Bathynomus Giganteus) Using Reflex Impairment

Florida State University Libraries

2016 Post-Release Mortality of Deep Sea Bycatch Species Brendan Suneel Talwar

Follow this and additional works at the FSU Digital Library. For more information, please contact [email protected] FLORIDA STATE UNIVERSITY

COLLEGE OF ARTS & SCIENCES

POST-RELEASE MORTALITY OF DEEP SEA BYCATCH SPECIES

BRENDAN SUNEEL TALWAR

A Thesis submitted to the Department of Biological Science in partial fulfillment of the requirements for the degree of Master of Science

2016

Brendan Suneel Talwar defended this thesis on March 31, 2016. The members of the supervisory committee were:

R. Dean Grubbs Professor Directing Thesis

Edward J. Brooks Committee Member

Don Levitan Committee Member

Joseph Travis Committee Member

The Graduate School has verified and approved the above-named committee members, and certifies that the thesis has been approved in accordance with university requirements.

This thesis is dedicated to Sydney and Sara.

iii

ACKNOWLEDGMENTS

This thesis represents the hard work, generosity, and support of countless friends, family

members, advisors, and students. Let me begin by acknowledging everyone that helped see it through, from heavily invested volunteers to strangers who donated the funds to get this work off the ground- this would not have been possible without you.

I offer my sincerest thanks to my advisor, Dean Grubbs, for his guidance, support, and incredible expertise. The independence I was afforded within the Grubbs Lab pushed me to become heavily invested in my work and ultimately gain more from my Master’s degree than I thought possible, a testament to his mentorship. I must also thank Edd Brooks and Travis Perry for the opportunities that they have given me since the beginning of my career. My limited success so far is largely attributable to their hard work. I also greatly appreciate Don Levitan,

John Mandelman, and Joseph Travis, who improved the quality of this work and act as exceptional role models to many early career scientists including me.

For contributing to my field work, lab work, and general well-being, thank you to numerous hard working volunteers and friends including I Boyoucos, K Durglo, E Van Eepoel,

R Fry, A Gokgoz, C Grossi, K Magnenat, J Mitchell, K Ontiveros, O O’Shea, Team EP, C

Raguse, C Seslar, M Violich, C Ward, and many others at the Cape Eleuthera Institute. I am also forever grateful to the students of The Island School Fall ’14 and Spring ’15 semesters, including

M Abouhamad, K Addams-Pilgrim, C Close, M Edie, S Gallagher, A Heher, N Henderson, A

Hoffman, H Lavelle, O Rask, M Rogers, and L Zachau for their field support and contagious enthusiasm. Coiling rope into a bucket wouldn’t have been the same with anyone else.

For financial and logistical support, I thank Experiment.com crowdfunding donors, the

PADI Foundation, the Cape Eleuthera Foundation, The Island School, the Guy Harvey Ocean

Foundation, the Sigma Xi Scientific Research Society, the New England Aquarium, and the

Florida State University Coastal & Marine Laboratory.

Lastly, I thank my Mom for throwing me into the deep end before I could walk and for sparking my interest in the natural world. I also thank my Dad for his constant encouragement and interest in my work.

TABLE OF CONTENTS

List of Tables ...... viii List of Figures ...... x Abstract ...... xii

1. AN ASSESSMENT OF POST-RELEASE MORTALITY FOR A COMMONLY DISCARDED DEEP-SEA ISOPOD (BATHYNOMUS GIGANTEUS) USING REFLEX IMPAIRMENT ...... 1 1.1 Introduction ...... 1 1.2 Materials and Methods ...... 3 1.2.1 Study Area ...... 3 1.2.2 Evaluating Reflexes ...... 3 1.2.3 Field Trials ...... 3 1.2.4 Data Analysis ...... 4 1.3 Results ...... 5 1.3.1 Capture Characteristics ...... 5 1.3.2 Reflex Action Mortality Predictors ...... 6 1.3.3 Factors Affecting Mortality ...... 6 1.3.4 Post-Capture Behavior ...... 6 1.3.5 Cumulative Effects of Capture and Cage Stress ...... 7 1.4 Discussion ...... 8 1.4.1 Reflex Action Mortality Predictors ...... 8 1.4.2 Factors Affecting Mortality ...... 8 1.4.3 Post-Capture Behavior and Cage Effects ...... 10 1.4.4 Conclusions ...... 11

2. STRESS, POST-RELEASE MORTALITY, AND RECOVERY OF COMMONLY DISCARDED DEEP-SEA SHARKS CAUGHT ON LONGLINES ...... 16 2.1 Introduction ...... 16 2.2 Materials and Methods ...... 17 2.2.1 Longline Sampling ...... 17 2.2.2 Blood Sampling ...... 18 2.2.3 Caging ...... 19 2.2.4 Post-Release Behavior ...... 20 2.2.5 Data Analysis ...... 20

2.3 Results ...... 22 2.3.1 Capture Characteristics ...... 22 2.3.2 Mortality and Blood Chemistry ...... 22 2.3.3 Predicting At-Vessel Blood Chemistry ...... 23 2.3.4 Predicting Post-Release Mortality ...... 23

2.3.5 Post-Release Behavior: S. cubensis ...... 24 2.3.6 Post-Release Behavior: Centrophorus sp...... 25

2.4 Discussion ...... 26 2.4.1 At-Vessel and Post-Release Mortality ...... 26 2.4.2 Stress and Behavior...... 28 2.4.3 Predicting Post-Release Mortality ...... 30 2.4.4 Limitations ...... 33 2.4.5 Conclusions ...... 34

APPENDICES ...... 48

A. ACUC LETTER OF APPROVAL ...... 48

REFERENCES ...... 50

BIOGRAPHICAL SKETCH ...... 61

vii

LIST OF TABLES

1 The reflexes identified for assessing condition of trap-caught Bathynomus giganteus. The test for each reflex is the action required to elicit a given response (i.e. positive or negative), listed in the same order as they were conducted in the field...... 12

2 Summary of negatively scored reflexes for emersion (n=50; 15 minutes) and control (n=50) groups of Bathynomus giganteus prior to caging. Proportions of negative responses for each reflex are shown as percentage of total negative responses for each group ...... 13

3 Full model and minimal adequate model from backwards stepwise GLM analysis to describe 5-day mortality in Bathynomus giganteus based on seven initial model parameters and two interaction terms...... 13

4 Akaike Information Criterion (AIC) values for each GLM used to describe the 5-day mortality of Bathynomus giganteus resulting from stepwise backwards elimination of nonsignificant terms...... 14

5 Description of vitality scores assigned to sharks placed in the post-release cage before being lowered to the sea floor...... 35

6 Correlation structure of primary stress physiology metrics, length measurements, and capture characteristics for S. cubensis...... 35

7 Capture composition and characteristics of sharks caught on deep-sea longlines throughout this study...... 35

8 At-vessel and 24 h post-release mortality rates for species caught in this study, calculated using Eqs. (1) and (2) ...... 36

9 Blood chemistry parameters and corresponding sample sizes for Squalus cubensis and Centrophorus sp. captured during this study ...... 36

10 Full model and minimal adequate model from backwards stepwise analysis to describe at- vessel blood pH in Squalus cubensis based on four initial model parameters and one interaction term ...... 36

11 Akaike Information Criterion (AIC) values for each model used to describe the at-vessel blood pH of Squalus cubensis resulting from stepwise backwards elimination of nonsignificant terms and evaluation of changes in AIC and deviance ...... 37

12 Full model and minimal adequate model from backwards stepwise analysis to describe at- vessel blood lactate in Squalus cubensis based on four initial model parameters and one interaction term ...... 37

viii

13 Akaike Information Criterion (AIC) values for each model used to describe the at-vessel blood lactate of Squalus cubensis resulting from stepwise backwards elimination of nonsignificant terms and evaluation of changes in AIC and deviance...... 38

14 Full model and minimal adequate model from backwards stepwise GLM analysis to describe 24 h mortality in Squalus cubensis based on five initial model parameters and five interaction terms ...... 38

15 Akaike Information Criterion (AIC) values for each GLM used to describe the 24 h mortality of Squalus cubensis resulting from stepwise backwards elimination of nonsignificant terms and evaluation of changes in AIC and deviance ...... 39

16 Significance levels for single variable fits to 24 h S. cubensis mortality data from GLMs with binomial distributions and a logit link function ...... 39

LIST OF FIGURES

1 Mortality of giant isopods as a function of reflex impairment score (sum of negative responses from RAMP assessments). Each point represents the calculated 5-day mortality rate for individuals with a given impairment score, pooled across treatment groups. The solid line shows the curved fit from logistic regression which represents the likelihood of mortality for an individual with a given reflex impairment score ...... 14

2 Number of Bathynomus giganteus with empty and non-empty stomachs grouped by cage result after five days in experimental enclosures at the seafloor ...... 15

3 Sum of negative responses for selected reflexes (mouth closure, leg retraction, telson flexion, and pereopod movement) before and after five days of caging at depth for Bathynomus giganteus that survived experimental trials (n=29) ...... 15

4 Relationship between pH values taken from a pH meter and the iStat point of care device, used to convert values from the iStat into their pH meter equivalents (R2=0.7039, p<0.05) ...... 40

5 Relationship between sea floor temperature (ºC) and depth (m) at our study site in Northeastern Exuma Sound, The Bahamas from July 2014 to July 2015 ...... 40

6 Time of death of Squalus cubensis and Centrophorus sp. that died within the 24 h video monitoring period in post-release cages at depth ...... 41

7 Squalus cubensis at-vessel blood pH as a function of capture depth (m) and total length (cm). Observed blood pH levels are shown with black dots, while the cross-hatched surface represents the values generated from the slope and intercept outputs from the most parsimonious linear model examined...... 41

8 Squalus cubensis at-vessel blood lactate as a function of capture depth (m) and maximum capture duration (min). Observed blood lactate levels are shown with black dots, while the cross-hatched surface represents the values generated from the slope and intercept outputs from the most parsimonious linear model examined ...... 42

9 Squalus cubensis survival probability as a result of total length (cm) and blood lactate (mmol/L). Predicted values were calculated according to the slope and intercept outputs from a GLM fit to mortality with significant predictors only. Lactate and total length inputs were chosen based on the minimum and maximum values reported in this study .42

10 Survival probability curve using at-vessel blood pH measurements for S. cubensis that either survived ( ) or died ( ) after 24 h post-capture. The solid line represents the probability of survival calculated using Eq. 4 and the dashed line represents the 50% chance of survival, calculated at a blood pH of 7.17 ...... 43

11 Post-release mortality of S. cubensis placed in 24 h post-release cages by vitality score (n caged). The distribution of sharks that went on to survive or die after 24 h post-release was significantly different between groups of sharks assigned excellent, fair, and poor vitality scores (Pearson’s x2=11.78, p=0.001, df=2) ...... 43

12 At-vessel S. cubensis blood pH by vitality score. Means that do not share a letter above each box are significantly different as determined by an ANOVA and Tukey’s test (p<0.05) ...... 44

13 Percent of all S. cubensis exhibiting swimming behavior during the 24 h video monitoring period in post-release cages at depth ...... 44

14 Squalus cubensis mean time swimming (calculated for the first minute of each video deployment only) for those that survived or died during the first 900 minutes following cage deployment. Mean time swimming of survivors increased significantly with duration in the cage (r2=.37, p<0.01) ...... 45

15 Squalus cubensis mean time swimming (calculated for the first minute of each video deployment only) for those that survived in either shallow (<625 m) or deep (>625 m) cages during the first 900 minutes following deployment. Mean time swimming increased significantly over time for both shallow and deep groups of survivors (shallow- r2=0.30, p<0.05; deep- r2=0.12, p=0.05), and the rate of increase was marginally higher for the sharks in shallow cages (ANCOVA interaction term, p=0.06) ...... 45

16 Squalus cubensis time of first swimming was significantly earlier for sharks in shallow post-release cages compared to those in deep cages (Mann Whitney U Test, p= <0.05)..46

17 At-vessel Squalus cubensis blood glucose levels were significantly greater for sharks that had a time of first swimming (TOFS) after 120 min post-capture (‘late’) compared to those that had a TOFS earlier than 120 min post-capture (‘early’; t=2.19, p<0.05) ...... 46

18 At-vessel S. cubensis blood lactate levels were significantly lower for active sharks compared to inactive sharks observed during the 24 h post-release caging period as calculated by percent time swimming (<20%- inactive, >20%-active; t=2.56, p<0.05) ...47

19 Squalus cubensis total length was significantly lower for sharks that had a time of death before 120 min post-caging compared to those that had a time of death after 120 min post-caging (Mann-Whitney U Test, p<0.05) ...... 47

ABSTRACT

Deep-sea organisms are increasingly subject to bycatch interactions worldwide. Recent studies have shown that discard mortality can lead to significant declines in deep sea fish stocks, and highlight the inherent vulnerability of deep sea organisms to overexploitation due to their shared suite of conservative life history characteristics. Estimating the post-release mortality (PRM) rates of these deep-sea organisms is a necessary step towards responsible fisheries management, particularly as PRM represents a substantial source of uncertainty when estimating total fishery mortality. The deep-sea giant isopod Bathynomus giganteus and its relatives are captured as bycatch in numerous fisheries, although knowledge is limited regarding their population trends or response to capture and release. In order to assess and predict PRM in B. giganteus, we used reflex action mortality predictors (RAMP) whereby the presence or absence of target reflexes was used to create a delayed mortality model, and considered factors affecting mortality. Mortality rates five days post-capture ranged from 50-100% and both RAMP scores and time at the surface were significant predictors of mortality, although our conclusions regarding the effect of surface time are limited. In-cage video documented little movement within the 24 h monitoring period following cage deployment, and it appeared that surviving individuals often fed within the holding period after cage deployment. Our results suggest that PRM in B. giganteus is common and that this unaccounted source of mortality should be quantified and investigated for other deep-sea crustaceans as well. Similarly, bycatch interactions with deep-sea elasmobranchs can lead to dramatic declines in abundance over short time scales. Sharks hooked in the deep sea could face a higher likelihood of severe physiological disturbance, at-vessel mortality, and PRM than their shallower counterparts. Unfortunately, robust PRM rates have not yet been estimated for deep-sea elasmobranchs and as such are not currently incorporated into total fishery mortality estimates or bycatch assessments, limiting the effectiveness of conservation or management initiatives. We empirically estimated PRM for two focal species of deep-sea shark, the Cuban dogfish Squalus cubensis and the gulper shark Centrophorus sp. using post-release cages deployed at-depth. We calculated 24 h PRM rates of 49.7% (± 8.5 SE) for S. cubensis and 83% (± 16 SE) for Centrophorus sp. and identified shark size (total length), blood lactate, blood pH, and vitality

xii scores as predictors of PRM in Squalus cubensis. We also observed all PRM within 11 h post- capture and demonstrated the effects of capture and recovery depth on stress and behavior. Our results suggest that PRM rates of deep-sea sharks are higher than previously assumed, and highlight the need for filling in this gap in fishery mortality estimates for other common deep-sea discards in the future.

xiii

CHAPTER ONE

AN ASSESSMENT OF POST-RELEASE MORTALITY FOR A COMMONLY DISCARDED DEEP-SEA ISOPOD (BATHYNOMUS GIGANTEUS) USING REFLEX IMPAIRMENT

This article has been accepted for publication in ICES Journal of Marine Science published by Oxford University Press. 1.1 Introduction

Bycatch, defined as non-target species that are discarded or unmanaged, makes up an estimated 40.4% of total catches for global marine fisheries (Davies et al., 2009), and represents a threat to the sustainable management of marine ecosystems (Crowder and Murawski 1998, Harrington et al., 2005, Kelleher, 2005). As a result of low economic value or harvest prohibitions (Harrington et al., 2005), many individuals caught as bycatch are discarded alive with unknown post-release mortality (PRM) rates. In the few instances where it has been estimated, PRM rates vary drastically depending on the fishery, gear type, and taxa in question (Davis, 2002). In general there is a lack of data regarding the contribution of discard mortality to total fishery mortality, which poses a serious challenge to marine fisheries management (Hall et al., 2000; Davis, 2002). Increasingly, commercial fisheries are expanding to the deep-sea (Morato et al., 2006, Watson and Morato, 2013) where species are generally highly susceptible to overexploitation due to their very conservative life histories (Large et al., 2003; Simpfendorfer and Kyne, 2009; Norse et al., 2012). Individuals captured at depth undergo a forced ascent from hundreds of meters deep to the surface facing extreme thermal, barometric, and photic stress (Brooks et al., 2015) and, if discarded alive, must return to depth after experiencing a suite of sub-lethal impairments that may increase the chance of post-release mortality or predation (Wilson et al., 2014). These sub-lethal effects can include physiological disturbance and/or physical injury, which can affect behavior, growth, reproduction and immune function, ultimately reducing fitness in the long term (reviewed by Wilson et al., 2014). In invertebrates, the magnitude of these sub-lethal effects often depends on gear type, capture duration, emersion time, and temperature (Giomi et al., 2005; Ridgeway et al., 2006; Haupt et al., 2006; Wilson et al., 2014).

In deep-water fisheries, the combined effects of high magnitude stressors experienced during capture, ascent, and descent is likely to result in higher PRM rates in these taxa than in their shallower counterparts. In order to predict the PRM rates of crustaceans, recent field studies have been successful in using reflex action mortality predictors (RAMP) (Stoner et al., 2008; Stoner, 2009; Stoner, 2012b; Urban, 2015) whereby the presence or absence of reflexes at-vessel can be used to quickly gauge an individual’s probability of survival. Reflex impairment scores (the sum of negative reflexes for a single assessment) are robust in that they reflect the cumulative effects of various types of stressors to multiple bodily systems (Stoner, 2012a). After calculating delayed mortality curves, which describe the relationship between reflex impairment scores and PRM rates (reviewed by Stoner, 2012a), RAMP can be employed to predict mortality without containment, tagging, or tracking. The deep dwelling giant isopod Bathynomus giganteus is the primary bycatch species in the trap fishery for golden crabs (Chaceon fenneri) (Perry et al., 1995; Harper et al., 2000) and common bycatch in the deep trawl fishery for rock shrimp (Sicyonia brevirostris) and royal red shrimp (Hymenopenaeus robustus) in the northern Gulf of Mexico (R.D. Grubbs, personal observation) as well as the monkfish (Lophius gastrophysus) gillnet fishery in southern Brazil (Perez and Wahrlich, 2005). It has also been documented as bycatch in Brazilian deep water shrimp trawl fisheries (Perez et al., 2013) while a close relative, Bathynomus doederleini, is common bycatch of Taiwanese hagfish (Eptatretus spp. / Paramyxine spp.) trap fisheries (Soong and Mok, 1994). Bathynomus giganteus has also been recently targeted in Japanese fisheries to fuel a new demand for isopod-infused rice crackers. The species inhabits depths from 359 to 1050 m off of the Yucatán Peninsula, Mexico (Barradas-Ortiz et al., 2003), 349 to 733 m in the southern Gulf of Mexico (Briones-Fourzán and Lozano-Alvarez, 1991), 594-1415m in northeastern Exuma Sound, The Bahamas (M. Violich, unpublished data) and to at least 1,735 m in the northern Gulf of Mexico (Grubbs, unpublished data). Whereas B. giganteus primarily scavenges on fish and squid remains (Barradas-Ortiz et al., 2003), active predation of a small squaloid shark has been observed in an experimental enclosure (B. Talwar, unpublished data) and stomach content analysis has suggested a rather wide diet for a strict scavenger (Briones- Fourzán and Lozano-Álvarez, 1991, Barradas-Ortiz et al., 2003).

Here, we assess the PRM of B. giganteus and used RAMP techniques (Stoner et al., 2008) to create a framework for rapidly predicting discard mortality for this common bycatch species in deep-sea fisheries. We also identify factors contributing to PRM and provide insight into the post-release behavior of individuals in experimental enclosures.

1.2 Materials and Methods 1.2.1 Study Area Field work was conducted May-June 2015 in northeastern Exuma Sound, approximately 2.5 km west of Powell Point in Eleuthera, The Bahamas (24.541°N, 76.121°W), where water in excess of 1000 m deep is accessible within 4 km of shore (Brooks et al., 2015).

1.2.2 Evaluating Reflexes Bathynomus giganteus were routinely captured in ongoing trap surveys and examined to identify reflexes that were lost over time while submerged in deck tanks. Target reflexes were chosen based on being 1) quickly and easily tested in the field, 2) consistent and stereotypic, 3) able to be scored as binary positive or negative responses, and 4) unambiguous across experimenters. Based on these criteria, six reflexes were selected for field trials (Table 1).

1.2.3. Field Trials Individuals were captured in a circular 2.5 m diameter trap made of 3.8 cm x 3.8 cm polyvinyl chloride (PVC) coated wire mesh. A rectangular bait cage filled with 1.4 kg of miscellaneous fish parts and/or little tunny (Euthynnus alletteratus) was suspended in the center of the trap and multiple layers of fine wire mesh prevented captured individuals from accessing the bait. A single trap per trial was attached to 1000 m of line and set for approximately 24 hours before being hauled to the surface at 0.3 m/s. Once on the boat, isopods were transferred to coolers filled with water at ambient sea surface temperature (26-28°C) in full sunlight. They were then randomly sorted into two treatment groups: those exposed to 15 minutes of air at 27-28°C and those retained in the coolers with water changes every five minutes (hereafter termed ‘controls’). During this period, we placed unique combinations of multi-colored 15cm zip ties on the pereopods of each individual

for identification. Zip ties were affixed with the minimal tightness required to avoid tag loss and tag ends were removed to allow for freedom of movement and pereopod extension. RAMP assessments were then conducted on haphazardly selected individuals, alternating between an animal exposed to air followed by a control individual. Once the six reflexes (Table 1) were assessed, each animal was assigned a carapace redness score from 0 (white) to 2 (red), which was previously hypothesized to vary with condition, and examined for physical injury. While tagging procedures and RAMP assessments were conducted, the trap was converted into a cage by sealing the two entrances with 1x1cm wire mesh. The bait cage was also opened to allow for access to the fish remains. Lastly, two programmable white LED lights (“Lanternfish”, Blue Turtle Engineering, Florida, USA) and a GoPro Hero 3 White Edition camera programmed with a Time Lapse Intervalometer (Cam-Do, USA) in a Scout Pro H3 deep- sea housing (Group B Distribution Inc, Florida, United States) were synced to record for four minutes every half hour for 24 hours and attached to the inside of the cage before deployment. Three 5 lb floats were attached via stainless steel longline snaps at 20, 60, and 80 m above the cage and an archival temperature and depth recorder (Lotek LAT1400, Newfoundland, Canada) inside of a PVC housing was attached just above the cage bridle prior to cage deployment at the capture location. Floats prevented the line from tangling with the cage or getting stuck on the bottom while the TDR recorded temperature and depth every 4 seconds for roughly 30 h allowing for the calculation of descent and ascent rates. After five days at the seafloor, cages were hauled to the surface and isopods were assessed for mortality and condition using the same RAMP assessments as described previously before being euthanized. Lastly, animals were inspected for physical injuries caused by the caging process and dissected to assess feeding.

1.2.4 Data Analysis Contingency analysis was used to model reflex presence/absence against mortality for animals pooled from both control and air exposed groups. Pearson’s Chi Square tests were then used to test the null hypothesis that the distribution of positive or negative reflexes was equal for groups of survivors and mortalities. Those reflexes that were distributed significantly differently than expected (using an alpha value of α<0.1) were used in all other analyses for predicting mortality.

A generalized linear model with a binomial probability distribution and a logit link function was fitted to the data using Firth adjusted maximum likelihood estimation (Firth, 1993) with total 5-day mortality as the binary response variable (either dead or alive) and nine possible explanatory variables included as predictors. Categorical variables included emersion, sex, carapace redness score, and physical injury. Continuous variables included surface time (the time interval between a trap reaching the boat and the associated cage being lowered to depth from the surface), reflex impairment score (i.e. sum of negative reflexes that were identified using contingency analysis), total length, and the interactions between both surface time and reflex impairment score and total length and reflex impairment score. The order of RAMP assessments across individuals was not included in the model as animals were chosen haphazardly and assessments were brief (<1min). A maximal model including all explanatory variables was fitted and nonsignificant factors were removed in stepwise fashion while evaluating the increases in deviance and Akaike Information Criterion (AIC) with each removal. The model was reduced until the minimal adequate model remained, which included only significant terms or terms that, once removed, caused a significant increase in AIC or deviance (see Crawley, 2007). Lastly, a Student’s t-test was used to look for differences between reflex impairment scores before and after caging. All of the analyses were performed using JMP 7.0.1 (SAS Institute, Cary, NC, USA) and R Programming Language (R Development Core Team, 2008) and the level of significance for the aforementioned tests was α<0.05.

1.3 Results

1.3.1. Capture Characteristics A total of 100 isopods of a mean total length of 28 cm (± 4.67 SD) were captured during four trials, including 77 males and 23 females. Traps were set at a mean depth and temperature of 845 m (±27.5 SE) and 7.49°C (± 3.74 SE) on a muddy bottom with no benthic structure. Physical injury was uncommon, occurring in only 13% of individuals prior to caging, and did not appear significant. Of these injuries, nine were broken pereopod or antennae tips, three were slight carapace breaks along the margins, and one was a small tear to a single pleopod. No at-vessel mortality was observed prior to caging.

1.3.2 Reflex Action Mortality Predictors Presence or absence of the antennae extension and pleopod movement reflexes could not differentiate between survivors and mortalities based on Contingency Analysis (x2= 2.336, p>0.1 for antennae extension; x2=1.597, p>0.1 for pleopod movement), and were thus not included in the reflexes used to predict mortality or in comparisons between control and treatment groups. Animals exposed to 15 minutes of air had a greater number of negative responses for target reflexes than those not exposed (Table 2), however there was no significant difference in 5-day mortality rates between these groups. Across subgroups, the most commonly lost (i.e. negatively scored) reflexes were pereopod movement and mouth closure, followed by telson flexion and leg retraction. Leg retraction was therefore the least sensitive reflex to capture stress and the combination of capture and emersion stress. Presence or absence of mouth closure (x2=2.908, p=0.08), leg retraction (x2=7.059, p=0.0079), telson flexion (x2=8.554, p=0.003), and pereopod movement (x2=10.005, p=0.001) could differentiate between survivors and mortalities and were thus used to create a mortality curve based on the sum (0-4) of lost reflexes (Figure 1).

1.3.3 Factors Affecting Mortality Logistic regression suggested a model including only surface time and reflex impairment score provided the best fit to binomial mortality data (AICfull model = 118.6, AICminimal model =107.7), while sex, physical injury, total length (TL), carapace redness score, and emersion were poor predictors of 5-day mortality (Table 3, 4). Mean surface time was 56 minutes, and ranged from 30-81 minutes across trials. While reflex impairment scores and surface time were correlated (r2=0.32, d.f. = 99, P<0.01) and total length and reflex impairment scores were weakly correlated (r2=0.05, d.f. = 99, P=0.01), the interactions between these terms were not significant and subsequently removed from the model.

1.3.4 Post-Capture Behavior Across the four cage trials, only seven individuals were observed moving during the 24 h video monitoring periods. The first movements were always either pleopod or pereopod twitching. In one trial, for instance, five animals started to move within two hours of reaching the sea floor, and one individual even began scavenging on a dead conspecific and on fish remains

shortly thereafter. The only other two isopods that exhibited any life were observed moving after 8.5 and 10 hours had elapsed during two other trials. Swimming behaviors were irregular, if they occurred at all, and were characterized by an inability to orient effectively before landing inverted on the cage floor after a short burst of activity. Furthermore, recovery from capture appeared very slow, with only 24% of surviving isopods exhibiting any sign of movement within the first 24 h following cage deployment. In all trials, small (<5 cm) isopods entered the cage through the mesh walls within the first hour and swam haphazardly throughout the cage, settling on both the bait cage and less frequently on unmoving B. giganteus. Live, non-caged B. giganteus also appeared on the outside of the cage within three hours and appeared highly active, regularly swimming or crawling on the cage walls with constant antennae movement and normal, highly oriented swimming behavior. Many were also observed excavating in the muddy benthos in an attempt to access fish remains or dead isopods from beneath the cage floor, at times concentrating around clusters of unmoving isopods within the cage. None showed signs of aversion to the lights. These high levels of activity are in contrast to caged individuals, which moved very little, if at all, during the 24 h post-capture observation period. Stomachs were present in 71% (n=71) of individuals after the 5-day period, while 29% (n=29) were completely devoid of stomachs or internal organs, suggesting that they were scavenged or preyed upon. Of the 71 isopods with stomachs, 42% (n=29) survived the trials and 64% (n=45) had stomach contents (Figure 2). Furthermore, of the 29 that survived, 93% (n=27) had non-empty stomachs.

1.3.5 Cumulative Effects of Capture and Cage Stress Reflex impairment scores were significantly higher after caging (mean = 1.9) than they were prior to caging (mean = 1.3) in those individuals that survived experimental trials (t = 2.39, d.f.= 28, p<0.05; Figure 3). After caging, the telson flexion reflex was lost most often followed by mouth closure, pereopod movement, and leg retraction. Pereopod movement was the only

reflex that was present more often after the trials than before (sum of negative responses = 15 before, 9 after).

1.4 Discussion

This study found that reflex impairment scores were successful in predicting mortality for the deep-sea giant isopod Bathynomus giganteus after five days post-capture, with observed PRM rates ranging between 50-100%. While reflex impairment scores generated from RAMP assessments are a common tool used to predict mortality across a wide range of marine species such as decapod crustaceans (Bergmann and Moore, 2001, Stoner et al., 2008, Stoner, 2011,Urban, 2015,), teleosts (Davis and Ottmar, 2006, Davis, 2007, Raby et al., 2012 ), and elasmobranchs (Danylchuk et al., 2013, Gallagher et al., 2014b), no methods to predict mortality had been previously tested for deep sea taxa or for marine isopods despite their prevalence as bycatch in emerging deep-sea fisheries (e.g. Perez and Wahrlich, 2005).

1.4.1 Reflex Action Mortality Predictors Only four of six reflexes that were identified in pre-trial assessments were able to differentiate between survivors and mortalities. While all six were stereotypic and repeatable, rough weather conditions during field trials likely contributed to the inaccurate assessment of responses for both pleopod movement and antennae extension. Of the four remaining reflexes, pereopod movement was the most sensitive to capture stress, followed by mouth closure, telson flexion, and leg retraction. Telson flexion and leg retraction were maintained most frequently, and are possible anti-predator responses for an impaired isopod attempting to protect its vulnerable ventral surface.

1.4.2 Factors Affecting Mortality Surface time was the only variable other than reflex impairment score that could predict mortality. Its effect was largely driven by the extremely high 5-day mortality rate (90%) for individuals that remained at the surface for 81 minutes, and it should be investigated more thoroughly using set increments in the future. Here, its inclusion in the sampling design was largely a result of field constraints on gear and personnel, and conclusions regarding its importance are limited. Surprisingly, emersion (air exposure) had no effect on mortality, although previous studies have shown that emersion stress, which increases with air temperature and duration, can

severely disrupt physiological function in decapod crustaceans (Ridgway et al., 2006). Ridgway et al. (2006) and Spicer et al. (1990) reported that acidemia due to an increase in L-lactate and haemolymph CO2 during emersion may suggest that the Norway lobster (Nephrops norvegicus) cannot maintain an adequate oxygen supply when exposed to air. Bathynomus giganteus, however, is particularly adapted to the low-oxygen environment characteristic of the deep ocean between 400-1000m (Childress and Seibel, 1998), and could thus be less sensitive to a limited oxygen supply. Perhaps, then, temperature is a more important factor to consider than air exposure or desiccation for B. giganteus. Indeed, the thermal gap experienced by an organism caught in the deep-sea in the summer subtropics is exceptionally high (19°C in the present study), and studies on other crustaceans have shown that thermal stress does affect survival. For a trawl-caught portunid crab Liocarcinus depurator, a thermal gap of 12-14°C resulted in 96% mortality whereas a thermal gap of 0-3°C resulted in 2% mortality (Giomi et al., 2008). Similarly, in the trawl-caught shrimp Crangon crangon, holding temperatures over 20°C resulted in high mortality rates whereas temperatures between 10-20°C resulted in no mortality (Gamito and Cabral, 2003). Further research on the effect of thermal stress is certainly warranted for deep-sea crustaceans. Along with thermal stress, deep-sea organisms undergo drastic changes in pressure and light levels during gear retrieval. Barometric stress is not typically considered relevant for crustaceans (Stoner 2012a), although it has not been examined for those residing in deep water. Basti et al. (2010), however, did show that increased hauling speed and increased capture depth negatively affected the survival of the American lobster (Homarus americanus) retrieved from <200 m. Photic stress can also have deleterious effects in crustaceans (Stoner 2012a). Gaten (1988), for instance, showed that deep-dwelling decapods can experience damage to the photoreceptor layer and are susceptible to morphological changes (e.g. cone shape) within the eye after excessive exposure to sunlight. Similar results have been reported by Loew (1976) for the Norway lobster (Nephrops norvegicus) and by Meyer-Rochow (1981) for deep-sea Antarctic amphipods. Furthermore, Herring et al. (1999) suggested that exposure to submersibles’ floodlights could blind shrimp associated with deep-sea vent communities and Nilsson and Lindstrom (1983) reported that exposure to light of 1,250 lux (approximately that of an overcast day) led to complete breakdown of the visual structures of the deep-sea isopod Natatolana

(Cirolana) borealis. Chamberlain et al. (1986) described the morphology of the compound eye of the giant isopod examined here (B. giganteus) and unsurprisingly suggested any exposure to daylight causes severe and irreversible damage to their photoreceptors. Given the low-light levels where giant isopod occur, one may predict that vision may not be an important sensory system for such deep-sea species, however the complexity of the compound eyes and their forward orientation suggests they are used not only to detect light but also to estimate distance (Chamberlain et al. 1986). Thus, damage to the visual systems of deep sea crustaceans and the resultant loss of orientation could be a major source of delayed post-release mortality. Due to the relatively benign traps used in this study, physical injury was rare after the initial capture event and did not affect mortality. Negative gear interactions would likely be more common in other capture gears (e.g. trawl or gillnet; Rose, 1999) and could result in elevated PRM rates compared to those reported here.

1.4.3 Post-Capture Behavior and Cage Effects While recovery was quite slow, a high percentage of surviving individuals had stomach contents, suggesting that many individuals recovered to the point of feeding after video monitoring stopped. Still, numerous individuals had non-empty stomachs and did not survive the caging process, suggesting that they had failed to evacuate food eaten prior to their initial capture or that they had recovered and eaten before they perished. Mortality, therefore, could have taken place late in the 5-day period as a result of confinement stress, delayed effects of capture, or predation from other small isopod species that could fit through the cage material. The caging process introduced a number of confounding factors in our estimation of PRM for each reflex impairment score. Caging could have increased the stress on ‘released’ individuals and artificially inflated the documented PRM rates as suggested by the increase in negative reflex scores after the second haul for surviving isopods. Alternatively, the PRM rate could have been deflated as the cages prevented predation from animals larger than the cage mesh. Given that the isopods exhibit extremely limited movement immediately after capture, and that isopods released at-vessel show little sign of swimming behavior, descending over 800m could take multiple hours. As such, numerous predators in the water column attracted by gear retrieval could easily prey on discarded B. giganteus. Although natural predators of B. giganteus are not well documented, a tiger shark has been found with a giant isopod in its gut (Briones-

Fourzán and Lozano-Álvarez, 1991) and presumably other large bodied elasmobranchs and teleosts would prey on impaired isopods in mid-water.

1.4.4 Conclusions Most importantly, this study showed that reflex impairment scores developed using a RAMP methodology can predict post-release mortality for discarded B. giganteus, which was estimated at greater than 50%. Furthermore, it suggests that a short emersion period may be of little consequence to the post-release survival of this species. It also highlights the need for additional research into the effect of surface time on PRM, including a better understanding of the relative contributions of thermal gap, pressure, and light changes on the mortality of discarded deep-sea crustaceans, particularly as fisheries continue to expand into the deep-sea and bycatch interactions increase. These results further underscore the prevalence of cryptic discard mortality and emphasize the need to better account for it when estimating bycatch mortality.

Table 1: The reflexes identified for assessing condition of trap-caught Bathynomus giganteus. The test for each reflex is the action required to elicit a given response (i.e. positive or negative), listed in the same order as they were conducted in the field.

Reflex Test Orientation Positive Response Negative Response Pereopods fall limply back Pereopods resist extension and/or Leg Extend first paired In water, into the start position and retract strongly to the start retraction pereopods ventral side up present no resistance to position extension Mandibles resist opening and, Open mandibles with a Mandibles fall limply into Mouth In water, upon removal of probe, close blunt probe, then the start position and show closure ventral side up quickly and tightly or open and remove the probe no resistance to opening close rapidly

Pleopods undulate or, if Brush the pleopods with Pleopods fall limply to rest Pleopod In water, contracted inward towards the a blunt probe towards flat against the body and movement ventral side up midline of the carapace, resist the mouth show no movement stimulation

Manually stimulate the Pereopod Out of water, Pereopods move spontaneously Pereopods are motionless pereopods with a blunt movement ventral side up when stimulated after stimulation probe

Manually extend both Antennae move after stimulation Antennae hang limply and Antennae Out of water, antennae straight out in or rapidly return to the start exhibit no response to extension ventral side up front of the body position stimulation

Pull the telson Telson resists the flattening Telson Out of water, Telson exhibits no resistance downward until it is at a motion and/or curls upward past flexion ventral side up or motion >180° angle to the body the original start postion

1312

Table 2. Summary of negatively scored reflexes for emersion (n=50; 15 minutes) and control (n=50) groups of Bathynomus giganteus prior to caging. Proportions of negative responses for each reflex are shown as percentage of total negative responses for each group.

Treatment type Air Emersion No Air Emersion % of No. of negative % of negative No. of negative negative Reflex responses responses responses responses Leg retraction 15 13.6 9 9.7 Telson flexion 26 23.6 23 24.7 Mouth closure 28 25.5 25 26.9 Pereopod movement 41 37.3 36 38.7 Totals 110 100 93 100.0

Table 3: Full model and minimal adequate model from backwards stepwise GLM analysis to describe 5-day mortality in Bathynomus giganteus based on seven initial model parameters and two interaction terms.

Parameter Estimate S.E. z-value Pr(>|z|) Full model Intercept -0.846 2.315 -0.365 0.715 Sex -0.372 0.657 -0.567 0.571 Physical Injury -0.550 0.853 -0.645 0.519 Emersion -0.391 0.560 -0.699 0.485 TL -0.039 0.107 -0.359 0.719 TL*Reflex impairment score -0.001 0.054 -0.019 0.985 Surface Time*Reflex impairment score 0.004 0.013 0.301 0.764 Carapace redness score 0.630 0.503 1.253 0.21 Surface Time 0.018 0.033 0.552 0.581 Reflex impairment score 0.446 1.092 0.409 0.683 Most parsimonious model Intercept -1.53 0.631 -2.426 ** Surface Time 0.028 0.013 2.15 * Reflex impairment score 0.505 2.42 2.089 * Estimates and standard errors (S.E.) are on a logit scale. *p<0.05, **p<0.02

Table 4. Akaike Information Criterion (AIC) values for each GLM used to describe the 5-day mortality of Bathynomus giganteus resulting from stepwise backwards elimination of nonsignificant terms.

Model AIC Full Model 118.56 - TL* Reflex impairment score 116.560 - Surface time * Reflex impairment score 114.660 - Sex 112.970 - Emersion 111.450 - Physical injury 109.670 - TL 108.720 - Carapace redness score 107.770 Most parsimonious model 107.770 - Reflex impairment score 110.490 - Surface time 110.660

Figure 1. Mortality of giant isopods as a function of reflex impairment score (sum of negative responses from RAMP assessments). Each point represents the calculated 5-day mortality rate for individuals with a given impairment score, pooled across treatment groups. The solid line shows the curved fit from logistic regression which represents the likelihood of mortality for an individual with a given reflex impairment score.

Non-empty stomach 30 Empty stomach 25

Number individuals of Number 5

0 Survivors Mortalities

Figure 2. Number of Bathynomus giganteus with empty and non-empty stomachs grouped by cage result after five days in experimental enclosures at the seafloor.

14 Before caging

12 After caging 10 8 6 4

Number individuals of Number 2 0 0 1 2 3 4 Reflex impairment score

Figure 3. Sum of negative responses for selected reflexes (mouth closure, leg retraction, telson flexion, and pereopod movement) before and after five days of caging at depth for Bathynomus giganteus that survived experimental trials (n=29).

CHAPTER TWO

STRESS, POST-RELEASE MORTALITY, AND RECOVERY OF COMMONLY DISCARDED DEEP SEA SHARKS CAUGHT ON LONGLINES

2.1 Introduction

In recent decades, commercial fisheries have expanded into the deep-sea (below 200 m) (Morato et al., 2006) due to advancements in fishing technology and declines in some coastal stocks (Cotton and Grubbs, 2015). Unfortunately, deep-sea fishes are highly susceptible to overexploitation due to their very conservative life histories (Large et al., 2003; Simpfendorfer & Kyne, 2009; Norse et al. 2012). Garcia et al. (2008) suggest that the average fishing pressure that it would take to drive a deep-sea species extinct is only 58% of that for a continental shelf species, and, as would be expected, rapid depletion and abandonment of deep-sea fish stocks has been documented repeatedly (Koslow et al., 2000, Graham et al., 2001, Jones et al., 2005, Devine et al., 2006, Norse et al., 2012). Deep-sea elasmobranchs are perhaps the least resilient fishes to exploitation as their maximum rates of population growth are at the lower end for chondrichthyans, making them among the lowest observed for any species (Kyne and Simpfendorfer, 2007; Norse et al., 2012). Furthermore, the greater a species’ capture depth, the more vulnerable it is to capture-induced stress as a result of decreased metabolic capacity, lowered energy stores (McClain et al., 2012; Koslow, 1996), and greater changes in temperature, pressure, and light levels experienced during the forced and rapid ascent to the surface. These and other factors can interact to increase the likelihood of at-vessel mortality or cryptic post-release mortality (PRM) after a capture event (Skomal and Mandelman, 2012; Brooks et al., 2015). At-vessel and PRM rates in elasmobranchs are species-specific and highly variable (Morgan and Burgess, 2007, Enever et al., 2009, Hale and Carlson, 2010, Braccini et al., 2012, Coelho et al., 2012, Gallagher et al., 2014a) depending on factors such as the gear type, capture

duration, respiratory mode, and metabolic capacity for the species in question (Davis, 2002, Mandelman and Skomal, 2009, Dapp et al., 2015). Similarly, the degree of physiological disturbance and/or physical injury experienced by a released individual can vary greatly, and may result in sub-lethal effects such as impaired behavior, growth, or immune function that ultimately lead to post-release predation or reduced fitness (Davis, 2002, Raby et al., 2014, Wilson et al., 2014). This variability in the sub-lethal and lethal effects of capture is currently not incorporated into fishery mortality assessments despite data that suggest that PRM of longline-caught deep- sea elasmobranch discards can be common (Brooks et al., 2015). Overall there remains a lack of empirically estimated PRM rates for discarded deep-sea sharks, many of which are thrown back due to low economic value or due to harvest prohibitions requiring their release. As deep-sea elasmobranchs are commonly caught as bycatch in fisheries targeting teleosts and crustaceans worldwide (Cotton & Grubbs, 2015), there is the potential that total fishery mortality estimates for these species are underestimated as a result of not accounting for discard mortality, or conversely overestimated by ignoring the potential for survivors, likely limiting the effectiveness of management efforts (Coggins et al., 2007; Molina and Cooke, 2012). Our goals for this study were to empirically estimate the 24 h PRM rates of deep-sea sharks, predict PRM using blood chemistry parameters and vitality scores, identify capture characteristics contributing to at-vessel stress, and shed light on the post-release behavior of individuals held in cages at the seafloor. Our primary species of interest was the Cuban dogfish (Squalus cubensis), the most commonly encountered squalid in the deep reef fish and tilefish fisheries of the northern Gulf of Mexico (National Marine Fisheries Service, unpublished data; Jones et al. 2013), where over 95% are discarded alive (Hale and Carlson, 2010, Gulak et al., 2012). Our secondary species of interest was the gulper shark (Centrophorus sp.), which is part of one of the most highly exploited species complexes of deep sea sharks to date (Kyne and Simpfendorfer, 2007; Kyne et al., 2012).

2.2 Materials and Methods 2.2.1 Longline Sampling We conducted our field work from July 2014 to June 2015 in northeastern Exuma Sound, approximately 2.5 km west of Powell Point on Eleuthera, The Bahamas (24.541°N, 76.121°W).

Standard demersal longlines were set in 450-900 m of water during daylight hours only. Mainline length was a minimum of 1.5 times the water depth to ensure that hooks were located on the bottom, resulting in lines of 1500-2000 m long. Longlines consisted of a grapnel anchor or weight to attach the mainline to the seafloor, twenty to thirty baited circle hooks (10/0 or 12/0) spaced 5-10 m apart, and an archival temperature and depth recorder (TDR) (Lotek LAT-1400, Newfoundland, Canada) placed 5 m from the last hook. The TDR recorded depth and temperature every four seconds, allowing for the calculation of longline descent and ascent rates, and longline depths and temperatures were recorded as the deepest and coldest points measured for a given dataset. Furthermore, the TDR allowed for the calculation of maximum capture duration for each individual, defined as the time between the line reaching its maximum depth and the time when an animal was unhooked. Hooks were baited with miscellaneous fish scraps and/or little tunny (Euthynnus alletteratus) and soak times were roughly 3.5 h. After the desired set duration, longlines were hauled using a commercially available electric pot hauler (Waterman Industries of Florida, Inc., Odessa, FL) at a rate of 0.3 m/s. Once at the boat, sharks were sequentially unhooked and placed in a large cooler to minimize air exposure for the remaining workup, during which sharks were measured for pre-caudal, fork, and total lengths and assessed for maturity based on external morphology and/or published size-at-maturity data. A small fin clip was then removed for genetic analyses. Fin clips were taken from a unique location to distinguish individuals while in the post-release cage during the subsequent 24 h of monitoring.

2.2.2 Blood Sampling Sharks were placed into tonic immobility and blood (~3 mL) was drawn by caudal venipuncture using a 25.4 mm, 22 gauge needle and either a 3 mL or 5 mL heparinized syringe. Roughly 95 μl of blood was then inserted into an iStat CG4+ cartridge, which was analyzed by an iStat blood gas analyzer (Heska Corporation, Fort Collins, CO, USA) thermoset to 37 ºC to determine blood lactate and pH levels (Mandelman and Skomal, 2009, Gallagher et al., 2010, Harter et al., 2015). Simultaneously, 1 mL of blood was transferred to a 1.5 mL eppendorf tube and analyzed by a waterproof pH meter (Hanna Instruments, Woonsocket, RI, USA) to determine blood temperature and pH. Immediately following these analyses, one drop of blood was placed directly onto an Accu-Chek glucose meter strip (Roche Diagnostics, Basel,

Switzerland) to determine blood glucose levels (validated by Cooke et al., 2008 for fish) and one drop was placed directly on a Lactate Plus Meter test strip (Nova Biomedical, Waltham, MA, USA) to determine blood lactate levels in the event of an iStat cartridge error (see Awruch et al., 2011 for validation of a similar unit for elasmobranchs). Blood chemistry analysis typically occurred within one minute following caudal venipuncture. The remaining blood sample was injected into a 10 mL vacutainer coated with lithium heparin (Becton, Dickinson and Co., Franklin Lakes, NJ, USA) and placed on ice until we reached the dock. In the lab, a micro-hematocrit tube (Drummond Scientific, Broomall, PA, USA) was filled with a small sample of whole blood and sealed with Critoseal (McCormick Scientific, St. Louis, MO, USA) before it was spun in a micro-hematocrit centrifuge (LW Scientific Zippocrit, Atlanta, GA, USA) at 4400 g for 4.5 min. Hematocrit was calculated as the percentage of total blood volume made up of red blood cells.

2.2.3. Caging Immediately after taking a fin clip, animals were placed into a circular post-release cage attached to the side of the boat and assigned a vitality score (Table 5). The cage was constructed of 3.8 cm x 3.8 cm polyvinyl chloride (PVC) coated wire mesh reinforced with PVC struts and measured roughly 2.5 m in diameter. After all individuals from a given longline set were added, the cage door was tied shut with a galvanic timed release (Neptune Marine Products, Port Townsend, WA, USA), allowing the cage door to fall open after 20-22 h so that surviving sharks could swim out. The cage was then lowered to the seafloor at a rate of 0.49 m/s as close to the capture location as possible. A TDR was attached 5 m above the cage bridle and two floats were attached to the mainline with stainless longline snaps at 50 m and 100 m from the cage to prevent the mainline from getting tangled with the cage material. Cage depth, temperature, ascent, and descent rates were calculated as discussed previously. Upon reaching the sea floor, the cage was pulled onto its side by the drag of the boat and buoys at the surface. It then remained untouched for 24 h before being hauled to the surface. Any surviving animals still in the cage either swam out during ascent or were released at the boat, while dead individuals were retained for dissection.

2.2.4. Post-Release Behavior Two programmable white LED lights (“Lanternfish”, Blue Turtle Engineering, Florida, USA) and a GoPro Hero 3 White Edition camera programmed with a Time Lapse Intervalometer (Cam-Do, USA) in a Scout Pro H3 deep-sea housing (Group B Distribution Inc, Florida, United States) were synced to record for four minutes every half hour for 24 hours and attached to the inside of the cage before deployment. Videos were later analyzed for time of first swimming (‘TOFS’, defined as time of first sustained forward movement) and time of death (defined as the last time an animal was observed ventilating) for each individual, and total seconds swimming was recorded for the first minute of each four minute video segment for each animal. Percent time swimming was then calculated for each animal by dividing its total time swimming across the first minute of all video segments in a given cage drop by the total time during which that animal was alive and observed. This metric was then binned into active (>20% swimming) and inactive (<20% swimming) categories for analysis. Similarly, time of death and TOFS were binned into early (<120 min post-capture) and late (>120 min post-capture) categories.

2.2.5. Data Analysis At-vessel mortality rates for all shark species were calculated as the percentage of the total catch of a species found to be dead upon first handling. Twenty-four hour PRM rates and standard errors (SE) were calculated using Eqs. (1) and (2) outlined in Pollock and Pine (2007), where M is the species-specific PRM rate and r is the number of cages.

(Eq. 1)

(Eq. 2)

All further statistical analyses were only conducted for Squalus cubensis due to low sample sizes for other species. Blood chemistry parameters were evaluated using the Shapiro- Wilk test for normality and outliers identified and removed using diagnostic plots in R

Programming Language (R Development Core Team, 2008). Blood chemistry parameters and total lengths were then re-scaled into measurements of deviation from the mean for use in generalized linear models (GLM). Blood pH values were taken directly from the waterproof pH meter, which measures pH as a function of blood temperature, or, if measurements were taken only from the iStat point of care device, converted from raw iStat values to their pH meter equivalents using an equation from linear regression relating values from both instruments when run simultaneously (Fig. 4). To predict PRM, a GLM with a binomial probability distribution and a logit link function was fitted to the data using maximum likelihood estimation. The model describes the relationship between 24 h mortality as a binary response variable and five potential explanatory variables as well as interaction terms. A random effect to account for cage deployment was initially included, but as it accounted for <1% of the variance in the full model it was removed. The possible main effects that were included were the continuous variables of blood pH, blood lactate, blood glucose, hematocrit, and total length as well as the interaction terms for lactate and total length, lactate and glucose, pH and lactate, pH and total length, and total length and glucose (see Table 6 for correlation structure). Nonsignificant factors were removed in stepwise fashion while evaluating the increases in deviance and Akaike Information Criterion (AIC) with each removal. The model was reduced until the minimal adequate model remained, which included only significant terms or terms that, once removed, caused an increase in AIC or deviance (see Crawley, 2007). Significant terms from the minimal model were then used to create a GLM predicting mortality as described above. Further, GLMs with only one explanatory variable were explored to identify a single blood chemistry parameter that provides the best fit to mortality. Linear models were then used to determine what capture characteristics best describe blood lactate and blood pH among maximum hooking duration, sea surface temperature, total length, capture depth, and the interaction between capture depth and total length. As capture temperature and depth were tightly correlated (R² = 0.9566, p<0.001; Fig. 5), only depth was included in these models, but biologically represents the cumulative effects of all abiotic variables associated with increased depth. Full models were fit to the response variable and terms removed as described previously. Pearson’s Chi Square tests were then used to test the null hypothesis that the distribution of survivors and mortalities was equal across vitality scores. Vitality scores were then examined

to identify blood chemistry parameters that differed between groups using one-way ANOVAs and Tukey’s tests. Further, the relationship between time post-caging and mean time swimming (pooled across all animals for each video segment) was examined with linear regressions and compared between groups of survivors and mortalities as well as within survivors for those at shallow (<625 m) and deep (>625 m) cage depths. The rate of increase in mean time swimming was compared among these groups using an ANCOVA. Lastly, binned swimming behaviors and times of death were compared with T-tests and/or Mann-Whitney U tests, as were blood chemistry parameters between species, depending on whether or not data were normally distributed. All analyses were performed using JMP 7.0.1 (SAS Institute, Cary, NC, USA) and R Programming Language (R Development Core Team, 2008) and the level of significance for all tests was α<0.05. Graphs were created using SigmaPlot (Sigmaplot, vers. 11.0, Systat Software Inc., Richmond, CA, USA).

2.3 Results 2.3.1 Capture Characteristics A total of 108 sharks from six species were captured over 72 longline sets fishing at a mean depth of 628 m and a mean temperature of 11.95 °C (Table 7). Seventy-seven of these individuals were caged over 37 trials, with post-release cages resting at a mean depth of 641 m and a mean temperature of 11.73 °C after descending at a mean rate of 0.49 m/s (range 0.28-0.59 m/s). The maximum number of sharks placed in a single cage was six, but was more commonly 1-3 animals/cage. The sea surface temperature over the study period ranged from 24 to 30 °C. Sharks were hooked in the jaw or soft palate except for one instance where an animal was hooked through the right spiracle. Physical injury at-vessel was documented in only one S. cubensis (broken jaw) and one M. canis insularis (secondary hooking in the pectoral fin). These individuals were not included in post-release caging trials. We saw no evidence of barotrauma.

2.3.2 Mortality and Blood Chemistry At-vessel mortality rates ranged from 8.33-100% across species while 24 h PRM rates for S. cubensis, Centrophorus sp., and Mustelus canis insularis ranged from 49.7 – 83%, although sample sizes were limited for Centrophorus sp. and M. canis insularis (Table 8). No individuals

were recaptured from this study; however, one S. cubensis was recaptured from a study by Brooks et al. (2015) after it was initially tagged in 2010. The mean time of death was 190 min (± 43.8 S.E.) post-capture for S. cubensis and 260 min (± 65.6 S.E.) post-capture for Centrophorus sp. that died within the 24 h caging period (Fig. 6). All mortalities were observed within 690 min post-caging. Blood glucose levels were significantly lower in Centrophorus sp. compared to those in S. cubensis (Student’s T-Test, p<0.05). There were no significant interspecific differences in blood pH, lactate or hematocrit levels (Table 9).

2.3.3 Predicting At-Vessel Blood Chemistry Analysis of linear models determined that a model including only capture depth and total length provided the best fit to at-vessel blood pH levels (AICFull Model = -54.08, AICReduced Model = -58.43). Both capture depth and total length were significant predictors of at-vessel blood pH (Table 10,11), which decreased with decreasing capture depth and total length values (Fig. 7). Analysis of linear models determined that a model including only capture depth and maximum hooking duration provided the best fit to at-vessel blood lactate levels (AICFull Model =

103.2, AICReduced Model =99.56). Only capture depth was a significant predictor of at-vessel blood lactate (Table 12, 13) in the reduced model, which increased with longer maximum hooking durations and shallower capture depths (Fig. 8).

2.3.4 Predicting Post-Release Mortality The GLM analysis determined that a model including glucose, hematocrit, lactate, total length, and the interaction between glucose and total length provided the best fit to binary 24 h mortality data (AICFull Model = 34.04, AICReduced Model = 26.80). Only lactate and total length, however, were significant predictors of mortality (Table 14, 15). A logistic model including only these significant terms was used to predict 24 h mortality for practical use in a fishery context. To obtain the probability of survival for an individual with known total length and blood lactate level, the maximum likelihood estimates (b0 = -1.84815, b1=

-0.29978, b2= 0.08717) for the survival curve were substituted into the response function in Eq. (3):

( ) = (Eq. 3) ( ) exp �0+�1�1+�2�2 Based�� on1+ thisexp model,�0 a+� probability1�1+�2 �of2 survival of 0.5 was found at a lactate value of 10.7 mmol/L for a shark of average total length (58 cm). The probability of survival then decreased with higher blood lactate levels and smaller total lengths (Fig. 9). The best standalone predictor of mortality was then identified (again for ease of practical application) as blood pH (Table 16), which was modeled to create a logistic survival curve using

the maximum likelihood estimates (b0 = -51.873, b1=7.234) substituted into the response function (Eq. 4): ( ) = (Eq. 4) ( ) exp �0+�1�1 ��Based on1+ thisexp model,�0 +�a probability1�1 of survival of 0.5 was found at an at-vessel blood pH value of 7.17. The probability of survival then decreased as pH declined (Fig. 10). Vitality scores were distributed differently than expected (Pearson’s x2=11.78, p=0.001, df=2), suggesting that these scores are good predictors of mortality. Of those individuals assigned a score of ‘excellent’, only 21% died whereas a score of ‘fair’ resulted in 42% mortality and a score of ‘poor’ resulted in 100% mortality (Fig. 11). A one-way ANOVA revealed differences in at-vessel blood pH between S. cubensis assigned vitality scores from poor to excellent (F2, 59= 8.42, p<0.01). Those assigned a vitality score of poor (M=7.01 ± 0.05) had a significantly lower at-vessel blood pH than those assigned a vitality score of excellent (M= 7.27 ± 0.04) or fair (M= 7.23 ± 0.03) according to Tukey’s test (Fig. 12).

2.3.5 Post-Release Behavior of S. cubensis Squalus cubensis swimming behaviors were normal (e.g. correct orientation, resting on the bottom, exploring the cage) and survivors often swam in circles around the cage perimeter. The mean time of first swimming (TOFS) was 113 min (± 17.8 S.E.) post-caging for S. cubensis that survived the 24 h caging period, while for those that died it was 172.5 min (± 74.8 S.E.) post-capture. All surviving sharks were documented swimming by the 420 min post-capture video segment while all sharks that died, yet still swam, did so by 400 min post-capture (Fig. 13). Only 19% of sharks that died swam.

The mean time swimming during the first minute of each video segment increased over time for surviving S. cubensis (r2=0.37, p<0.01) from roughly 5 seconds during the 30 min post- capture video segment to over 20 s during the 900 min post-capture segment (Fig. 14). For those that died, mean time swimming reached a peak during the 210 min post-capture segment before activity ceased entirely by 330 min post-capture. All S. cubensis that died did so by 690 min post-capture (Fig. 6, 14). Among those S. cubensis that survived, mean time swimming was higher among sharks in post-release cages shallower than 625 m compared to those in cages greater than 625 m deep (t= 3.72, p<0.05). The rate of increase in behavior (i.e. recovery) was marginally faster for sharks

in shallower cages compared to deeper cages (slopeshallow= 0.02, slopedeep= 0.01; ANCOVA interaction term, p=0.06; Fig. 15). Further, the TOFS was significantly earlier for sharks in shallow cages (<625 m; M=73.13 ± 12.84 SE) compared to deep cages (>625 m; M= 167.50 ±

32.71 SE) (Mann Whitney U Test, nshallow= 16, ndeep= 12, p= <0.05; Fig. 16). At-vessel blood glucose levels were significantly greater for S. cubensis that had a TOFS more than 120 min post-caging (late, M= 5.22 ± 0.19 SE) compared to those that had a TOFS less than 120 min post-caging (early, M=4.66 ± 0.17 SE; t=2.19, p<0.05; Fig. 17). There were no other differences between early and late TOFS when examining its relationship with other at- vessel blood chemistry metrics (e.g. blood lactate, blood pH). However, at-vessel S. cubensis blood lactate levels were significantly lower for active (M= 6.7 ± 1.73 SE) compared to inactive (M=10.42 ± 0.98 SE) sharks observed during the 24 h post-release caging period as calculated by percent time swimming (<20%- inactive, >20%-active; t=2.56, p<0.05; Fig. 18). Total length was significantly greater for S. cubensis that had a time of death after 120 min post-capture (late, M= 61.99 ± 3.39 SE) compared to those that had a time of death before 120 min post-capture (early, M= 50.32 ± 3.56 SE; Mann-Whitney U Test, p<0.05; Fig. 19). No other variables (e.g. cage depth, blood chemistry metrics) could differentiate between groups of sharks with early and late times of death.

2.3.6 Post-Release Behavior: Centrophorus sp. While only one Centrophorus sp. survived the entire 24 h caging period, some individuals did survive for multiple hours after capture and were monitored for post-release behavior. None of these sharks, including the individual that survived, exhibited normal

swimming behaviors. Instead, even when alive, they hovered inverted while respiring primarily through their spiracles and only moved with very brief (<5 s), irregular, slow tail movements. These periods of activity were often the result of scavenging isopods climbing onto an animal and eliciting a twitch or single tail beat. Centrophorus sp. often hovered motionless for hours at a time, appearing neutrally buoyant mid-cage. The mean time of death was 260 min post-capture (± 65.6 SE) for this species.

2.4 Discussion

Although PRM rates have been estimated for multiple shark species captured in coastal waters (e.g. Heupel and Simpfendorfer, 2002, Morgan and Burgess, 2007), this is the first study to estimate a PRM rate (49.7%, S. cubensis) for any deep sea shark despite their prevalence as fisheries bycatch (Cotton and Grubbs, 2015). We found that PRM rates for longline-caught deep sea sharks are higher than previously assumed, ranging from 49.7 – 83%. Furthermore, we identified drivers of capture stress, predicted PRM using blood lactate, pH and vitality scores, and established links between at-vessel blood chemistry and post-release behaviors. Lastly, we documented post-release behavior for S. cubensis and Centrophorus sp., and highlighted the effects of cage depth on recovery and activity levels.

2.4.1 At-Vessel and Post-Release Mortality At-vessel mortality rates reported here are nearly identical to those reported by Brooks et al. (2015) at the same study site for S. cubensis (9%), Centrophorus sp. (29%), and Hexanchus griseus (0%), although we documented higher rates of mortality for Hexanchus nakamurai (7% vs. 16% here) and Mustelus canis insularis (0% vs. 7% here). Hale et al. (2010) reported slightly lower at-vessel mortality rates for S. cubensis (2.9%) in the bottom longline fishery targeting sharks in the Southeastern USA whereas Gulak et al. (2013) reported a 9% dead discard rate for this species in the shark and reef fish bottom longline fisheries in that region. We also documented a 100% at-vessel mortality rate, albeit based only on two individuals, for the sharpnose sevengill shark Heptranchias perlo, for which there are few data available on fishery interactions or otherwise in the Western Atlantic.

Post-release mortality rates were previously unknown for S. cubensis although it is regularly caught as bycatch in the northern Gulf of Mexico (National Marine Fisheries Service, unpublished data; Jones et al., 2013). Past research suggested that this species is able to survive longline capture and release based on occasional recaptures of tagged individuals after extended periods (579 days in one instance; Brooks et al., 2015). We also recaptured one male S. cubensis that was tagged by Brooks et al. (2015) in 2010 after roughly four years at liberty. Our data show that PRM takes place quickly, with 100% of both Centrophorus sp. and S. cubensis that die doing so by 690 min post-capture. Past research, however, has shown that mortality can take place up to 14 days post-capture in Atlantic herring (Suuronen et al., 1996) and up to 30 days post-capture in Pacific halibut (Davis and Olla, 2001). Whereas our estimates exclude mortality past 24 hours, in-cage behavioral data show that S. cubensis swimming activity increased within our observation period, implying some degree of recovery. As such, these estimates likely account for the majority of PRM based on immediate capture-associated insults alone (excluding factors such as post-release predation, Raby et al. 2014). Conclusions regarding our estimated 24 h PRM rates for the Antillean smoothhound shark M. canis insularis are limited by small sample size. A deepwater relative of the dusky smoothhound shark Mustelus canis canis, the Antillean smoothhound occurs throughout the Caribbean (Heemstra, 1997) and is captured in trawls (Kiraly et al., 2003). Our preliminary data suggest that PRM rates could be high for this species (75%) when it is captured at the deeper extent of its depth range. Data from our caging experiments also suggest that PRM rates for Centrophorus sp. are high (83%) and agree with previous satellite telemetry data for Centrophorus sp., where, out of eleven tags, eight did not report and three suggested immediate predation in a study at the same location (Brooks et al., 2015). Conversely, Daley et al. (2015) reported extremely high survivorship for longline-caught Centrophorus zeehani in temperate waters in southern Australia. An important note on these discrepancies is that animals in Daley et al. (2015) were captured with the intent to maximize survivorship by fishing during cool winter nights (sea surface temperatures between 15-25 ºC), whereas both this study and Brooks et al. (2015) took place when sea surface temperatures reached 30 ºC during the day. Interestingly, Centrophorus sp. in both our study and those in southern Australia responded to caging by hovering upside down during the post-release monitoring period (R Daley, personal communication). However,

when released without a cage in the cooler surface waters of southern Australia, Centrophorus sp. swam away with little behavioral impairment (R Daley, personal communication), whereas those in warmer Bahamian waters exhibited little movement upon release at the surface. As such, temperature and/or cage effects on this species group could be substantial, although why these sharks hover upside down in the cages is unknown. One potential mechanism could be related to the high oil content of Centrophorus livers (Deprez et al., 1990) which plays a role in buoyancy control (Weatherbee and Nichols, 2000), as liver oil can be sensitive to changes in pressure (Phlegar, 1998; Pethybridge et al., 2010; Daley et al., 2015). This remains to be explored. Our findings also suggest that post-release predation could be high for deep sea sharks, particularly in regions where large predators are abundant (such as near a shelf edge, Brooks et al., 2015), although including it into PRM estimates is logistically challenging (Raby et al., 2014). Loss of orientation and a lack of strong swimming behavior categorized Centrophorus sp. and many S. cubensis and H. nakamurai released at the surface during this study. As such, these relatively small deep sea sharks could be highly vulnerable to predation immediately following release, particularly when considering the propensity for pelagic sharks to circle the longline and cage during retrieval, a common occurrence in commercial fisheries (Stevens et al., 2000; Raby et al., 2014). Post-release predation was also observed for a single S. cubensis released at the surface, which was eaten by a silky shark (Carcharhinus falciformis) as it approached 130 m, as determined by a trailing TDR. We also documented a single successful predation event and multiple unsuccessful attempts made by various isopod species (Bathynomus giganteus, Booralana spp.) on both live and dead sharks in the post-release cages. Predators were also observed during video monitoring at the seafloor, including H. griseus and unidentified deep sea groupers which attempted to access the cage occupants. Ultimately, post-release predation during an individual’s descent or recovery at the seafloor could inflate the PRM rates reported here.

2.4.2 Stress and Behavior Recovery depth has implications for post-release behavior and the possibility of predation as well. Our data show that swimming activity increased throughout the first 900 minutes post- capture for S. cubensis, and that the rate of increase was slower at deeper depths, presumably due to lower metabolic rates associated with colder temperatures (up to a 6 ºC difference between shallow and deep cages). Similarly, time of first swimming post-capture was significantly earlier

for sharks in cages <625 m deep compared to those in cages >625 m deep. With a depressed metabolic rate, the return to physiological homeostasis following capture-induced perturbations in acid-base balance, among other things, could be delayed at deeper depths, particularly as metabolic demands are increased and energy is reallocated towards the systems most necessary for survival (e.g. cardiac functioning; Barton and Iwama, 1991; Skomal and Mandelman, 2012). Temperature (and depth) could play a similar role in determining a shark’s behavior and condition while hooked. While on-hook behavioral metrics are unavailable from this dataset, at- vessel blood lactate levels were lower and blood pH levels higher in S. cubensis the deeper and colder the capture site. Blood lactate concentrations and pH levels are often associated with exhaustive exercise, during which lactic acid disassociates into lactate anions − + (CH3CH(OH)CO2 ) and hydrogen protons (H ), contributing to a drop in blood pH known as metabolic acidosis (Hoffmayer and Parsons, 2001, Mandelman and Farrington, 2007, Mandelman and Skomal, 2009, reviewed by Skomal and Mandelman, 2012). While a buildup of pCO2 can also be the primary driver of blood acidosis as a result of limited ventilation (Skomal and Mandelman, 2012), S. cubensis is able to buccal pump as well as respire through its spiracles, suggesting that the metabolic pathway is more likely. Although blood chemistry metrics for unstressed individuals are unavailable, higher blood lactate concentrations and lower blood pH generally imply greater physiological disturbance (Skomal and Mandelman, 2012). These results suggest that the behavioral response to capture may have been subdued at greater capture depths (i.e. they struggled less) leading to reduced physiological impairment. Indeed, warmer water temperatures can increase metabolic rates, leading to a lower anaerobic threshold and greater magnitude of capture stress in sharks (Skomal and Bernal, 2010; Hoffmayer et al., 2012; Gallagher et al., 2014a, Danylchuk et al., 2014, Hyatt et al., 2016). Past research has shown that just a 6 ºC increase in water temperature (similar to the difference between the deepest and shallowest depths here) can result in a three-fold increase in the metabolic rate of a bat ray Myliobatis californica (Hopkins and Cech, 1994). As such, it is possible that sharks at shallower depths had higher blood lactate concentrations simply due to an increase in metabolic rate (Guida et al., 2016). Elevated capture temperatures have also been linked to higher at-vessel mortality in blue, dusky, night, and silky sharks (Gallagher et al., 2014a) as well as PRM in numerous species in a shark fishery off of Australia (Braccini, Van Rijn, and Frick, 2012). Gallagher et al. (2014a) also found that survival increased at deeper hooking depths for multiple

shark species, and suggested that higher oxygen levels in cooler waters can limit oxygen deficits in captured sharks. This is unlikely for the range of depths and temperatures discussed here, although dissolved oxygen levels at 500-1000 m deep in Exuma Sound are unknown. Blood lactate levels in S. cubensis also increased with maximum capture duration, a rough metric for how long an individual could have been hooked prior to blood sampling. Long fight times and lactate accumulation as a result of anaerobic glycolysis (Skomal and Bernal, 2010) can significantly affect the magnitude of physiological disturbance in sharks (Morgan and Burgess, 2007; Frick et al., 2010) and can contribute to high levels of stress and/or mortality (Danylchuk et al., 2014, Gallagher et al., 2014b). This was also reflected in the post-release swimming activity of S. cubensis, which was reduced in individuals with elevated at-vessel blood lactate levels. Further, at-vessel blood glucose levels were significantly higher for sharks that had a late (>120 min) time of first swimming in the post-release cages. Elevated blood glucose (hyperglycemia) is a common response to capture stress as hormones mobilize hepatic glycogen, which is then transported through the blood to fuel active muscle tissue (McDonald & Milligan, 1992; Hoffmayer and Parsons, 2001; Mandelman and Farrington, 2007; Bernal and Skomal, 2010). Similar results have been reported for finfish, where the intensity and duration of a capture event can affect an individual’s ability to recover post-capture and evade predators (Olla et al., 1992; Olla et al., 1995, Kieffer et al., 1995; Ryer et al., 2004; Wilson et al., 2014). Capture stress can also lead to other sub-lethal effects including impaired immune function (Lupes et al., 2006) and the disruption of normal feeding and mating behaviors, which can negatively affect fitness in discards (Wilson et al., 2014). Lastly, physiological disturbance was implicated in the assignment of vitality scores (i.e. release condition scores). Individuals assigned a score of ‘poor’ had lower blood pH than those assigned a score of fair or excellent, which was apparent in their whole-animal condition as measured by lethargy on the boat. Hyatt et al. (2016) reported similar results for bonnethead Sphyrna tiburo and bull sharks Carcharhinus leucas, where behavioral release condition scores

partially reflected blood pH, lactate, and pCO2 levels.

2.4.3 Predicting Post-Release Mortality Post-release mortality estimates are typically limited in their accuracy when applied across fisheries as many factors vary greatly from one fishery to another (e.g. capture

characteristics, fishing and handling procedures, environmental conditions; Davis, 2002). As such, being able to predict fishery-specific PRM rates is of vital importance. While Renshaw et al. (2012) highlight the limitations of blood chemistry in forecasting long-term discard mortality in elasmobranchs, other studies have successfully used both blood chemistry (Moyes et al. 2006, Skomal, 2006, Skomal, 2007, Heberer et al., 2010, Gallagher et al., 2014b, Hutchinson et al., 2015) and reflex impairment indices (reviewed by Davis, 2010) to predict PRM in sharks and teleost fishes. Certainly the incorporation of physical injury is recommended in the future (Ranshaw et al., 2012); however in our study injury was very uncommon and thus ignored. Instead, we focused on blood chemistry and a rough metric of condition (vitality scores), to provide fisheries management with multiple methods to predict S. cubensis PRM. As blood chemistry can reflect the magnitude of stress experienced by an individual regardless of the type of stressor, these predictive models should prove useful for other capture scenarios than our own. For ease of application, we simplified models to include only one or two predictors of mortality. Only at-vessel blood pH was included in one model while total length and at-vessel blood lactate concentrations were included in the other. The first suggested a greater than 50% probability of mortality at a blood pH of 7.17 for S. cubensis, which is below most published pH levels for longline caught elasmobranchs (Gallagher et al., 2014b; Brooks et al., 2011, Brooks et al., 2012). Some sharks with blood pH values below this threshold survived the fishing event, however, including one individual with a blood pH of 7.01, while some with a blood pH over this threshold died, including one with a blood pH of 7.36. It is also important to note that most pH values reported in the literature are often the product of temperature correction formulas applied to pH values taken from an iStat blood analyzer, whereas the values and model presented here are based on pH measurements from a pH meter that automatically corrects to blood temperature. The relationship between these values and raw iStat values is predictable (see Fig. 4), but conversions are necessary to accurately compare across studies or apply this model elsewhere. The second model included total length and blood lactate concentrations. This model showed that the probability of 24 h PRM increased with higher lactate levels and smaller total length. Linking lactate to stress and mortality has been documented elsewhere (Moyes et al., 2006, Hight et al., 2007, Gallagher et al., 2014b, Hutchinson et al., 2015), and our values are comparable to those reported in past studies. Our predictive model identified a greater than 50%

probability of mortality at blood lactate concentrations of greater than 10.7 mmol/L for an individual of average total length (58 cm). Lactate concentrations for moribund sharks have been reported as 15 mmol/L for blue sharks Prionace glauca, 19 mmol/L for thresher sharks Alopias vulpinus, and 20 mmol/L for shortfin mako Isurus oxyrinchus sharks (Hight et al., 2007), while unstressed or ‘control’ values were less than 5 mmol/L in sandbar sharks Carcharhinus plumbeus, dusky sharks Carcharhinus obscurus, Caribbean reef sharks Carcharhinus perezi, and spiny dogfish Squalus acanthias (Spargo, 2001, Cliff and Thurman, 1984, Mandelman and Farrington, 2007, Brooks et al., 2012). For comparison, the highest lactate level in S. cubensis that we recorded was 17.5 mmol/L for a shark that did not survive the 24 h caging period. Total length was also a significant predictor of mortality and smaller sharks died earlier than larger sharks post-capture (as in Davis and Parker, 2004). Larger S. cubensis also had a higher blood pH across capture depths than their smaller counterparts. Clearly large size imparts some advantage by limiting the behavioral response to capture and/or allowing sharks to cope with physiological insults, ultimately improving their chances of survival. Other studies have indeed shown that the probability of mortality is higher for small size classes of haddock, whiting, vendace, and Atlantic herring after trawl capture, attributed to greater fatigue and injury in small individuals (Suuronen et al., 1995; Suuronen et al., 1996; Sangster et al., 1996; Davis, 2002). For longline discards, size can have a similar effect (Neilson et al., 1989; Milliken et al., 1999; Davis, 2002), although Morgan and Carlson (2010) documented the opposite relationship for C. plumbeus, suggesting that fight intensity could be greater for larger animals. Large size could also act as a thermal buffer to the drastic temperature changes experienced by a deep sea shark during capture (over 15 ºC here). The core temperature of smaller sharks may warm faster than large sharks, leading to greater thermal stress as seen in sablefish and Pacific halibut, among others (Spigarelli et al., 1977; Davis et al., 2001; Davis and Olla, 2001; Davis and Olla, 2002; Davis, 2002). Large sharks such as H. griseus do tend to have lower at-vessel muscle temperatures than smaller sharks like Squalus spp. when captured from similar depths (RD Grubbs, unpublished data). While this study could not establish a link between thermal stress and mortality, previous studies have for marine fishes (Muoneke and Childress, 1994, Davis and Olla, 2001). Vitality scores also predicted mortality for S. cubensis. While scores of release condition are very rough and can be subjective (Benoît et al., 2010), they have been useful in predicting

PRM for sharks (Manire et al., 2001, Heuter et al., 2006) and are shown here to relate, at some level, with blood chemistry, although there was overlap between pH levels across ‘excellent’ and ‘fair’ groupings. These results are similar to those reported in Hyatt et al. (2016) for Carcharhinus leucas and Sphyrna tiburo, where a similar scoring system could differentiate some levels of condition but not all. Reflex impairment indices may be a better choice to provide a robust assessment of PRM in the future (as in Braccini, Van Rijn, and Frick, 2012, Danylchuk et al., 2014; Gallagher et al., 2014b). Still, using metrics like these are far less expensive and far easier to implement than blood chemistry analysis and could prove useful to fisheries management despite their limitations in providing fine scale PRM estimates.

2.4.4 Limitations Post-release cages are commonly used to estimate PRM in a field setting (St. John and Syers, 2005, Mandelman and Farrington, 2007, Stewart, 2008, Mandelman et al., 2013, Weltersbach and Srehlow, 2013, Campbell et al., 2014), although some studies have pointed out the flaws in this experimental design (Muoneke and Childress, 1994, Pollock and Pine, 2007). Concerns are mainly focused on the semi-artificial holding conditions that don’t mimic true post- release conditions (e.g. exclusion of predators, altered descent rate, and unnatural behavior; Weltersbach and Strehlow, 2013). Cages can also impart additive stress and/or cause physical injury, although in our study we saw no evidence of either. We also found no effect of cage density on mortality or behavior in exploratory GLMs, and the circular cage design allowed for sharks to swim continuously. Including control groups to separate out the effects of caging from capture and release could alleviate some of these concerns (Pollock and Pine, 2007; Weltersbach and Strehlow, 2013), but true controls would have been logistically impossible for this study given the constraints of working in deep water. The PRM rates reported here could also be underestimated due to differences in handling practices between our study and those in a commercial setting. As S. cubensis and Centrophorus sp. have dorsal fin spines, fishermen often discard them immediately to avoid personal injury, but in doing so they sometimes break an animal’s jaw during the de-hooking process (S Gulak, personal communication). Although we tried to mimic commercial sorting practices by limiting air exposure and releasing animals quickly, we intentionally minimized animal injury by removing hooks by hand.

2.4.5 Conclusions Considering the current trends for fishing deeper (Morato et al., 2006), the magnitude of commercial discards worldwide (estimated at 25% of total catches; Pascoe, 1997, Davis, 2002, Kelleher, 2005), and the disproportionate contribution of elasmobranchs to these figures (Cahmi 2009, Molina and Cooke, 2012), data deficient and highly vulnerable deep sea sharks are likely at greater risk for bycatch-induced population declines than ever before, particularly when considering the high PRM rates reported here. Based on our findings, we hypothesize that retaining, utilizing, and accounting for bycatch-caught deep-sea elasmobranchs may improve the future management of deep-sea fisheries.

Table 5. Description of vitality scores assigned to sharks placed in the post-release cage before being lowered to the sea floor.

Vitality Score Description

Vigorous body movement, no apparent injuries, strong Excellent swimming behavior

Inconsistent body movement, responds to stimulation, Fair possible minor injuries, moderate swimming behavior Weak body movement, little response to stimulation, Poor possible minor injuries, weak or absent swimming behavior

Table 6. Correlation structure of primary stress physiology metrics, length measurements, and capture characteristics for S. cubensis.

Total Max. Capture Lactate Glucose Parameters length capture pH Hem. (%) depth (m) (mmol/L) (mmol/L) (cm) duration 1 Total length (cm) Capture depth (m) -0.1629 1

Max. capture 0.1128 0.4374 1 duration -0.0312 -0.3183 0.1311 1 Lactate (mmol/L) 0.3855 0.3007 0.1028 -0.5795 1 pH 0.1232 0.3619 0.5263 0.2275 0.1384 1 Glucose (mmol/L) Hematocrit (%) -0.2704 0.4457 0.3706 -0.217 0.0905 0.1289 1

Table 7. Capture composition and characteristics of sharks caught on deep-sea longlines throughout this study.

Mean Total Mean Capture Mean Capture Species N M F Length ± SD Temperature ± SD Depth ± SD (m) (cm) (°C) Squalus cubensis 72 25 47 58.11 ± 11.64 622.26 ± 71.86 12.03 ± 1.75 Centrophorus sp. 13 5 8 88.7 ± 11.5 769.5 ± 42.8 8.8 ± 0.79 Mustelus canis insularis 14 7 7 88.62 ± 11.06 539.60 ± 64.95 14.06 ± 2.03 Hexanchus nakamurai 6 3 3 129.3 ± 12.0 582.1 ± 43.9 13.1 ± 1.2 Heptranchias perlo 2 2 0 75 ± 17.0 631.5 ± 34.7 12 ± 0.36 Hexanchus griseus 1 1 0 227 772 10

Table 8. At-vessel and 24 h post-release mortality rates for species caught in this study, calculated using Eqs. (1) and (2).

At-vessel 24h post-release Species N Captured N Caged mortality rate % mortality rate %

Squalus cubensis 71 8.33 54 49.7 ± 8.5 SE Centrophorus sp. 13 30.77 8 83 ± 16 SE Mustelus canis insularis 14 7.14 4 75 ± 25 SE Hexanchus nakamurai 6 16.66 0 N/A Heptranchias perlo 2 100 0 N/A Hexanchus griseus 1 0 0 N/A

Table 9. Blood chemistry parameters and corresponding sample sizes for Squalus cubensis and Centrophorus sp. captured during this study.

Blood Chemistry Squalus cubensis Centrophorus sp. Parameter N Mean ± SE N Mean ± SE pH 51 7.20 ± 0.17 8 7.13 ± 0.06 Lactate (mmol/L) 41 9.80 ± 0.59 4 9.29 ± 1.90 Hematocrit (%) 53 25.4 ± 0.52 9 25 ± 1.64 Glucose (mmol/L) 51 4.72 ± 0.12 9 3.39 ± 4.69

Table 10. Full model and minimal adequate model from backwards stepwise analysis to describe at-vessel blood pH in Squalus cubensis based on four initial model parameters and one interaction term.

Parameter Estimate S.E. t Ratio Pr(>|t|) Full Model Intercept 6.450 0.435 14.820 ** Maximum hooking duration -0.001 0.001 -1.200 0.233 Sea surface temperature -0.005 0.012 -0.380 0.708 Capture depth : Total length 0.000 0.000 -0.130 0.900 Total length 0.007 0.001 3.900 ** Capture depth 0.001 0.003 3.140 ** Most parsimonious model Intercept 6.273 0.203 30.840 ** Capture depth 0.001 0.000 3.390 ** Total length 0.007 0.002 3.970 ** *p≤0.05, **p≤0.01

Table 11. Akaike Information Criterion (AIC) values for each model used to describe the at- vessel blood pH of Squalus cubensis resulting from stepwise backwards elimination of nonsignificant terms and evaluation of changes in AIC and deviance.

Model AIC Full Model -54.08 - Capture depth : Total length -56.07 - Sea surface temperature -57.93 - Maximum hooking duration -58.43 Most parsimonious model -58.43 - Capture depth -51.18 - Total length -45.77

Table 12: Full model and minimal adequate model from backwards stepwise analysis to describe at-vessel blood lactate in Squalus cubensis based on four initial model parameters and one interaction term.

Parameter Estimate S.E. t Ratio Pr(>|t|) Full Model Intercept -3.550 31.070 -0.114 0.91 Capture depth -0.009 0.052 -0.179 0.86 Total length 0.109 0.521 0.208 0.84 Capture depth : Total length 0.000 0.001 -0.287 0.776 Sea surface temperature 0.434 0.354 1.225 0.229 Maximum hooking duration 0.041 0.019 2.144 * Most parsimonious model Intercept 18.693 5.839 3.202 ** Maximum hooking duration 0.031 0.017 1.866 0.07 Capture depth -0.027 0.010 -2.681 ** *p≤0.05, **p≤0.01

Table 13: Akaike Information Criterion (AIC) values for each model used to describe the at- vessel blood lactate of Squalus cubensis resulting from stepwise backwards elimination of nonsignificant terms and evaluation of changes in AIC and deviance.

Model AIC Full Model 103.2 - Capture depth : Total length 101.3 - Total length 100.97 - Sea surface temperature 99.56 Most parsimonious model 99.56 - Maximum hooking duration 101.16 - Capture depth 104.67

Table 14. Full model and minimal adequate model from backwards stepwise GLM analysis to describe 24 h mortality in Squalus cubensis based on five initial model parameters and five interaction terms.

Parameter Estimate S.E. z-value Pr(>|z|) Full Model Intercept 2.92 4.34 0.67 0.50 Lactate : Total length 0.20 3.78 0.05 0.96 pH 0.16 1.83 0.09 0.93 Lactate : Glucose -0.57 3.67 -0.15 0.88 Hematocrit -0.27 1.75 -0.16 0.88 pH : lactate 1.55 2.58 0.60 0.55 pH : Total length 1.73 2.16 0.80 0.42 Glucose : Total length -2.38 2.60 -0.91 0.36 Total length 3.89 3.98 0.98 0.33 Glucose 1.75 1.76 1.00 0.32 Lactate -4.75 3.65 -1.30 0.19 Most parsimonious model Intercept 1.69 1.17 1.45 0.15 Hematocrit 0.62 0.86 0.72 0.47 Glucose 1.41 1.00 1.41 0.16 Glucose : Total length -1.94 1.18 -1.64 0.10 Lactate -4.13 2.08 -1.98 * Total length 3.34 1.62 2.06 * Estimates and standard errors (S.E.) are on a logit scale. *p≤0.05

Table 15. Akaike Information Criterion (AIC) values for each GLM used to describe the 24 h mortality of Squalus cubensis resulting from stepwise backwards elimination of nonsignificant terms and evaluation of changes in AIC and deviance.

Model AIC Full Model 34.035 - Lactate : Total length 32.080 - pH 30.045 - Lactate : Glucose 28.104 - pH : Total length 26.875 - Lactate : pH 26.795 Most parsimonious model 26.795 - Glucose 27.800 - Glucose : Total length 29.000 - Hematocrit 30.900 - Total length 37.100 - Lactate 63.100

Table 16. Significance levels for single variable fits to 24 h S. cubensis mortality data from GLMs with binomial distributions and a logit link function.

Residual Possible Predictors of Mortality Pr(>|z|) Null Deviance Deviance Glucose 0.26 45.5 44.3 Hematocrit 0.37 45.5 44.6 Total length 0.03 45.5 40.2 Lactate 0.01 45.5 37 pH <0.01 45.5 30.03

7.15 7.1 7.05

7 6.95

iStat pH 6.9 6.85 6.8 6.75 6.8 6.9 7 7.1 7.2 7.3 7.4 7.5 7.6 pH meter

Figure 4. Relationship between pH values taken from a pH meter and the iStat point of care device, used to convert values from the iStat into their pH meter equivalents (R2=0.7039, p<0.05).

Figure 5. Relationship between sea floor temperature (ºC) and depth (m) at our study site in Northeastern Exuma Sound, The Bahamas from July 2014 to July 2015.

Figure 6. Time of death of Squalus cubensis and Centrophorus sp. that died within the 24 h video monitoring period in post-release cages at depth.

Figure 7. Squalus cubensis at-vessel blood pH as a function of capture depth (m) and total length (cm). Observed blood pH levels are shown with black dots, while the cross-hatched surface represents the values generated from the slope and intercept outputs from the most parsimonious linear model examined.

Figure 8. Squalus cubensis at-vessel blood lactate as a function of capture depth (m) and maximum capture duration (min). Observed blood lactate levels are shown with black dots, while the cross-hatched surface represents the values generated from the slope and intercept outputs from the most parsimonious linear model examined.

Figure 9. Squalus cubensis survival probability as a result of total length (cm) and blood lactate (mmol/L). Predicted values were calculated according to the slope and intercept outputs from a GLM fit to mortality with significant predictors only. Lactate and total length inputs were chosen based on the minimum and maximum values reported in this study. 42

Figure 10. Survival probability curve using at-vessel blood pH measurements for S. cubensis that either survived ( ) or died ( ) after 24 h post-capture. The solid line represents the probability of survival calculated using Eq. 4 and the dashed line represents the 50% chance of survival, calculated at a blood pH of 7.17.

Figure 11. Post-release mortality of S. cubensis placed in 24 h post-release cages by vitality score (n caged). The distribution of sharks that went on to survive or die after 24 h post-release was significantly different between groups of sharks assigned excellent, fair, and poor vitality scores (Pearson’s x2=11.78, p=0.001, df=2).

Figure 12. At-vessel S. cubensis blood pH by vitality score. Means that do not share a letter above each box are significantly different as determined by an ANOVA and Tukey’s test (p<0.05).

Figure 13. Percent of all S. cubensis exhibiting swimming behavior during the 24 h video monitoring period in post-release cages at depth.

Figure 14. Squalus cubensis mean time swimming (calculated for the first minute of each video deployment only) for those that survived or died during the first 900 minutes following cage deployment. Mean time swimming of survivors increased significantly with duration in the cage (r2=.37, p<0.01).

Figure 15. Squalus cubensis mean time swimming (calculated for the first minute of each video deployment only) for those that survived in either shallow (<625 m) or deep (>625 m) cages during the first 900 minutes following deployment. Mean time swimming increased significantly over time for both shallow and deep groups of survivors (shallow- r2=0.30, p<0.05; deep- r2=0.12, p=0.05), and the rate of increase was marginally higher for the sharks in shallow cages (ANCOVA interaction term, p=0.06).

Figure 16. Squalus cubensis time of first swimming was significantly earlier for sharks in shallow post-release cages compared to those in deep cages (Mann Whitney U Test, p= <0.05).

Figure 17. At-vessel Squalus cubensis blood glucose levels were significantly greater for sharks that had a time of first swimming (TOFS) after 120 min post-capture (‘late’) compared to those that had a TOFS earlier than 120 min post-capture (‘early’; t=2.19, p<0.05).

Figure 18. At-vessel Squalus cubensis blood lactate levels were significantly lower for active sharks compared to inactive sharks observed during the 24 h post-release caging period as calculated by percent time swimming (<20%- inactive, >20%-active; t=2.56, p<0.05).

Figure 19. Squalus cubensis total length was significantly lower for sharks that had a time of death before 120 min post-caging compared to those that had a time of death after 120 min post- caging (Mann-Whitney U Test, p<0.05).

APPENDIX A

ACUC APPROVAL

ANIMAL CARE AND USE COMMITTEE [ACUC] 101 BIOMEDICAL RESEARCH FACILITY TALLAHASSEE, FL 32306-4341 TELEPHONE: 644-4262 FAX: 644-5570 MAIL CODE: 4341

March 3, 2016

The Graduate School Florida State University

To Whom It May Concern:

Concerning the thesis/dissertation submitted to the Graduate School by:

Graduate Student: Brendan Talwar Thesis/Dissertation Title: Post-Release Mortality of Deep Sea Bycatch Species Department: Biological Science Major Professor: Dr. R. Dean Grubbs

The above named graduate student has provided assurance to the FSU Animal Care and Use Committee that all animal procedures utilized in the work resulting in this thesis/dissertation are described in FSU ACUC Protocol(s):

Protocol Date ACUC Title Number Approval Community structure, life histories, trophic structure and stress physiology of deepwater fishes in the Gulf 1412 February 24, 2014 of Mexico and assessment of their exposure to petroleum-based pollutants

The Animal Care and Use Committee has confirmed that this student was included as a project member during the period covering their thesis/dissertation work. This institution has an Animal Welfare Assurance on file with the Office for Laboratory Animal Welfare. The Assurance Number is A3854-01.

Sincerely,

Attending Veterinarian FSU Animal Care and Use Committee

KMH/kjj cc: Brendan Talwar Dr. R. Dean Grubbs

REFERENCES

Awruch, C. A., Simpfendorfer, C., & Pankhurst, N. W. 2011. Evaluation and use of a portable field kit for measuring whole-blood lactate in sharks. Marine and Freshwater Research 62(6): 694-699.

Barradas-Ortiz, C., Briones-Fourzán, P., and Lozano-Álvarez, E. 2003. Seasonal reproduction and feeding ecology of giant isopods Bathynomus giganteus from the continental slope of the Yucatan peninsula. Deep Sea Research Part I: Oceanographic Research Papers, 50: 495-513.

Barton, B. A., & Iwama, G. K. 1991. Physiological changes in fish from stress in aquaculture with emphasis on the response and effects of corticosteroids. Annual Review of Fish Diseases 1: 3-26.

Basti, D., Bricknell, I., Hoyt, K., Chang, E. S., Halteman, W., and Bouchard, D. 2010. Factors affecting post-capture survivability of lobster Homarus americanus. Diseases of Aquatic Organisms, 90: 153.

Benoît, H. P., Hurlbut, T., & Chassé, J. 2010. Assessing the factors influencing discard mortality of demersal fishes using a semi-quantitative indicator of survival potential. Fisheries Research 106(3): 436-447.

Braccini, M., Van Rijn, J., & Frick, L. 2012. High post-capture survival for sharks, rays and chimaeras discarded in the main shark fishery of Australia.PloS one, 7(2), e32547.

Briones-Fourzán, P., and Lozano-Alvarez, E. 1991. Aspects of the biology of the giant isopod Bathynomus giganteus A. Milne-Edwards, 1879 (Flabellifera: Cirolanidae) off the Yucatán Peninsula. Journal of Crustacean Biology, 11: 375–385.

Brooks, Edward J., Annabelle ML Brooks, Sean Williams, Lance KB Jordan, Debra Abercrombie, Demian D. Chapman, Lucy A. Howey-Jordan, and R. Dean Grubbs. 2015. First description of deep-water elasmobranch assemblages in the Exuma Sound, The Bahamas. Deep Sea Research Part II: Topical Studies in Oceanography 115: 81-91.

Brooks, Edward J., Katherine A. Sloman, Stephanie Liss, Laila Hassan-Hassanein, Andy J. Danylchuk, Steven J. Cooke, John W. Mandelman, Gregory B. Skomal, David W. Sims, and Cory D. Suski. 2011. The stress physiology of extended duration tonic immobility in the juvenile lemon shark, Negaprion brevirostris (Poey 1868). Journal of experimental marine biology and ecology 409(1): 351-360.

Cahmi, M.D., 2009. Conservation status of pelagic elasmobranchs. In: Cahmi, M.D., Pikitch, E.K., Babcock, E.A. (Eds.), Sharks of The Open Ocean: Biology, Fisheries and Conservation. Blackwell Publishing Ltd., Oxford, UK.

Campana, S. E., Joyce, W., & Manning, M. J. 2009. Bycatch and discard mortality in commercially caught blue sharks Prionace glauca assessed using archival satellite pop-up tags. Marine Ecology Progress Series 387: 241-253.

Campbell, M. J., McLennan, M. F., & Sumpton, W. D. 2014. Short-term survival of discarded pearl perch (Glaucosoma scapulare Ramsay, 1881) caught by hook-and-line in Queensland, Australia. Fisheries Research 151: 206-212.

Chamberlain, S. C., Meyer‐Rochow, V. B., and Dossert, W. P. 1986. Morphology of the compound eye of the giant deep‐sea isopod Bathynomus giganteus. Journal of Morphology, 189: 145-156.

Childress, J. J., and Seibel, B. A. 1998. Life at stable low oxygen levels: adaptations of animals to oceanic oxygen minimum layers. Journal of Experimental Biology, 201: 1223-1232.

Cliff, G., & Thurman, G. D. 1984. Pathological and physiological effects of stress during capture and transport in the juvenile dusky shark, Carcharhinus obscurus. Comparative Biochemistry and Physiology Part A: Physiology 78(1): 167-173.

Coelho, R., Fernandez-Carvalho, J., Lino, P. G., & Santos, M. N. 2012. An overview of the hooking mortality of elasmobranchs caught in a swordfish pelagic longline fishery in the Atlantic Ocean. Aquatic Living Resources 25(4): 311-319.

Coggins L. G., Catalano M. J., Allen M. S., Pine W. E., Walters C. J. 2007. Effects of cryptic mortality and the hidden costs of using length limits in fishery management. Fish Fish 8:196–210.

Cooke, S., Suski, C., Danylchuk, S., Danylchuk, A., Donaldson, M., Pullen, C., Bulté, G., O'Toole, A., Murchie, K., Koppelman, J., Shultz, A., Brooks, E., Goldberg, T., 2008. Effects of different capture techniques on the physiological condition of bonefish (Albula vulpes) evaluated using field diagnostic tools. Journal of Fish Biology 73: 1351–1375.

Cotton, C. F., & Grubbs, R. D. 2015. Biology of deep-water chondrichthyans: Introduction. Deep Sea Research Part II: Topical Studies in Oceanography 115: 1-10.

Crawley, M. J. 2007. The R Book. John Wiley & Sons: Chichester.

Crowder, L. B., and Murawski, S. A. 1998. Fisheries bycatch: implications for Management. Fisheries 23: 8-17.

Danylchuk, A. J., Suski, C. D., Mandelman, J. W., Murchie, K. J., Haak, C. R., Brooks, A. M., & Cooke, S. J. 2014. Hooking injury, physiological status and short-term mortality of juvenile lemon sharks (Negaprion bevirostris) following catch-and-release recreational angling. Conservation Physiology 2(1): cot036.

Dapp, D. R., Walker, T. I., Huveneers, C., & Reina, R. D. 2015. Respiratory mode and gear type are important determinants of elasmobranch immediate and post‐release mortality. Fish and Fisheries.

Davies, R. W. D., Cripps, S. J., Nickson, A., and Porter, G. 2009. Defining and estimating global marine fisheries bycatch. Marine Policy 33: 661-672.

Davis, M. W. 2002. Key principles for understanding fish bycatch discard mortality. Canadian Journal of Fisheries and Aquatic Sciences 59(11): 1834-1843.

Davis, M. W. 2010. Fish stress and mortality can be predicted using reflex impairment. Fish and Fisheries 11(1): 1-11.

Davis, M. W. 2007. Simulated fishing experiments for predicting delayed mortality rates using reflex impairment in restrained fish. ICES Journal of Marine Science 64: 1535-1542.

Davis, M. W., & Parker, S. J. 2004. Fish size and exposure to air: potential effects on behavioral impairment and mortality rates in discarded sablefish. North American Journal of Fisheries Management 24(2): 518-524.

Davis, M. W., and Ottmar, M. L. 2006. Wounding and reflex impairment may be predictors for mortality in discarded or escaped fish. Fisheries Research, 82: 1-6.

Davis, M.W., and Olla, B.L. 2001. Stress and delayed mortality induced in Pacific halibut Hippoglossus stenolepis by exposure to hooking, net towing, elevated sea water temperature and air: implications for management of bycatch. N. Am. J. Fish. Manag. 21: 725–732.

Davis, M.W., Olla, B.L., and Schreck, C.B. 2001. Stress induced by hooking, net towing, elevated sea water temperature and air in sablefish: lack of concordance between mortality and physiological measures of stress. Journal of Fish Biology 58: 1–15.

Deprez, P. P., Volkman, J. K., & Davenport, S. R. 1990. Squalene content and neutral lipis composition of Livers from Deep-sea sharks caught in Tasmanian waters. Marine and Freshwater Research 41(3): 375-387.

Devine, J. A., Baker, K. D., & Haedrich, R. L. 2006. Fisheries: deep-sea fishes qualify as endangered. Nature 439(7072): 29-29.

Enever, R., Catchpole, T. L., Ellis, J. R., & Grant, A. 2009. The survival of skates (Rajidae) caught by demersal trawlers fishing in UK waters. Fisheries Research 97(1): 72-76.

Firth, D. 1993. Bias reduction of maximum likelihood estimates. Biometrika 80: 27-38.

Frick, L. H., Reina, R. D., & Walker, T. I. 2010. Stress related physiological changes and post- release survival of Port Jackson sharks (Heterodontus portusjacksoni) and gummy sharks (Mustelus antarcticus) following gill-net and longline capture in captivity. Journal of Experimental Marine Biology and Ecology 385(1): 29-37.

Gallagher, A. J., Frick, L. H., Bushnell, P. G., Brill, R. W., & Mandelman, J. W. 2010. Blood gas, oxygen saturation, pH, and lactate values in elasmobranch blood measured with a commercially available portable clinical analyzer and standard laboratory instruments. Journal of Aquatic Animal Health 22(4): 229-234.

Gallagher, A. J., Orbesen, E. S., Hammerschlag, N., & Serafy, J. E. 2014a. Vulnerability of oceanic sharks as pelagic longline bycatch. Global Ecology and Conservation 1: 50-59.

Gallagher, A. J., Serafy, J. E., Cooke, S. J., & Hammerschlag, N. 2014b. Physiological stress response, reflex impairment, and survival of five sympatric shark species following experimental capture and release. Marine Ecology Progress Series 496: 207-218.

Gamito, R., and Cabral, H. 2003. Mortality of brown-shrimp discards from the beam trawl fishery in the Tagus estuary, Portugal. Fisheries Research 63: 423-427.

García, V.B., Lucifora, L.O. and R.A. Myers. 2008. The importance of habitat and life history to extinction risk in sharks, skates, rays and chimaeras. Proceedings of the Royal Society, B 275: 83-89.

Gaten, E. 1988. Light‐induced damage to the dioptric apparatus of Nephrops norvegicus (L.) and the quantitative assessment of the damage. Marine & Freshwater Behaviour & Physiology 13: 169-183.

Giomi, F., Raicevich, S., Giovanardi, O., Pranovi, F., Di Muro, P., and Beltramini, M. 2008. Catch me in winter! Seasonal variation in air temperature severely enhances physiological stress and mortality of species subjected to sorting operations and discarded during annual fishing activities. In Challenges to Marine Ecosystems. Ed. By J. Davenport, G.M. Burnell, T. Cross, M. Emmerson, R. McAllen, R. Ramsay, E. Rogan pp. 195-202. Springer. Netherlands. 211 pp.

Graham, K.J., Andrew, N.L. and Hodgson, K.E. 2001. Changes in relative abundance of sharks and rays on Australian South East Fishery trawl grounds after twenty years of fishing. Marine and Freshwater Research 52: 549-561.

Guida, L., Walker, T. I., & Reina, R. D. 2016. Temperature Insensitivity and Behavioural Reduction of the Physiological Stress Response to Longline Capture by the Gummy Shark, Mustelus antarcticus. PloS one 11(2): e0148829.

Gulak S.J.B., M.P. Enzenauer, and J.K. Carlson. 2013. Characterization of the shark and reef fish bottom longline fisheries: 2012. NOAA Technical Memorandum NMFS-SEFSC-652, 42 p.

Gulak, S.J.B., M.S. Passerotti, J.K. Carlson. 2012. Catch and Bycatch in U.S. Southeast Gillnet Fisheries, 2011. NOAA Technical Memorandum NMFS-SEFSC-629. 28 p.

Hale, L. F., S.J.B. Gulak, and J.K. Carlson. 2010. Characterization of the shark bottom longline fishery: 2009. NOAA Technical Memorandum NMFS-SEFSC-586, 26 pp.

Hall, M. A., Alverson, D. L., and Metuzals, K. I. 2000. By-catch: problems and solutions. Marine Pollution Bulletin 41: 204-219.

Harper, D. E., Eyo, P. B., and Scott, G. P. 2000. Updated golden crab fishery trends and production model analysis based on trip report logbook and trip interview data collection programs. National Marine Fisheries Service (USA), Southeast Fisheries Center, Contribution Number PRD-99/00-12, 15pp.

Harrington, J. M., Myers, R. A., and Rosenberg, A. A. 2005. Wasted fishery resources: discarded by‐catch in the USA. Fish and Fisheries 6: 350-361.

Harter, T. S., Morrison, P. R., Mandelman, J. W., Rummer, J. L., Farrell, A. P., Brill, R. W., & Brauner, C. J. 2015. Validation of the i-STAT system for the analysis of blood gases and acid–base status in juvenile sandbar shark (Carcharhinus plumbeus). Conservation Physiology 3(1): cov002.

Haupt, P., Brouwer, S. L., Branch, G. M., and Gäde, G. 2006. Effects of exposure to air on the escape behaviour and haemolymph chemistry of the South African Cape lobster, Jasus lalandii. Fisheries Research 81: 210-218.

Heemstra, P. C. 1997. A review of the smooth-hound sharks (genus Mustelus, family Triakidae) of the western Atlantic Ocean, with descriptions of two new species and a new subspecies. Bulletin of Marine Science 60(3): 894-928.

Herring, P. J., Gaten, E., & Shelton, P. M. 1999. Are vent shrimps blinded by science? Nature 398: 116-116.

Heupel, M. R., & Simpfendorfer, C. A. 2002. Estimation of mortality of juvenile blacktip sharks, Carcharhinus limbatus, within a nursery area using telemetry data. Canadian Journal of Fisheries and Aquatic Sciences 59(4): 624-632.

Hoffmayer, E. R., Hendon, J. M., & Parsons, G. R. 2012. Seasonal modulation in the secondary stress response of a carcharhinid shark, Rhizoprionodon terraenovae. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology 162(2): 81-87.

Hoffmayer, E.R., Parsons, G.R., 2001. The physiological response to capture and handling stress in the Atlantic sharpnose shark, Rhizoprionodon terraenovae. Fish Physiol. Biochem. 25, 277–285.

Hopkins, T. E., and Cech, J. J. 1994. Effect of temperature on oxygen-consumption of the bat ray, Mylioobatis californica (Chondrichthyes, Myliobatididae). Copeia (2):529–32.

Hueter, R. E., Manire, C. A., Tyminski, J. P., Hoenig, J. M., and Hepworth, D. A. 2006. Assessing mortality of released or discarded fish using a logistic model of relative survival derived from tagging data. Transactions of the American Fisheries Society 135(2): 500–508.

Hutchinson, M. R., Itano, D. G., Muir, J. A., & Holland, K. N. 2015. Post-release survival of juvenile silky sharks captured in a tropical tuna purse seine fishery. Marine Ecology Progress Series 521: 143-154.

Hyatt, M. W., Anderson, P. A., & O'Donnell, P. M. 2016. Behavioral Release Condition Score of Bull and Bonnethead Sharks as a Coarse Indicator of Stress. Journal of Coastal Research.

Jones, E., Beare, D., Dobby, H., Trinkler, N., Burns, F., Peach, K., & Blasdale, T. 2005. The potential impact of commercial fishing activity on the ecology of deepwater chondrichthyans from the west of Scotland. ICES CM, (16).

Jones, L. M., Driggers III, W. B., Hoffmayer, E. R., Hannan, K. M., & Mathers, A. N. 2013. Reproductive Biology of the Cuban Dogfish in the Northern Gulf of Mexico. Marine and Coastal Fisheries 5(1): 152-158.

Kelleher, K. 2005. Discards in the world’s marine fisheries: an update. FAO Fisheries Technical Paper No. 470, Food and Agriculture Organization of the United Nations, Rome, Italy, 131 pp.

Kieffer, J. D., Kubacki, M. R., Phelan, F. J. S., Philipp, D. P., & Tufts, B. L. 1995. Effects of catch-and-release angling on nesting male smallmouth bass. Transactions of the American Fisheries Society 124(1): 70-76.

Kiraly, S. J., Moore, J. A., & Jasinski, P. H. 2003. Deepwater and other sharks of the US Atlantic ocean exclusive economic zone. Marine Fisheries Review 65(4): 1-20.

Koslow, J.A. 1996. Energetic and life-history patterns of deep-sea benthic, benthopelagic and seamount- associated fish. Journal of Fish Biology 49: 54-74.

Koslow, J.A., Boehlert, G.W., Gordon, J.D.M., Haedrich, R.L., Lorance, P., Parin, N. 2000. Continental slope and deep-sea fisheries: implications for a fragile ecosystem. ICES Journal of Marine Science 57: 548-557.

Kyne P. M., Simpfendorfer C. A. A Collation and Summarization of Available Data on Deepwater Chondrichthyans: biodiversity, Life History and Fisheries. Bellevue, WA: Marine Conservation Biology Institute; 2007. 137 p.

Kyne, P. M., Simpfendorfer,C. A. 2010. Deepwater Chondrichthyans. In: Carrier, J.C., Musick, J.A., Heithaus, M.R.(Eds.), Sharks and Their Relatives, II: Biodiversity, Adaptive Physiology, and Conservation. CRC Press, Boca Raton, Florida, pp. 37–113.

Kyne, P. M., Carlson, J. K., Ebert, D. A., Fordham, S. V., Bizzarro, J. J., Graham, R. T., Kulka, D. W., Tewes, E. E., Harrison, L. R., and Dulvy, N. K. (eds). 2012. The Conservation Status of North American, Central American, and Caribbean Chondrichthyans. IUCN Species Survival Commission Shark Specialist Group, Vancouver, Canada.

Large P.A., Hammer, C., Bergstad, O.A., Gordon, J.D.M., Lorance, P. 2003. Deep-water fisheries of the Northeast Atlantic: II assessment and management approaches. Journal of Northwest Atlantic Fisheries Science 31: 151–63.

Loew, E. R. 1976. Light, and photoreceptor degeneration in the Norway lobster, Nephrops norvegicus (L.). Proceedings of the Royal Society of London B: Biological Sciences, 193: 31-44.

Lupes, S.C., Davis, M.W., Olla, B.L., Schreck, C.B. 2006. Capture-related stressors impair immune system function in sablefish. Transactions of the American Fisheries Society 135: 129–138.

Mandelman, J. W., & Skomal, G. B. 2009. Differential sensitivity to capture stress assessed by blood acid–base status in five carcharhinid sharks. Journal of Comparative Physiology B 179(3): 267-277.

Mandelman, J. W., Cicia, A. M., Ingram, G. W., Driggers, W. B., Coutre, K. M., & Sulikowski, J. A. 2013. Short-term post-release mortality of skates (family Rajidae) discarded in a western North Atlantic commercial otter trawl fishery. Fisheries Research 139: 76-84.

Mandelman, J. M., Farrington, M. A. 2007. The physiological status and mortality associated with otter trawl capture, transport, and captivity of an exploited elasmobranch Squalus acanthias. ICES Journal of Marine Science: 64.

Manire, C.; Hueter, R.; Hull, E., and Spieler, R., 2001. Serological changes associated with gillnet capture and restraint in three species of sharks. Transactions of the American fisheries Society 130(6): 1038–1048.

McClain, C.R., Allen, A.P., Tittensor, D.P., Rex, M.A. 2012. Energetics of life on the deep seafloor. Proceedings of the National Academy of Sciences USA 109: 15366-15371.

Meyer-Rochow, V. B. 1981. The eye of Orchomene sp. cf. O. rossi, an amphipod living under the Ross Ice Shelf (Antarctica). Proceedings of the Royal Society of London B: Biological Sciences 212: 93-111.

Milliken, H.O., Farrington, M., Carr, H.A., and Lent, E. 1999. Survival of Atlantic cod (Gadus morhua) in the Northwest Atlantic longline fishery. Marine Technology Society Journal 33: 19–24.

Molina, J. M., & Cooke, S. J. 2012. Trends in shark bycatch research: current status and research needs. Reviews in Fish Biology and Fisheries 22(3): 719-737.

Morato, T., Watson, R., Pitcher, T. J., and Pauly, D. 2006. Fishing down the deep. Fish and Fisheries 7: 24-34.

Morgan, A., & Burgess G. H. 2007. At-vessel fishing mortality for six species of sharks caught in the Northwest Atlantic and Gulf of Mexico. Gulf and Caribbean Research 19:123–129.

Morgan, A., & Carlson, J. K. 2010. Capture time, size and hooking mortality of bottom longline- caught sharks. Fisheries Research 101(1): 32-37.

Moyes, C. D., Fragoso, N., Musyl, M. K., & Brill, R. W. 2006. Predicting postrelease survival in large pelagic fish. Transactions of the American Fisheries Society 135(5): 1389-1397.

Muoneke, M.I., and Childress, W.M. 1994. Hooking mortality: a review for recreational fisheries. Reviews in Fisheries Science 2: 123–156.

Neilson, J.D., Waiwood, K.G., and Smith, S.J. 1989. Survival of Atlantic halibut (Hippoglossus hippoglossus) caught by longline and otter trawl gear. Canadian Journal of Fisheries and Aquatic Sciences 46: 887–897.

Nilsson, H.L., and M. Lindstrom. 1983. Retinal damage and sensitivity loss of a light-sensitive crustacean compound eye (Cirolana borealis): Electron microscopy and electrophysiology. Journal of Experimental Biology 207: 277-292.

Norse, E. A., Brooke, S., Cheung, W. W., Clark, M. R., Ekeland, I., Froese, R., Gjerde, K. M., Haedrich, R. L., Heppell, S. S., Morato, T. and Morgan, L. E. 2012. Sustainability of deep-sea fisheries. Marine Policy 36(2): 307-320.

Olla, B. L., Davis, M. W., & Schreck, C. B. 1992. Notes: Comparison of predator avoidance capabilities with corticosteroid levels induced by stress in juvenile coho salmon. Transactions of the American Fisheries Society 121(4): 544-547.

Olla, B. L., Davis, M. W., & Schreck, C. B. 1995. Stress‐induced impairment of predator evasion and non‐predator mortality in Pacific salmon.Aquaculture Research 26(6): 393-398.

Perez, J. A. A., and Wahrlich, R. 2005. A bycatch assessment of the gillnet monkfish Lophius gastrophysus fishery off southern Brazil. Fisheries Research 72: 81-95.

Perez, J. A. A., Pereira, B. N., Pereira, D. A., and Schroeder, R. 2013. Composition and diversity patterns of megafauna discards in the deep‐water shrimp trawl fishery off Brazil. Journal of Fish Biology 83: 804-825.

Perry, H., Waller, R., Trigg, C., McBee, J., Erdman, R., and Blake, N. 2015. A note on bycatch associated with deepwater trapping of Chaceon in the northcentral Gulf of Mexico. Gulf and Caribbean Research 9: 139-142.

Pethybridge, H., Daley, R., Virtue, P., & Nichols, P. 2010. Lipid composition and partitioning of deepwater chondrichthyans: inferences of feeding ecology and distribution. Marine Biology 157(6): 1367-1384.

Phleger, C. F. 1998. Buoyancy in marine fishes: direct and indirect role of lipids. American Zoologist 38(2): 321-330.

Pollock, K. H., & Pine, W. E. 2007. The design and analysis of field studies to estimate catch‐and‐release mortality. Fisheries Management and Ecology 14(2): 123-130.

Raby, G. D., Donaldson, M. R., Hinch, S. G., Patterson, D. A., Lotto, A. G., Robichaud, D., English, K. K., Willmore, W. G., Farrell, A. P., Davis, M. W. & Cooke, S. J. 2012. Validation of reflex indicators for measuring vitality and predicting the delayed mortality of wild coho salmon bycatch released from fishing gears. Journal of Applied Ecology 49: 90-98.

Raby, G. D., Packer, J. R., Danylchuk, A. J., & Cooke, S. J. 2014. The understudied and underappreciated role of predation in the mortality of fish released from fishing gears. Fish and Fisheries 15(3): 489-505.

Ridgway, I. D., Taylor, A. C., Atkinson, R. J. A., Chang, E. S., & Neil, D. M. 2006. Impact of capture method and trawl duration on the health status of the Norway lobster, Nephrops norvegicus. Journal of Experimental Marine Biology and Ecology 339: 135-147.

Rollin, B. E., and Kessel, M. L. 1998. Guidelines for the treatment of animals in behavioural research and teaching. Animal Behavior, 55: 251-257.

Rose, C. S. 1999. Injury rates of red king crab, Paralithodes camtschaticus, passing under bottom-trawl footropes. Marine Fisheries Review 61: 72-76.

Ryer, C. H. 2004. Laboratory evidence for behavioural impairment of fish escaping trawls: a review. ICES Journal of Marine Science: Journal du Conseil 61(7): 1157-1164.

Sangster, G.I., Lehmann, K., and Breen, M. 1996. Commercial fishing experiments to assess the survival of haddock and whiting after escape from four sizes of diamond mesh cod-ends. Fisheries Research 25: 323–345.

Simpfendorfer, C. A., and Kyne, P. 2009. Limited potential to recover from overfishing raises concerns for deep-sea sharks, rays and chimaeras. Environmental Conservation, 36: 97-103.

Skomal G. 2006. The Physiological Effects of Capture Stress on Post-release Survivorship of Sharks, Tunas, and Marlin. PhD Dissertation, Boston, MD, USA: Boston University, 304 pp.

Skomal, G. B. 2007. Evaluating the physiological and physical consequences of capture on post‐release survivorship in large pelagic fishes. Fisheries Management and Ecology 14(2): 81-89.

Skomal, G. B., & Mandelman, J. W. 2012. The physiological response to anthropogenic stressors in marine elasmobranch fishes: a review with a focus on the secondary response. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology 162(2): 146-155.

Skomal, G., & Bernal, D. 2010. Physiological responses to stress in sharks. Sharks and Their Relatives II: Biodiversity, Adaptive Physiology, and Conservation. CRC Press, Boca Raton, 459-490.

Soong, K., & Mok, H. K., 1994. Size and maturity stage observations of the deep-sea isopod Bathynomus doederleini Ortmann, 1894 (Flabellifera: Cirolanidae), in Eastern Taiwan. Journal of Crustacean Biology 14: 72–79.

Spargo, A., 2001. The physiological effects of catch and release angling on the post release survivorship of juvenile sandbar sharks (Carcharhinus plumbeus). Master's Thesis University of Rhode Island.

Spigarelli, S.A.,Thommes, M.M., and Beitinger, T.L. 1977. The influence of body weight on heating and cooling of selected Lake Michigan fishes. Comp. Biochem. Physiol. A Comp. Physiol. 56: 51–57.

St John, J., & Syers, C. J. 2005. Mortality of the demersal West Australian dhufish, Glaucosoma hebraicum (Richardson 1845) following catch and release: the influence of capture depth, venting and hook type. Fisheries Research 76(1): 106-116.

Stevens, J. D., Bonfil, R., Dulvy, N. K., & Walker, P. A. 2000. The effects of fishing on sharks, rays, and chimaeras (chondrichthyans), and the implications for marine ecosystems. ICES Journal of Marine Science: Journal du Conseil 57(3): 476-494.

Stewart, J. 2008. Capture depth related mortality of discarded snapper (Pagrus auratus) and implications for management. Fisheries research 90(1): 289-295.

Stoner, A. W. 2009. Prediction of discard mortality for Alaskan crabs after exposure to freezing temperatures, based on a reflex impairment index. Fishery Bulletin 107: 351-362.

Stoner, A. W. 2012a. Assessing stress and predicting mortality in economically significant crustaceans. Reviews in Fisheries Science 20: 111-135.

Stoner, A. W. 2012b. Evaluating vitality and predicting mortality in spot prawn, Pandalus platyceros, using reflex behaviors. Fisheries Research 119: 108-114.

Stoner, A. W., Rose, C. S., Munk, J. E., Hammond, C. F., and Davis, M. W. 2008. An assessment of discard mortality for two Alaskan crab species, Tanner crab (Chionoecetes bairdi) and snow crab (C. opilio), based on reflex impairment. Fishery Bulletin 106: 337- 347.

Stoner, A.W. 2012. Assessing stress and predicting mortality in economically significant crustaceans. Reviews in Fisheries Science 20: 111–125.

Suuronen, P., Perez-Comas, J.A., Lehtonen, E., and Tschernij, V. 1996. Size-related mortality of herring (Clupea harengus L.) escaping through a rigid sorting grid and trawl codend meshes. ICES Journal of Marine Science 53: 691–700.

Suuronen, P., Turunen, T., Kiviniemi, M., and Karjalainen, J. 1995. Survival of vendace (Coregonus albula) escaping from a trawl cod end. Canadian Journal of Fisheries and Aquatic Sciences 52: 2527–2533.

Urban, J. D. 2015. Discard mortality rates in the Bering Sea snow crab, Chionoecetes opilio, fishery. ICES Journal of Marine Science 72: 1525-1529.

Watson, R. A., & Morato, T. 2013. Fishing down the deep: Accounting for within-species changes in depth of fishing. Fisheries Research 140: 63-65.

Weltersbach, M. S., & Strehlow, H. V. 2013. Dead or alive—estimating post-release mortality of Atlantic cod in the recreational fishery. ICES Journal of Marine Science: Journal du Conseil 70(4): 864-872.

Wetherbee, B. M., & Nichols, P. D. 2000. Lipid composition of the liver oil of deep-sea sharks from the Chatham Rise, New Zealand. Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology 125(4): 511-521.

Wilson, S. M., Raby, G. D., Burnett, N. J., Hinch, S. G., & Cooke, S. J. 2014. Looking beyond the mortality of bycatch: sublethal effects of incidental capture on marine animals. Biological Conservation 171: 61-72.

BIOGRAPHICAL SKETCH

Curriculum Vitae

BRENDAN TALWAR

EDUCATION

Florida State University, M.Sc. Biological Science, Advisor: Dr. Dean Grubbs (expected May 2016) Furman University, B.Sc. Biology, Magna cum laude Undergraduate Field Programs: School for Field Studies, Marine Resource Management, Turks and Caicos Islands, Fall 2010 Wilderness Semester (Conservation Biology and Management), New Mexico, Fall 2009

RESEARCH

U.S. Oceans Southeast, Subcontractor, The Pew Charitable Trusts, March 2016: • Identified gaps in fishery-independent survey coverage of the Gulf of Mexico region

Fisheries Ecology Laboratory, Graduate Assistant, Florida State University, August 2013-Present: • Primary investigator for projects estimating the post-release survivorship of Cuban dogfish and gulper sharks & the use of reflex impairment to predict the discard mortality of giant isopods • Field assistant on fishery independent surveys, including DeepC Consortium cruises investigating the impacts of the Deepwater Horizon oil spill on deep-sea fishes & coastal GULFSPAN surveys • Research assistantship- smalltooth sawfish habitat use • Mentor to 15 undergraduates and 5 gap year students

Shark Research & Conservation Program, Research Associate, Cape Eleuthera Institute, The Bahamas, August 2012-July 2013: • Field manager for 4-8 person research team, coordinator for visiting researchers • Director of shark educational programs, intern hiring coordinator, advisor for undergraduate thesis students, interns, and research assistants, dormitory manager • Assisted with all research programs (Flats, Sea Turtle, Lionfish, etc.), assisted in the operation and maintenance of a remote field station

Shark Bay Ecosystem Research Project, Research Assistant, Australia, Summer 2012 • Field work on the role of herbivorous fish in trophic cascades, maintenance of tiger shark abundance surveys & prey transects, sea turtle habitat use / foraging with Crittercam

Lionfish Research & Education Program/ Shark Research and Conservation Program, Intern, Cape Eleuthera Institute, The Bahamas, Spring 2012 • Field work on lionfish recruitment rates to patch reefs, long term fish abundance surveys, deep water elasmobranch surveys using longlines and deep water cameras, coastal shark surveys • Gear maintenance, conducting outreach, assisting students, and supporting community events

School For Field Studies, Student Researcher, Lionfish habitat use / prey selection, TCI, Fall 2010

Furman University Puma Project, Research Assistant, Puma ecology, New Mexico, Summer 2010

Bermuda Institute of Ocean Sciences, Student Researcher, Coral health, Bermuda, July 2009

Belize Marine Research & Education Center, Student Researcher, Fish behavior, Belize, May 2009 61

TEACHING

Monmouth University, West Long Branch, NJ, January 2015 & January 2016 Co-Instructor: Tropical Island Ecology (Bahamas field course focused on field research techniques)

Sea-to-Sea, Tallahassee, FL, October 2015-May 2016 Marine Science Instructor: Grades 2-8

Florida State University, Tallahassee, FL, August 2013-May 2016 Teaching Assistant (no. courses taught): Eukaryotic Diversity (1), Biology of Fishes (1, intensive field course), Investigations in Biology (6)

The Island School, Eleuthera, The Bahamas, Fall 2012, Spring 2013, Fall 2014, Spring 2015 Applied Research Instructor (Lead): Grades 10-12 • Experiential study abroad program focused on the scientific method, data analysis, scientific writing, public speaking, field skills, and course-specific materials (e.g. stress physiology, fisheries science) • Field expedition & classroom components • Courses resulted in published posters with the Fisheries Conservation Foundation o Spring 2015: Post-release survivorship of the Cuban dogfish and gulper shark o Fall 2014: Post-release Survivorship of the Cuban dogfish (Squalus cubensis) o Spring 2013: Investigating The Interspecific Stress Response of Sharks to Longline Capture o Fall 2012: Physiological and Behavioral Stress Response of Nurse Sharks to Longline Capture

Newfound Harbor Marine Institute, Big Pine Key, FL, Fall 2011 School Coordinator and Marine Science Instructor

Seacamp Association, Inc., Big Pine Key, FL, Summer 2011 Marine Science Instructor & Camp Counselor (Ages 12-15) • Fisheries Science, Florida Keys Flora and Fauna, SCUBA II (Coral Reef Ecology), SCUBA III (NAUI Master Diver Certification Course, Marine Ecosystems)

Furman University, Greenville, SC, Spring 2011 Teaching Assistant, Ecology

Marjorie Basden High School, South Caicos, Turks and Caicos Islands, Fall 2010 Volunteer Intermediate Science Teacher, Grade 11

OUTREACH

Film: • Novel observations of an opportunistic predation event by apex predatory sharks • The Story of Sharks o French Federation of Film/Video Special Jury Award, Festival Mondial de L’Image Sous Marine o Finalist- Beneath the Waves Film Festival, BLUE Ocean Film Festival, Animal Film Festival in Suncheon, Beneath the Waves Film Festival (2012-2014), Miami Underwater Festival • Eating the Enemy: Death to Lionfish

Event Organization: • Wild & Scenic Film Festival, Lead organizer, Tallahassee, FL 2014 o Raised $2000 to bring students to the FSU Marine Lab for free marine science field excursions • Shark Education Day, Cape Eleuthera Institute Organizer, Nassau, The Bahamas 2013 o Engaged over 300 middle school students (presentations, lesson plans, and snorkel trips) • Lionfest, Eleuthera, The Bahamas 2012 o Promoted lionfish as a sustainable food choice through filleting/cooking demos

Media and Public Involvement: • Citizen science: o Over 320 international students engaged in my M.Sc. field research from 2014-2015. o Crowdfunding/blog writing: 105 donors engaged in the scientific process monthly. o Cape Eleuthera Institute blog writer. • Educational presentations: o Over 40 volunteer presentations (elementary–undergraduate) on marine science (2012-2015)

PUBLICATIONS AND PRESENTATIONS

Publications: Talwar B., Brooks E.J., Grubbs R.D. In press. An assessment of post-release mortality for a commonly discarded deep-sea isopod using reflex impairment. ICES Journal of Marine Science.

Talwar B., Harborne A., Brooks E.J. 2015. The conservation implications of spatial and temporal variability in the use of Bahamian tidal mangrove creeks by transient predatory fishes. Extended abstract. In Proceedings of the Mangroves as Fish Habitat Symposium. American Fisheries Society.

O’Shea O. R., Mandelman J., Talwar B., Brooks E.J. 2015. Novel observations of an opportunistic predation event by four apex predatory sharks. Marine & Freshwater Behavior and Physiology.

Harborne A., Talwar B., Brooks E.J. 2015. The conservation implications of spatial and temporal variability in the diurnal use of Bahamian tidal mangrove creeks by transient predatory fishes. Aquatic Conservation: Marine and Freshwater Ecosystems.

Publications (In prep): Shipley O., Talwar B., Grubbs R.D., Brooks E.J. In prep. Isopods associated with deepwater sharks Squalus cubensis and Heptranchias perlo from The Bahamas.

Talwar B., Bouyoucos I., Shipley O., Mandelman J., Brooks E.J., Grubbs R.D. In prep. Validating a portable blood pH analyzer in two elasmobranch species.

Talwar B., Mandelman J., Brooks E.J., Grubbs R.D. In prep. Stress, post-release mortality, and recovery of commonly discarded deep sea sharks caught on longlines.

Talwar B., Mandelman J., Brownscombe J., Bouyoucos I., Cooke S.J., Brooks E.J. In prep. Investigating the interspecific behavioral responses to longline capture using accelerometer and video analysis.

Conference and Seminar Presentations: Talwar B, Mandelman J, Brooks EJ, Grubbs RD. Post-release behavior and recovery of deep sea sharks after longline capture. Southeastern Ecology and Evolution Conference; March 2016; Tallahassee, FL.

Talwar B, Mandelman J, Brooks EJ, Grubbs RD. Predicting the fate of deep sea fisheries bycatch. FSU Department of Biological Science, Ecology & Evolution seminar. December 2015; Tallahassee, FL.

Talwar B, Mandelman J, Brooks EJ, Gokgoz A, Grubbs RD. Stress and post-release survivorship of Cuban dogfish and gulper sharks caught on longlines. American Elasmobranch Society; July 2015; Reno, NV.

Talwar B, Brooks EJ, Grubbs RD. Post-release survivorship of deep sea sharks in Exuma Sound, Bahamas. FSU Coastal & Marine Lab Symposium; March 2015; Tallahassee, FL.

Talwar B, Brooks EJ, Grubbs RD. Deep sea sharks in Exuma Sound: Diversity & Vulnerability. Bahamas Conservation Symposium; March 2015; Exuma, The Bahamas.

O’Shea O, Brooks EJ, Talwar B, Mandelman J. Hierarchical predation behavior in four marine apex predators. American Elasmobranch Society Meeting; July 2014; Chatanooga, TN.

Talwar B, Harborne A, Brooks EJ. Conservation implications of variability in the use of mangrove creeks by transient predatory fishes. AFS Western Division Meeting; April 2014; Mazatlán, Mexico.

Brooks EJ, Suski CD, Talwar B, LaBella RP, Sanford TH, Cooke SJ, Mandelman JW. Does on hook behavior modulate the physiological status of longline-caught sharks? American Elasmobranch Society; July 2013; Albuquerque, NM.

Talwar, B and Perry, T. Pumas in narrow riparian habitats of New Mexico. Cape Eleuthera Institute Seminar Series; February 2013; Eleuthera, The Bahamas.

Talwar, Brendan S. Invasive lionfish habitat use and prey selection in the Turks and Caicos. Department of Environment and Coastal Resources Conference for Sustainability; December 2010; South Caicos.

Scientific Posters: Talwar, B and Perry, T. Metapopulation implications of resident cougars in narrow riparian habitats in New Mexico. Association of Southeastern Biologists; April 2011; Huntsville, AL.

Perry T, Anderson A, Pittman M, Talwar B, Upton B. Exclusive Use of Riparian Habitat by a Population of Puma in Central New Mexico. The 10th Mountain Lion Workshop- Cougars: Conservation, Connectivity, and Population Management; May 2011; Bozeman, MT.

Talwar, Brendan. The Landscape of Fear: Distribution of Puma Prey by Habitat Type. Furman Engage Symposium; April 2011; Greenville, SC.

GRANTS & AWARDS

Graduate Awards 2015: • Florida State University: (1) Inducted to the Florida State University Fellows Society, (2) FSU Congress of Graduate Students Presentation Grant: $200, (3) Coastal & Marine Laboratory Scholarship, $1000 • Cape Eleuthera Institute Awards: (1) Excellence in Research: $2500, (2) Educational Opportunity Grant: $1000, (3) Forrestal International Travel: $1500 • American Elasmobranch Society Travel Award: $500 • Sigma Xi Grant in Aid: $1000 • Guy Harvey Scholarship Award: $5000 • PADI Foundation Grant: $6000 2014: • Experiment.com crowdfunding grant: $6,545 • FSU Congress of Graduate Students Presentation Grant: $500 • Fisheries Conservation Foundation Conference Presentation Grant: $100 • Cape Eleuthera Institute Conference Presentation Grant: $200 • New England Aquarium Research Grant: $2000

Undergraduate Awards • Dean’s List, 2007-2011 • Achiever Scholarship, four year total: $35,000 • Honor Scholar, four year total: $22,500 • National Science and Mathematics Access to Retain Talent Grant, January 2010: $2,000 • School For Field Studies General Scholarship, August 2010: $3,500 • Furman Advantage Research Program, June 2010: $3,000 • Furman Advantage Research Program, July 2009: $5,000

CERTIFICATIONS & SKILLS

Certifications: • PADI Divemaster, NAUI Skin Diving Instructor, AAUS Scientific Diver • CPR/AED, First Aid, O2 Admin, Waterfront Lifeguarding (American Red Cross)

Field Research Techniques: • Abundance Surveys: o Long lining/drum lining, seine netting, underwater video surveys, quadrat surveys o REEF fish surveys (SCUBA, SNUBA, and snorkel) • General Skills: o Training as a small craft captain on oceanographic research vessels o Operating boat trailers o Marine species identification, dissection, and husbandry o Coral health and bleaching surveys o PIT, satellite, dorsal, and dart tagging of sharks o Sea turtle and fish handling and tagging o GPS and radio telemetry o GIS

MISCELLANEOUS

Professional Affiliations • American Elasmobranch Society 2014-2016, Florida Marine Science Educator’s Association 2012- 2014 Extracurricular Activities • Furman & FSU Club Swimming (2007-2016) • Furman University Fish and Fly Club (Founder and President 2007-2010, Treasurer 2011) Language Proficiency • English (fluent), Spanish (conversational)