Gulf of Mexico. Ecology Letters, 8(10), 1095–1104. Shortis, M. R., Harvey, E. S. (1998). Design and calibration of an underwater stereo-video system for the monitoring of marine fauna populations. International Archives Photogrammetry and Remote Sensing, 32(5), 792–799. Sieben, K., Rippen, A. D., Eriksson, B. K. (2011). Cascading effects from predator removal depend on resource availability in a benthic food web. Marine Biology, 158(2), 391-400. Simmons, G., Bremner, G., Whittaker, H., Clarke, P., Teh, L., Zylich, K., Zeller, D., Pauly, D., Stringer, C., Torkington, B., Haworth, N. (2016). Reconstruction of marine fisheries catches for New Zealand (1950-2010). Sea Around Us, Global Fisheries Cluster, Institute for the Oceans and Fisheries, University of British Columbia, Vancouver, Canada. Simpfendorfer, C. A., Heupel, M. R., White, W. T., Dulvy, N. K. (2011). The importance of research and public opinion to conservation management of sharks and rays: a synthesis. Marine and Freshwater Research, 62(1), 518–527. Sims, D., Nash, J., Morritt, D. (2001). Movements and activity of male and female dogfish in a tidal sea lough: Alternative behavioural strategies and apparent sexual segregation. Marine Biology, 139(6), 1165–1175. Siqueira, A. C., Quimbayo, J. P., Cantor, M., Silveira, R. B., Daura-Jorge, F. G. (2017). Estimating population parameters of longsnout seahorses, Hippocampus reidi (Teleostei: Syngnathidae) through mark-recapture. Neotropical Ichthyology, 15(4), 1–8. Skomal, G. B. (2007). Evaluating the physiological and physical consequences of capture on post-release survivorship in large pelagic fishes. Fisheries Management and Ecology, 14(2), 81–89. Skomal, G. B., Mandelman, J. W. (2012). The physiological response to anthropogenic stressors in marine elasmobranch fishes: A review with a focus on the secondary response. Comparative Biochemistry and Physiology - A Molecular and Integrative Physiology, 162(2), 146–155. Soldat, V. T. (1982). Age and Size of Spiny Dogfish, Squalus acanthias, in the Northwest Atlantic, NAFO Scientific Council Studies, 3, 47–52. Speed, C. W., Meekan, M. G., Rowat, D., Pierce, S. J., Marshall, A. D., Bradshaw, C. J. A. (2008). Scarring patterns and relative mortality rates of Indian Ocean whale sharks. Journal of Fish Biology, 72(6), 1488–1503.

93 Stehfest, K. M., Patterson, T. A., Barnett, A., Semmens, J. M. (2014). Intraspecific differences in movement, dive behaviour and vertical habitat preferences of a key marine apex predator. Marine Ecology Progress Series, 495, 249–262. Stevens, J. (2000). The effects of fishing on sharks, rays, and chimaeras (chondrichthyans), and the implications for marine ecosystems. ICES Journal of Marine Science, 57(3), 476–494. Stevick, P. T., Neves, M. C., Johansen, F., Engel, M. H., Allen, J., Marcondes, M. C. C., Carlson, C. (2011). A quarter of a world away: female humpback whale moves 10,000km between breeding areas. Biology Letters, 7(2), 299–302. Stevick, P. T., Palsbøll, P. J., Smith, T. D., Bravington, M. V, Hammond, P. S. (2001). Errors in identification using natural markings: rates, sources, and effects on capture– recapture estimates of abundance. Canadian Journal of Fisheries and Aquatic Sciences, 58(9), 1861–1870. Strong, D. R. (1992). Are Trophic Cascades All Wet? Differentiation and Donor-Control in Speciose Ecosystems. Ecology, 73(3), 747–754. Sulikowski, J. A., Wheeler, C. R., Gallagher, A. J., Prohaska, B. K., Langan, J. A., Hammerschlag, N. (2016). Seasonal and life-stage variation in the reproductive ecology of a marine apex predator, the tiger shark Galeocerdo cuvier, at a protected female- dominated site. Aquatic Biology, 24(3), 175–184. Symonds, M. R. E., Moussalli, A. (2011). A brief guide to model selection, multimodel inference and model averaging in behavioural ecology using Akaike’s information criterion. Behavioral Ecology and Sociobiology, 65(1), 13–21. Taylor, I. G., Gertseva, V., Matson, S. E. (2013). Spine-based ageing methods in the spiny dogfish shark , Squalus suckleyi: How they measure up. Fisheries Research, 147, 83– 92. Taylor, M. D., Baker, J., Suthers, I. M. (2013). Tidal currents, sampling effort and baited remote underwater video (BRUV) surveys: Are we drawing the right conclusions? Fisheries Research, 140, 96–104. Templeton, W. (1954). Migrations of spiny dogfish tagged in Newfoundland Waters. Journal of the Fisheries Research Board of Canada, 11(4), 351–354. Tricas, T., Le Feuvre, E. (1985). Mating in the white tip reef shark Trianodon obesus. Marine Biology, 237(84), 233–237. Urian, K., Gorgone, A., Read, A., Balmer, B., Wells, R. S., Berggren, P., Durban, J., Eguchi, T., Rayment, W., Hammond, P. S. (2015). Recommendations for photo-identification methods used in capture-recapture models with cetaceans. Marine Mammal Science, 31(1), 298–321. Van Tienhoven, A. M., Den Hartog, J. E., Reijns, R. A., Peddemors, V. M. (2007). A computer-aided program for pattern-matching of natural marks on the spotted raggedtooth shark Carcharias taurus. Journal of Applied Ecology, 44(2), 273–280. Vandenabeele, S. P., Grundy, E., Friswell, M. I., Grogan, A., Votier, S. C., Wilson, R. P. (2014). Excess baggage for birds: Inappropriate placement of tags on gannets changes flight patterns. PLoS ONE, 9(3). Vas, P. (1990). The abundance of the blue shark, Prionace glauca, in the western English

94 Channel. Environmental Biology of Fishes, 29(3), 209–225. Wallach, A. D., Ripple, W. J., Carroll, S. P. (2015). Novel trophic cascades: Apex predators enable coexistence. Trends in Ecology and Evolution, 30(3), 146–153. Williams, G. D., Andrews, K. S., Farrer, D. A., Bargmann, G. G., Levin, P. S. (2011). Occurrence and biological characteristics of broadnose sevengill sharks (Notorynchus cepedianus) in Pacific Northwest coastal estuaries. Environmental Biology of Fishes, 91(4), 379–388. Williams, G. D., Andrews, K. S., Katz, S. L., Moser, M. L., Tolimieri, N., Farrer, D. A., Levin, P. S. (2012). Scale and pattern of broadnose sevengill shark Notorynchus cepedianus movement in estuarine embayments. Journal of Fish Biology, 80(5), 1380– 1400. Wourms, J. P. (1977). Reproduction and development in chondrichthyan fishes. American Zoologist, 17(2), 379-410. Würsig, B., Würsig, M. (1977). The photographic determination of group size, composition, and stability of coastal porpoises (Tursiops truncatus) hyperalgesia induced by naloxone follows diurnal rhythm in responsivity to painful stimuli. Science, 198, 755–756. Young, J.Z. (1962). The life of vertebrates. 2nd ed. Oxford: Oxford University Press.