UNIVERSITY of CALIFORNIA Santa Barbara Resetting Predator

Home , Save Our Seas Foundation

UNIVERSITY OF CALIFORNIA

Santa Barbara

Resetting predator baselines in coral reef ecosystems: population dynamics, behavior, and ecological characteristics

A dissertation submitted in partial satisfaction of the

requirements for the degree Doctor of Philosophy

in Environmental Science and Management

Darcy E. Bradley

Committee in charge:

Professor Steven Gaines, Co-Chair

Professor Bruce Kendall, Co-Chair

Professor Robert Warner

Dr. Jennifer Caselle

September 2016 The dissertation of Darcy Bradley is approved.

______Jennifer Caselle

______Robert Warner

______Bruce Kendall, Committee Co-Chair

______Steven Gaines, Committee Co-Chair

September 2016

Resetting predator baselines in coral reef ecosystems: population dynamics, behavior, and ecological characteristics

Darcy Bradley

iii

ACKNOWLEDGEMENTS

I am truly indebted to my advisors, Steve Gaines and Bruce Kendall, and my committee members, Jenn Caselle and Bob Warner, for their support throughout the past five years. Steve, thank you for granting me the freedom to pursue ideas that I thought were interesting, and thank you for giving me the necessary insights to make those ideas into publishable science. Bruce, thank you for always holding me to a very high standard, particularly in regards to math and methods. Jenn, I am truly indebted to you for the unparalleled opportunities you gave me in the field, for inviting me to participate in your ongoing research projects, and for not just teaching me, but also showing me, how to be a better scientist. Bob, thank you for generously offering your time, ear, and discerning eye to talk and read through each of my chapters – your comments and suggestions have profoundly improved this dissertation.

All of the work presented here was conducted in collaboration with multiple individuals, and none of it would have been possible without their participation. In particular, I want to thank Yannis Papastamatiou for his contribution to each of these chapters, and for giving me numerous opportunities to collaborate on other projects. I also want to thank Eric Conklin, Doug McCauley, Kydd Pollock, and Amanda Pollock for their significant contributions to this work. Several field assistants helped collect the data used for my first two chapters: Jay Calhoun, Ryan Carr, Jacob Eurich, Alex Filous, Jonatha Giddens,

Melanie Hutchinson, Scott Larned, Rebecca Most, Rich Pollock, Kosta Stamoulis, Jessica

Schem, Mike Shepard, Roxy Sylva, Yuuki Watanabe, and Tim White. My third chapter never would have been completed without field assistance from Katie Davis and Peter

Carlson, and lab assistance from Wade Sedgwick and Lindsay Hornsby.

Fellowship funding was provided by a National Science Foundation Graduate

Research Fellowship (DGE-114408), a Toyota Environmental Fellowship, and a Dr. Daniel

Vapnek Fellowship. Field research was made possible by support from an AAUS Kathy

Johnston English Scholarship, the National Geographic/Waitt Institute for Discovery, Save

Our Seas Foundation, and the UCSB Coastal Fund.

I would not have gotten through this work without my labmates, my dear friends, and my family – thank you all for your unwavering support. Thank you to my brother, Ryan

Bradley, for encouraging me to be a marine biologist over ten years ago so we could “go on adventures for the rest of our lives and get paid.” Most importantly, I would like to thank my husband, Klaus Fastenmeier, for encouraging and supporting me unconditionally and unfathomably.

This dissertation is dedicated to my grandparents, Joan and Larry Bradley, who introduced me to the wonders of the underwater world and who continue to inspire me to live a life full of great adventures.

CURRICULUM VITAE

DARCY BRADLEY September 2016

EDUCATION Ph.D. University of California, Santa Barbara (UCSB) – Bren School Environmental Science & Management (2016)

B.A. (cum laude) Princeton University Philosophy (2007)

FELLOWSHIPS & SCHOLARSHIPS 2013-16 NSF Graduate Research Fellowship 2013 AAUS Kathy Johnston English Scholarship 2012 Dr. Daniel Vapnek Award, UCSB Bren School 2011-12 Toyota Environmental Fellowship, UCSB Bren School 2007-08 Princeton in Asia Postgraduate Fellowship 2006 Princeton University International Internship Scholarship

GRANTS 2016 UCSB Earth Research Institute, “Combatting illegal fishing, one high-value fishery at a time.” ($1000)

2016 UCSB Academic Senate Doctoral Student Travel Grant ($685)

2016 UCSB Bren School Conference Travel Award ($745)

2015 UCSB Bren School Conference Travel Award ($420)

2015-16 Save Our Seas Foundation, “Accounting for shark behaviour in estimates of reef shark population size and abundance.” ($4800)

2015-16 UCSB Coastal Fund, “Baited remote underwater video for monitoring and citizen science in California.” ($10200)

2014-15 National Geographic Society/Waitt Institute for Discovery (with J. Caselle and Y. Papastamatiou), “Exploration of deep water ecosystems and apex predators in a pristine atoll.” ($10900)

2010 Marine Conservation Action Fund, “Conservation of the threatened whale shark: public education and community monitoring via photo-identification.” ($6550)

AWARDS 2016 Best Oral Presentation (UCSB Bren School PhD Student Symposium) 2015 Gruber Award (best student oral presentation, American Elasmobranch Society) 2015 Carrier Award (best student poster, American Elasmobranch Society)

TEACHING EXPERIENCE 2014 Ecology Teaching Assistant (TA) for Dr. Dave Tilman 2014 Advanced Statistics Lecturer & TA for Dr. Allison Horst 2012-14 Statistics & Data Analysis TA for Dr. Allison Horst

PUBLICATIONS 2016 Dee, L. E., S. Miller, L. Peavey, D. Bradley, et al. Functional diversity of catch mitigates negative effects of temperature variability on fisheries yields. Proceedings of the Royal Society B. 283:20161435. doi: 10.1098/rspb.2016.1435

2015 Papastamatiou, Y. P., Y. Y. Watanabe, D. Bradley, et al. Drivers of Daily Routines in an Ectothermic Marine Predator: Hunt Warm, Rest Warmer? PLoS ONE. 10(6):e0127807. doi: 10.1371/journal.pone.0127807

2015 Aceves, E., A. Adeyemi, D. Bradley et al. Citizen Science as a Tool for Overcoming Insufficient Monitoring and Inadequate Stakeholder Buy-in in Adaptive Management: Criteria and Evidence. Ecosystems. doi:

10.1007/s10021-015-9842-4

vii

2014 Bradley, D. and S. D. Gaines. Counting the cost of overfishing on sharks and rays. eLife 3: e02199.

2014 Papastamatiou, Y. P., C. L. Wood, D. Bradley, D. J. McCauley, A. L. Pollock, J. E. Caselle. First records of the sicklefin lemon shark, Negaprion acutidens, at Palmyra Atoll, central Pacific: a recent colonization event? Marine Biodiversity

Records. 7(e114).

MANUSCRIPTS IN REVIEW In review Bradley, D. et al. Resetting predator baselines in coral reef ecosystems.

In review Bradley, D., Y. P. Papastamatiou, J. E. Caselle. No persistent behavioral effects of scuba diving on marine predators

In review Bradley, D. et al. Growth and life history variability of the grey reef shark (Carcharhinus amblyrhynchos) across its range.

In review Davis, K., P. M. Carlson, D. Bradley, R. R. Warner, J. E. Caselle. Predation risk influences feeding rates but not space use for a Pacific parrotfish.

In review Papastamatiou, Y. P., T. W. Bodey, A. Friedlander, C. Lowe, D. Bradley, et al. Competition shapes reef shark distribution.

In review Papastamatiou, Y. P., Y. Y. Watanabe, U. Demsar, V. Leos-Barajas, D. Bradley, et al. Activity seascapes highlight central place foraging strategies in marine predators that do not require a home.

PROFESSIONAL PRESENTATIONS 2016 13th International Coral Reef Symposium, “Low but stable reef shark population abundance and density at an unfished coral reef.” Honolulu, Hawaii, June 2016, oral presentation.

viii

2016 Gordon Research Conference on Predator-Prey Interactions, “Do SCUBA divers affect the long term spatial behavior of reef sharks?” Ventura, California, January 2016, poster presentation.

2015 31st Annual Meeting of the American Elasmobranch Society (Joint Meeting of Ichthyologists and Herpetologists), “Managing a moving target: a spatially explicit capture-recapture reef shark population density estimate at an unfished coral reef” Reno, Nevada, July 2015, oral presentation.

2015 31st Annual Meeting of the American Elasmobranch Society (Joint Meeting of Ichthyologists and Herpetologists), “Do non-extractive human impacts have quantifiable behavioral effects on Palmyra’s reef sharks?” Reno, Nevada, July 2015, poster presentation.

2012 Ecological Society of America 97th Annual Meeting, “The Tradeoff Between Biodiversity and Management Effectiveness in Predicting Fisheries Production and Stability.” Portland, Oregon. August 2012, poster presentation.

2012 Gordon Research Conference in Industrial Ecology, “Effectiveness of Management Regime Better Predicts Fisheries Food Provisioning Services than Biodiversity at the EEZ Scale.” Les Diablerets, Switzerland, June 2012, poster presentation.

2010 Whale Shark International Conference, “Whale Shark Community Monitoring: understanding a threatened species.” Isla Mujeres, Mexico, July 2010, invited talk.

LEADERSHIP External Advisor, Master’s Group Theses Projects (2014-present) Undergraduate Mentor, UCSB Women in STEM (2016-present) Founder & Mentor, Master’s Group Project Mentorship Program (2014-2016)

Founder & Member, PhD Student Symposium Committee (2014-2016) PhD Student Representative, Dean’s Advisory Council (2013-2016) Co-advisor, Undergraduate Honors Thesis project (2014-2015) PhD Student Representative, PhD Program Committee (2013-2015)

SERVICE Scientific Peer Review ICES Journal of Marine Science PLoS ONE Ecological Modeling Aquatic Living Resources

Society Membership American Academy of Underwater Sciences (AAUS) American Association for the Advancement of Science American Elasmobranch Society (AES) International Society for Reef Studies (ISRS) Society for Conservation Biology (SCB)

ABSTRACT

Resetting predator baselines in coral reef ecosystems:

population dynamics, behavior, and ecological characteristics

Darcy Bradley

What did coral reef ecosystems look before human impacts became pervasive?

Although we cannot effectively manage the marine environment without knowledge of what it looked like in a pristine state, human impacts have wreaked such havoc on the ocean that it is often impossible to accurately characterize marine populations pre-exploitation. Lacking a baseline is particularly critical for species of conservation concern such as sharks, which have experienced severe population declines at local and global scales. I use the unique setting of Palmyra Atoll, a remote, U.S. National Wildlife Refuge in the central Pacific

Ocean, to describe the population dynamics, ecological characteristics, and behavior of reef sharks absent significant human impacts. With a combination of extensive field surveys and experiments and spatially and temporally explicit analytical tools, I attempt to resolve a set of ecological debates regarding the trophic structure of coral reef communities, the drivers of life history variability, and the ecological effects of non-consumptive human impacts on marine predators. Ultimately, I find that the trophic structure of an unexploited reef fish community is not inverted and that even healthy top predator populations may be considerably smaller, and more precarious, than previously thought (Chapter 1). At the same time, I show that there is substantial but not predictable variability in life history traits

observed for the grey reef sharks across its range, suggesting that regional management may be necessary to set sustainable harvest targets and to recover this shark species globally

(Chapter 2). Finally, I demonstrate that humans can interact with reef sharks without persistent behavioral impacts, and that well-regulated shark diving can capture the economic benefits of tourism without undermining conservation goals (Chapter 3). Throughout this dissertation, I focus on coral reef associated shark species (Fig. A), but the challenges I consider and the methods I employ are important for many systems and many predators, which tend to move more than their prey. We often find ourselves managing a moving target, and so counting things correctly, assessing ecological variability, and understanding behavioral responses due to human interactions with wildlife can profoundly change the way we understand and manage both terrestrial and marine systems.

Figure A. A word cloud highlighting key words used in this dissertation

xii

CHAPTER 1

RESETTING PREDATOR BASELINES IN CORAL REEF ECOSYSTEMS

Darcy Bradley1, Eric Conklin2, Yannis P. Papastamatiou3, Douglas J. McCauley4,5,

Kydd Pollock2, Amanda Pollock6, Bruce E. Kendall1, Steven D. Gaines1, Jennifer E. Caselle5

1 Bren School of Environmental Science and Management, University of California, Santa

Barbara, CA 93106, U.S.A.

2 The Nature Conservancy, Hawai’i, 923 Nu’uanu Avenue, Honolulu, HI 96817, U.S.A.

3 Department of Biological Sciences, Florida International University, North Miami, FL 33181,

U.S.A.

4 Department of Ecology, Evolution, and Marine Biology, University of California, Santa

Barbara, CA 93106, U.S.A.

5 Marine Science Institute, University of California, Santa Barbara, CA 93106, U.S.A.

6 U.S. Fish and Wildlife Service, 300 Ala Moana Blvd, Honolulu, HI 96850, U.S.A.

Abstract

Accurate estimates of population size and density are key to successful wildlife management.

While the human footprint is ubiquitous on land, the ocean still has wild places with limited human impacts. Marine conservation has a unique opportunity unavailable to terrestrial managers to use unexploited habitats to provide invaluable descriptions of historical populations and community dynamics. Previous efforts to reconstruct historical baselines for coral reef ecosystems provided the foundation for the concept of inverted trophic biomass pyramids in near-pristine coral reefs, suggesting that ocean management had largely failed globally and significant ocean restoration was required. The validity of the inverted trophic pyramid has been questioned, but not yet resolved empirically. Here, we use extensive field surveys and spatially explicit, capture-recapture models to accurately assess the population size and density of key top predators at an unfished coral reef, focusing on the most abundant predator in terms of biomass, the grey reef shark (Carcharhinus amblyrhynchos). Our density estimate is at least an order of magnitude lower than previous estimates from the same location, which suggests that the trophic structure of an unexploited reef fish community is less top-heavy than previously assumed.

Effective marine conservation requires historical reference points to set appropriate management and recovery targets. The insights from our study suggest that present reef shark recovery targets are likely greatly overestimated and therefore may be impossible to achieve. At the same time, harvest quotas may warrant similar downgrading to prevent continued overexploitation of reef sharks.

Introduction

Wildlife management and conservation is fundamentally concerned with assessing the status of wild populations today and forecasting their trends into the future. Two bounding points of reference can play critical roles for different decisions – historical abundances prior to human impacts and population extinction. On land, human impacts have been ubiquitous and enduring for centuries (Vitousek et al. 1997, Sanderson et al. 2002, Haberl et al. 2007, Estes et al. 2011), and habitats are so altered that recovering species to historical levels is generally unfeasible and rarely even imagined. Instead, terrestrial conservation biologists are disproportionately concerned with the other end of the abundance spectrum – reducing risks of extinction (Ceballos et al. 2015). By contrast, in the sea extinctions are still comparatively rare (IUCN 2014), but thousands of species are harvested from wild populations, often at rates that are currently unsustainable (Jackson 2008, Costello et al. 2016). Setting harvest targets that maintain vibrant populations of targeted species and the broader structure of the ecosystem in which they thrive requires a benchmark understanding of marine ecosystems (Carlton 1998, Clark et al. 2001,

Lotze and Worm 2009). Reconstructing what a pristine ocean looked like is therefore critical to set appropriate management targets for fished species and to recover threatened species (Carlton

1998, Roberts 2003).

Unsurprisingly, a great deal of effort has gone into estimating historical baselines of ocean abundance. Most estimates have come from indirect measures, such as reconstructing fisheries landings trajectories to hindcast pre-exploitation abundances (e.g. Baum and Myers

2004, Scott Baker and Clapham 2004, Rosenberg et al. 2005, Lotze and Worm 2009) or using patterns of genetic diversity to infer historical population sizes (Alter et al. 2012, Ruegg et al.

2013). However, humans began aggressively removing species from the ocean well before we

began studying them, making it generally impossible to establish an accurate historical baseline of population abundance for most marine species solely using data from a post-industrial fishing world (Pauly 1995). A third method of estimating historical population baselines hinges upon the fact that -- unlike on land -- there are still places in the ocean that are relatively pristine. Efforts to reconstruct baseline information about marine populations and ecosystem dynamics have focused on the remote central Pacific Ocean (Cox et al. 2002, Friedlander and DeMartini 2002,

Stevenson et al. 2007, Sandin et al. 2008), where the Pacific Remote Islands are protected under

U.S. jurisdiction and provide a potential window into the historical ocean. Early studies of these systems documented large abundances of top predators, which was the genesis of a surprising and controversial idea – historical marine food webs may have had an inverted trophic pyramid with larger biomass at the top of the food web than at trophic levels below (Stevenson et al.

2007, Sandin et al. 2008).

If this vision of past ocean ecosystems was correct, and if it applied to other ocean geographies and habitats, the implications for ocean management and the need for ocean restoration would be profound. Not surprisingly, this rather radical scientific revision to our understanding of trophic structure and biomass organization in coral reef ecosystems has also evoked considerable scientific skepticism. Theoretical examinations of biomass distribution and energy flow in ecosystems (Trebilco et al. 2013) questioned the potential validity of the empirical findings. Such conceptual challenges raised the specter of empirical errors, especially with respect to sampling highly mobile but spatially patchy predators with classical approaches

(Ward-Paige et al. 2010b, Nadon et al. 2012). One path to resolution is to solve the empirical sampling challenges and see if better estimates from pristine habitats in the ocean support a different view of unexploited ocean food webs.

To this end, we used survey techniques and analytical methods that can be applied over large spatial areas while accounting explicitly for movement on and off of study areas by target species as well as heterogeneous probability of detection (Efford 2004, Royle and Young 2008,

Royle et al. 2009). With spatially explicit tag-recapture data, spatial capture-recapture (SCR) models estimate both population abundance and individual movement using precise capture- recapture locations and predictions based on standard detection models (Appendix 1.1). SCR models also give ecological meaning to density estimates: density is estimated as a function of the area actually occupied by individuals in a population, not by area surveyed +/- an arbitrary buffer (Royle et al. 2014). Here, we use SCR models to produce the first spatially and temporally explicit island-wide baseline population abundance and density estimates for a dominant reef shark at an unexploited coral reef. We focus on the grey reef shark (Carcharhinus amblyrhynchos), which is one of the most abundant reef sharks on Indo-Pacific coral reefs

(McCauley et al. 2012, Nadon et al. 2012, Rizzari et al. 2014). Given the propensity of grey reef sharks to associate with shallow coastal/reef zones, this species is acutely susceptible to overfishing (Dulvy et al. 2014), and it is listed as near threatened on the IUCN Red List of

Threatened species (Smale 2009). We conducted our study at Palmyra Atoll, a U.S. National

Wildlife Refuge in the central Pacific, which has a long history of protection and is one of the few remaining unfished coral reef ecosystems on the planet.

To implement our SCR model, tag-recapture field sampling was conducted during 12 separate sampling occasions over 88 days of fishing and 8 years, with 1356 individuals captured

(887 female, 469 male, 1 unrecorded) and 389 individuals recaptured (Appendix 1.1; Fig. A1.1).

We sampled sharks from offshore vessels, and estimated SCR models with methods from

Thompson et al. (2012) and Russell et al. (2012). In their modification to traditional SCR models

-- which use traps set at fixed locations -- they impose a spatial sampling grid on the survey area, and grid cells are used as conceptual traps with capture locations assigned to the center point of the grid cell in which the individual was captured (Fig. 1.1). SCR models differ from traditional capture-recapture models in that they introduce a random effect, which corresponds to the location of an individual’s center of activity during the sampling occasion. Individuals that reside primarily in non-sampled areas will therefore have a low probability of capture, while individuals with activity centers close to sampled areas will have a higher probability of capture; the distance between sampled areas (a grid cell center) and an individual’s activity center ultimately determines an the probability of capture. Importantly, SCR models make the assumption that an individual is more likely to be captured in areas closest to its activity center, which is assumed to be fixed through the study period (for a full list of SCR model assumptions see Appendix 1.1). To test the validity of this assumption, and to ground truth SCR model activity space estimates, we examined residency patterns by estimating activity space independent of capture-recapture effort using long-term passive acoustic monitoring data available for a subset of the population (Appendix 1.1; Fig. A1.2).

Within the SCR modeling framework, the activity centers of all individuals are estimated as the realization of a point process, and the number of activity centers in a given region represents the population size of that region. We estimated the density of sharks by dividing the area occupied by the number of sharks in the sampled area, where area occupied is a function of the activity space of individuals in the population. Changes in population abundance and density were also assessed through time (Appendix 1.1; Fig. A1.3) to evaluate whether the population was at quasi-equilibrium (required for it to be representative of unexploited densities). We used a Gaussian hazard model that assumes a circular bivariate normal activity space to model the

effect of distance on detection probability (Efford 2004) and to estimate activity space (Royle et al. 2014). We ran several models (details reported in Appendix 1.2), introducing four covariates we expected would influence detection probability: distance from activity center, effort (fishing days), sex, and size (fork length).

Methods

Field sampling

Minimally invasive tags with unique number IDs were applied to individual grey reef sharks during 12 separate sampling occasions in the spring, summer, and/or fall from October

2006 through October 2014 (for tagging and sampling details see Appendix 1.1 and Fig. A1.2).

Sequential sampling occasions were a minimum of 58 days apart to decrease the likelihood of behavioral effects (e.g. attraction, habituation, aversion to fishing activity). Sampling was unstructured, and fishing locations were chosen opportunistically to cover Palmyra’s forereef habitat and select channel, lagoon, and backreef habitats (Fig. 1.1). Chum (tuna, wahoo, and/or mackerel) was used to attract sharks to the boat, where they were caught using hand lines baited with barbless circle hooks and restrained at the side of the boat. For each individual, we recorded length (precaudal, fork, and total), sex, and capture/recapture location (latitude and longitude).

Given the highly variable oceanographic conditions at individual locations around the atoll, we were unable to accurately estimate the area of attraction.

Spatial Capture-Recapture Models

We carried out a Bayesian analysis of our spatial capture-recapture (SCR) models to relate the known observation process (capture-recapture encounter data) to the unknown

ecological process (population size, population density, movement) (Royle et al. 2014) and allow for temporary emigration (Kendall et al. 1997). Utilizing methods from Thompson et al. (2012) and Russell et al. (2012), we imposed a spatial sampling grid on our survey area (Thompson et al. 2012), and used grid cells as conceptual traps with capture locations assigned to the center point of the grid cell in which each individual was captured (Russell et al. 2012) (Fig. 1.1;

Appendix 1.1). Encounter histories take the form yijk for an individual i, grid-cell j, and sampling occasion k, where yijk = 1 if individual i was encountered in grid-cell j during sampling occasion k, otherwise yijk = 0. Encounter probabilities are a function of distance and they are grid-cell specific where �!"# = Pr �!"# = 1 = 1 − exp (−�!!�!"#) (Gardner et al. 2010).

We used a Gaussian hazard model that assumes a circular bivariate normal activity space

! ! to model the effect of distance on detection probability as �!"# = exp(−�!"/� ) (Efford 2004).

The distance between an individual’s activity center and grid-cell center d is Euclidean, and � is a scaling parameter (Gardner et al. 2010). An advantage of the bivariate normal encounter model is that it allows direct estimation of activity space as a function of � (Royle et al. 2014). We included four parameters we expected would influence baseline encounter probability: distance from activity center, effort (fishing days), sex, and size (fork length). As sampling intensity varied across sampling occasions, we included fishing effort (days) as a covariate (Appendix

1.1). We ran several models, introducing our covariates as linear effects on the linear predictor for detection probability

log �!!"# = �! + �! ∗ ��! + �! ∗ SEX! + �! ∗ (SIZE!) [1]

�!!"# is the expected number of captures in a grid cell given the individual’s activity center.

All sampled grid-cells have a constant baseline encounter rate �! for the average fishing effort.

Both sex and size are unknown for unobserved individuals. They were therefore modeled as latent variables and were assigned uninformative priors (Appendix 1.1). We define sexfemale = 0 and sexmale = 1 and report �!"# as an estimate of the proportion of the population that is male

(Appendix 1.2). To estimate population abundance, we defined a state space S that encompassed our sampling area with individual potential activity centers si assigned a prior uniform distribution �! ~ Uniform(S) (Efford 2004, Royle and Young 2008, Gardner et al. 2010). We estimated the density of sharks at Palmyra by dividing the area occupied by the number of sharks in the sampled area, where area occupied is a function of the activity space of individuals in the population. Changes in population abundance and density were also assessed through time using a subset of data from a region of the island (the western forereef) that was consistently sampled across all sampling occasions (Fig. A1.3). Population time series data were analyzed for the first

9 sampling occasions (2006-2013), because sampling effort was comparable during these occasions (Fig. A1.1). For all SCR models, convergence was assessed by examining traceplots and the Gelman-Rubin � statistic where values <1.1 indicate convergence (Gelman and Hill

2007). Model implementation and code are provided in the Appendix 1.1.

Passive Acoustic Telemetry

To test the validity of the SCR model assumption that activity centers are stationary through the study period, and to ground-truth SCR model activity space estimates, we examined residency patterns by estimating activity space independent of capture-recapture effort using long-term passive acoustic monitoring data available for a subset of the population. Passive

acoustic telemetry was used to calculate 99% bivariate normal kernel utilization distributions

(Worton 1995) for 37 grey reef sharks (for details see Appendix 1.1 and Table A1.2).

Results and Discussion

We estimate that grey reef shark density at a near-pristine coral reef is 17-22 sharks/km2

(depending on model assumptions), which translates to a total population size of 6669-8344 sharks (Table 1.1). These results suggest that unexploited predator density and abundance are substantially lower than has been previously reported. Previous underwater diver surveys at

Palmyra yielded grey reef shark density estimates that range from ~200 sharks/km2 (towed-diver survey; Nadon et al. 2012 and SCUBA belt transects; McCauley et al. 2012) to over 1000 sharks/km2 (SCUBA belt-transect survey (Nadon et al. 2012) calculated from shark biomass reported in Sandin et al. 2008). It is worth noting that these differences in density estimates

(more than an order of magnitude) far surpass even the diver survey bias correction suggested by

Ward-Paige et al. (2010). In terms of biomass, numerous coral reef top predators are moderately to highly mobile; therefore, overestimation of predator abundance resulting from methodological errors associated with spatially variable movement may be especially common at the top trophic levels.

Our results provide clear evidence that the concept of an inverted trophic biomass pyramid inaccurately represents coral reef ecosystems as a result of overestimated apex predator densities likely due to sampling biases (Ward-Paige et al. 2010b, Nadon et al. 2012, Trebilco et al. 2013). If all species of shark at Palmyra were similarly overestimated by traditional underwater diver surveys, but less vagile predators (e.g. groupers, snappers) were accurately assessed, our estimate would reduce total top predator biomass by nearly 56% compared to the

numbers used as evidence of an inverted biomass pyramid at Palmyra (Sandin et al. 2008). This substantial biomass reduction reshapes the previously proposed trophic structure into a non- inverted pyramid (Fig. 1.2), and aligns the biomass profile with known size-based constraints on trophic pyramids (Trebilco et al. 2013). For species with such complicated behaviors as grey reef sharks, an accurate representation of the coral reef ecosystem and its spatial variability is required to set appropriate management and conservation targets.

We also found no significant trend in grey reef shark population density or abundance through time (Fig. 1.3), indicating that the population is temporally stable. Grey reef sharks have

-1 a reported intrinsic rebound potential of 0.05 yr (r2M), which equates to a doubling time of 12.8 years (Smith et al. 1998). The observed population stability may therefore indicate true population recovery from potential fishing pressure prior to the establishment of the marine refuge in 2001, validating the use of our population estimate as representative of a historical baseline.

An additional important observation to emerge from our spatially extensive sampling design was that grey reef sharks are not uniformly distributed around Palmyra (Fig. 1.4). Rather, density hotspots were discovered on the eastern and western forereefs, with shark densities up to an order of magnitude lower on the north and south forereefs, on the backreefs, and within the lagoons (Fig. 1.4). This spatial non-uniformity would likely bias spatially limited population surveys or surveys conducted with non-random sampling, further highlighting the necessity of comprehensive spatial sampling to establish historical baselines accurately. Interestingly, while shark density is variable across space, grey reef sharks displayed strong residency, consistent with previous findings for relatively isolated populations (Field et al. 2011, Barnett et al. 2012,

Vianna et al. 2013). Activity space estimates from our spatial capture-recapture model were

similar to mean 99% kernel utilization distribution estimates from grey reef shark passive acoustic monitoring (25.2-28.4km2 and 28.8km2, respectively; Appendix 1.2; Table A1.2). We also found that male sharks had slightly larger activity spaces than females (Table A1.1), similarly consistent with published results (Heupel and Simpfendorfer 2014, Espinoza et al.

2014); however, all of our acoustically monitored sharks made occasional excursions all or part of the way around Palmyra’s forereefs and/or into backreef and lagoon habitat (J. Caselle & Y.

Papastamatiou, unpublished data). The drivers of these infrequent excursions are currently unknown and merit further investigation, but island-wide movement has important implications for the spatial management of reef sharks and could explain why small marine reserves may be insufficient to maintain and recover reef shark populations (Robbins et al. 2006, McCook et al.

2010, Espinoza et al. 2014).

From a conservation perspective, our lower than expected reef shark density and abundance estimates are both troubling and encouraging. If healthy coral reef ecosystems tend to support far fewer sharks than previously thought, then recovering reef shark populations may not be an insurmountable goal. Decline rates for reef shark species (e.g. over 90% reported in

Robbins et al. 2006, Graham et al. 2010) may have been greatly overestimated for some species of reef shark, if they were based on equally overstated baseline population numbers. This could be good news for sharks and shark conservation. On the other hand, if stable shark populations are indeed smaller than expected as our study has shown, they are consequently more precarious when faced with a given level of harvest. Even with our low intensity, single hook, hand-line fishing, we estimate that we interacted with 16-22% of all grey reef sharks at Palmyra; a short visit by a commercial longline vessel could reset population trajectories for this species in a matter of days. In fact, it has long been reported that fisheries are able to expediently devastate

reef shark populations (Graham et al. 2010), and over the last decade global shark landings have declined due to population reductions caused by overfishing (Davidson et al. 2015). These fishing-induced population declines may be the result of inflated harvest quotas and insufficient protection status resulting from an overestimated baseline, as has been previously speculated

(Ward-Paige et al. 2010b). It is not uncommon for findings from fisheries independent underwater visual census to be translated into management targets for shark populations (e.g.

McCook et al. 2010), which could be devastating for these species. To be broadly applicable, the shark population information reported here will need to be contextualized to the broader biotic and abiotic features of Palmyra; however, our finding that an unexploited location with high quality habitat has surprisingly low, but stable, shark abundance implies that population size limitations resulting from density dependence may be as important in explaining shark vulnerability to overexploitation as the slow life history characteristics that are primarily cited.

Marine wildlife management has an opportunity unparalleled in terrestrial systems to reconstruct historical baselines for many species using the few remaining large unexploited marine ecosystems. However, monitoring mobile species in complex ecosystems presents a suite of technical challenges that demand the application of appropriately sophisticated analytics. By employing an assessment method that directly addresses the potential biases associated with counting mobile species, we find that the conclusions that were drawn from methodologically and analytically simplistic methods can be strikingly at odds with reality. Ultimately, overestimates of shark density and biomass represent a double-edged sword that can weaken shark conservation: our far lower estimates of shark abundance and density suggest that shark recovery targets should be adjusted. Present targets are likely unrealistic and therefore

impossible to achieve; at the same time, harvest quotas should be similarly downgraded greatly to prevent continued overexploitation of these important predators.

Acknowledgements

We thank The Nature Conservancy staff for support at the Palmyra Atoll research station, the

Palmyra Atoll Research Consortium, our volunteer field assistants including J. Calhoun, R. Carr,

J. Eurich, A. Filous, J. Giddens, M. Hutchinson, S. Larned, R. Most, R. Pollock, K. Stamoulis, J.

Schem, M. Shepard, R. Sylva, Y. Watanabe, and T. White. Thanks to K. Weng for use of

Palmyra’s passive acoustic receiver array. Funding was provided by The Nature Conservancy,

Hawai’i, a Marisla Foundation grant (JEC), and the Benioff Ocean Initiative (DJM); the Center for Scientific Computing from the CNSI/MRL at UCSB provided access to a high-performance computer cluster (NSF CNS-0960316 and MRSEC DMR-1121053). DB was supported by a

NSF Graduate Research Fellowship (DGE-114408), a Dr. Daniel Vapnek Fellowship (UCSB

Bren School), and an AAUS Kathy Johnston English Scholarship. This is publication number

[TBD] from the Palmyra Atoll Research Consortium (PARC).

Table 1.1. Mean abundance and density estimates of an unfished population of reef sharks.

Mean Mean density Density

Modela abundance (SD) (sharks/km2) (SD) 95% CIb �c

Distance + effort + sex + size 8344 (738) 21.3 (1.9) 17.8, 24.7 1.08

Distance + effort + sex 8485 (768) 21.6 (2.0) 17.9, 25.3 1.09

Distance + effort + size 6905 (496) 17.6 (1.3) 15.4, 20.2 1.01

Distance + sex + size 8214 (782) 21.0 (2.0) 17.0, 24.1 1.07

Distance + effort 6671 (399) 17.0 (1.0) 15.1, 18.9 1.01

Distance + sex 8244 (729) 21.0 (1.9) 17.8, 24.6 1.09

Distance + size 7028 (523) 17.9 (1.3) 15.6, 20.9 1.01

Distance 6669 (467) 17.0 (1.2) 14.8, 19.4 1.01 a Spatial capture-recapture models were estimated with the covariates distance, fishing effort (days), sex, and size (fork length, cm) on detection probability. b Lower and upper 95% Bayesian credible intervals (95% CI). c Gelman-Rubin r statistics (values <1.1 indicate model convergence).

Figure 1.1. A map of Palmyra Atoll, a U.S. Wildlife Refuge with a ban on extractive fishing, showing forereef/offshore, backreef, channel, and lagoon habitats, all of which were sampled during our 8-year capture-recapture study. Capture and recapture locations are shown for all sampling periods (x). Circles (¢) denote the 2km sampling grid used for the spatial capture- recapture models.

Figure 1.2. The structural consequences of a 56% reduction to total top predator biomass (area shown in red) based on our density estimates as compared to the numbers used to motivate the claim of an inverted biomass pyramid.

Figure 1.3. Grey reef shark density (black) and abundance (grey) estimates (values reported for

2008-2013 from 2006-2013 capture-recapture data).

Figure 1.4. Spatially explicit density estimates for grey reef shark adults and sub-adults at an unfished coral fish (mean 21.3 sharks/km2; 95% CI 17.8-24.7). Density is color-coded (red = highest and green = lowest).

Figure 1.1

Figure 1.2

Figure 1.3

19 656

level Figure 1.4 0.0010 0.0015 656000 0.0020

●●●●●●●●● ●●●●●●● ●●●●●●● ●●●●●●●●●●●●●●●●● ● ●●● ● ● ●● ●●● ●● ● 0.0025 ●●● ● ●●● ●●●● ●●●●● ●●●●● ● ●●●●●●● ●●●● ●●● ●● ●●●●● ●●●●●●●●●● ●●● ●●●●● ● ● ●●●● ●●●● ● ●●● ●●●● ●●● ● ●●●●●● ●●●●●● ●● ●● ●●● ●●●●●●●●● ●●●●● ● ●● ● ● ●●●●●●● ●●● ●● ● ● ●●●●●● ●●●●● ● ● ● ● ● ●●●● ●● ● ● ● ● ● ● ●●● ● ● ● ● ● ●● ●●●●●●● 652 ● ●●●●●●●●● ●●● ● ●●●●● ●● ● ● ● ●●●●● ● ● ●●●● ●●●● ● ● ●●● ●●●●●●● ● ● ●●● ●●●●●●●●●●● ● ●●●●● ●●●●●●● ●● ● ● ●● ● ●● ● ● ●● ● ● ●● ● ● ●●●●●● ●●●●● ●●● ● ●●● ●●●●● ● ●●●●●● ●●● ● ● ●● ● ●● ● ● ●●●●●●●●●●●●●●●●● ●●●●●●●●● ● ● ●●● ●●●●●● ●●●●●●●●●●● ●● ● ● ●● ●● ● ●● ● ●● ●●● ● ●● ●● ● ●●●●● ●●●●●●●●● ●● ●● ● ●● ● ●● ● ● ● ● ● ●●● ●●●●●● ●●●●●● ● ● ●●● ● ● ●● ●● ●●● ●●●●● ●●●●●●●●●●●● ● ●●●●● ●●●● ● 0.0030 ● ● ● ● ● ● ●●● ●●●●●●●●● ● ● ●●● ● ●● ●●● ●●●●●●●● ● ●● ●● ●●●●●●●●●●●●● ●●● ● ●● ●●●●●●●● ● ●● ● ●● ●●●●● ●●●●●● ●●●●●●●●● ●● ●● ● ● ●●● ●● ●● ●●●●●●● ● ●●●●●●●●●● ● ●●● ●● ●● ● ●●● ●● ●●●●●● ●●●●●● ● ●●●●●●●● ● ● ●●●●●● ●●● ●● ●●●●● ●● ●●●● ●● ● ● ●●● ● ●●● ●●●●●●● ●● ● ● ●● ●● ● ● ● ● ●● ● ●● ●●●●●●● ●●● ●● ●●● ● ● ● ●● ●●● ●●● ●●● ●●● ●●● ● ●●● ●●●● ●●●●● ●● ●●● ●●● ●● ●●● ●●●● ●●●●● ●● ●●●●● ● ●●●●●● ●●● ● ●●●● ● ●● ●● ● ●●● ●● ● ●●●● ● ●● ●● ●●●●●●● ●● ●● ●●● ● ●●●● ● ●●●●● ●● ●● ●●●●●●●●● ● ● ● ● ● ● ●● ●●●●● ●● ●●● ● ●● ●●● ●●● ●● ●●● ●●● ● ● ● ●●● ● ● ●●● ●●●●● ●●●● ●●● ● ●● ● ●● ●●●● ● ●●●● ●● ●● ●●● ●●●●●● ●●●●● ●● ●●● ●●●●● ● ●●●●●●●● ● ●●●● ●●● ●●● ●●●●●●●● ●● ● ● ●●●●●●●●● ●●●●● ●● ● -2! ●●●● ● ●●●●●●●● ● ●●● ●● ●●● ●● ● ● ●● ●●●●●●● ●●●●●●● ● ● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ● ●● ●●●●● ●● ●● ●●● ●● ●●● Sharks km ●● ● ● ●●●●●● ●●●● ● ● ● ●●●●●● ●●●●●●●●●●●●●●●●● ● ●●●●●●● ● ●●● ●●●●●● ● ●●●●●● ●●●●● ●● ●● ●●●●● ●● ● ● ●●●● ●●● ● ●● level ● ● ● ●●●●●● ●●●● ●●●● ● ●● ● ●●●● ●●●● ●●●● ●● ● ● ●●●● ●● ●●● ●●●●●●● ●● ●● level ●● ● ●● ●● ●●●●●●●●●●●●●●●●●●●●●● ● ●● ●● ●● ●●●●●●●●●●●● ●●●● ●● ●●●●●●● ●● ●●●● ●● ●●●●●●●●●● ●●● ● ●●●● ●●● ●●●●● ● ●●● ●● ● ●● ● ●● ●●●● ●●● ●● ●●● ●●●●●●●●●● ●●●● ● ●● ● ●●●●● ●●●●●●●● ●● ●● ● ●●●●●●●●●● ●● ● ●●●● ●●●●●● ●● ●● ●●● ●●● ●●●● ● ● ●● ●● ●●●● ● ●●●●●●● ●● ●● ●●● ●●●●●●●●● ●● ●●● ● ●●●● ●●●●●●● 0.0030 ● ● ●●● ●●●●●● ● ●●● ● ●● ●● ● 30! ● ●●● ●● ●●● ●● ●● ● ●●● ●●● ●●●● ● ●●●● ●● ● ●● ● ●●●●● ● ●●●●● ●● ●●●●●●●●●● ●● ●●● ● ● ●●●●●●●● ● ●●●● ●● ● ●●●●●●●●●●●●●●●●●● ●● ●●●● ●● ●● ●● ●●● 652000 ● ●●●●●●●●●●●●●●●●●●●●● ●●●● ●● ●●● ● ●●●● ●●●● ●●● ●● ●●●● ● ● ●●● ●● 3.0e−09 ●● ●● ● ● ● ● ●●● ●● ●●● ● ●●● ●●● ● ● ●● ●●● ● ●●● ● ● ● ●●●●●●●●●●●●●●●●●●●●●●●●●●● ● ● ●● ● ● ●● ●●●●●● ●●● ● ● ● ● ● ●● ●●●●●●●● ● ● ●●●●●●● ● ● ●●● ●●●●●● ● ●●●● ● ● ●●● ●● ●●● ●● ● ●●●●●●●● ●● ●●●●●●●● ●● ●●● ● ● ●●● ●● ● 0.0025 ●●● ● ●●●●● ●● ● ● ● ● ● ● ● 25! ●●●●● ●●● ● ● ●●● 2.5e−09 ●●● ●● ●● ● ●● ● ● ●● ● ●● ●●●● ● ● ●●● ● ●● ●●● ● ● ●●● ● ● ● ●●● ●●● ● ●●●●●●●●●●●● ● ●● ● ●● ● ●●●● ● ●●●● ● ●● ● ●●●●●●●●●● ● ●● ● ● ●●●●●●● ●● ●● ●● ●●●● ●●●● ●●●●● ● ●● ●● ●●● ●● ●●● ●●● ● ●●●●●● ● ●●● ● ● ● ● ● ●● ●●●● ●●● ●●●● ●●●● ●● ● ● ● ●● ●● ●● ●●●●● ● ● ●● ●●●●●● ●●●●●● ●●●●●●●●●● ●● ● ● ●● ●● ● ●●●●●●● ●● ●●●●●● ●●●●●● ●● ●●●●●●● ●●●● ●●● ● ●●● ●●●●●● ● ● ●●●● ●●● ● ●● ● ●● ●●● ●● ● ● ●●● ●●●● ● ●●●●● ● ● ●●● ●● ●●● ●● ● ●●●●● ●● ● ● ● ●●● ● ● ● ●●●● ●● ● ●● ● 2.0e−09 ● ● ● ● ● ●●● ●● ●●● ●●●●●●●●● ● ● ●●●●●● ●●● ●●●●● ●● ● ●● ● ● ●● ● ● ●● ●●●●●●●●● ●● ●●● ●●●●●● ● ●●● ● ●●●●●●●●● ● ●●●● ●●●●●●● ●●● ● ●● ● ●●● ● ●● ●● ●● ●● ● ●●●●●●●● ●● ●●● ● ●●● ●● ●●● ●●● ●● ●● ● ● ●● ●●●● ● ●●●●● ●●● ●●● ●●●●●● ● ●●● ●●● ●● ●● ●● ● ●● ●● ●● ● ●●●● 0.0020 ●● ● ● ●●● ●● ●● ● ●● 20! ●●● ●● ● ● ●●●●●●● ● ●●●● ●●●●●●●●●●●●●●●●●● ● ● ● ●● ● ● ● ●●● ●●●● ● ●● ●●● ●●●● ●● ●●●● ● ●●● ● ●●●●●● ●●●●●● ● ●●●●●●● ●●●● ● ● ●● ● ●●●●● ●●●●●●●●●●● ●● ●●●● ● ● ●●● ●●●●●● ●●● ●● ●●●●●●●●●●●●●●●●●● ●● ●●●●●● ● ●●●●●● ● ● ●● ●●●● ● ●●● ● ●●● ●● ●●●●● ●●●● ●● ● ●●●●● ● ●●● ● ●●●●●●●● ●● ●●● ●● ●●●●● ● ●●●●●●●● ●●●●●●● ● ●● ● ● ● ●● ●●●●●● ● ● ●●● ●● ●● ●● ●●●●● ● ●●●●● ● ●●● ● ●● ●●● ●●●● ●●●●●●● ● ● ●● ●● ●● ● ● ●●● ● ●●●●●●● ●●●●●●●●●●●● ●● ● ●●●● ●● ● ●●●●●● ●●●●●●● ● ● ●●● ●● ●●●●●●● ● ● ● ● ●● ●●● ●●●●●●● ●● ●●●●● ●● ●●●●●●●● ●●●●● ●●●●●● ●●● ●●● ● ●● ●● ● ●●●●● ● ●●●● ● ●● ● ●●●●●●●● ●● ● ● ● ● ● ●● ●●●●●● ●● 1.5e−09 ● ● ●●●●●●● ●●● ● ●● ● ●●●● ● ●●●● ●●●●●●●●●● ● ●●●●●● ●●●●●●●● ●●● ● ●●●● ● ●●● ●●●● ●●●●●●●● ● ●●●●● ●●● ● ● ● ●● ● ●●● ●● ●●●●●●●●● ●●● ●●● ● ●●● ●●●● ●●●●●●●● ●●● ●●●● ●● ●●●●●●●● ●●●●●● ● ●●● ●●●● ●●●● ●●● ●●●●● ● ●●● ●●●● ●●● ● ● 15! ● ●●● 0.0015 ●●●●● ●●●●●●● ●●●●● ●● ● ●●●● ●●●●● ●● ●● ●● ●●● ●●●● 0.001010

648000

648 5km

810000 820000 830000

810 820 830

Appendices

Appendix 1.1. Supplementary Materials & Methods

Study Site. Palmyra atoll is a U.S. incorporated territory that was established as a Fish and

Wildlife Refuge in 2001 and is currently managed by The Nature Conservancy and the U.S.

Fish and Wildlife Service. Palmyra is located in the central Pacific (5o54’N; 162o05’W); the associated marine refuge and marine national monument, within which commercial fishing is banned, extends out 50 nautical miles. Shark fishing within the refuge is scientific and non-extractive.

Statement on Animal Subjects. This project was certified by the Institutional Animal Care and Use Committee (IACUC), University of California, Santa Barbara, Protocol no. 856

(date of IACUC approval: 5/31/2012).

Field Sampling. Uniquely numbered dart tags and roto tags were applied to individual C. amblyrhynchos in the spring, summer, and/or fall of 2006-2014. Between 2006 and 2012, tag application was primarily via use of roto tags (also called fin tags), which are applied with an applicator through a hole punched in the leading edge of the first dorsal fin. Starting in 2010, 15cm stainless steel head dart tags (Hallprint Co.) were introduced and applied to several individual C. amblyrhynchos using stainless steel tag applicators that were attached to a wooden tagging pole. The head of the dart tag fits into a notch at the point of the applicator needle and the tag is held in place prior to application using a rubber band to hold the tag to the tagging pole. Individual tags were implanted in the dorsal musculature near the base of the first dorsal fin. Tagging by stainless steel dart tag was the only method of

tagging used during the four sampling occasions in 2013 and 2014 (n=1072 individuals).

Tag loss has been reported for dart tags used on reef sharks (Chin et al. 2013), which can significantly bias capture-recapture estimates. To determine rate of tag loss, we double- tagged 79 individuals (from all sampled habitats: forereef, backreef, lagoons), applying a uniquely numbered dart tag to each side of the animal near the first dorsal fin. Of these double-tagged individuals, 18 animals were recaptured at least once in a subsequent sampling period (a minimum of 70 days and a maximum of 495 days after initial tagging).

All recaptured individuals that were initially double-tagged had both tags intact and in good condition upon recapture, indicating that dart tags had minimal tag loss during the study period.

Fishing effort (days spent scientific fishing) varied in intensity at each sampling occasion (Fig. S1). Fishing effort was included in the analysis as a covariate because it was expected to have a positive relationship with capture probability.

Model Sampling Grid Specification. Given our spatially opportunistic survey method and lack of fixed capture/recapture locations, we estimated SCR models with methods from

Thompson et al. (2012) and Russell et al. (2012). In their modification to traditional SCR models -- which use traps set at fixed locations -- they impose a spatial sampling grid on the survey area (Thompson et al. 2012), and grid cells are used as conceptual traps with capture locations assigned to the center point of the grid cell in which the individual was captured

(Russell et al. 2012). Grid cell size was selected to allow adequate modeling of spatial heterogeneity due to habitat features and the location of individuals within those habitats

(Thompson et al. 2012). Although grey reef sharks have a relatively high movement

capacity, Palmyra has highly variable habitat (Fig. 1). Therefore, we used a grid that encompassed our study area with 84, 2km x 2km grid-cells (Fig. 1). If no research fishing occurred in a grid cell during a sampling occasion, we imposed the constraint that the probability of encountering an individual was necessarily 0. We also transformed capture/recapture data into binary encounters for each individual during each sampling occasion to minimize the effects of spatial autocorrelation between grid cells given our relatively small grid cell size. Encounter histories are grid cell specific and take the form yijk for an individual i, grid-cell j, and sampling occasion k, where yijk = 1 if individual i was encountered in grid-cell j during sampling occasion k, otherwise yijk = 0.

Model Covariates. We included four covariates we expected would influence baseline encounter probability: distance from activity center, effort (fishing days), sex, and size (fork length). As sampling intensity varied across sampling occasions, we included fishing effort

(days) as a covariate. Sex and size were also included, because they may impact baseline encounter probability and distance from activity center (via activity space; modeled for sex only), and size was centered. We ran several models, introducing our covariates as linear effects on the linear predictor for detection probability

log �!!"# = �! + �! ∗ ��! + �! ∗ SEX! + �! ∗ (SIZE!) [S1]

�!!"# is the expected number of captures in a grid cell given the individual’s activity center.

All sampled grid-cells have a constant baseline encounter rate �! for the average fishing effort.

SCR Model Assumptions. Spatial capture-recapture (SCR) models – like all capture- recapture models – make a number of assumptions. The first of these is the operational assumption that no tags are lost and all tags are correctly identified (see above). The next set of assumptions relate to statistical aspects of the probability model and the model parameters, and include: (a) demographic closure, (b) no permanent immigration/emigration

(but temporary emigration is allowed), (c) activity centers that are randomly distributed and stationary during the sampling period, (d) encounters among individuals and for any given individual across the sampling space are independent, (e) detection is a declining function of distance from an individual’s activity center (Royle et al. 2014).

Assumptions (a) & (b). The assumptions of demographic and geographic closure (i.e. no births, deaths, immigration, or emigration) are routinely violated and our study is no exception. However, provided demographic and geographic changes to the population occur randomly, we can expect a loss of precision, but estimates should remain unbiased (Kendall

1999). Given the slow life history of the grey reef shark – including late maturity, low fecundity, and an average lifespan of at least 25 years (significantly longer than the study period) (Wetherbee et al. 1997, Robbins 2006) – we would expect limited loss of precision in parameter estimates due to violations of the demographic closure assumption over the course of our study. Regarding geographic closure, five individual shark tags were recovered post fishing-mortality on Teraina and Tabuaeran Islands (230km and 380km away from Palmyra atoll) and these individuals were removed from the analysis. It is possible that other tagged individuals emigrated from Palmyra; however, we have reason to believe that

permanent emigration is limited and sufficiently random to prevent bias in parameter estimates as all acoustically tagged individuals were detected on the Palmyra array from initial tagging (2011/2012) through 2015.

Assumptions (c) & (d). Although SCR models are reasonably robust to violations of assumptions (a) and (b) (Royle et al. 2014), we are confident that Assumptions (c) and (d) are reasonable assumptions for our sampling population. The acoustic telemetry analysis revealed relatively small home range sizes for individual sharks (mean 19.5km2), indicating relatively stationary activity centers. Additionally, we transformed encounter histories into binary encounters for each individual to ensure independence in encounter probability for a given individual during a sampling occasion. Processing time (from capture to release), for each tagged individual was <4 minutes on average, indicating minimal effect of capture of one individual on the probability of capturing other individuals.

Assumption (e). Passive acoustic telemetry data was used to ground truth SCR model activity space estimates; no significant differences were found between 99% activity space utilization distributions estimated from the SCR models or the passive acoustic telemetry data.

SCR Model Implementation. We constructed all SCR models using data augmentation

(Royle et al. 2007) with Markov chain Monte Carlo (MCMC) sampling in the rjags package

(Plummer 2014) in R (Team 2014). We ran the MCMC algorithm 12,000 times with three chains and discarded the first 2,000 runs as burn-in. Model convergence was assessed by

examining traceplots and the Gelman-Rubin � statistic where values <1.1 indicate convergence (Gelman and Hill 2007).

SCR Model Code

model { sigma[1] ~ dunif(0, 40) # sigma female sigma[2] ~ dunif(0, 40) # sigma male lambda0 ~ dnorm(0,.1) beta2 ~ dnorm(0,.1) ## effort beta3 ~ dnorm(0,.1) ## sex beta4 ~ dnorm(0,.1) ## size

psi~dunif(0,1) psi.sex ~dunif(0,1)

for(i in 1:M){ # individual level variables; M is the data augmentation parameter z[i] ~ dbern(psi) # individuals in the population sex[i] ~ dbern(psi.sex) # prior for unobserved individuals sex2[i] <- sex[i]+1 size[i] ~ dnorm(-1,1) # prior for unobserved individuals s[i,1]~dunif(xlim[1],xlim[2]) s[i,2]~dunif(ylim[1],ylim[2])

for(j in 1:J){ # distance function d2[i,j]<- pow(s[i,1]-x[j,1],2) + pow(s[i,2]-x[j,2],2)

for(k in 1:K){ # the likelihood logit(p0[i,j,k])<- ifelse(samplingGrid[k,j]==0,0, lambda0+beta1*sex[i]+beta2*effort[k]+ beta3*size[i]) p[i,j,k]<- z[i]*p0[i,j,k]*exp(-d2[i,j]/(2*sigma[sex2[i]]*sigma[sex2[i]])) y[i,j,k] ~ dbin(p[i,j,k],K) } } } N<-sum(z[]) }

Passive Acoustic Telemetry. SCR models make the assumption that activity centers are stationary through the study period. To test the validity of this assumption, and to ground truth SCR model activity space estimates, we examined residency patterns by estimating activity space independent of capture-recapture effort using long-term passive acoustic monitoring data available for a subset of the population. In July and August of 2011 and

2012, 45 grey reef sharks were surgically implanted with acoustic transmitters (69 kHz,

V16, Vemco Ltd., Nova Scotia, Canada; for surgical implantation methods see

Papastamatiou et al. 2009). Passive acoustic telemetry was used to monitor the movement of grey reef sharks via an array of over 70 individual underwater acoustic receivers (VR2W,

Vemco) that cover all of Palmyra’s unique forereef and backreef habitats (Fig. S2). We calculated 99% bivariate normal kernel utilization distributions (KUDs) (Worton 1995) for each acoustically tagged individual with >100 detections and a minimum of 10 months of location data (n=37) in the package adehabitatHR (Calenge 2006) in R (Team 2014).

Appendix 1.2. Supplementary Results

Spatial Capture-Recapture Model Parameters and Fit. Male and female sharks exhibited differences in �, indicating a sharper decline in the detection function for female sharks at increasing distances from activity centers; this also indicates that male sharks had larger activity spaces than females (28.4 and 25.2 km2 maximum 99% activity space kernel utilization distribution, male and female respectively) (Table A1.1). Activity space estimates from the spatial capture-recapture models and the mean bivariate normal kernel utilization distribution estimated from passive acoustic monitoring of 37 individual grey reef sharks at

Palmyra between 2011 and 2015 were comparable at the 99% level (25.2-28.4km2 (95% CI

20.5-31.0km2) and 28.8km2 (95% CI 19.8-37.8km2), respectively) (Table A1.2). However, there was a significant amount of individual heterogeneity in 99% KUDs estimated from the acoustic data (minimum <1 km2, maximum 117.4 km2) (Table A1.2). Mean estimates of the proportion of male grey reef sharks in the population (�!"#) were consistent across all models (0.41 to 0.46) (Table A1.1). Gelman-Rubin convergence diagnostics reported � values <1.1 for all �, �, and � parameters.

Table A1.1. Key parameter estimates from spatial capture-recapture models of grey reef shark abundance and density at an unfished coral reef.

a b b b e Model Female σ Male σ Both σ β!""#$% β!"# β!"#$ ψ!"# Distance + effort + sex + size Mean (SD) 2.71 (0.25) 2.98 (0.13) 0 0.16 (0.005) 0.80 (0.26) 0.15 (0.06) 0.44 (0.04) 95% CId 2.26, 3.24 2.73, 3.25 0 0.15, 0.17 0.29, 1.27 0.04, 0.27 0.35, 0.53 Distance + effort + sex Mean (SD) 2.7 (0.26) 2.98 (0.13) 0 0.16 (0.005) 0.93 (0.22) 0 0.41 (0.04) 95% CI 2.15, 3.19 2.74, 3.25 0 0.15, 0.17 0.41, 1.34 0 0.34, 0.48 Distance + effort + size Mean (SD) 0 0 2.93 (0.12) 0.16 (0.005) 0 0.26 (0.06) 0 95% CI 0 0 2.71, 3.17 0.15, 0.17 0 0.15, 0.39 0 Distance + sex + size Mean (SD) 2.64 (0.25) 2.98 (0.13) 0 0 0.66 (0.20) 0.15 (0.06) 0.46 (0.04) 95% CI 2.18, 3.16 2.75, 3.25 0 0 0.26, 1.04 0.04, 0.27 0.39, 0.55 Distance + effort Mean (SD) 0 0 2.92 (0.12) 0.16 (0.005) 0 0 0 95% CI 0 0 2.71, 3.16 0.15, 0.17 0 0 0 Distance + sex Mean (SD) 2.70 (0.28) 2.97 (0.13) 0 0 0.89 (0.26) 0 0.42 (0.04) 95% CI 2.20, 3.24 2.72, 3.24 0 0 0.38, 1.40 0 0.33, 0.50 Distance + size Mean (SD) 0 0 2.92 (0.12) 0 0 0.27 (0.06) 0 95% CI 0 0 2.68, 3.15 0 0 0.15, 0.39 0 Distance Mean (SD) 0 0 2.90 (0.12) 0 0 0 0 95% CI 0 0 2.68, 3.14 0 0 0 0 a σ is a scaling factor on the effect of distance in the detection function (σ is estimated separately for male and female sharks in models that contain a covariate for sex). b β parameters denote the effect of survey effort (days), shark sex, and shark size (fork length, cm) on the probability of detection. c ψ!"# is the probability a grey reef shark in the population is male. d 95% Bayesian credible intervals.

Table A1.2. 99% bivariate normal kernel utilization distributions (KUDs) estimated for grey reef sharks to validate SCR model assumptions (n=37).

Transmitter IDa 99% KUD (km2) A69.1303.33561 60.56 A69.1303.33562 26.32 A69.1303.43870 9.65 A69.1303.43876 1.88 A69.1303.61334 34.46 A69.1303.61335 14.01 A69.1303.61336 6.33 A69.1303.61337 38.09 A69.1303.61338 5.23 A69.1303.61339 8.18 A69.1303.61340 12.91 A69.1303.61341 6.20 A69.1303.61342 7.75 A69.1303.61343 58.78 A69.9001.30702 6.61 A69.9001.30703 70.34 A69.9001.30705 3.06 A69.9001.30706 49.75 A69.9001.30707 15.79 A69.9001.30710 20.38 A69.9001.30711 24.50 A69.9001.30714 8.21 A69.9002.13982 12.55 A69.9002.13984 23.00 A69.9002.13988 73.22 A69.9002.13990 41.39 A69.9002.13992 70.73 A69.9002.13994 117.39 A69.9002.13998 17.67 A69.9002.14000 73.82 A69.9002.15320 58.89 A69.9002.15322 15.39 A69.9002.15324 35.17 A69.9002.15326 29.02 A69.9002.15328 8.76 A69.9001.30713 0.02 A69.9001.30708 0.11 a All individuals had >100 detections and >10 months of data.

Figure A1.1. Fishing effort (days, grey bars, left vertical axis) and sharks captured

(individuals, red line, right vertical axis) at each of the 12 sampling occasions between

2006-2014.

Figure A1.2. Passive acoustic telemetry was used to monitor the movement of C. amblyrhynchos via an array of 76 individual underwater acoustic receivers (red triangles,

VR2W, Vemco) that cover all of Palmyra’s unique forereef, backreef, and lagoon habitats.

Figure A1.3. Capture-recapture data from the western forereef and backreef (area contained within the yellow box) was used to estimate C. amblyrhynchos abundance and density through time. 774 individual sharks were captured in the region (578 female, 195 male, 1 unknown).

Figure A1.1

Figure A1.2

AcousticAcoustic receiverreceiver! ForereefForereef! Backreef/terracesBackreef/Terraces! Lagoons/channelLagoons/Channel! LandLand! 648 651 654

815 820 825 830

Figure A1.3

● ● ● ● ● Western Forereefsubset! Forereef! Backreef/TerracesBackreef/terraces! ● ● ● ● ● Lagoons/ChannelLagoons/channel! Land!

● ● ● ● ● Capture/RecaptureCapture/recapture! ● TrappingSampling Grid grid ● ● ● ● ● ● ● ● ● ●

● ● ● ● ● ● ● ● ● ●

● ● ● ● ● ● ● ● ● ● 647 651 655

● ● ● 810 ● ● 820 830

● ● ● ● ●

CHAPTER 2

GROWTH AND LIFE HISTORY VARIABILITY OF THE GREY REEF SHARK

(CARCHARHINUS AMBLYRHYNCHOS) ACROSS ITS RANGE

Darcy Bradley1, Eric Conklin2, Yannis P. Papastamatiou3, Douglas J. McCauley4,5,

Kydd Pollock2, Bruce E. Kendall1, Steven D. Gaines1, Jennifer E. Caselle5

1 Bren School of Environmental Science and Management, University of California, Santa

Barbara, CA 93106, U.S.A.

2 The Nature Conservancy, Hawai’i, 923 Nu’uanu Avenue, Honolulu, HI 96817, U.S.A.

3 Department of Biological Sciences, Florida International University, North Miami, FL

33181, U.S.A.

4,5 Department of Ecology, Evolution, and Marine Biology, University of California, Santa

Barbara, CA 93106, U.S.A.

5 Marine Science Institute, University of California, Santa Barbara, CA 93106, U.S.A.

Abstract

For broadly distributed, overexploited species such as elasmobranchs (sharks and rays), conservation management would benefit from understanding how life history traits predictably change in response to local environmental and ecological factors. However, fishing obfuscates this objective by causing complex and often mixed effects on the life histories of target species. Disentangling the many drivers of life history variability requires knowledge of elasmobranch populations absent fishing. Here, we synthesize published information on the life history of the grey reef shark (Carcharhinus amblyrhynchos) from across its geographic range. We then describe the growth, maximum size, sex ratios, size at maturity, and survival of an unfished population of C. amblyrhynchos using data from a 9- year tag-recapture study in which 1399 sharks were caught at Palmyra atoll. For Palmyra’s unfished C. amblyrhynchos, the von Bertalanffy growth coefficient k was 0.05 and maximum asymptotic length �!was 163.3 cm total length (TL). Maximum size was 175.5 cm TL from a female shark, length at maturity was estimated at 116.7-120.9 cm TL for males and 126.1 cm TL for females, maximum lifespan was 15.1 years for males and 16.1 years for female sharks, and annual survival was 0.74 year-1. We conclude that Palmyra’s unfished C. amblyrhynchos had different life history traits than what has been reported from other parts of its geographic range, with different ecological contexts, and varying human impacts. However, we found no obvious patterns in growth or length for the species as a function of biogeography or fishing pressure. Ultimately, the substantial, but not predictable variability in life history traits observed for C. amblyrhynchos suggests that regional management may be necessary to set sustainable harvest targets and to recover this and other shark species globally.

Introduction

Life history traits are the realization of internal trade-offs that are increasingly affected by human impacts. Many elasmobranchs (sharks and rays) have broad, even circumglobal, distributions and therefore experience regional differences in selection pressures and anthropogenic stressors. Geographic variation has been observed in many life history traits (e.g. growth, maximum size, size at maturity, fecundity) for a number of elasmobranch species (Sminkey and Musick 1995, Carlson and Baremore 2003, Lombardi-

Carlson et al. 2003, Driggers et al. 2004, Neer and Thompson 2005, Carlson et al. 2006,

Walker 2007, Farrell et al. 2010, Mourier et al. 2013, Holmes et al. 2015, Smart et al. 2015), but the drivers of this variability are often unclear. Shark populations have been severely depleted by fishing and are globally considered overfished (Jackson et al. 2001, Myers and

Worm 2003, Baum and Myers 2004, Davidson et al. 2015), making it generally impossible to disentangle changes to population parameters due to geographic variability, ecological variability, and anthropogenic impacts.

For broadly distributed, overexploited species such as elasmobranchs, conservation management would benefit from understanding how life history traits predictably change in response to local environmental and ecological contexts. Generally such information is lacking. In fact, nearly half of the chondrichthyan species (sharks, rays, and chimaeras) are considered ‘Data Deficient’, that is, lacking requisite information to assess their status by

Red List Categories and Criteria of the International Union for the Conservation of Nature

(Dulvy et al. 2014). Problems of discerning context-specific effects on life history traits are further complicated by the fact that fishing causes complex and often mixed effects on the

life histories of target species. For example, fishing can both increase fecundity via density dependent effects (Ellis and Keable 2008) and depress fecundity via size selective harvest

(Stevens et al. 2000). Fishing can have similarly complicated effects on mortality rates

(Wood et al. 1979), maximum size (Sminkey and Musick 1995, Carlson and Baremore

2003), size and age at maturity (Walker et al. 1998, Carlson and Baremore 2003), and growth rates (Sminkey and Musick 1995, Stevens et al. 2000, Carlson and Baremore 2003).

Correcting life history traits for fishing effects requires knowledge of populations absent fishing, which is rarely possible.

Like most fishes, sharks grow continuously and asymptotically and the von

Bertalanffy growth function (VBGF; von Bertalanffy 1938) is thought to accurately describe their growth (Beverton and Holt 1959, Cailliet et al. 2006). VBGF parameters are also used as proxies in estimations of life history parameters, including natural mortality, in fisheries stock assessments (Pauly 1980, Pardo et al. 2013). A biased estimate of growth can therefore bias life history estimates, resulting in inaccurate assessments of stock status

(Pardo et al. 2013). This is problematic, because stock assessments depend on accurate life history information to set harvest targets. Given the importance of having accurate data to set proper management goals for at risk elasmobranchs, a fundamental goal of shark biology should be to synthesize life history information to understand the range of variability in a species’ life history traits across its geographic range and across a gradient of human impacts.

Although no area of the world’s ocean is unaffected by human influence (Halpern et al. 2008), Palmyra Atoll in the northern Line Islands is considered a little-disturbed ecological reference site that provides an opportunity to study a coral reef ecosystem without

significant human impacts, including extractive fishing pressure. Palmyra is a remote, U.S.

National Wildlife Refuge in the central Pacific Ocean that supports a large and temporally stable population of grey reef sharks (Carcharhinus amblyrhynchos) (Bradley et al. in review).

Carcharhinus amblyrhynchos is a reef-associated shark species, broadly distributed throughout the Indian, and western and central Pacific Oceans (Compagno 1984, Wetherbee et al. 1997). While adult female grey reef sharks are generally, but not universally, larger than males, there is a high level of variability in reported maximum total length (TL). In 13 published studies with a minimum sample size of 25 sharks, adult C. amblyrhynchos maximum TL varies by region from 152.5 cm in the Line Islands (male; Wass 1971) to 200 cm in the Marshall Islands (sex not reported; McKibben and Nelson 1986) (Table 2.1), with a maximum reported size for the species of 255 cm TL (Garrick 1982). Thorough life history analyses have been conducted on C. amblyrhynchos in the Northwestern Hawaiian

Islands (NWHI), the northern and central Great Barrier Reef (GBR), and Papua New Guinea

(PNG), encompassing a gradient of environmental, ecological, and anthropogenic impacts.

Including Palmyra, these regions span 2-28° latitude and have different ecological characteristics. C. amblyrhynchos is the most abundant reef associated predator in terms of biomass at Palmyra (McCauley et al. 2012) and the GBR (Robbins 2006, Rizzari et al. 2014) but not in the NWHI (Nadon et al. 2012) or PNG (based on commercial fisheries landings data Kumoru 2003). Competition and predation are therefore expected to exert different selection pressures on C. amblyrhynchos in each location. However, fishing pressure also varies regionally, with high fishing mortality reported for C. amblyrhynchos in PNG (Smart et al. 2016a), substantial mortality in the GBR (Robbins et al. 2006, Heupel et al. 2009),

minimal fishing mortality in the NWHI (Friedlander and DeMartini 2002), and no fishing in

Palmyra. Unsurprisingly, there is significant regional variation in nearly all life history parameters reported for C. amblyrhynchos (Table 2.1), and sharks from the GBR have a considerably lower reported VBGF growth constant k and higher asymptotic length ��!

(Robbins 2006) than those in the NWHI (Radtke and Cailliet 1984) and PNG (Smart et al.

2016a) (Table 2.2). The key challenge is disentangling the contributions of fishing intensity from geographical variation in the biological and physical characteristics of these ecosystems. One barrier to this effort is the absence of data from sites with no history of fishing.

Here, we describe the growth, maximum size, sex ratios, size at maturity, and natural mortality and survival of an unfished population of C. amblyrhynchos using data from a 9- year tag-recapture study. We propose a set of hypotheses to predict how this set of life history characteristics may have evolved in the biotic and abiotic context of Palmyra, and how this context differs from other C. amblyrhynchos populations from locations with varying environmental, ecological, and human impacts.

Methods

Ethics Statement

This project has been certified by the Institutional Animal Care and Use Committee

(IACUC), University of California, Santa Barbara, Protocol no. 856 (date of IACUC approval: 5/31/2012). Sharks were captured at Palmyra Atoll, which has been a U.S.

National Wildlife Refuge since 2001 and part of the Pacific Remote Islands Marine National

Monument since 2009, under U.S. Fish and Wildlife Service special use permits (Permit

numbers #12533-14011, #12533-13011, #12533-12011, #12533-11007, #12533-10011,

#12533-09010, #12533-08011, and #12533-07006).

Data collection

From October 2006 to October 2014, we captured and tagged C. amblyrhynchos in the forereef, backreef, lagoon, and channel habitats on Palmyra. Sharks were caught using hand lines baited with yellowfin tuna (Thunnus albacares), wahoo (Acanthocybium solandri), and/or mackerel (Scomber scombras and Decapterus macrosoma) on barbless circle hooks; once captured, individuals were restrained at the side of the boat using a tail rope. Up to three length measurements were recorded for each shark: precaudal length

(PCL), fork length (FL), and total length (TL). Sex was determined by the presence/absence of claspers, which were measured, and clasper state was assessed to determine maturity for all male individuals (calcified, partially calcified, not calcified). Uniquely numbered tags were then affixed to individual sharks. Early in the study, rototags were applied with an applicator through a hole punched in the leading edge of the first dorsal fin. Starting in 2010, minimally invasive HallprintTM dart tags with stainless steel heads were applied using stainless steel tag applicators, with the tag head implanted in the epaxial muscle near the base of the first dorsal fin. Handling time was <4 minutes on average for an individual shark and tag loss was negligible (see Bradley et al. in review for details).

Statistical analyses

We used t-tests to compare the average size of captured female and male C. amblyrhynchos, and chi-squared tests to assess whether the observed sex ratio was

significantly different from 1:1. Length at maturity TLm for male sharks was estimated by logistic regression, solving for the total length at which 50% of males had calcified claspers

(odds of calcified clasper = 1). Individuals with partially calcified claspers were not included in the maturity analysis. For comparison, length at maturity was also estimated for both sexes using the published elasmobranch-wide linear relationship between TLm and maximum length TLmax of a captured individual (TLmax = 175.5 cm) (Frisk et al. 2001) (Appendix 2.1).

Maximum lifespan Tmax for C. amblyrhynchos was estimated as a function of Tm (Frisk et al.

2001) (Appendix 2.1). In addition, length-to-length conversion formulae (i.e. TL-PCL, TL-

FL, FL-PCL, etc.) were estimated using linear regression analyses. Including gender in the slopes and intercepts of length-to-length conversion models decreased model fit (lowered

AIC), and so we report only models with sexes combined.

Using our capture-recapture data, we estimated the Francis (1988b) formulation of the von Bertalanffy growth function (VBGF) for female sharks and with both sexes combined (due to our limited sample of recaptured male sharks). Previous studies have found nearly identical growth curves in male and female requiem sharks (Simpfendorfer

2000, Meyer et al. 2014) including C. amblyrhynchos (Robbins 2006, Smart et al. 2016a).

The Francis (1988b) model includes parameters �! and �!, which are the mean annual growth increments of a species at the arbitrary reference lengths � and �, which can be used to estimate the conventional VBGF parameters �! and k (Appendix 2.2). Francis (1988b) model �! and k values estimated from tagging data have the same biological meaning as

VBGF parameters derived from age-length data (Francis 1988a): �! (cm) is the asymptotic mean length at age (i.e. the average length of an “old” shark), and k (year-1) is a growth coefficient that describes the rate at which growth approaches �!. The Francis (1988b)

model is also more flexible than other VBGF methods for tagging data in that it allows the inclusion of additional parameters that can affect model fit: growth variability (v), mean (m) and standard deviation (s) of measurement error, outlier contamination (p), and seasonal variation (u, w). Several combinations of additional parameters were considered, but we only present the best fitting models. Only recaptured individuals with minimum time-at- large of 150 days were included in the growth analysis. Previous studies of C. amblyrhynchos used TL (Radtke and Cailliet 1984, Robbins 2006, Smart et al. 2016a) and

PCL (De Crosta et al. 1984) to generate growth estimates; PCL was not measured at every sampling occasion in this study and so TL was used instead. However, ��! and ��! were estimated from ��! using the length-to-length conversions described above. All Francis

(1988b) model parameters were estimated using the R package fishmethods (Nelson 2014).

Model selection was performed using likelihood ratio tests (improved model fit indicated by a likelihood value >1.92 for one additional parameter and >3.0 for two additional parameters Francis 1988b) and by comparing AIC values. We also visually assessed model fit by plotting model residuals (observed-expected growth) against length-at- release and predicted growth; residual deviation was expected to decrease as length at release increases (L1), because the likelihood function assumes an allometric relationship between individual growth variation and mean growth, and the latter declines with length

(McGarvey et al. 1999). Residual deviation was also expected to positively correlate with predicted growth based on the assumed relationship for individual growth variability

(Appendix 2.2). We used C. amblyrhynchos reference lengths � (100 cm TL) and � (130 cm

TL), which included the majority of captured individuals, to estimate k and ��!. Ranges of

� (95-105 cm) and � (125-145 cm) values were examined, but parameter estimates were

insensitive to values within these ranges. These k and ��! parameters were then used to generate a growth curve for unfished grey reef sharks at Palmyra Atoll (sexes combined and assuming a size at birth of 60cm (Radtke and Cailliet 1984, Wetherbee et al. 1997)).

Finally, we estimated survival � and total mortality Z for C. amblyrhynchos at

Palmyra Atoll, with the expectation that Z = M (natural mortality) given the absence of fishing. Tag-recapture data was used to formulate a restricted dynamic occupancy version of the Jolly-Seber (JS) model (Royle and Dorazio 2008) to estimate probability of annual survival (Appendix 2.3). We used a Bayesian analysis and specified a uniform prior U(0,1) on our survival parameter. Our model was estimated using data augmentation with Markov chain Monte Carlo (MCMC) sampling in the R package rjags (Plummer 2014). We ran 3 chains for 25,000 iterations and discarded the first 5,000 runs as burn-in. We examined traceplots and the Gelman-Rubin � statistic (values <1.1 indicate convergence) to assess model performance. Total mortality was estimated using the Hoenig (1983) equation (Z as a function of Tmax; Appendix 2.4) parameterized separately for teleost fishes and for cetaceans, which have demographic characteristics more similar to sharks than teleosts (Smith et al.

1998). All statistical analyses were performed in R (Team 2014).

Results

We captured 1399 individual C. amblyrhynchos by research fishing on Palmyra between 2006-2014; of these, 1356 individuals were tagged with either a dart and/or roto tag. The sample was slightly female biased (male:female ratio = 0.52, �!=135.8, p<0.001), and female sharks were significantly larger than male sharks (Fig. 2.1): the average captured female was 146.0 cm TL (SD=16.6 cm) while the average captured male was 138.7 cm TL

(SD=14.2 cm) (t1109.8=8.6; p<0.001). The largest captured shark was a 175.5 cm TL female, and the smallest captured individuals were four 66 cm TL females. Length-to-length relationships had conversion estimates with R2 > 0.90 (Table 2.3). The smallest captured mature (calcified claspers) male shark was 112.5 cm TL and the largest immature

(uncalcified claspers) male was 146 cm TL (Fig. 2.2A); male C. amblyrhynchos had a 50% chance of reaching maturity at 120.9 cm TL (logistic regression, p<0.001) (Fig. 2.2B).

Length at maturity TLm estimated as a function of TLmax (Frisk et al. 2001) was 116.7 cm for male sharks and 126.1 cm for female sharks. Maximum lifespan Tmax was 15.1 years for male sharks, and 16.1 years for female sharks.

Of the 1356 captured and tagged C. amblyrhynchos, 118 individuals were recaptured a minimum of 150 days after initial capture (females = 100; males = 18). Median time at large was 2.0 years and the longest time between release and recapture was 7.2 years. The maximum total growth recorded for a shark was 46.5 cm from a female captured at 103 cm

TL and recaptured 6.8 years later at 149.5 cm TL. The assumptions of the Francis (1988b) model were confirmed by plotting residuals against length at release and predicted growth.

As expected, residual deviance increased over predicted growth and decreased over length at release, reinforcing the expectation of the von Bertalanffy growth function (VBGF) that growth variability should decline at increasing length (Fig. 2.3).

The female-only and combined-sexes growth models produced consistent results, indicating that including male sharks did not bias model estimates or increase standard errors. The growth model with the best fit to the VBGF for our capture-recapture data included �! (�!""), �! (�!"#), standard deviation of measurement error, and growth variability. The inclusion of parameters for mean of measurement error, seasonal variability,

and outliers did not significantly improve model fit (Table 2.4). Growth at size was estimated at 3.33 cm/year (�!"") and 1.75 cm/year (�!"#) for our best model (Table 2.4).

The growth coefficient k was 0.054 and ��!was 163.3 cm (Table 2.4). A size at birth of 60 cm, reported in (Radtke and Cailliet 1984, Wetherbee et al. 1997), was used to fit a von

Bertalanffy (von Bertalanffy 1938) growth curve to the data (Fig. 2.4). Annual survival rate estimated from tag-recapture data was 0.739 year-1 (95% CI 0.703-0.775; �= 1.04). Total mortality Z from the Hoenig (Hoenig 1983) equation was 0.226 year-1 and 0.260 year-1 (for teleost and cetacean parameterizations, respectively).

Discussion

At an unfished Pacific coral reef, C. amblyrhynchos had relatively slow growth and small asymptotic length compared to most previously studied populations. Small individuals grew 3.3 cm/year (�!"") and larger individuals grew just 1.8 cm/year (�!"#), equating to an estimated growth coefficient k of 0.05, substantially lower than the maximum k of 0.29 reported for C. amblyrhynchos in the Northwest Hawaiian Islands (NWHI), despite a small difference in asymptotic length ��! between the two locations (Table 2.2). Asymptotic length ��!was equivalent and relatively small in Palmyra and Papua New Guinea (PNG)

(��!=163 cm), followed by the NWHI (��! = 173.3 cm; Radtke and Cailliet 1984), and then the Great Barrier Reef (GBR) (��! = 229.2 cm; Robbins 2006) (Table 2.2). Female sharks were larger than males at Palmyra (Fig. 2.1), consistent with most other studies

(Table 2.1). However, C. amblyrhynchos in the largest size classes that have been reported elsewhere (>180cm; Wetherbee et al. 1997) were not encountered in our sample. Very large

and very small C. amblyrhynchos are also rarely observed by SCUBA divers in any of

Palmyra’s forereef, backreef, or lagoon habitats (D. Bradley, personal observation).

One possible interpretation for the differences in growth and length observed in

Palmyra compared to C. amblyrhynchos in other parts of its range is that there is no fishing in Palmyra. However, although unregulated, fishing pressure was minimal in the NWHI at the time of the De Crosta et al. (1984) study (Kittinger et al. 2010), likely due to the remoteness of the island chain. Nevertheless, observed C. amblyrhynchos life history characteristics were dramatically different in Palmyra and the NWHI (Table 2.2), suggesting that biotic and abiotic factors likely exert significant selective pressures on C. amblyrhynchos life history traits. Both the NWHI and Palmyra have high predator densities

(Friedlander and DeMartini 2002, Nadon et al. 2012), but in Palmyra, where growth rate and asymptotic length were relatively low, C. amblyrhynchos is the competitively predominant shark species in forereef habitat (Papastamatiou et al. in review), and conspecifics do not comprise a known component of their diet (Wetherbee et al. 1997). In contrast, in the

NWHI, where growth rate is comparatively high and asymptotic length is larger than

Palmyra, C. amblyrhynchos are outnumbered by the larger Galapagos shark (C. galapagensis) (Nadon et al. 2012), which regularly consume other elasmobranchs

(Wetherbee et al. 1996), likely including juvenile C. amblyrhynchos. While extremely rare in Palmyra, tiger sharks (Galeocerdo cuvier) are also common in the NWHI where elasmobranchs make up 20% of their diet (Lowe et al. 1996). We hypothesize that the faster growth observed in the NWHI was driven by juveniles employing a strategy to reduce predation risk, which has been suggested for other elasmobranchs (Simpfendorfer et al.

2008, Meyer et al. 2014). There is an ontonogenic shift in the composition and diversity of

prey consumed by many elasmobranchs including C. amblyrhynchos (Wetherbee et al.

1997), suggesting that juvenile sharks threatened with predation in the NWHI could target different prey species to fuel their faster growth, which has been observed in other elasmobranch species (e.g. Simpfendorfer et al. 2008, Meyer et al. 2014).

Carcharhinus amblyrhynchos are likely at carrying capacity at Palmyra (Bradley et al. in review), and food resources may therefore be limited due to high levels of intraspecific competition. Consumptive competition across all size and age classes is expected to reduce overall resource acquisition, diminishing growth and preventing sharks from reaching the largest sizes (Stevens 1984, Papastamatiou et al. 2009a), consistent with our findings (Table

2.2). Ultimately, we hypothesize that resource limitations due to intraspecific competition may limit growth and maximum size of C. amblyrhynchos in Palmyra, while risk of predation in the NWHI may favor a different resource acquisition strategy that results in faster growth and larger maximum size.

The growth coefficient we estimated for Palmyra C. amblyrhynchos was identical to that reported for the Great Barrier Reef (GBR) population (k=0.05; Robbins 2006) (Table

2.2), where commercial fisheries and poaching activities have substantially reduced C. amblyrhynchos abundance (Robbins et al. 2006, Heupel et al. 2009, McCook et al. 2010). If exploited sharks responded to reduced competition, we would expect higher growth in the fished GBR compared to unfished Palmyra, assuming that the prey resource base was not equally affected by fishing, which could be expected given the targeted poaching of sharks thought to occur in the GBR (Robbins et al. 2006, McCook et al. 2010). Instead the low growth observed for C. amblyrhynchos in both locations indicates that a density-dependent growth response from reduced competition may not be operating in this shark species. In

general, shark fishing is also expected to decrease maximum size (Stevens et al. 2000), yet

GBR sharks had the largest reported �! (��! = 229 cm; Robbins 2006) (Table 2.2). At the same time, intermediate growth and small asymptotic length were reported for heavily fished C. amblyrhynchos at PNG (k=0.15; ��! = 163 cm; Smart et al. 2016a) (Table 2.2);

PNG C. amblyrhynchos have been harvested as a target species since the early 1980’s

(Kumoru 2003), indicating that a density-dependent compensatory growth response could be possible given substantial harvest pressure and sufficient time to allow the population to respond.

Geographic variation is also thought to affect population parameters in elasmobranchs, with asymptotic length and growth generally increasing along a latitudinal cline (Lombardi-Carlson et al. 2003, Kimura 2008, Dureuil and Worm 2015). The growth constant k was in fact largest in the highest latitude NWHI (23.1-28.2° N; De Crosta et al.

1984), but k was identical in the GBR (14.7-19.3° S; Robbins 2006) and Palmyra (5.8-5.9°

N), which are located at substantially different latitudes with minimum sea surface temperatures that range from 19.0 C° in the NWHI to 27.3 C° in Palmyra (Gove et al. 2013).

Growth coefficients from PNG (2-11° S) and Palmyra were also not equivalent, despite being located as similar latitudes. In addition, expected geographic trends in �! were not observed, with the highest ��! reported from the GBR, which was not the highest latitude location. There was also no latitudinal pattern in maximum reported length for C. amblyrhynchos from studies across the species’ Indo-west and central Pacific range (Table

2.1). We offer the hypothesis that locally influences environmental factors that do not correlate with latitude may be more important in exerting selection pressure on C. amblyrhynchos life history traits than those that do follow a latitudinal cline. There are also

substantial differences in habitat availability, complexity, and connectivity in Palmyra, the

NWHI, PNG, and the GBR. Palmyra is a small (12 km2), isolated atoll; the NWHI is a relatively more dense chain of atolls, islands, and seamounts of varying sizes; the Bismarck and Solomon Seas are surrounded by large islands (e.g. PNG) and archipelagos; and the central GBR is comprised of a large network of mid-shelf reefs with near continuous coral reef habitat. The physical structure of coral reefs plays a fundamental role in determining the strength of biological interactions and the allocation of available energy to life history traits

(Almany 2004).

The sex ratio of captured C. amblyrhynchos at Palmyra was skewed towards females.

This is consistent with observations from northeastern Australia (Barnett et al. 2012), and

Palau (Vianna et al. 2013); however, sex ratios were skewed towards males in the Main and

Northwest Hawaiian Islands (Wetherbee et al. 1997) and in PNG (Smart et al. 2016a). This likely reflects a bias in sampling location and corresponding sampling method and not an actual skewed sex ratio. Sexual segregation is a common feature observed in marine species

(Wearmouth and Sims 2008) including reef-associated sharks (Papastamatiou et al. 2009a,

Knip et al. 2012, Chin et al. 2013); coastal aggregations of C. amblyrhynchos tend to be dominated by female sharks (Stevens 1984, McKibben and Nelson 1986, Economakis and

Lobel 1998, Speed et al. 2011, Vianna et al. 2013), while male individuals have been observed dispersing farther and exhibiting lower site fidelity (Heupel et al. 2010, Espinoza et al. 2014). It therefore is not surprising that near-shore handline fishing was used in all cases where female sex ratios were observed, and offshore or deeper water longline fishing was the primary method of sampling that revealed a male biased sex ratio. When a variety of fishing methods were employed, sex ratios were not significantly different from unity in

both northern Australia (Stevens and McLoughlin 1991) and Madagascar (Robinson and

Sauer 2013). Method of capture and capture location should therefore be an important consideration in future elasmobranch studies.

Survival and natural mortality rates were within range previously reported values for

C. amblyrhynchos. Survival estimated from tag-recapture data was 0.74 year-1 (95% CI

0.70-0.78 year-1). Using the relationship between maximum size of a captured individual and total mortality reported in Hoenig (1983), natural mortality rate M (assuming that total mortality=natural mortality in an unfished system) for C. amblyrhynchos at Palmyra was

-1 -1 0.23-0.26 year , compared to M = 0.04-0.17 year (GBR (Hisano et al. 2011)) and M = 0.25 year-1 (NWHI (Smith et al. 1998); using data from De Crosta et al. 1984, Radtke and Cailliet

1984).

We note that previous authors have suggested a multimodel approach to estimate growth for sharks (Dureuil and Worm 2015, Smart et al. 2016b), particularly for tag- recapture samples (Dureuil and Worm 2015). For this analysis, other growth methods were considered (e.g. Gulland and Holt (1959), Fabens (1965), Fabens with a fixed asymptote

(James 1991)) but none were deemed as appropriate for our data as the Francis (1988b) method. The Fabens (1965) method is known to produce biased estimates of k and �! when individual growth is variable (Sainsbury 1980, Francis 1988a, Maller and DeBoer 1988,

Kimura et al. 1993) or measurement error is high (Eveson et al. 2007), indicating that the

Fabens (1965) model would likely overestimate length at time for our data (Sainsbury

1980). Previous studies have found that the Gulland and Holt (1959) method tends to produce growth estimates that are similar to or less biologically realistic than the Francis

(1988b) method (Sainsbury 1980, Skomal and Natanson 2003, Kneebone et al. 2008, Meyer

et al. 2014). The Francis (1988b) method is certainly not free from bias, but it is generally the least biased of the commonly used growth models (Dureuil and Worm 2015), and it has been found to be a better fit than other methods for tag-recapture data for other requiem shark species (Skomal and Natanson 2003, Meyer et al. 2014). McAuley et al. (2006) suggested that Francis (1988b) method bias was primarily due to variability in growth when time at liberty was short, sample size was small, and when data contained outliers. We used a minimum time at liberty of 150 days, our sample size is larger than many tag-recapture growth studies (e.g. Skomal and Natanson 2003, Meyer et al. 2014, Dureuil and Worm

2015), and outliers did not significantly contaminate our model fit. Other studies found that the Francis (1988b) method can variably bias k and �! in samples where old sharks were absent (Natanson et al. 2002, Skomal and Natanson 2003); however, given that our size range is slightly skewed towards larger individuals (Fig. 2.1), we are confident that the oldest sharks are well represented in our sample. Ultimately, the Francis (1988b) method was preferred over other methods, because our best fit model included parameters for individual growth variability and measurement error, indicating the importance of using a model that can capture these sources of bias that were present in our data.

Ultimately, we found that Palmyra’s unfished C. amblyrhynchos had different life history traits than what has been reported from other parts of its geographic range, with different ecological contexts, and varying human impacts. However, we found no obvious patterns in growth or length for the species across a latitudinal cline. Sharks in the heavily fished PNG had a growth constant slightly higher than those in Palmyra, suggesting that a C. amblyrhynchos may experience a density-dependent growth response to intensive, prolonged exploitation. By comparison, GBR sharks had low rates of growth despite

reported interactions with commercial and illegal fisheries (Robbins et al. 2006, Heupel et al. 2009, McCook et al. 2010). Critically, C. amblyrhynchos from the near-pristine NWHI and unfished Palmyra had substantial differences in all reported life history traits. We hypothesize that these differences may be driven by differences in ecological context, with resource limitations due to intraspecific competition resulting in slowed growth and depressed size in Palmyra, and risk of predation favoring a strategy of fast growth through the juvenile stage and larger asymptotic size. As managers must accelerate their efforts to recover overexploited shark populations around the world, it will be increasingly important to synthesize available biological information and consider how local environmental factors, ecological context, and anthropogenic impacts exert varying selection pressures on demographic parameters. In the meantime, the substantial, but ultimately unpredictable variability in life history traits observed for C. amblyrhynchos across its geographic range suggests that regional management may be necessary to set sustainable harvest targets and to recover this and other shark species globally.

Acknowledgements

We thank The Nature Conservancy staff for support at the Palmyra Atoll research station and our volunteer field assistants J. Calhoun, R. Carr, J. Eurich, A. Filous, J. Giddens, M.

Hutchinson, S. Larned, R. Most, R. Pollock, K. Stamoulis, J. Schem, M. Shepard, R. Sylva,

Y. Watanabe, and T. White. Funding was provided by The Nature Conservancy, Hawai’i, and a Marisla Foundation grant (JEC). DB was supported by a NSF Graduate Research

Fellowship (DGE-114408), a Dr. Daniel Vapnek Fellowship (UCSB Bren School), and an

AAUS Kathy Johnston English Scholarship.

Table 2.1. Life history characteristics for Carcharhinus amblyrhynchos from across its geographic range (only studies with a minimum of N=25 sharks were included).

Length Region Max length Maturity at birth Litter (~latitude)* (cm) [sex] (cm) [sex] (cm) size N Sample Reference

Non Australia ------extractive (central GBR) 142 FL 40 scientific (Espinoza (18.5-19.0° S) [female] fishing et al. 2014)

118 TL Commercial Australia [male]; and scientific (GBR) 170 TL 130-142 TL 54-61 TL 1-4 199 reef-line (Robbins (14.5-19.0° S) [female] [female] fisheries 2006)

Non Australia ------27 extractive (northeast) 182 TL scientific (Barnett et (14.0° S) [female] fishing al. 2012)

Commercial (Stevens Australia 137-140 TL 63 TL 2-3 94 gillnet fishery and (northern) 178 TL [female] and research McLoughli (10.0-20.0° S) [female] cruises n 1991)

Australia Non (Heupel (southern extractive and GBR) 150 FL ------28 scientific Simpfendor (23.5° S) [female] fishing fer 2014)

Non ------25 extractive Johnston Atoll 135 FL scientific (Skomal et (17.0° N) [male] fishing al. 2007)

Cooperative Johnston ------25 shark research Atolla 147.4 TL and control (Wass (17.0° N) [female]b program 1971) (Robinson Madagascar >170 TL Commercial and Sauer (12.0° S) [both] ------134 gillnet fishery 2013)

Marshall Non Islands extractive (McKibben (Enewetak) 200 TL [not ------31 scientific and Nelson (11.0-12.0° N) reported] fishing 1986) Marshall 152.5 TL Islands [female]b 85-90 PCL Cooperative (Wass

(Enewetak)a [male]; shark research 1971) (11.0-12.0° N) 96-97 PCL -- -- 76 and control [female] program

Cooperative Northern Line ------79 shark research Islandsa 152.5 TL and control (Wass (4.0-6.0° N) [male]b program 1971)

Non ------39 extractive Palau 158 TL scientific (Vianna et (7.0° N) [female] fishing al. 2013)

117-121 TL Non [male]; -- -- 1399 extractive Palmyra Atoll 175.5 TL 126 TL scientific (6.0° N) [female] [female] fishing This study

123 TL Papua New [male]; Commercial Guinea 182 TL 136 TL 71-73 TL -- 133 longline (Smart et (2.0-12.0° S) [male] [female] fishing al. 2016a)

120-140 TL Cooperative USA (MHI [male]; shark research and NWHI) 190 TL 125 TL 60 TL 3-6 367 and control (Wetherbee (18.5-28.5° N) [female] [female] program et al. 1997)

168.8 TL (Radtke USA (NWHI) [not ------59 Longline and Cailliet (23.0-28.5° N) reported] fishing 1984)

Cooperative -- -- 3-6 274 shark research USA (Hawaii) 187 TL and control (Tester (18.5-22.0° N) [female] program 1969)

Cooperative USA 100 PCL 105 PCL 3-6 28 shark research (Hawaii)b 171.8 TL and control (Wass (18.5-22.0° N) [male]c program 1971)

a Latitude rounded to the nearest 0.5° b as Carcharhinus menisorrah c PCL reported; regression of PCL on TL in (Wass 1971): �� = 0.78�� − 3.022

Table 2.2. Parameter values estimated for growth models for Carcharhinus amblyrhynchos.

Location ��! (cm) ��! (cm) ��! (cm) k N Method Reference

GBR Australia (14.7-19.3° S) 229.2 188.61a 173.33a 0.05 89 VBGFb (Robbins 2006) Papua New Guinea (Smart et al. (2.0-12.0° S)c 163 --d -- 0.15 133 VBGF 2016a) NW Hawaiian Islands (De Crosta et al. (23.1-28.2° N) 173.3e -- 134 0.29 62 VBGF 1984) NW Hawaiian Islands (Radtke and (23.1-28.2° N) 187.9 ------59 VBGF Cailliet 1984)

Palmyra Atoll Francis (5.8-5.9° N) 163.3 137.7f 122.7f 0.05 118 (1988b) This study a Calculated using length-to-length relationship reported in (Robbins 2006) b VBGF is the von Bertalanffy growth function (von Bertalanffy 1938). c Minimum and maximum values for Bismarck and Solomon Seas; no latitude values reported in (Smart et al. 2016a) d “--" indicates an unavailable value e Calculated using length-to-length relationship reported for Hawaiian C. amblyrhynchos in (Wetherbee et al. 1997): TL = 4.146 + 1.262PCL f Calculated using length-to-length relationships reported in this study

Table 2.3. Length to length relationships for Carcharhinus amblyrhynchos (sexes combined) captured by research fishing in Palmyra 2006-2014.

2 x y b0 b1 r df p

TL FL -1.12(0.71) 0.85(0.00) 0.96 1390 <0.001

TL PCL -6.37(0.96) 0.79(0.01) 0.93 1115 <0.001

FL TL 7.67(0.79) 1.13(0.01) 0.96 1390 <0.001

FL PCL -1.74(0.88) 0.91(0.01) 0.93 1114 <0.001

PCL TL 17.88(1.06) 1.17(0.01) 0.93 1115 <0.001

PCL FL 9.78(0.89) 1.03(0.01) 0.93 1114 <0.001

Linear regression coefficients for the model yi = b0 + b1xi. Numbers in parentheses are standard errors. PCL, pre-caudal length (cm); FL, fork length (cm); TL, total length (cm).

Table 2.4. Parameter values from the best fit estimated Francis (Francis 1988b) growth models for Carcharhinus amblyrhynchos at Palmyra Atoll.

a Parameters held fixed b Best model

Figure 2.1. Total length frequency of female and male C. amblyrhynchos captured by research fishing in Palmyra 2006-2014.

Figure 2.2. Grey reef male maturity. (A) Clasper size (cm) as a function of TL; NC = not calcified (N=35); PC = partially calcified (N=33); C = calcified (N=361). (B) Only calcified

(C=1) and not calcified (NC=0) individuals were included in the logistic regression.

Figure 2.3. Residual plots to assess Francis (Francis 1988b) model fit.

Figure 2.4. Von Bertalanffy growth curves for Palmyra C. amblyrhynchos compared to other regions (NWHI, GBR, PNG); both sexes combined. Size at birth based on what is reported in each corresponding reference.

Figure 2.1

4040 ! Female (N=913)! Male (N=479)

3030 !

2020 ! number of sharksnumber Numbersharksof

1010 !

00 60-80 81-100 101-120 121-140 141-160 161-180 total length Total length (cm)

Figure 2.2

2020! ● A ● ! ● ● ● ● ● ● ●●● ● ●●●●●●●●● ●●●●●● ● ●● ●●● ●● ●● ● ● ● ●●●● ●●●●●●●●●●●●●●●●●●●●●●●●● ●●●● ● ● ●●●●●● ● ●● ● ● ● ● ● ● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●● ●● ● ● ●● ●●●● ● ●●●●● ● ● 1515! ● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●● ●● ●●●● ● ●●● ● ● ● ● ●●●●●●● ●●●● ● ● ●● ● ●● ● ● ● ● ● ●

● 1010! ● Clasperlength (cm)

C! 55 NC! PC

100 120 140 160 100 120 140 160 TL (cm)

1.001.00! B ●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

0.750.75! logit(p) = 0.30(TL) - 37.52

0.500.50!

0.250.25! Probabilitymaturity(calciﬁedof 1)= 0.000.00 ● ● ●● ●●●●●●●●●●●●●●●●●●● ● ●

75 100 125 150 175 75 100 125 150 175 TL (cm)

Figure 2.3

A ● 2525 !

2020 !

1515 !

1010 ! ●

Residual ● ● ● ● ● ● ● Residual ● ● ●

5 ● 5! ●● ● ● ● ●● ●●● ●● ● ● ● ● ● ● ● ●●●●● ●● ● ● ●●●●● ● ● ● ● 0 ●●●●●●●●● 0! ●●● ●●●● ●● ● ● ●● ●●● ● ● ●●● ● ● ●● ● ● ● ●●● ● ● ● 5 ●●● ● ● -5− ● ● ●

0 5 1010 1515 2020 Predicted growth (TL cm/yr) Predicted growth (TL cm yr-1)

●

25!25 B

20!20

15!15

10!10 ●

Residual ● ● ● ● ● ● ● ● ● ● ● 5 Residual 5! ● ● ● ● ● ● ● ● ●● ● ● ●●● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●●● 0!0 ● ●● ● ●●●●●● ●●● ●● ● ●●● ● ● ●●● ● ●● ● ●● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ●● 5 ● ● ● ● ●

-5 − ● ● ●

100 110 120 130 140 150 160 170 100 110 120 130 140 150 160 170 Length at release (TL cm) Length at release (TL cm)

Figure 2.4

Appendices

Appendix 2.1. Frisk, Miller, and Fogarty (2001) elasmobranch empirical relationships

Frisk, Miller, and Fogarty (Frisk et al. 2001) quantified the relationship between body size

(total length) and length at maturity and age at maturity for 150 elasmobranch species including requiem sharks. Length at maturity Lm was significantly related to maximum length Lmax

�! = 0.70 �!"# + 3.29. (1.1)

They also found a significant relationship between age at maturity Tm and total maximum length Lmax for requiem sharks:

�! = 5.92 �� !"# − 23.25. (1.2)

The relationship between age at maturity Tm and maximum lifespan Tmax was reported as

�! = 7.20 ��(�!"#) − 12.68. (1.3)

Appendix 2.2. Francis (1988) growth model

The Francis (Francis 1988b) formulation of the von Bertalanffy growth function (VBGF) for tag-recapture data describes the expected growth from a fish of initial length L over some time period Δ�:

!!!! !!! !!! !! ! ! ∆� = !! , (2.1) !!! !! !! !! !!!

where �! and �! are the mean annual growth increments of a species at reference lengths � and � (which should be chosen to include a substantial proportion of by the tagging data within their range). We set Δ�=1 and standardized growth to an annual timestep. Parameters

�! and �! can be used to estimate the conventional parameters �! and k of the VBGF by the equations

!!!! !!! �! = ; (2.2) !!! !!

! ! ! � = −�� 1 + ! ! . (2.3) !!!

The Francis model is flexible in that it allows the addition of additional parameters.

Assuming that the growth of a shark of length L over some time period is normally distributed with mean � and standard deviation �, then growth variability can be described using a single parameter v where

� = � �. (2.4)

If this mean-variance relationship results in inadequate model fit, then additional parameters can be introduced (Francis 1988b), but this was not necessary for our data. Outliers can also

bias growth model parameters, but may represent true values that should not necessarily be discarded. The contamination probability p can be added to ensure that extreme data points have minimal effect on growth parameters (as long as outliers are somewhat rare). Finally, mean m and standard deviation s of measurement error in ∆� can be modeled, and the log likelihood function can be rewritten as

! � = !!! �� 1 − � �! + �/� , (2.5)

! ! ! !!.! ∆!!! !!!! /(!! ! ! ) where �! = exp ! ! !.! (2.6) [!! !! ! ! ]

R is the range of observed growth increments ��! and the likelihood is summed over all observed growth increments. We estimated the model using the grotag function with limited memory, bound-constrained BFGS maximization in the fishmethods package (Nelson 2014) to find the set of parameters that maximizes �.

Appendix 2.3. Jolly-Seber annual survival (�)

Royle and Dorazio (2008) formulated the Jolly-Seber (JS) for capture-recapture data as a restricted dynamic occupancy model where individuals can be in one of three states: “not yet entered”, “alive”, “dead” (Kéry and Schaub 2012). Transitions between these states are determined by the ecological processes entry and survival, which are estimated. We were interested in the probability of annual survival �, and so we estimated a model with an annual time step where the state of an individual i in the first year is

�!,! ~ ��(�!), (3.1)

where � is the probability that a “not yet entered” individual enters the population, and

�!,! = 1 if an individual is “alive” and present, and �!,! = 0 if an individual is “dead” or has

“not yet entered” the population (Kéry and Schaub 2012). Subsequent states of each individual are determined by survival for live individuals already in the population (�!,! = 1) or by recruitment to the population for a new individual (�!,! = 0) such that

! �!,!!! | �!,!, … , �!,! ~ �� !,!�!,! + �!!! !!!(1 − �!,!) , (3.2)

where �!,! is the probability of survival for individual i between year t and t + 1. The observation process conditions on the above state process as

�!,! | �!,! ~ ��(�!,!�!,!), (3.3)

where p is the probability of capture. We used a Bayesian analysis and specified uniform priors U(0,1) for all estimated parameters (�, �, p) to express our ignorance about their values (Kéry and Schaub 2012). The model was formulated in the JAGS language with

Markov chain Monte Carlo (MCMC) sampling available in the R package rjags (Plummer

2014).

Appendix 2.4. Hoenig (1983) total mortality (Z)

The Hoenig (1983) method of estimating total mortality (Z) is parameterized around the observed relationship between longevity (Tmax) and mortality. The equation takes the form

�� = � + � ��(�!"#), (4.1)

where a and b are fitted parameters, and Tmax is the maximum observed age in the catch. The equation is parameterized separately for teleost fishes (a = 1.46, b = -1.01) and cetaceans (a

= 0.941, b = -0.873), both of which have been used for sharks (Smith et al. 1998, Hisano et al. 2011). We assumed that Z was equal to natural mortality M given the absence of fishing at Palmyra. Tmax was estimated from equation 1.3.

CHAPTER 3

NO PERSISTENT BEHAVIORAL EFFECTS OF SCUBA DIVING ON

MARINE PREDATORS

Darcy Bradley1, Yannis P. Papastamatiou2, Jennifer E. Caselle3

1 Bren School of Environmental Science and Management, University of California, Santa

Barbara, CA 93106, U.S.A.

2 Department of Biological Sciences, Florida International University, North Miami, FL

33181, U.S.A.

3 Marine Science Institute, University of California, Santa Barbara, CA 93106, U.S.A.

Abstract

Despite rapid growth in the marine tourism sector, the impacts of recreation on the marine environment are generally not well understood. Most existing studies of marine recreation ecology have focused on behavioral changes resulting from direct interactions between humans and wildlife including provisioning. However, non-consumptive, non- provisioning human impacts may also result in persistent behavioral impacts to shark populations. In this study, we examined differences in residency, abundance, and behavior of reef sharks in response to scuba diving intensity using a combination of survey techniques including baited remote underwater video systems and multi-year passive acoustic monitoring at Palmyra Atoll. In most locations with recreational diving operations, some level of human impact is pervasive, but on Palmyra, extractive fishing is prohibited, and scientific diving activities are concentrated on just a few sites that house long-term monitoring projects. These sites experience relatively intensive diving, while the majority of the island is entirely undived. Evidence from elsewhere has shown that sharks behaviorally respond to people in the water over short time scales, but our results indicate that this response may not persist. We did not detect differences in reef shark abundance or behavior between heavily dived and undived locations, nor were there differences in shark residency patterns at dived and undived sites in a year with substantial diving activity and a year without any diving. Our results suggest that humans can interact with reef sharks without persistent behavioral impacts and that well-regulated shark diving tourism can be accomplished without undermining conservation goals.

Introduction

To combat pervasive overexploitation of reef sharks (Baum and Myers 2004,

Davidson et al. 2015, Ferretti et al. 2008, Robbins et al. 2006, Ward-Paige et al. 2010), several countries have taken legislative action to ban the harvest and sale of sharks and shark products (Daves and Nammack 1998, Cortes and Neer 2006, Sybersma 2015) while encouraging shark-watching tourism (Topelko and Dearden 2005, Dobson 2006). Shark tourism has contributed millions of dollars annually to many economies (reviewed in

Topelko and Dearden 2005, Gallagher and Hammerschlag 2011), and shark diving can generate higher economic returns than shark fishing (Anderson and Ahmed 1993, Gallagher and Hammerschlag 2011, Vianna et al. 2012, Cisneros-Montemayor et al. 2013). Many shark species are relatively long-lived, so even with a high discount rate, revenue from a live animal can accrue over many years.

Although studies have lauded non-consumptive tourism and recreation for its positive benefits to ocean ecosystems (Halpern et al. 2012), there is also a negative side to ecotourism that can degrade habitats, disrupt ecosystems, and alter animal behavior, even reducing fitness in some cases (reviewed in Boyle and Samson 1985, Boo 1991, Krüger

2005). Tourism activities on land have reduced breeding success (Ellenberg et al. 2006), decreased foraging and feeding rates (Müllner et al. 2004, Yasué 2005), and lowered juvenile survival rates (Müllner et al. 2004) in several bird species. Similarly, human disturbance has altered activity budgets and led to increased vigilance and decreased foraging in elk (Naylor et al. 2009, Ciuti et al. 2012). While the field of recreation ecology is well developed in the terrestrial literature, impacts of recreation on the marine environment are far less well understood, despite rapid growth in the marine tourism sector and

promotion of ecotourism as justification for increased marine protection. However, there is a growing interest in documenting non-consumptive human impacts on marine systems

(reviewed in Madin et al. 2015). For example, there is evidence that human recreational activities can disrupt activity budgets of seals and orcas (Kovacs and Innes 1990, Williams et al. 2006, Stafford-Bell et al. 2012), reduce social interactions in southern right whales

(Vermeulen et al. 2012), and cause acute behavioral changes and chronic spatial displacement in bottlenose dolphins (Scarpaci et al. 2000, Constantine and Zealand 2001,

Constantine et al. 2004, Lusseau 2004). Following intense tourism activity and provisioning, the naturally solitary southern stingray developed shoaling behavior, changed feeding habits, and experienced a higher rate of injury and larger parasite load resulting in an overall lower body condition than stingrays at non-tourist sites (Shackley 1998, Semeniuk and Rothley

2008). Hawksbill turtles may also decrease time spent eating and breathing in response to an approaching scuba diver (Hayes et al. 2016), and reef fish may forgo cleaning opportunities in the presence of divers (Titus et al. 2015).

Despite significant resources going towards establishing shark sanctuaries and promoting shark diving tourism, the impact of recreational scuba diving on shark species is largely unknown (Vianna et al. 2012). This question is pressing, because disturbing apex predator populations can induce cascading changes to marine communities (Heithaus et al.

2008, Babcock et al. 2010), and indirect behavioral effects resulting from human impacts are likely to disrupt marine ecosystems (Madin et al. 2015). Instead, the shark tourism literature has largely focused on provisioning, and how it may result in behavioral changes that alter the ecological function of top predators. For example, shark feeding caused changes in daily vertical space use, activity budgets, and metabolic rate of whitetip reef sharks (Triaenodon

obesus; Fitzpatrick et al. 2011, Barnett et al. 2016), increased residency in sicklefin lemon sharks (Negaprion acutidens; Clua et al. 2010), and increased local abundance in bull sharks

(Carcharhinus leucas; Brunnschweiler and Baensch 2011), all of which are likely to directly and indirectly affect other species inhabiting the reef. However, over longer time scales, provisioning had no effect on residency and space use in Caribbean reef sharks

(Carcharhinus perezii; Maljković and Côté 2011) and tiger sharks (Galeocerdo cuvier;

Hammerschlag et al. 2012), had only minimal behavioural effects that decreased with time in white sharks (Carcharodon carcharias; Laroche et al. 2007), but resulted in significant changes to the relative abundance of several shark species at a multi-species shark feeding site in Fiji (Brunnschweiler et al. 2014) and to shark communities in Hawaii (Meyer et al.

2009). There is some evidence that the presence of scuba divers absent provisioning can also lead to changes in shark behaviour in the short term, with sharks both evading and approaching groups of recreational divers (Quiros 2007, Cubero-Pardo et al. 2011) and changing activity budgets while in the presence of scuba divers (Smith et al. 2010, Baronio

2012). Scuba divers also caused the spatial displacement of grey nurse sharks over short time scales (Carcharias taurus; Baronio 2012).

There are two major concerns that emerge from these findings. First, instantaneous changes in shark behavior in response to scuba divers in the water can bias abundance estimates from diver-based, non-instantaneous underwater visual surveys (UVS; Watson et al. 1995, Willis et al. 2000, Ward-Paige et al. 2010b). If sharks avoid divers (e.g. Quiros

2007, Cubero-Pardo et al. 2011), UVS may underestimate shark density. Conversely, if sharks approach divers (e.g. Cubero-Pardo et al. 2011), density estimates will be biased high. Failing to account for acute diver impacts on shark behavior may be problematic,

because in most locations it is impossible to differentiate declines in estimated population abundance due to behavioral modifications and those resulting from fishing (e.g. Robbins et al. 2006, McCook et al. 2010). Second, non-consumptive, non-provisioning human impacts may result in chronic spatial displacement, as has been observed in large terrestrial and marine mammals, which shifted space use away from human activities (Constantine and

Zealand 2001, George and Crooks 2006). If sharks permanently shift their space use away from reefs frequented by recreational divers, then human presence could undermine spatial protection strategies, such as marine refuges and shark sanctuaries, by forcing sharks out of protected areas and into zones open to fishing.

In this study, we set out to resolve this second concern by assessing chronic differences in the space use, abundance, and behavior of reef sharks at heavily dived and undived sites at Palmyra Atoll. Palmyra is a remote, historically uninhabited, U.S. National

Wildlife Refuge in the central Pacific Ocean that supports a large population of blacktip reef

(Carcharhinus melanopterus) and grey reef sharks (Carcharhinus amblyrhynchos).

Differences in space use, abundance, and behavior in response to scuba diving intensity were assessed using a combination of survey techniques including baited remote underwater video systems (BRUVs) and passive acoustic monitoring. In most locations with recreational diving operations, some level of human impact is pervasive, but on Palmyra, scientific diving activities are concentrated on just a few sites that house long-term monitoring projects. These sites experience relatively intensive research diving, particularly during May-August, while the majority of the island is entirely undived. In 2015, logistical constraints caused there to be no research diving activity in Palmyra prior to our study in

September; the atoll had therefore been undived for an entire year, providing an opportunity

to examine the time scales of human presence effects on the abundance, behavior, and spatial distribution of reef sharks. We then assessed multi-year residency patterns of reef sharks at dived and undived sites to compare shark residency in contrasting years -- one with substantial scuba diving activities (2014), the other without diving (2015). We hypothesize that intensive scuba diving activity results in chronic spatial displacement of reef sharks.

Methods

Study site

Palmyra Atoll is located in the central Pacific (5o54’N; 162o05’W) and was established as a U.S. Fish and Wildlife Refuge in 2001. The Marine Refuge and Marine

National Monument extends out 50 nautical miles and are managed by The Nature

Conservancy and the U.S. Fish and Wildlife Service. Commercial fishing and extractive recreational fishing are banned within the refuge. Sharks were captured and released under

U.S. Fish and Wildlife permits #12533-14011 (2014) and #12533-15011 (2015).

Baited remote underwater video systems (BRUVs)

BRUVS have become the standard tool for monitoring large bodied, potentially cautious reef fish including sharks (Meekan and Cappo 2004, Malcolm et al. 2007). They are non-invasive, repeatable, and allow the accurate collection of data on the relative abundance and distribution of the marine faunal community (Harvey et al. 1996, 2001), particularly for motile fauna. The use of bait with the BRUV system serves to attract motile predators to the camera unit; however, while bait increases the abundance of generalist

carnivores in the area immediately surrounding the BRUV system, it does not influence the abundance or distribution of herbivorous fish (Watson et al. 2005, Harvey et al. 2007).

During August and September 2015, we conducted 50 BRUV surveys on Palmyra’s forereef habitat at four research (dived) and four non-research (undived) sites (Fig. 3.1).

Research sites (Penguin Spit, FR5, FR7, FR9) surveyed were the most dived sites on

Palmyra’s forereefs (Fig. 3.1). Research sites were originally selected based on accessibility and coverage of both north and south forereefs; site selection was not based on habitat features or presence of particular species, but site usage was maintained to (a) minimize disturbance within the marine refuge, and (b) maximize information collection via data sharing and long-term monitoring. Dived sites are visited 100s of times each year, with particularly high diving pressure during the summer “research” season (May-August; TNC

DSO, personal communication). Diving activities often include deployment and removal of research equipment (e.g. instruments, settlement plates, markers, cages, moorings) and underwater visual surveys (UVS). Sites herein referred to as “undived” have been visited 1-

3 times over the last 15 years by NOAA’s Coral Reef Ecosystem Program divers, or are dived once annually for ~15 minutes by a pair of divers to collect and redeploy an acoustic receiver (FR5NR, FR7NR). Undived sites were a minimum of 1km from dived sites (Fig.

3.1) and selected for similarities in their ecological characteristics to adjacent dived sites

(e.g. habitat, species composition).

BRUVs were deployed for 120 minutes in depths between 8-30 m. Individual sites were surveyed 6 to 7 times at slightly different locations (within 100m of each other) and at various depths with a minimum of 24 hr between repeat sampling (Fig. 3.1). All BRUVs were deployed during daylight hours (08:00 h – 17:00 h), and each location was sampled at

multiple times of day to prevent bias with respect to diel influences on fish assemblages

(Birt et al. 2012). We used GoPro Hero cameras (4 Silver), mounted on PVC or metal frames (BRUV systems), which we lowered to the reef floor from the boat using a surface line (in areas without live coral) or diver-deployed (in areas with high coral cover, as required by our U.S. National Wildlife Refuge permit) by swimming the system to the reef and attaching it to non-living substrate using lashing straps. All BRUVs were baited with standardized amounts of mackerel (0.5 kg), which appeared in the video frame. BRUV surveys were conducted simultaneously (within 30 minutes of each other) in a paired design so that one system was at a dived site and one was in an adjacent undived site. To precisely assess the distance travelled by a bait plume, Cappo et al. (2004) developed an equation that considers current speeds and tidal movements; however, currents on Palmyra’s forereefs are highly variable and unpredictable and we were unable to use this equation to inform our study design. Previous studies demonstrated that distances of 100m (Ellis and

Demartini 1995) and 450m (Cappo et al. 2004) between simultaneously deployed BRUV systems was adequate to ensure independence, indicating that a distance of 1 km between

BRUV systems is sufficiently large to be considered independent. However, if scuba diving activities cause the spatial displacement of reef sharks over scales greater than 1 km, our

BRUV surveys would be unable to detect this effect.

All video footage was analyzed using the SeaGIS software EventMeasure (Version

4.4). We analyzed the first 90 minutes of video following deployment where start time was the moment the BRUV landed on the bottom for lowered systems, and the time at which the diver was no longer visible in the video frame for diver-deployed BRUVs. For each BRUV survey, we recorded maximum number of individuals in frame (MaxN), time of MaxN, and

time of first sighting (Tarrive) for every species of predator (for a full species list see Table

A1). MaxN is commonly used as a metric of relative abundance; in order to avoid double- counting individuals, MaxN counts only the maximum number of fish recorded in a single video frame. MaxN positively correlates with fish abundance and Tarrive, because fish are more likely to arrive at the bait quickly if their abundance is high (Ellis and Demartini 1995,

Willis and Babcock 2000). Estimates of MaxN are considered a conservative representation of fish abundance, particularly when fish density is high (Cappo et al. 2004, 2007). For each shark species, we visually examined patterns in MaxN and Tarrive at paired dived and undived sites (Fig. A3.1). Given limited sample sizes at individual sites (n=6 or 7), dived and undived sites were pooled and t-tests were used to examine differences in MaxN, time of

MaxN, and Tarrive . If scuba diving activity spatially displaced reef sharks, we predict MaxN will be higher at undived than dived site, and time of MaxN, and Tarrive will be earlier at undived than dived sites.

To ensure that there were no significant differences in the structure of faunal assemblages between sites -- and that sites were directly comparable -- we constructed a similarity matrix using the Bray-Curtis similarity coefficient. A non-metric multidimensional scale (nMDS) ordination plot of relative abundance estimates (MaxN) was examined to identify patterns in predatory fish assemblages between surveyed sites, and an ordered one-way analysis of similarities (ANOSIM) was used to test for differences in assemblage structure between sites. One-way ANOVAs and Tukey HSD post-hoc tests were used to compare predator diversity (species richness) across sites, and a t-test was used to assess differences in species richness at dived and undived sites. We also assessed differences in MaxN, time of MaxN, and Tarrive between deployment methods and over the

course of the study for each reef shark species to ensure that no bias was introduced via method of deployment or repeated sampling.

Acoustic Monitoring

In the summers of 2011 and 2012, acoustic transmitters (69kHz, V16, Vemco, Ltd) were surgically implanted in 45 grey reef sharks on Palmyra’s forereefs (for surgery details see Papastamatiou et al. 2009). Shark movement was passively tracked using stationary acoustic receivers (Vemco, Ltd. VR2). Part of Palmyra’s large acoustic array includes receivers at 4 of the sites surveyed by BRUVs: FR9 (dived), Penguin Spit (dived), FR5NR

(undived), FR7NR (undived) (Fig. 3.1). Detection radii for VR2 receivers at Palmyra is approximately 300 m (Papastamatiou 2008). Data from each VR2 were downloaded annually.

To test the hypothesis that scuba diving activity spatially affects reef sharks, we examined patterns of site residency and presence during a year with scuba divers present

(2014) and one without diving (2015). From May-August 2014, 789 scuba dives were completed at research sites on Palmyra’s forereefs, and zero dives were conducted during

May-August 2015 (TNC DSO, personal communication). We assessed shark residency patterns by determining the number of days each shark was detected on dived and undived sites from January-August in 2014 and 2015. Only sharks detected anywhere on the receiver array throughout the year were included in the analysis. For each month and year, we calculated a site residency index for each shark at FR9, Penguin Spit, FR5NR, and FR7NR.

A shark was considered present at a receiver if it was detected more than once in the same day (Carlson et al. 2008). The site residency index was then defined as the number of days

each individual was present at a receiver in a given month divided by the number of days in the month. Residency indices are reported as proportions (with a range between 0 to 1), where values closer to 1 indicate a higher degree of site residency. Grey reef sharks at

Palmyra have 28.8 km2 activity spaces on average (95% CI 19.8-37.8 km2; 99% kernel utilization distribution) (Bradley et al. in review), which far exceed the detection range of each VR2. However, we were interested in whether residency patterns differed between dived and undived sites in a year with diving (2014), and whether residency patterns during periods with high diving intensity (May-August 2014) were significantly different from residency patterns during a year without diving (2015).

Mixed effect models were used to assess the effects of seasonal (month) and human use (dived and undived sites and years) drivers on residency patterns of grey reef sharks.

Site and individual were included as random effects in the model, and a logit link and the binomial distribution were used, because the response variable, residency, is reported as a proportion. Models were constructed using the glmer function in the lme4 package in R

(Bates et al. 2015), and Aikake’s information criterion (AIC) was used to assess model performance against a null model (random effects only).

Results

Baited remote underwater video systems

In total, 30 unique species of predatory fishes were recorded from 6 distinct families: sharks (Carcharhinidae), trevallies (Carangidae), snappers (Lutjanidae), groupers

(Serranidae), emperors (Lethrinidae), and eels (Muraenidae) (Fig. 3.2). All three species of reef shark (blacktip, grey reef, whitetip) were present in 8 of the 50 surveys (6 dived, 2

undived); at least one species of reef shark was observed in every survey. Blacktip reef sharks were observed in the most surveys, followed by grey reef sharks, and then whitetip reef sharks (98%, 62%, 22%, respectively). Given the relative rarity of whitetip reef sharks, differences in relative abundance (MaxN), time of MaxN, and time of first sighting (Tarrive) at dived and undived sites were only examined for blacktip and grey reef sharks.

No significant differences were found in MaxN, time of MaxN, or Tarrive between dived and undived sites for blacktip or grey reef sharks (Table 3.1; Fig. 3.3). Blacktip reef sharks had higher relative abundances and earlier time of first sighting and time of MaxN than grey reef sharks across dived and undived locations. We were able to rule out sources of bias from camera deployment method (lowered from boat vs. diver-deployed) and repeated samples: no differences were found in MaxN, time of MaxN, or Tarrive as a function of camera deployment method or as a function of time for any of the species of reef shark examined (Fig. A3.2-3).

Maximum number of predatory species observed in a single 90 min survey was 15

(FR7NR) and minimum species richness was 6 (FR9NR). Only one site, FR5, had significantly lower species richness than all other sites (p<0.001; Fig. A3.4). However, no differences were found in predator species richness (t(47.7)=-0.80, p=0.43) or faunal community assemblage between dived and undived sites (ANOSIM, R=-0.07, p=0.66) (Fig.

3.4).

Acoustic Monitoring

Of the 45 acoustically tagged grey reef sharks, 36 individuals were detected on the acoustic receivers located at Penguin Spit, FR9, FR5NR, and/or FR7NR in 2014, and 16

individuals were detected in 2015. Sharks acoustically tagged in 2011 were only detected on the Palmyra array for the first few months of 2015 (likely due to expired battery life), and so were excluded from the analysis. A minimum of 1 and a maximum of 11 sharks were present at each site, each month and year. There were no differences in residency across any month within a year (p>>0.05), between months in a dived and an undived year (p>>0.05), or between dived and undived sites (p>>0.05) (Fig. 3.5A,C). The null model (random effects only) outperformed the model including effects for month, year, and dived and undived site

(AIC=201.1 and 215.8, respectively). Sharks had significantly higher residency at only one site, Penguin Spit, which is a dived site, but this was consistent throughout both years (Fig.

3.5A,C). Residency at Penguin Spit ranged from 0.03 to 0.97 in 2014 with an average of 8 individuals present, and 0.03 to 0.39 in 2015 with an average of 3 individuals present.

Number of sharks present at dived and undived sites varied across months and between years, but patterns appeared biogeographical rather than driven by scuba diving activity. In 2014, number of sharks present peaked in May and decreased through August at southern dived and undived sites, while shark presence remained high throughout the year at the undived northern site (FR7NR) and was highly variable at the dived northern site (FR9)

(Fig. 3.5B,D). Overall, fewer sharks were present at each site in 2015; however, with the exception of FR7NR, number of sharks present was relatively consistent throughout the year.

Discussion

Our study did not support the hypothesis that scuba diving activity results in the chronic spatial displacement of reef sharks. While previous evidence has shown that sharks

behaviorally respond to people in the water over short time scales (e.g. Smith et al. 2010,

Cubero-Pardo et al. 2011, Baronio 2012), our results from a combination of survey techniques including baited remote underwater video (BRUV) systems and passive acoustic monitoring indicate that this response may not persist. Following a year without any human diving activity at the sites surveyed, we did not detect differences in reef shark abundance or behavior between heavily dived and undived locations (Table 3.1; Fig. 3.3), nor were there differences in their faunal communities (Fig. 3.4). Similarly, there were no differences in shark residency patterns at dived and undived sites in a year with substantial diving activity and a year without any diving (Fig. 3.5).

While Palmyra Atoll is unique in that it allows us to isolate scuba diving impacts on marine ecosystems absent fishing, it should be noted that diving impacts on Palmyra are relatively low compared to many heavily dived tourist destinations. Between 2012-2014, there were an average of 1392 dives completed at Palmyra each year, primarily during the summer months (May-August). However, given Palmyra’s history as a mostly uninhabited island, there was minimal diving activity of any kind prior to the establishment of the refuge in 2001. Palmyra’s shark population was therefore largely naïve to human disturbance until very recently. Despite low levels of diving on Palmyra, different underwater diver-based survey methods revealed opposing trends in shark abundance on Palmyra’s forereefs between 2002-2008 (NOAA CREP, 2015), indicating that naïve sharks may have modified their behavior in response to scuba divers in the water. Despite evidence that Palmyra’s grey reef shark population is stable and may be at carrying capacity (Bradley et al. in review), grey reef shark abundance appeared to be declining according to belt transect surveys, and increasing based on information from surveys conducted by towed divers (Fig. A3.5).

Although our results suggest that the acute responses by sharks to the presence of scuba divers that have been observed elsewhere do not necessarily result in chronic changes to the spatial distribution of reef sharks, more intense diving, year-round diving, and or unregulated diving in terms of species interaction guidelines could elicit stronger, potentially persistent, behavioral responses by reef sharks. On the other hand, sharks may become habituated to the presence of divers after persistent exposure. Ultimately, behavioral responses to scuba divers are likely context-dependent and may correlate with patterns of human use in complex ways (e.g. habituation to divers at regularly dived areas versus investigation of divers by naïve sharks; Ayling and Choat 2008).

The economic benefits of shark diving tourism can significantly outperform shark exploitation (Anderson and Ahmed 1993, Gallagher and Hammerschlag 2011, Vianna et al.

2012, Cisneros-Montemayor et al. 2013), but realizing these benefits over long time scales requires minimizing negative ecological impacts to shark populations (Topelko and Dearden

2005). At the same time, overly restrictive regulation can hinder potential economic gains from tourism activities. Well-managed shark diving industries use standards and regulations to inform best-practice tourism and species interaction guidelines (e.g. Quiros 2007, Smith et al. 2014). However, the scientific basis for these regulations is often tenuous (Vannuccini

1999), likely because most studies have focused on short-term behavioral effects of human interventions often via in situ visual surveys (e.g. Quiros 2007, Smith et al. 2010, Cubero-

Pardo et al. 2011).

Long-term consequences of non-consumptive scuba diving tourism on sharks are not well understood, nor are they easy to understand due to the unique set of challenges associated with studying highly mobile marine predators (Austin et al. 2004; but see

Hammerschlag et al. 2012). Using a combination of monitoring tools, our results suggest that short-term impacts by divers on shark behavior described in other studies may not persist or translate to long-term effects. Unlike terrestrial and marine mammals, which alter space use away from recreational activity (e.g. Constantine and Zealand 2001, George and

Crooks 2006), we found no evidence of long-term spatial displacement in reef sharks in response to scuba diving.

Although scuba divers may not spatially displace reef sharks over long time scales, the inherent mobility of sharks and other marine predators has been shown to independently undermine the efficacy of marine reserves designed to protect them from fishing (Robbins et al. 2006, McCook et al. 2010, Espinoza et al. 2014). Roughly 40% of coral reef marine protected areas (MPAs) are less than 2 km2 (Mora et al. 2006), far smaller than the activity spaces of many sharks and other vagile predatory reef fishes (e.g. 28km2 reported for grey reef sharks; Bradley et al. in review), which are often key fisheries’ targets. Therefore, over short time scales, even moderate space use changes by sharks, such as temporary avoidance of divers (e.g. Quiros 2007, Cubero-Pardo et al. 2011, Baronio 2012), could be enough to cause a shark to leave a small MPA. It is also important to note that relatively small behavioral changes in terms of space use might have been undetectable with our methods; both the detection range of a VR2 and the attraction radius of the bait plume over our 90 min

BRUV surveys could potentially detect and attract sharks from 100s of meters away. The value of sharks in shark diving tourism depends on tourists actually seeing them, and even small-scale movements away from divers could have therefore have significant economic, as well as ecological, repercussions. However, our results indicate that humans can interact

with reef sharks without persistent behavioral impacts, and that well-regulated shark diving tourism can likely be accomplished without undermining conservation goals.

Acknowledgments

We thank The Nature Conservancy staff for support at the Palmyra Atoll research station, K.

Davis and P. Carlson for field assistance, L. Hornsby and W. Sedgwick for lab work, and A.

Friedlander, C. Lowe, and K. Weng for the setup and maintenance of Palmyra’s acoustic array. Funding was provided by Save Our Seas Foundation (DB) and a Marisla Foundation grant (JEC). DB was supported by a NSF Graduate Research Fellowship (DGE-114408) and an AAUS Kathy Johnston English Scholarship.

Table 3.1. No differences were found for relative abundance (MaxN), time of MaxN, or arrival time (Tarrive) from BRUV surveys for grey reef sharks (Carcharhinus amblyrhynchos) and blacktip reef sharks (Carcharhinus melanopterus) at dived and undived sites at Palmyra

Atoll. T-test results are reported with corresponding p-values, mean, and standard error (in parentheses).

Species MaxN Time MaxN (min) Tarrive (min)

C. amblyrhynchos t(48.0)=-0.40; p=0.69 t(48.8)=-0.44; p=0.66 t(28.8)=1.37; p=0.18

Mean (SE) Diveddd 1.08(0.22) 32.3(4.5) 14.2(4.5)

Mean (SE) Undived 0.96(0.21) 29.2(5.3) 23.1(4.7)

C. melanopterus t(48.0)=1.49; p=0.14 t(47.3)=-0.69; p=0.50 t(47.0)=-0.48; p=0.63

Mean (SE) Diveddd 2.04(0.25) 38.8(5.0) 13.4(2.8)

Mean (SE) Undived 2.56(0.25) 34.2(4.4) 11.4(2.9)

Figure 3.1. Location of baited remote underwater video (BRUV) surveys on Palmyra’s forereef habitat at dived (N=25) and undived (N=25) sites.

Figure 3.2. Average relative abundance (MaxN) of all predators at Palmyra Atoll in dived

(green) and undived (blue) sites. Error bars represent +/-1 standard error. Note: Lutjanus bohar not pictured because MaxN was significantly higher than for all other species (9.5 and

7.5, dived and undived, respectively).

Figure 3.3. MaxN (A), time of MaxN (B), and Tarrive (C) for blacktip reef sharks (C. melanopterus) and grey reef sharks (C. amblyrhynchos) at dived and undived sites at

Palmyra Atoll.

Figure 3.4. Relative abundance nMDS plot for overall predator faunal community assemblage in dived and undived sites at Palmyra Atoll. Dived sites (green): Penguin Spit

(PengSpit), FR5, FR7, FR9. Undived sites (blue): Penguin Spit Non-Research

(PengSpitNR), FR5NR, FR7NR, FR9NR. Black letters are species codes (full species list available in Table A3.1).

Figure 3.5. Grey reef shark residency (with standard errors) in January-August, 2014 (A) and 2015 (C) at two dived (Penguin Spit, FR9) and two undived sites (FR7NR, FR5NR).

Number of sharks present at each site each month in 2014 (B) and 2015 (D).

Figure 3.1

653000653000 ! FR9 FR9NR FR7 FR7NR Undived! 652000652000 ! Dived

factor(Research) 651000651000 ! 0

lat_m 1 Latitude

650000650000 !

FR5NR FR5 649000649000 Penguin Spit NR Penguin Spit

815000 820000 825000 830000 815000!! ! 820000!!!long_m 825000!! ! ! 830000 Longitude

Figure 3.2

Lethrinidae Lutjanidae Serranidae Carangidae Muraenidae 3 Carcharhinidae Undived! Dived!

Site Type 2 Research Non−research MaxN

0 ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! louti furca lysan argus fulvus gibbus miniata obesus tauvina urodeta ignobilis howlandi javanicus olivaceus virescens leopardus bipinnulata spilotoceps grandoculis xanthochilus melampygus aureolineatus Variola louti Variola melanopterus melanostigma albomarginata polyphekadion orthogrammus amblyrhynchos Lutjanusfulvus Aphareusfurca Lutjanusgibbus Caranxignobilis Aprionvirescens Triaenodon obesus Triaenodon Lethrinusolivaceus Elegatisbipinnulata Scomberoideslysan Epinephelustauvina Cephalopholisargus Caranxmelampygus Epinephelushowlandi Gracilaalbomarginata Cephalopholisminiata Monotaxisgrandoculis Lethrinusxanthochilus Epinephelusspiloceps Cephalopholisurodeta Gymnothoraxjavanicus Cephalopholisleopardus Epinephelusmelanostigma Epinepheluspolyphekadion Carcharhinusmelanopterus Carangoidesorthogrammus Gnathodentexaureolineatus Carcharhinusamblyrhynchos

Figure 3.3

5!5 A Undived! Dived 4!4 ●

3!3 ●

MaxN 2!2

1!1

0 0 CAAB CAML

● 8080! B ●

●

6060!

4040!

2020! Time MaxN (min) Time

0 0 CAAB CAML

6060! !C ● ●

● 4040! !

2020! Tarrive (min) Tarrive !

0 0 CAAB CAML C. amblyrhynchos C. melanopterus

Figure 3.4

2D Stress: 0.058

0.6 EPHO

GNAU 0.4

0.2 PengSpitNR FR9NR LUFU EPPO EPTA

NMDS2 GRALPengSpitAPFU CAMLGYJA CAOR MOGRLUGICEARLUBO 0.0 VALO CAIGCEURCAMECAABFR7NR FR7TROB FR9 FR5LEXA ELBI FR5NR SCLY● 0.2 − CEMI LEOL 0.4 − CELE●

−0.4 −0.2 0.0 0.2 0.4 0.6 0.8 NMDS1

Figure 3.5

0.6!0.6 ● ● A FR9! B PengSpit! ● ● ● ●

FR7NR! 9!9 ● ● ● ● ● FR5NR ● ● ● ●

0.4 0.4! ● ● ● ●

6 ● ● ● ● ● FR9 6! ● FR9 ● Heart_Palm ● Heart_Palm

● inds ● ● Kaula ● Kaula residencymean ● Peng_Spit_Forereef ● ● Peng_Spit_Forereef ● ● ● ● ● ● ● ● ● ● 0.2 Residency 0.2! ● ● ● ● Individuals ● 3 ● 3! ●

● ● ● ● ● ● ● ● ● ● ● ● ● ● 0.0 0.0 0 0

2 4 6 8 2 4 6 8 month_num month_num Jan Feb Mar Apr May Jun Jul Aug Jan Feb Mar Apr May Jun Jul Aug

0.6!0.6 C D

9!9

●

0.4 0.4! ● ●

6 ● ● FR9 6! ● FR9 ● Heart_Palm ● Heart_Palm

● inds ● Kaula ● Kaula residencymean ● Peng_Spit_Forereef ● Peng_Spit_Forereef

● ● ● ●

0.2 Residency 0.2! ● ● Individuals 3!3 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

0.0 0.0 0 0

2 4 6 8 2 4 6 8 month_num month_num Jan Feb Mar Apr May Jun Jul Aug Jan Feb Mar Apr May Jun Jul Aug

Appendices

Table A3.1. Species list. All predator species observed in baited remote underwater video

(BRUV) surveys at Palmyra Atoll.

Family Genus Species Species Code Carangidae Carangoides orthogrammus CAOR Carangidae Caranx ignobilis CAIG Carangidae Caranx melampygus CAME Carangidae Elagatis bipinnulata ELBI Carangidae Scomberoides lysan SCLY Carcharhinidae Carcharhinus amblyrhynchos CAAB Carcharhinidae Carcharhinus melanopterus CAML Carcharhinidae Triaenodon obesus TROB Chanidae Chanos chanos CHCH Lethrinidae Gnathodentex aureolineatus GNAU Lethrinidae Lethrinus olivaceus LEOL Lethrinidae Lethrinus xanthochilus LEXA Lethrinidae Monotaxis grandoculis MOGR Lutjanidae Aphareus furca APFU Lutjanidae Aprion virescens APVI Lutjanidae Lutjanus bohar LUBO Lutjanidae Lutjanus fulvus LUFU Lutjanidae Lutjanus gibbus LUGI Muraenidae Gymnothorax javanicus GYJA Serranidae Cephalopholis argus CEAR Serranidae Cephalopholis leopardus CELE Serranidae Cephalopholis miniata CEMI Serranidae Cephalopholis urodeta CEUR Serranidae Epinephelus howlandi EPHO Serranidae Epinephelus melanostigma EPML Serranidae Epinephelus polyphekadion EPPO Serranidae Epinephelus spilotoceps EPSL Serranidae Epinephelus tauvina EPTA Serranidae Gracila albomarginata GRAL Serranidae Variola louti VALO

Figure A3.1. Paired BRUV survey results. MaxN, time of MaxN, and Tarrive for grey reef and blacktip reef sharks at paired dived and undived sites at Palmyra Atoll.

Figure A3.2. Baited remote underwater video system (BRUVS) deployment methods.

Camera deployment method (lowered from boat vs. diver-deployed) did not affect Tarrive or

MaxN for any of the three species of reef shark examined. Tarrive shown for different methods of deployment (No Diver and Diver deployed) for (A) blacktip reef sharks, (B), grey reef sharks, and (C) whitetip reef sharks.

Figure A3.3. Baited remote underwater video system (BRUVS) repeat sampling. MaxN, time of MaxN, and Tarrive for grey reef sharks are shown for each day at each surveyed site.

Figure A3.4. Predator species richness. Only one site, FR5, had significantly lower species richness than all other sites (one-way ANOVA, p<0.001). Different letters (A, B, C) indicate a significant difference (Tukey HSD, p<0.05).

Figure A3.5. Grey reef shark underwater visual surveys. Grey reef sharks per belt transect survey (left axis) and towed diver survey (right axis) on Palmyra’s forereefs between 2002-2008. Error bars represent +/- 1 SE. Data provided by NOAA CREP 2015

Figure A3.1

A grey reef 4! 4 B blacktip reef Dived! Undived!

2 2! 3! 3

2! 2 MaxN 1! 1

1! 1

0 0 0 0

PengSpit PengSpitNR FR5 FR5NR FR7 FR7NR FR9 FR9NR PengSpit PengSpitNR FR5 FR5NR FR7 FR7NR FR9 FR9NR PS PS FR5 FR7 FR9 FR5 FR7 FR9 PSNR PSNR FR5NR FR7NR FR9NR FR5NR FR7NR FR9NR grey reef blacktip reef C 60!60 D 60!60

40 40!40 40!

20 20!20 20! Time MaxN (min) Time

0 0 0 0

PengSpit PengSpitNR FR5 FR5NR FR7 FR7NR FR9 FR9NR PengSpit PengSpitNR FR5 FR5NR FR7 FR7NR FR9 FR9NR PS PS FR5 FR7 FR9 FR5 FR9 FR7 PSNR PSNR FR5NR FR7NR FR9NR FR5NR FR7NR FR9NR

40!40 E grey reef F blacktip reef 30!30

30!30

20!20

10!10 Tarrive (min) Tarrive 10!10

0 0 0 0

PengSpit PengSpitNR FR5 FR5NR FR7 FR7NR FR9 FR9NR PengSpit PengSpitNR FR5 FR5NR FR7 FR7NR FR9 FR9NR PS PS FR7 FR5 FR9 FR5 FR9 FR7 PSNR PSNR FR9NR FR5NR FR7NR FR5NR FR7NR FR9NR

Figure A3.2

5 ! 5 ! A blacktip reef B grey reef

4 ! 4 !

3 ! 3 !

2 ! 2 ! MaxN

1 ! 1 !

0 0 NoDropped Diver DiverSet NoDropped Diver DiverSet

80! 80! 80 C 80 D

60! 60! 60 60

40! 40! 40 40

20! 20! 20 20 Time MaxN (min) Time

0 0

NoDropped Diver DiverSet NoDropped Diver DiverSet

60! 60! 60 60 ● F 50! E 50! 50 50 40! 40! 40 ● 40 ●

30! ● 30! 30 30 20! 20! 20 20

Tarrive (min) Tarrive 10! 10! 10 10

0 0

NoDropped Diver DiverSet NoDropped Diver DiverSet

Figure A3.3

4!4 4 4 4 Penguin Spit Penguin Spit NR FR5 FR5NR 3!3 3 3 3

2!2 2 2 2

1!1 1 1 1

0 0 0 0 0 8312015 9022015 9032015 9102015 9112015 9132015 9032015 9042015 9082015 9112015 9132015 9142015 9022015 9042015 9062015 9082015 9102015 9132015 8302015 9022015 9042015 9062015 9082015 9092015 9132015 8/31 9/2 9/3 9/10 9/11 9/13 9/3 9/4 9/8 9/11 9/13 9/14 9/2 9/4 9/6 9/8 9/10 9/13 8/30 9/2 9/4 9/6 9/8 9/9 9/13 4!4 4 4 4 MaxN FR7 FR7NR FR9 FR9NR 3!3 3 3 3

2!2 2 2 2

1!1 1 1 1

0 0 0 0 0 8312015 9022015 9082015 9102015 9112015 9132015 8312015 9032015 9042015 9082015 9102015 9112015 9132015 9022015 9042015 9062015 9092015 9102015 9132015 8302015 9022015 9042015 9062015 9102015 9132015 8/31 9/2 9/8 9/10 9/11 9/13 8/31 9/3 9/4 9/8 9/10 9/11 9/13 9/2 9/4 9/6 9/9 9/10 9/13 8/30 9/2 9/4 9/6 9/10 9/13

8080! 80 80 80 Penguin Spit Penguin Spit NR FR5 FR5NR 6060! 60 60 60

4040! 40 40 40

2020! 20 20 20

9/2 9/4 9/6 9/8 9/10 9/13 0 0 0 0 0 8312015 9022015 9032015 9102015 9112015 9132015 9032015 9042015 9082015 9112015 9132015 9142015 9022015 9042015 9062015 9082015 9102015 9132015 8302015 9022015 9042015 9062015 9082015 9092015 9132015 8/31 9/2 9/3 9/10 9/11 9/13 9/3 9/4 9/8 9/11 9/13 9/14 9/2 9/4 9/6 9/8 9/10 9/13 8/30 9/2 9/4 9/6 9/8 9/9 9/13 8080! 80 80 80 FR7 FR7NR FR9 FR9NR 6060! 60 60 60

4040! 40 40 40 Time MaxN (min) Time 2020! 20 20 20

6060! 60 60 60 Penguin Spit Penguin Spit NR FR5 FR5NR 4040! 40 40 40

2020! 20 20 20

0 0 0 0 0 8312015 9022015 9102015 9132015 9042015 9082015 9142015 9022015 9042015 9062015 9082015 9102015 9132015 8302015 9062015 9082015 8/31 9/2 9/10 9/13 9/4 9/8 9/14 9/2 9/4 9/6 9/8 9/10 9/13 8/30 9/6 9/8 6060! 60 60 60 FR7 FR7NR FR9 FR9NR 4040! 40 40 40 Tarrive (min) Tarrive 2020! 20 20 20

0 0 0 0 0 8312015 9022015 9102015 9112015 8312015 9042015 9132015 9022015 9042015 9062015 9092015 9132015 8302015 9102015 9132015 8/31 9/2 9/10 9/11 8/31 9/4 9/13 9/2 9/4 9/6 9/9 9/13 8/30 9/10 9/13

9 8

Figure A3.4

A A A A,B A B A,B

1010! C

5!5 Predatorspecies richness

0 0

PengSpit PengSpitNR FR5 FR5NR FR7 FR7NR FR9 FR9NR FR7 FR9 FR5 FR5NR FR7NR FR9NR PengSpit

PengSpitNR

Figure A3.5

100

REFERENCES

Almany, G. R. 2004. Differential effects of habitat complexity, predators and competitors on abundance of juvenile and adult coral reef fishes. Oecologia 141:105–113.

Alter, S. E., S. D. Newsome, and S. R. Palumbi. 2012. Pre-whaling genetic diversity and population ecology in eastern pacific gray whales: Insights from ancient DNA and stable isotopes. PLoS ONE 7: e35039.

Anderson, R., and H. Ahmed. 1993. The Shark Fisheries of the Maldives. MOFA Malé and FAO Rome: 73-84.

Austin, D., W. D. Bowen, and J. I. McMillan. 2004. Intraspecific variation in movement patterns: modeling individual behaviour in a large marine predator. Oikos 105:15–30.

Ayling, A. M., and J. H. Choat. 2008. Abundance patterns of reef sharks and predatory fishes on differently zoned reefs in the offshore Townsville region: Final report to the Great Barrier Reef Marine Park Authority, Research Publication No. 91:32pp.

Babcock, R. C., N. T. Shears, A. C. Alcala, N. S. Barrett, G. J. Edgar, K. D. Lafferty, T. R. McClanahan, and G. R. Russ. 2010. Decadal trends in marine reserves reveal differential rates of change in direct and indirect effects. Proceedings of the National Academy of Sciences of the United States of America 107:18256–61.

Barnett, A., K. G. Abrantes, J. Seymour, and R. Fitzpatrick. 2012. Residency and spatial use by reef sharks of an isolated seamount and its implications for conservation. PLoS ONE 7:e36574.

Barnett, A., N. L. Payne, J. M. Semmens, and R. Fitzpatrick. 2016. Ecotourism increases the field metabolic rate of whitetip reef sharks. Biological Conservation 199:132–136.

Baronio, M. 2012. The use of a micro remotely operated vehicle as a tool for studies of shark behaviour and diver impact. PhD thesis. Southern Cross University:166pp.

Baum, J. K., and R. A. Myers. 2004. Shifting baselines and the decline of pelagic sharks in the Gulf of Mexico. Ecology Letters 7:135–145.

Von Bertalanffy, L. 1938. A quantitative theory of organic growth (inquiries on growth laws. II). Human Biology 10:181–213.

Beverton, R. J. H., and S. J. Holt. 1959. A review on the lifespans and mortality rates of fish in nature, and their relation to growth and other physiological characteristics. CIBA Foundation Colloquia on Ageing. 5:142–180.

101

Birt, M. J., E. S. Harvey, and T. J. Langlois. 2012. Within and between day variability in temperate reef fish assemblages: Learned response to baited video. Journal of Experimental Marine Biology and Ecology 416-417:92–100.

Boo, E. 1991. Ecotourism: Potentials and Pitfalls. The Professional Geographer 43:536–537.

Boyle, S. A., and F. B. Samson. 1985. Effects of nonconsumptive recreation on wildlife: a review. Wildlife Society Bulletin 13:110–116.

Brunnschweiler, J. M., K. G. Abrantes, and A. Barnett. 2014. Long-term changes in species composition and relative abundances of sharks at a provisioning site. PLoS ONE 9:e86682.

Brunnschweiler, J. M., and H. Baensch. 2011. Seasonal and long-term changes in relative abundance of bull sharks from a tourist shark feeding site in Fiji. PLoS ONE 6:e16597.

Cailliet, G. M., W. D. Smith, H. F. Mollet, and K. J. Goldman. 2006. Age and growth studies of chondrichthyan fishes: the need for consistency in terminology, verification, validation, and growth function fitting. Environmental Biology of Fishes 77:211–228.

Calenge, C. 2006. The package “adehabitat” for the R software: A tool for the analysis of space and habitat use by animals. Ecological Modelling 197:516–519.

Cappo, M., G. De, and P. Speare. 2007. Inter-reef vertebrate communities of the Great Barrier Reef Marine Park determined by baited remote underwater video stations. Marine Ecology Progress Series 350:209–221.

Cappo, M., P. Speare, and G. De’ath. 2004. Comparison of baited remote underwater video stations (BRUVS) and prawn (shrimp) trawls for assessments of fish biodiversity in inter-reefal areas of the Great Barrier Reef Marine Park. Journal of Experimental Marine Biology and Ecology 302:123–152.

Carlson, J. K., and I. E. Baremore. 2003. Changes in biological parameters of Atlantic sharpnose shark Rhizoprionodon terraenovae in the Gulf of Mexico: Evidence for density-dependent growth and maturity? Marine and Freshwater Research 54:227–234.

Carlson, J. K., M. R. Heupel, D. M. Bethea, and L. D. Hollensead. 2008. Coastal habitat use and residency of juvenile Atlantic sharpnose sharks (Rhizoprionodon terraenovae). Estuaries and Coasts 31:931–940.

Carlson, J. K., J. R. Sulikowski, and I. E. Baremore. 2006. Do differences in life history exist for blacktip sharks, Carcharhinus limbatus, from the United States South Atlantic Bight and Eastern Gulf of Mexico? Environmental Biology of Fishes 77:279–292.

Carlton, J. T. 1998. Apostrophe to the ocean. Conservation Biology 12:1165–1167.

102

Ceballos, G., P. R. Ehrlich, A. D. Barnosky, A. García, R. M. Pringle, and T. M. Palmer. 2015. Accelerated modern human – induced species losses: entering the sixth mass extinction. Sciences Advances 1:1–5.

Chin, A., A. J. Tobin, M. R. Heupel, and C. A. Simpfendorfer. 2013a. Population structure and residency patterns of the blacktip reef shark Carcharhinus melanopterus in turbid coastal environments. Journal of Fish Biology 82:1192–1210.

Cisneros-Montemayor, A. M., M. Barnes-Mauthe, D. Al-Abdulrazzak, E. Navarro-Holm, and U. R. Sumaila. 2013. Global economic value of shark ecotourism: Implications for conservation. Oryx 47:381–388.

Ciuti, S., J. M. Northrup, T. B. Muhly, S. Simi, M. Musiani, J. A. Pitt, and M. S. Boyce. 2012. Effects of humans on behaviour of wildlife exceed those of natural predators in a landscape of fear. PLoS ONE 7:e50611.

Clark, J. S., S. R. Carpenter, M. Barber, S. L. Collins, A. P. Dobson, J. a Foley, D. M. Lodge, M. Pascual, R. Pielke, W. Pizer, C. M. Pringle, W. V Reid, K. a Rose, O. Sala, W. H. Schlesinger, D. H. Wall, and D. Wear. 2001. Ecological forecasts: an emerging imperative. Science 293:657–60.

Clua, E., N. Buray, P. Legendre, J. Mourier, and S. Planes. 2010. Behavioural response of sicklefin lemon sharks negaprion acutidens to underwater feeding for ecotourism purposes. Marine Ecology Progress Series 414:257–266.

Compagno, L. J. V. 1984. FAO Species catalogue. Vol. 4. Sharks of the world. An annotated and illustrated cataloge of shark species known to date. Part 2. Carcharhiniformes. FAO Fisheries Synopsis 4:251–655.

Constantine, R., D. H. Brunton, and T. Dennis. 2004. Dolphin-watching tour boats change bottlenose dolphin (Tursiops truncatus) behaviour. Biological Conservation 117:299– 307.

Constantine, R., and N. Zealand. 2001. Increased avoidance of swimmers by wild Bottlenose Dolphins (Tursiops truncatus) due to long-term exposure to swim-with- dolphin tourism. Marine Mammal Science 17:689–702.

Cortes, E., and J. A. Neer. 2006. Preliminary reassessment of the validity of the 5% fin to carcass weight ratio for sharks. Col. Vol. Sci. Pap. ICCAT 59:1025–1036.

Coral Reef Ecosystem Program; Pacific Islands Fisheries Science Center. 2015. Ecosystem Monitoring & Assessment: Rapid Ecological Assessments of Fish Belt Transect Survey in the Pacific Ocean from 2000 to 2009. NOAA National Centers for Environmental Information. Unpublished Dataset. [09/16/2015] https://inport.nmfs.noaa.gov/inport/item/5565

103

Coral Reef Ecosystem Program; Pacific Islands Fisheries Science Center. 2015. National Coral Reef Monitoring Program: Towed-diver Surveys of Large-bodied Fishes of the U.S. Pacific Reefs since 2000 NOAA National Centers for Environmental Information. Unpublished Dataset. [09/16/2015] https://inport.nmfs.noaa.gov/inport/item/5568

Costello, C., D. Ovando, T. Clavelle, C. Strauss, R. Hilborn, M. Melnychuk, T. A. Branch, S. D. Gaines, C. Szuwalski, R. Cabrai, D. Rader, and A. Leland. 2016. Global fishery futures under contrasting management regimes. Proceedings of the National Academy of Sciences of the United States of America. 113:5125-5129.

Cox, S. P., T. E. Essington, J. F. Kitchell, S. J. D. Martell, C. J. Walters, C. Boggs, and I. Kaplan. 2002. Reconstructing ecosystem dynamics in the central Pacific Ocean, 1952- 1998. II. A preliminary assessment of the trophic impacts of fishing and effects on tuna dynamics. Canadian Journal of Fisheries and Aquatic Sciences 59:1736–1747.

De Crosta, M. A., L. R. J. Taylor, and J. D. Parrish. 1984. Age determination , growth, and energetics of three species of carcharhinid sharks in Hawaii. Proceedings of the 2nd Symposium on Resource Investigations in the NWHI, Vol 2 UNIHI-SEA-GRANT- MR-84–01. Honolulu: University of Hawaii Sea Grant.:75–95.

Cubero-Pardo, P., P. Herrón, and F. González-Pérez. 2011. Shark reactions to scuba divers in two marine protected areas of the Eastern Tropical Pacific. Aquatic Conservation: Marine and Freshwater Ecosystems 21:239–246.

Daves, N. K., and M. F. Nammack. 1998. US and International mechanisms for protecting and managing shark resources. Fisheries Research. 223–228.

Davidson, L. N. K., M. A. Krawchuk, and N. K. Dulvy. 2015. Why have global shark and ray landings declined: improved management or overfishing? Fish and Fisheries.

Dobson, J. 2006. Sharks, Wildlife Tourism, and State Regulation. Tourism in Marine Environments 3:15–23.

Driggers, W. B., J. K. Carlson, B. Cullum, J. M. Dean, D. Oakley, and G. Ulrich. 2004. Age and growth of the blacknose shark, Carcharhinus acronotus in the western North Atlantic Ocean with comments on regional variation in growth rates. Environmental Biology of Fishes 71:171–178.

Dulvy, N. K., S. L. Fowler, J. A. Musick, R. D. Cavanagh, P. M. Kyne, L. R. Harrison, J. K. Carlson, L. N. Davidson, S. V Fordham, M. P. Francis, C. M. Pollock, C. a Simpfendorfer, G. H. Burgess, K. E. Carpenter, L. J. Compagno, D. A. Ebert, C. Gibson, M. R. Heupel, S. R. Livingstone, J. C. Sanciangco, J. D. Stevens, S. Valenti, and W. T. White. 2014. Extinction risk and conservation of the world’s sharks and rays. eLife 3:e00590.

104

Dureuil, M., and B. Worm. 2015. Estimating growth from tagging data: An application to north-east Atlantic tope shark Galeorhinus galeus. Journal of Fish Biology 87:1389– 1410.

Economakis, A. E., and P. S. Lobel. 1998. Aggregation behavior of the grey reef shark, Carcharhinus amblyrhynchos, at Johnston Atoll, Central Pacific Ocean. Environmental Biology of Fishes 51:129–139.

Efford, M. 2004. Density estimation in live-trapping studies. Oikos 106:598–610.

Ellenberg, U., T. Mattern, P. J. Seddon, and G. L. Jorquera. 2006. Physiological and reproductive consequences of human disturbance in Humboldt penguins: The need for species-specific visitor management. Biological Conservation 133:95–106.

Ellis, D. M., and E. E. Demartini. 1995. Evaluation of a video camera technique for indexing abundances of juvenile pink snapper Pristipomoides filamentosus, and other Hawaiian insular shelf fishes. Fishery Bulletin 93:67–77.

Ellis, J. R., and J. Keable. 2008. Fecundity of Northeast Atlantic spurdog (Squalus acanthias). ICES Journal of Marine Science 65:979–981.

Espinoza, M., M. R. Heupel, A. J. Tobin, and C. A. Simpfendorfer. 2014. Residency patterns and movements of grey reef sharks (Carcharhinus amblyrhynchos) in semi- isolated coral reef habitats. Marine Biology 162:343-358.

Estes, J. A., J. Terborgh, J. S. Brashares, M. E. Power, J. Berger, W. J. Bond, S. R. Carpenter, T. E. Essington, R. D. Holt, J. B. C. Jackson, R. J. Marquis, L. Oksanen, T. Oksanen, R. T. Paine, E. K. Pikitch, W. J. Ripple, S. A. Sandin, M. Scheffer, T. W. Schoener, J. B. Shurin, A. R. E. Sinclair, M. E. Soulé, R. Virtanen, and D. A. Wardle. 2011. Trophic downgrading of planet Earth. Science 333:301–6.

Eveson, J. P., T. Polacheck, and G. M. Laslett. 2007. Consequences of assuming an incorrect error structure in von Bertalanffy growth models: a simulation study. Canadian Journal of Fisheries and Aquatic Sciences 64:602–617.

Fabens, A. J. 1965. Properties and fitting of the Von Bertalanffy growth curve. Growth 29:265–289.

Farrell, E. D., S. Mariani, and M. W. Clarke. 2010. Reproductive biology of the starry smooth-hound shark Mustelus asterias: Geographic variation and implications for sustainable exploitation. Journal of Fish Biology 77:1505–1525.

Ferretti, F., R. A. Myers, F. Serena, and H. K. Lotze. 2008. Loss of large predatory sharks from the Mediterranean Sea. Conservation Biology 22:952–64.

105

Field, I. C., M. G. Meekan, C. W. Speed, W. White, and C. J. A. Bradshaw. 2011. Quantifying movement patterns for shark conservation at remote coral atolls in the Indian Ocean. Coral Reefs 30:61–71.

Fitzpatrick, R., K. G. Abrantes, J. Seymour, and A. Barnett. 2011. Variation in depth of whitetip reef sharks: Does provisioning ecotourism change their behaviour? Coral Reefs 30:569–577.

Francis, R. I. C. C. 1988a. Are Growth Parameters Estimated from Tagging and Age-Length Data Comparable? Canadian Journal of Fisheries and Aquatic Sciences 45:936–942.

Francis, R. I. C. C. 1988b. Maximum likelihood estimation of growth and growth variability from tagging data. New Zealand Journal of Marine and Freshwater Research 22:43–51.

Friedlander, A. M., and E. E. DeMartini. 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.

Frisk, M. G., T. J. Miller, and M. J. Fogarty. 2001. Estimation and analysis of biological parameters in elasmobranch fishes: A comparative life history study. Canadian Journal of Fisheries and Aquatic Sciences 58:969–981.

Gallagher, A. J., and N. Hammerschlag. 2011. Global shark currency: the distribution, frequency, and economic value of shark ecotourism. Current Issues in Tourism 14:797– 812.

Gardner, B., J. A. Royle, M. T. Wegan, R. E. Rainbolt, and P. D. Curtis. 2010. Estimating Black Bear Density Using DNA Data From Hair Snares. Journal of Wildlife Management 74:318–325.

Garrick, J. A. F. 1982. Sharks of the Genus Carcharhinus National Oceanic and Atmospheric Administration. NOAA Technical Report:194pp.

Gelman, A., and J. Hill. 2007. Data analysis using regression and multilevel/hierarchical models. Cambridge University Press, New York, New York, USA.

George, S. L., and K. R. Crooks. 2006. Recreation and large mammal activity in an urban nature reserve. Biological Conservation 133:107–117.

Gove, J. M., G. J. Williams, M. A. McManus, S. F. Heron, S. A. Sandin, O. J. Vetter, and D. G. Foley. 2013. Quantifying Climatological Ranges and Anomalies for Pacific Coral Reef Ecosystems. PLoS ONE 8:e61974.

Graham, N. A. J., M. D. Spalding, and C. R. C. Sheppard. 2010. Reef shark declines in remote atolls highlight the need for multi-faceted conservation action. Aquatic Conservation: Marine and Freshwater Ecosystems 20:543–548.

106

Gulland, J. A., and S. J. Holt. 1959. Estimation of growth parameters for data at unequal time intervals. Journal de Conseil Conseil International pour l’Exploration de la Mer 25:47–49.

Haberl, H., K. H. Erb, F. Krausmann, V. Gaube, A. Bondeau, C. Plutzar, S. Gingrich, W. Lucht, and M. Fischer-Kowalski. 2007. Quantifying and mapping the human appropriation of net primary production in earth’s terrestrial ecosystems. Proceedings of the National Academy of Sciences 104:12942–12947.

Halpern, B. S., C. Longo, D. Hardy, K. L. McLeod, J. F. Samhouri, S. K. Katona, K. Kleisner, S. E. Lester, J. O’Leary, M. Ranelletti, A. a Rosenberg, C. Scarborough, E. R. Selig, B. D. Best, D. R. Brumbaugh, F. S. Chapin, L. B. Crowder, K. L. Daly, S. C. Doney, C. Elfes, M. J. Fogarty, S. D. Gaines, K. I. Jacobsen, L. B. Karrer, H. M. Leslie, E. Neeley, D. Pauly, S. Polasky, B. Ris, K. St Martin, G. S. Stone, U. R. Sumaila, and D. Zeller. 2012. An index to assess the health and benefits of the global ocean. Nature 488:615–20.

Halpern, B. S., S. Walbridge, K. A. Selkoe, C. V Kappel, F. Micheli, C. D’Agrosa, J. F. Bruno, K. S. Casey, C. Ebert, H. E. Fox, R. Fujita, D. Heinemann, H. S. Lenihan, E. M. P. Madin, M. T. Perry, E. R. Selig, M. Spalding, R. Steneck, and R. Watson. 2008. A global map of human impact on marine ecosystems. Science 319:948–52.

Hammerschlag, N., A. J. Gallagher, J. Wester, J. Luo, and J. S. Ault. 2012. Don’t bite the hand that feeds: assessing ecological impacts of provisioning ecotourism on an apex marine predator. Functional Ecology 26:567–576.

Harvey, E., M. Cappo, J. Butler, N. Hall, and G. Kendrick. 2007. Bait attraction affects the performance of remote underwater video stations in assessment of demersal fish community structure. Marine Ecology Progress Series 350:245–254.

Harvey, E., D. Fletcher, and M. Shortis. 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.

Harvey, E., M. Shortis, S. Robson, M. Stadler, and M. Cappo. 1996. A system for stereo- video measurement of sub-tidal organisms: Implications for assessments of reef fish stocks. Marine Technology Society Journal 29:10–22.

Hayes, C. T., D. S. Baumbach, D. Juma, and S. G. Dunbar. 2016. Impacts of recreational diving on hawksbill sea turtle (Eretmochelys imbricata) behaviour in a marine protected area. Journal of Sustainable Tourism 9582:1–17.

Heithaus, M. R., A. Frid, A. J. Wirsing, and B. Worm. 2008. Predicting ecological consequences of marine top predator declines. Trends in Ecology & Evolution 23:202– 10.

107

Heupel, M. R., and C. A. Simpfendorfer. 2014. Importance of environmental and biological drivers in the presence and space use of a reef-associated shark. Marine Ecology Progress Series 496:47–57.

Heupel, M. R., C. A. Simpfendorfer, and R. Fitzpatrick. 2010. Large-scale movement and reef fidelity of grey reef sharks. PLoS ONE 5:e9650.

Heupel, M. R., A. J. Williams, D. J. Welch, A. Ballagh, B. D. Mapstone, G. Carlos, C. Davies, and C. A. Simpfendorfer. 2009. Effects of fishing on tropical reef associated shark populations on the Great Barrier Reef. Fisheries Research 95:350–361.

Hisano, M., S. R. Connolly, and W. D. Robbins. 2011. Population growth rates of reef sharks with and without fishing on the great barrier reef: robust estimation with multiple models. PLoS ONE 6:e25028.

Hoenig, J. 1983. Empirical use of longevity data to estimate mortality-rates. Fishery Bulletin 82:898-903.

Holmes, B. J., V. M. Peddemors, a. N. Gutteridge, P. T. Geraghty, R. W. K. Chan, I. R. Tibbetts, and M. B. Bennett. 2015. Age and growth of the tiger shark Galeocerdo cuvier off the east coast of Australia. Journal of Fish Biology 87:422–448.

IUCN. 2014. IUCN Red List of Threatened Species.

Jackson, J. B. C. 2008. Ecological extinction and evolution in the brave new ocean. Proceedings of the National Academy of Sciences of the United States of America 105:11458–11465.

Jackson, J. B., M. X. Kirby, W. H. Berger, K. A. Bjorndal, L. W. Botsford, B. J. Bourque, R. H. Bradbury, R. Cooke, J. Erlandson, J. A. Estes, T. P. Hughes, S. Kidwell, C. B. Lange, H. S. Lenihan, J. M. Pandolfi, C. H. Peterson, R. S. Steneck, M. J. Tegner, and R. R. Warner. 2001. Historical overfishing and the recent collapse of coastal ecosystems. Science 293:629–37.

James, I. R. 1991. Estimation of von Bertalanffy growth curve parameters from recapture data. Biometrics 47:1519–1530.

Kendall, W. L. 1999. Robustness of closed capture-recapture methods to violations of the closure assumption. Ecology 80:2517–2525.

Kendall, W. L., J. D. Nichols, and J. E. Hines. 1997. Estimating temporary emigration using capture – recapture data with Pollock’s robust design. Ecology 78:563–578.

Kéry, M., and M. Schaub. 2012. Bayesian Population Analysis using WinBUGS. Academic Press, Waltham, Massachusetts.

108

Kimura, D. K. 2008. Extending the von Bertalanffy growth model using explanatory variables. Canadian Journal of Fisheries and Aquatic Sciences 65:1879–1891.

Kimura, D. K., A. M. Shimada, and S. A. Lowe. 1993. Estimating von Bertalanffy growth parameters of sablefish Anoplopoma fimbria and Pacific cod Gadus macrocephalus using tag-recapture data. Fishery Bulletin 91:271-280.

Kittinger, J. N., K. N. Duin, and B. A. Wilcox. 2010. Commercial fishing, conservation and compatibility in the Northwestern Hawaiian Islands. Marine Policy 34:208–217.

Kneebone, J., L. J. Natanson, A. H. Andrews, and W. H. Howell. 2008. Using bomb radiocarbon analyses to validate age and growth estimates for the tiger shark, Galeocerdo cuvier, in the western North Atlantic. Marine Biology 154:423–434.

Knip, D. M., M. R. Heupel, and C. A. Simpfendorfer. 2012. Habitat use and spatial segregation of adult spottail sharks Carcharhinus sorrah in tropical nearshore waters. Journal of Fish Biology 80:767–84.

Kovacs, K. M., and S. Innes. 1990. The impact of tourism on harp seals (Phoca groenlandica) in the Gulf of St. Lawrence, Canada. Applied Animal Behaviour Science 26:15–26.

Krüger, O. 2005. The role of ecotourism in conservation: Panacea or Pandora’s box? Biodiversity and Conservation 14:579–600.

Kumoru, L. 2003. The shark longline fishery in Papua New Guinea. Billfish and By-catch Research Group, at the 176th Meeting of the Standing Committee on Tuna and Billfish, Mooloolaba, Australia.

Laroche, R. K., A. A. Kock, L. M. Dill, and W. H. Oosthuizen. 2007. Effects of provisioning ecotourism activity on the behaviour of white sharks Carcharodon carcharias. Marine Ecology Progress Series 338:199–209.

Lombardi-Carlson, L., E. Cortés, G. Parsons, and C. Manire. 2003. Latitudinal variation in life-history traits of bonnethead sharks, Sphyrna tiburo, (Carcharhiniformes: Sphyrnidae) from the eastern Gulf of Mexico. Marine & Freshwater Research 54:875– 883.

Lotze, H. K., and B. Worm. 2009. Historical baselines for large marine animals. Trends in Ecology & Evolution 24:254–62.

Lowe, C. G., B. M. Wetherbee, G. L. Crow, and A. L. Tester. 1996. Ontogenetic dietary shifts and feeding behavior of the tiger shark, Galeocerdo cuvier, in Hawaiian waters. Environmental Biology of Fishes 47:203–211.

109

Lusseau, D. 2004. The hidden cost of tourism: Detecting long-term effects of tourism using behavioral information. Ecology and Society 9.

Madin, E. M. P., L. M. Dill, A. D. Ridlon, M. R. Heithaus, and R. R. Warner. 2015. Human activities change marine ecosystems by altering predation risk. Global Change Biology 22:44-60.

Malcolm, H., W. Gladstone, S. Lindfield, J. Wraith, and T. Lynch. 2007. Spatial and temporal variation in reef fish assemblages of marine parks in New South Wales, Australia—baited video observations. Marine Ecology Progress Series 350:277–290.

Maljković, A., and I. M. Côté. 2011. Effects of tourism-related provisioning on the trophic signatures and movement patterns of an apex predator, the Caribbean reef shark. Biological Conservation 144:859–865.

Maller, R. A., and E. S. DeBoer. 1988. An analysis of two methods of fitting the von Bertalanffy curve to capture-recapture data. Australian Journal of Freshwater Research 39:459.

McAuley, R. B., C. A. Simpfendorfer, G. A. Hyndes, R. R. Allison, J. A. Chidlow, S. J. Newman, and R. C. J. Lenanton. 2006. Validated age and growth of the sandbar shark, Carcharhinus plumbeus (Nardo 1827) in the waters off Western Australia. Environmental Biology of Fishes 77:385–400.

McCauley, D. J., K. A. McLean, J. Bauer, H. S. Young, and F. Micheli. 2012. Evaluating the performance of methods for estimating the abundance of rapidly declining coastal shark populations. Ecological Applications 22:385–392.

McCook, L. J., T. Ayling, M. Cappo, J. H. Choat, R. D. Evans, D. M. De Freitas, M. Heupel, T. P. Hughes, G. P. Jones, B. Mapstone, H. Marsh, M. Mills, F. J. Molloy, C. R. Pitcher, R. L. Pressey, G. R. Russ, S. Sutton, H. Sweatman, R. Tobin, D. R. Wachenfeld, and D. H. Williamson. 2010. Adaptive management of the Great Barrier Reef: a globally significant demonstration of the benefits of networks of marine reserves. Proceedings of the National Academy of Sciences of the United States of America 107:18278–85.

McGarvey, R., G. Ferguson, and J. Prescott. 1999. Spatial variation in mean growth rates at size of southern rock lobster, Jasus edwardsii, in South Australian waters. Marine and Freshwater Research 50:332–342.

McKibben, J. N., and D. R. Nelson. 1986. Patterns of movement and grouping of gray reef sharks, Carcharhinus amblyrhynchos, at Enewetak, Marshall Islands. Bulletin of Marine Science 38:89-110.

110

Meekan, M., and M. Cappo. 2004. Non-destructive Techniques for Rapid Assessment of Shark Abundance in Northern Australia. Australian Government Department of Agriculture, Fisheries and Forestry.

Meyer, C. G., J. J. Dale, Y. P. Papastamatiou, N. M. Whitney, and K. N. Holland. 2009. Seasonal cycles and long-term trends in abundance and species composition of sharks associated with cage diving ecotourism activities in Hawaii. Environmental Conservation 36:104.

Meyer, C. G., J. M. O’Malley, Y. P. Papastamatiou, J. J. Dale, M. R. Hutchinson, J. M. Anderson, M. a Royer, and K. N. Holland. 2014. Growth and maximum size of tiger sharks (Galeocerdo cuvier) in Hawaii. PLoS ONE 9:e84799.

Mora, C., S. Andrèfouët, M. J. Costello, C. Kranenburg, A. Rollo, J. Veron, K. J. Gaston, and R. A. Myers. 2006. Coral Reefs and the Global Network of Marine Protected Areas. Science 312:1750–1751.

Mourier, J., S. C. Mills, and S. Planes. 2013. Population structure, spatial distribution and life-history traits of blacktip reef sharks Carcharhinus melanopterus. Journal of Fish Biology 82:979–993.

Müllner, A., K. Eduard Linsenmair, and M. Wikelski. 2004. Exposure to ecotourism reduces survival and affects stress response in hoatzin chicks (Opisthocomus hoazin). Biological Conservation 118:549–558.

Myers, R. A., and B. Worm. 2003. Rapid worldwide depletion of predatory fish communities. Nature 423:280–283.

Nadon, M. O., J. K. Baum, I. D. Williams, J. M. McPherson, B. J. Zgliczynski, B. L. Richards, R. E. Schroeder, and R. E. Brainard. 2012. Re-creating missing population baselines for Pacific reef sharks. Conservation Biology 26:493–503.

Natanson, L. J., J. J. Mello, and S. E. Campana. 2002. Validated age and growth of the porbeagle shark (Lamna nasus) in the western North Atlantic Ocean. Fishery Bulletin 100:266–278.

Naylor, L. M., M. J. Wisdom, and R. G. Anthony. 2009. Behavioral Responses of North American Elk to Recreational Activity. Journal of Wildlife Management 73:328–338.

Neer, J. A., and B. A. Thompson. 2005. Life history of the cownose ray, Rhinoptera bonasus, in the northern Gulf of Mexico, with comments on geographic variability in life history traits. Environmental Biology of Fishes 73:321–331.

Nelson, G. A. 2014. fishmethods: Fishery science methods and models in R. R package version 1.6-0.

111

Papastamatiou, Y. P. 2008. Movement patterns, foraging ecology, and digestive physiology of blacktip reef sharks, Carcharinus melanopterus, at Palmyra Atoll: A predator dominated ecosystem. PhD thesis, University of Hawaii.

Papastamatiou, Y. P., J. E. Caselle, A. M. Friedlander, and C. G. Lowe. 2009a. Distribution, size frequency, and sex ratios of blacktip reef sharks Carcharhinus melanopterus at Palmyra Atoll: a predator-dominated ecosystem. Journal of Fish Biology 75:647–54.

Papastamatiou, Y. P., C. G. Lowe, J. E. Caselle, and A. M. Friedlander. 2009b. Scale- dependent effects of habitat on movements and path structure of reef sharks at a predator-dominated atoll. Ecology 90:996–1008.

Pardo, S. A., A. B. Cooper, and N. K. Dulvy. 2013. Avoiding fishy growth curves. Methods in Ecology and Evolution 4:353–360.

Pauly, D. 1980. On the interrelationships between natural mortality , growth parameters , and mean environmental temperature in 175 fish stocks. Journal du Conseil International pour l’Exploration de laMer 39:175–192.

Pauly, D. 1995. Anecdotes and the shifting baseline syndrome of fisheries. Trends in Ecology & Evolution 10:430.

Plummer, M. M. 2014. rjags: Bayesian graphical models using MCMC. R package version 3-14.

Quiros, A. L. 2007. Tourist compliance to a Code of Conduct and the resulting effects on whale shark (Rhincodon typus) behavior in Donsol, Philippines. Fisheries Research 84:102–108.

Radtke, R. L., and G. M. Cailliet. 1984. Age estimation and growth of the gray reef shark CarCharhinus amblyrhynchos from the Northwestern Hawaiian Islands. Proceedings of the 2nd Symposium on Resource Investigations in the NWHI, UNIHI-SEA-GRANT- MR-84–01. Honolulu: University of Hawaii Sea Grant. 2:121–127.

Rizzari, J. R., A. J. Frisch, and K. A. Magnenat. 2014. Diversity, abundance, and distribution of reef sharks on outer-shelf reefs of the Great Barrier Reef, Australia. Marine Biology 161:2847–2855.

Robbins, W. D. 2006. Abundance, demography and population structure of the grey reef shark (Carcharhinus amblyrhynchos) and the white tip reef shark (Triaenodon obesus) (Fam. Charcharhinidae). PhD thesis, James Cook University.

Robbins, W. D., M. Hisano, S. R. Connolly, and J. H. Choat. 2006. Ongoing collapse of coral-reef shark populations. Current Biology 16:2314–9.

Roberts, C. M. 2003. Our shifting perspectives on the oceans. Oryx 37:166–177.

112

Robinson, L., and W. Sauer. 2013. A first description of the artisanal shark fishery in northern Madagascar: implications for management. African Journal of Marine Science 35:9–15.

Rosenberg, A. A., W. J. Bolster, K. E. Alexander, W. B. Leavenworth, A. B. Cooper, and M. G. McKenzie. 2005. The history of ocean resources: Modeling cod biomass using historical records. Frontiers in Ecology and the Environment 3:84–90.

Royle, J. A., R. B. Chandler, R. Sollmann, and B. Gardner. 2014. Spatial Capture-recapture. Academic Press, Waltham, Massachusetts.

Royle, J. A., and R. M. Dorazio. 2008. Hierarchical modeling and inference in ecology: the analysis of data from populations, metapopulations and communities. Academic Press. San Diego, California:464pp.

Royle, J. A., R. M. Dorazio, and W. A. Link. 2007. Analysis of Multinomial Models With Unknown Index Using Data Augmentation. Journal of Computational and Graphical Statistics 16:67–85.

Royle, J. A., K. U. Karanth, A. M. Gopalaswamy, and N. S. Kumar. 2009. Bayesian inference in camera trapping studies for a class of spatial capture-recapture models. Ecology 90:3233–3244.

Royle, J. A., and K. V. Young. 2008. A hierarchical model for spatial capture-recapture data. Ecology 89:2281–2289.

Ruegg, K., H. C. Rosenbaum, E. C. Anderson, M. Engel, A. Rothschild, C. S. Baker, and S. R. Palumbi. 2013. Long-term population size of the North Atlantic humpback whale within the context of worldwide population structure. Conservation Genetics 14:103– 114.

Russell, R. E., J. A. Royle, R. Desimone, M. K. Schwartz, V. L. Edwards, K. P. Pilgrim, and K. S. Mckelvey. 2012. Estimating abundance of mountain lions from unstructured spatial sampling. The Journal of Wildlife Management 76:1551–1561.

Sainsbury, K. J. 1980. Effect of individual variability on the von Bertalanffy growth equation. Canadian Journal of Fisheries and Aquatic Sciences 37:241–247.

Sanderson, E. W., M. Jaiteh, M. a. Levy, K. H. Redford, A. V. Wannebo, and G. Woolmer. 2002. The human footprint and the last of the wild. BioScience 52:891.

Sandin, S. A., J. E. Smith, E. E. Demartini, E. A. Dinsdale, S. D. Donner, A. M. Friedlander, T. Konotchick, M. Malay, J. E. Maragos, D. Obura, O. Pantos, G. Paulay, M. Richie, F. Rohwer, R. E. Schroeder, S. Walsh, J. B. C. Jackson, N. Knowlton, and E. Sala. 2008. Baselines and degradation of coral reefs in the Northern Line Islands. PLoS ONE 3:e1548.

113

Scarpaci, C., S. W. Bigger, P. J. Corkeron, and D. Nugegoda. 2000. Bottlenose dolphins (Tursiops truncatus) increase whistling in the presence of “swim-with-dolphin” tour operations. Journal of Cetacean Research and Management 2:183–185.

Scott Baker, C., and P. J. Clapham. 2004. Modelling the past and future of whales and whaling. Trends in Ecology & Evolution 19:365–71.

Semeniuk, C. A. D., and K. D. Rothley. 2008. Costs of group-living for a normally solitary forager: Effects of provisioning tourism on southern stingrays Dasyatis americana. Marine Ecology Progress Series 357:271–282.

Shackley, M. 1998. “Stingray City”- Managing the impact of underwater tourism in the Cayman Islands. Journal of Sustainable Tourism 6:328–338.

Simpfendorfer, C. A. 2000. Growth rates of juvenile dusky sharks , Carcharhinus obscurus (Lesueur , 1818), from southwestern Australia estimated from tag-recapture data. Fishery Bulletin 98:811–822.

Simpfendorfer, C. A., A. B. Goodreid, and R. B. Mcauley. 2001. Size, sex and geographic variation in the diet of the tiger shark, Galeocerdo cuvier, from Western Australian waters. Environmental Biology of Fishes 61:37–46.

Simpfendorfer, C. A., G. R. Poulakis, P. M. O’Donnell, and T. R. Wiley. 2008. Growth rates of juvenile smalltooth sawfish Pristis pectinata Latham in the western Atlantic. Journal of Fish Biology 72:711–723.

Skomal, G., P. S. Lobel, and G. Marshall. 2007. The use of animal-borne imaging to assess post-release behavior as it relates to capture stress in grey reef sharks, Carcharhinus amblyrhynchos. Marine Technology Society Journal 41:44–48.

Skomal, G., and L. Natanson. 2003. Age and growth of the blue shark (Prionace glauca) in the North Atlantic Ocean. Fishery Bulletin 101:627–639.

Smale, A. 2009. Carcharhinus amblyrhynchos. The IUCN Red List of Threatened Species 2009.

Smart, J., A. Chin, A. Tobin, C. Simpfendorfer, and W. White. 2015. Age and growth of the common blacktip shark Carcharhinus limbatus from Indonesia, incorporating an improved approach to comparing regional population growth rates. African Journal of Marine Science 37:177–188.

Smart, J. J., A. Chin, L. Baje, M. E. Green, S. A. Appleyard, A. J. Tobin, C. A. Simpfendorfer, and W. T. White. 2016a. Effects of including misidentified sharks in life history analyses: A case study on the grey reef shark Carcharhinus amblyrhynchos from Papua New Guinea. PLoS ONE 11:e0153116.

114

Smart, J. J., A. Chin, A. J. Tobin, and C. A. Simpfendorfer. 2016b. Multimodel approaches in shark and ray growth studies: Strengths, weaknesses and the future. Fish and Fisheries.

Sminkey, T. R., and J. A. Musick. 1995. Age and growth of the sandbar shark, Carcharhinus plumbeus, before and after population depletion. Copeia:871–883.

Smith, K. R., C. Scarpaci, M. J. Scarr, and N. M. Otway. 2014. Scuba diving tourism with critically endangered grey nurse sharks (Carcharias taurus) off eastern Australia: Tourist demographics, shark behaviour and diver compliance. Tourism Management 45:211–225.

Smith, K., M. Scarr, and C. Scarpaci. 2010. Grey nurse shark (Carcharias taurus) diving tourism: Tourist compliance and shark behaviour at Fish Rock, Australia. Environmental Management 46:699–710.

Smith, S. E., D. W. Au, and C. Show. 1998. Intrinsic rebound potentials of 26 species of Pacific sharks. Marine and Freshwater Research 49:663.

Speed, C. W., M. G. Meekan, I. C. Field, C. R. McMahon, J. D. Stevens, F. McGregor, C. Huveneers, Y. Berger, and C. J. A. Bradshaw. 2011. Spatial and temporal movement patterns of a multi-species coastal reef shark aggregation. Marine Ecology Progress Series 429:261–275.

Stafford-Bell, R., M. Scarr, and C. Scarpaci. 2012. Behavioural responses of the Australian Fur Seal (arctocephalus pusillus doriferus) to vessel traffic and presence of swimmers in Port Phillip Bay, Victoria, Australia. Aquatic Mammals 38:241–249.

Stevens, J., R. Bonfil, N. K. Dulvy, and P. A. Walker. 2000. The effects of fishing on sharks, rays, and chimaeras (chondrichthyans), and the implications for marine ecosystems. ICES Journal of Marine Science 57:476–494.

Stevens, J. D. 1984. Life-History and Ecology of Sharks at Aldabra Atoll, Indian Ocean. Proceedings of the Royal Society of London. Series B, Biological Sciences 222:79– 106.

Stevens, J. D., and K. J. McLoughlin. 1991. Distribution, size and sex composition, reproductive biology and diet of sharks from Northern Australia. Australian Journal of Marine and Freshwater Research 42:151–199.

Stevenson, C., L. S. Katz, F. Micheli, B. Block, K. W. Heiman, C. Perle, K. Weng, R. Dunbar, and J. Witting. 2007. High apex predator biomass on remote Pacific islands. Coral Reefs 26:47–51.

Sybersma, S. 2015. Review of shark legislation in Canada as a conservation tool. Marine Policy 61:121–126.

115

Team, R. 2014. R Development Core Team. R: A Language and Environment for Statistical Computing.

Tester, A. L. 1969. Cooperative shark research and control program; final report 1967- 69. Department of Zoology, University of Hawaii, Honolulu.

Thompson, C. M., J. A. Royle, and J. D. Garner. 2012. A framework for inference about carnivore density from unstructured spatial sampling of scat using detector dogs. The Journal of Wildlife Management 76:863–871.

Titus, B. M., M. Daly, and D. A. Exton. 2015. Do reef fish habituate to diver presence? Evidence from two reef sites with contrasting historical levels of SCUBA intensity in the Bay Islands, Honduras. PLoS ONE 10:e0119645.

Topelko, K. N., and P. Dearden. 2005. The Shark Watching Industry and its Potential Contribution to Shark Conservation. Journal of Ecotourism 4:108–128.

Trebilco, R., J. K. Baum, A. K. Salomon, and N. K. Dulvy. 2013. Ecosystem ecology: Size- based constraints on the pyramids of life. Trends in Ecology & Evolution 28:423–431.

Vannuccini, S. 1999. Shark Utilization, Marketing and Trade. FAO. Fisheries Technical Paper.

Vermeulen, E., A. Cammareri, and L. Holsbeek. 2012. Alteration of southern right whale (Eubalaena australis) behaviour by human-induced disturbance in Bahía San Antonio, Patagonia, Argentina. Aquatic Mammals 38:56–64.

Vianna, G. M. S., M. G. Meekan, J. J. Meeuwig, and C. W. Speed. 2013. Environmental influences on patterns of vertical movement and site fidelity of grey reef sharks (Carcharhinus amblyrhynchos) at aggregation sites. PLoS ONE 8:e60331.

Vianna, G. M. S., M. G. Meekan, D. J. Pannell, S. P. Marsh, and J. J. Meeuwig. 2012. Socio-economic value and community benefits from shark-diving tourism in Palau: A sustainable use of reef shark populations. Biological Conservation 145:267–277.

Vitousek, P. M., H. A. Mooney, J. Lubchenco, and J. M. Melillo. 1997. Human domination of Earth’s ecosystems. Science 277:494–499.

Walker, T. I. 2007. Spatial and temporal variation in the reproductive biology of gummy shark Mustelus antarcticus (Chondrichthyes: Triakidae) harvested off southern Australia. Marine and Freshwater Research 58:67–97.

Walker, T. I., B. L. Taylor, R. J. Hudson, and J. P. Cottier. 1998. The phenomenon of apparent change of growth rate in gummy shark (Mustelus antarcticus) harvested off southern Australia. Fisheries Research: 139–163.

116

Ward-Paige, C. A., C. Mora, H. K. Lotze, C. Pattengill-Semmens, L. McClenachan, E. Arias-Castro, and R. A. Myers. 2010a. Large-scale absence of sharks on reefs in the greater-Caribbean: a footprint of human pressures. PLoS ONE 5:e11968.

Ward-Paige, C., J. Mills Flemming, and H. K. Lotze. 2010b. Overestimating fish counts by non-instantaneous visual censuses: consequences for population and community descriptions. PLoS ONE 5:e11722.

Wass, R. C. 1971. A comparative study of the life history, distribution, and ecology of the sandbar shark and the grey reef shark in Hawaii. PhD thesis, University of Hawaii.

Watson, D. L., E. S. Harvey, M. J. Anderson, and G. A. Kendrick. 2005. A comparison of temperate reef fish assemblages recorded by three underwater stereo-video techniques. Marine Biology 148:415–425.

Watson, R. A., G. M. Carlos, and M. A. Samoilys. 1995. Bias introduced by the non-random movement of fish in visual transect surveys. Ecological Modelling 77:205–214.

Wearmouth, V. J., and D. W. Sims. 2008. Sexual segregation in marine fish, reptiles, birds and mammals behaviour patterns, mechanisms and conservation implications. Advances in Marine Biology 54:107–70.

Wetherbee, B. M., G. L. Crow, and C. G. Lowe. 1996. Biology of the Galapagos shark, Carcharhinus galapagensis, in Hawai’i. Environmental Biology of Fishes 45:299–310.

Wetherbee, B. M., G. L. Crow, and C. G. Lowe. 1997. Distribution, reproduction and diet of the gray reef shark Carcharhinus amblyrhynchos in Hawaii. Marine Ecology Progress Series 151:181–189.

Williams, R., D. Lusseau, and P. S. Hammond. 2006. Estimating relative energetic costs of human disturbance to killer whales (Orcinus orca). Biological Conservation 133:301– 311.

Willis, T., and R. Babcock. 2000. A baited underwater video system for the determination of relative density of carnivorous reef fish. Marine and Freshwater Research 51:655–63.

Willis, T. J., R. B. Millar, and R. C. Babcock. 2000. Detection of spatial variability in relative density of fishes: comparison of visual census, angling, and baited underwater video. Marine Ecology Progress Series 198:249–260.

Wood, C. C., K. S. Ketchen, and R. J. Beamish. 1979. Population Dynamics of Spiny Dogfish (Squalus acanthias) in British Columbia Waters Age-Structure. J. Fish. Res. Board Can 36:647–656.

Worton, B. J. 1995. Using Monte Carlo Simulation to Evaluate Kernel-Based Home Range Estimators. Journal of Wildlife Management 59:794–800.

117

Yasué, M. 2005. The effects of human presence, flock size and prey density on shorebird foraging rates. Journal of Ethology 23:199–204.

118