Stress physiology of scalloped and great hammerhead sharks from a bottom longline fishery

Abstract

The scalloped hammerhead Sphyrna lewini and the great hammerhead S. mokarran are large, coastal to semi-oceanic shark species common to waters of the US east coast where they are regularly taken in commercial and recreational fisheries, particularly the bottom longline fishery. High rates of hooking mortality and low rates of population growth are believed to have caused severe declines in the US Atlantic populations of these species. The objective of this study was to determine the physiological stress induced by bottom longline capture in both S. lewini and S. mokarran. Physiological stress was quantified using the blood biochemical indicators glucose, lactate, pH, haematocrit, sodium, potassium, calcium, chloride and magnesium, which have been demonstrated to indicate physiological stress in elasmobranchs. Each shark captured was assigned a condition factor, which was compared with the stress parameters and time on hook to quantify stress induced by different longline hook times. In S. lewini, the physiological stress parameters lactate, pH, sodium and chloride scaled with hook time, whereas in S. mokarran, only lactate was affected by hook time. In both species, water temperature affected lactate and glucose levels, as well as sodium and pH levels in S. lewini and magnesium levels in S. mokarran. These data will be useful for estimating post-release mortality of S. lewini and S. mokarran from measurements taken at the time of capture, and quantifying the physiological stress response to longline capture in both species to the Atlantic bottom longline fishery.

Introduction

The scalloped hammerhead S. lewini and great hammerhead S. mokarran are large, coastal to semi-oceanic species that are distributed throughout warm temperate and tropical oceans of the world, including the nearshore and pelagic waters of the US east coast (Compagno, 1984; Castro, 2011). Because of their broad range of habitats in US Atlantic waters, these species are regularly caught in both inshore and offshore fisheries in this region, such as the Atlantic directed shark bottom longline fishery (Morgan et al., 2009) and the US Atlantic pelagic longline fishery (Miller et al., 2013). Previous stock assessments suggest that these fisheries may have contributed to significant declines in the northwest Atlantic populations of both species (Hayes et al., 2009; Jiao et al., 2011). For example, the US Atlantic S. lewini population is believed to have been depleted by over 80% of their virgin stock biomass since the early 1980s (Hayes et al., 2009). Comparable declines have also been reported for US Atlantic S. mokarran populations, but are less certain because of species misidentification (Beerkircher et al., 2002, 2004) and naturally lower densities of this species resulting in low sample sizes. The most recent stock assessments for the US waters concluded S. lewini is not currently overfished or experiencing overfishing, though the stock had experienced overfishing in the past; S. mokarran is currently overfished but overfishing is no longer occurring (SEDAR, 2024).

Because of high fishery exposure, it is important to obtain information on capture-induced stress and post-release mortality rates to understand the full effect of fisheries on these vulnerable species. Both species experience high at-vessel mortality (AVM) in commercial fisheries (Morgan & Burgess, 2007; Morgan et al., 2009; Gulak et al., 2015). For example, Morgan et al. (2009) reported >98% total mortality rate for both S. lewini and S. mokarran based on fishery observer data collected aboard commercial longline vessels targeting sharks. Hooking mortality rates for a variety of species including hammerheads also tend to increase with time spent hooked (Morgan et al., 2009; Gulak et al., 2015), which suggests that limiting gear soak times may be an effective way to reduce hooking mortality in sharks. However, the probability of survival for S. mokarran and S. lewini released alive from fishing gear is unknown.

The magnitude of physiological stress from capture experienced in elasmobranchs (sharks and rays) is thought to be most influenced by the capture method, hook time (HT) and the metabolic scope of the species (i.e. low metabolic scope linked to benthic/sluggish species vs. high metabolic scope linked to pelagic/continuously swimming species) (Skomal, 2006, 2007; Mandelman & Skomal, 2009; Skomal & Mandelman, 2012). Differences in stress induced by capture method can result from varying degrees of physical trauma and respiratory inhibition (Skomal & Mandelman, 2012). Longer HTs are generally related to increased stress and mortality (Morgan et al., 2009; Morgan & Carlson, 2010), although some species have recovered after release even with long HTs (Brooks et al., 2012; Marshall et al., 2012; Whitney et al.,, 2021). Stress is examined in most wild animals by quantifying stress hormones; however, at the time of this study, there was not a validated assay to quantify the primary stress hormone in elasmobranchs (Anderson, 2012). An alternative way to examine stress in these organisms is to quantify the secondary stress response, which can include investigating blood glucose, pCO₂, lactate, bicarbonate, pH, sodium, potassium, calcium, chloride, magnesium and haematocrit. Skomal and Mandelman (2012) provide a detailed review of the secondary stress response in marine elasmobranchs.

The purpose of this study was to examine longline capture-induced physiological stress in both S. lewini and S. mokarran. Our specific objectives were to (i) quantify how secondary stress parameters fluctuate in relation to species, time on the hook, shark size and the water temperature at capture and (ii) identify characteristic secondary stress parameters in relation to release condition.

Materials and Methods

Survey

S. lewini and S. mokarran were captured and sampled during an ongoing Florida State University (FSU) fishery-independent longline survey, which targets elasmobranchs. This work was conducted under FSU Institutional Animal Care and Use Committee (IACUC) protocol 1718. Separately and concurrently, S. lewini and S. mokarran were also captured and sampled through contracted bottom longline fishing efforts with commercial fishermen under Mote Marine Laboratory IACUC protocol 13–11-NW2. The FSU survey longline consisted of a 4.0-mm monofilament mainline anchored on each end and marked with a surface buoy bearing the permit numbers. Each mainline set was approximately 750 m long. A standard set included 50 gangions consisting of a stainless-steel tuna clip with an 8/0 stainless-steel swivel attached to 2.5 m of 360 kg monofilament that was doubled in the terminal 25 cm and attached to a 16/0 barbed circle hook. Hooks were baited with ladyfish Elops saurus or Spanish mackerel Scomberomorus maculatus. Each gangion included an inline HT-600 HTr (Lindgren-Pitman, Pompano Beach, FL). Depth (m), turbidity (cm), water temperature (°C), salinity and dissolved oxygen (mg/L) (YSI Pro 2030, Yellow Springs, OH, USA) were recorded from the surface to the bottom for all sets made in depths less than 10 m. Additionally, bottom water temperature (°C) was recorded for sets deeper than 10 m using a Lotek LAT2000 (Seattle, WA) archival temperature depth recorder programmed to record every 10 s. Two sets were typically soaked concurrently. Soak times for the first set were 1 h to minimize mortality, and all lines were set during daylight hours. Soak times for the second set varied depending on the haul duration of the first set but were typically less than 3 h. Additional fishery independent samples were also collected from a South Carolina Department of Natural Resources (SCDNR) longline vessel, with methods standardized to the FSU survey.

The concurrent contracted fishing efforts were done in collaboration with commercial bottom longline fishermen aboard their vessels. Gear specifications were similar to the FSU longline, with the following differences: up to 260 gangions were suspended on 3–6 nm of 4.0 mm monofilament mainline. Each gangion was approximately 3 m of 3.5-mm monofilament attached to an 18/0 barbed circle hook, also integrated with HT-600 HTrs (Lindgren-Pitman, Pompano Beach, FL, USA). Soak times ranged 2–18 h. Oceanographic conditions were measured at each fishing location using a hand-held meter (YSI model Pro Plus, Yellow Springs, OH, USA).

For both fishing efforts, the line was hauled in the order and direction it was set, and hammerheads were sampled as they were caught during retrieval. Sharks were captured and sampled in the nearshore and coastal waters off South Carolina, the Atlantic side of the Florida Keys from Key West to Islamorada, inside Everglades National Park in Florida Bay, and state and federal waters off of the southeast coast of Florida (specifically near Madeira Beach and Key West, FL, USA) (Fig. 1). Fishing activities were permitted by Special Activity Licence 1345 and 12–0041-SRP from the Florida Fish and Wildlife Conservation Commission, and permit SHK-EFP-1310 issued by the National Marine Fisheries Service Highly Migratory Species Management Division.

Sampling

Hooked sharks were brought alongside the vessel and, in most cases, were brought onboard the deck of the boat or a transom on the back of the boat level with the water where they were restrained by hand. Gills were irrigated with a seawater hose for sharks that were brought on deck for > 1 min (e.g. Whitney et al., 2021). As soon as a shark was restrained, a 1–5 ml blood sample was collected (within ~ 30–120 s of being restrained), using a 16–18-gauge needle attached to a heparinized syringe (Lithium heparin #374858, Sigma-Aldrich, St. Louis, MO, USA). Blood samples were obtained via caudal venipuncture either from the ventral or lateral surface of the caudal region (Lawrence et al., 2020). The lateral caudal puncture was employed more often throughout the study as a means to reduce on-deck time, because the blood sample and tagging could happen concurrently (with shark lying on ventral side). The lateral puncture technique proved to be quick, and involved inserting the needle from the side of the individual’s caudal region, targeting the hemal arch. Given that all blood samples (from lateral or ventral side) were from the hemal arch and not another region of the body, it is not expected this sampling approach would affect blood values.

After blood sampling, sharks were measured (precaudal length (PCL), fork length (FL), stretch total length (STL)), sex was determined and sharks were externally tagged with an identification tag or one of various electronic tags for a separate study (either an acceleration data-logger attached through two holes in the dorsal fin, or a pop-up satellite archival tag attached via dart to the dorsal musculature, e.g. Whitney et al., 2021). Blood sampling, measurement and tagging generally took between 1 and 5 min. Upon release, each shark was assigned a condition score using a 5-point scale: 1 = vigorous, excellent condition, 2 = normal swimming, good condition, 3 = laboured or disoriented swimming, fair condition, 4 = nictitating membrane response, slow movement, poor condition and 5 = at-AVM or moribund (Hueter et al., 2006; Whitney et al., 2021).

Blood analyses

For all sharks sampled, pH and lactate were assessed by running a small aliquot of blood immediately (within 15 min of the blood draw) in a VetScan i-STAT 1 point of care device (Zoetis, Parsippany, NJ) using CG4+ cartridges, which have been validated for use in elasmobranchs (Mandelman and Farrington, 2007; Mandelman and Skomal, 2009; Gallagher et al., 2010). Because of variability in additional blood gas data, pCO₂ and bicarbonate data are not reported here (Harter et al., 2015). For sharks caught during FSU research longline surveys, blood glucose was measured onboard using an Accu-Chek glucose meter (Roche Diagnostics, Basel, Switzerland), which has been validated for use on fishes (Cooke et al., 2008). The remaining FSU blood sample was then placed on ice (4°C) for an average of 6 h before further processing in the laboratory, where the blood was gently inverted multiple times to account for any separation that had occurred, and an aliquot of each sample was spun in a haematocrit centrifuge at 15 000 g for 5 min. Haematocrit levels were determined by calculating the red blood cell percentage of the whole blood volume.

Alternatively, for sharks caught during collaborative efforts with commercial fishermen, the blood samples were centrifuged entirely onboard the fishing vessels after i-STAT assessment of pH and lactate. These samples were first spun (n = 4 microcapillary tubes per individual) in a portable haematocrit centrifuge (Zipocrit, LW Scientific, Lawrenceville, GA, USA) for 5 min. The remaining blood for these samples was then centrifuged so that plasma could be separated from the red blood cell pellet, and both the plasma and pellet were immediately frozen using a liquid nitrogen dry shipper for later laboratory analyses of glucose, and ion (potassium, sodium, chloride, magnesium and calcium) concentrations using benchtop Critical Care Xpress and pHOx blood analysers (Nova Biomedical, Waltham, MA, USA).

Statistical analyses

Stress physiology data for pH were temperature-corrected to bottom water temperature measurements at the time of capture (Mandelman & Skomal, 2009; Gallagher et al., 2010). Because haematocrit is represented as a percentage, these data were arcsine-transformed prior to analyses.

Two sets of linear models were used to identify patterns between blood stress parameters and HT, bottom water temperature and/or FL, and between species for sharks that were alive when brought to the vessel. Models were constructed in R (v. 3.3.3; R Foundation for Statistical Computing, Vienna, Austria) using base functions. First, to investigate whether there were overall differences in blood indicators between species, a full global linear model was fit for each blood stress response, which included species, HT, FL and bottom water temperature as fixed predictor variables, as well as interactions between all variables. Models with all possible combinations of these predictors and interactions were compared using the ‘dredge’ function in the MuMIn package (v. 1.43.17; Bartón, 2020), with the best-fit model selected using the Aikake’s Information Criterion (AIC) corrected for small sample size and parsimony (i.e. the model with the least degrees of freedom within ΔAIC < 2 of the top model was chosen). If species was maintained as a fixed predictor within the best-fit model, this was taken as indication of significant overall differences in that blood parameter between S. mokarran and S. lewini, given variation in the other predictor variables. Subsequently, to more clearly define which predictive factors influenced blood stress parameter values in each species, a set of species-specific models were constructed, with HT, FL, water temperature and interactions between these variables included as potential fixed predictors of each blood stress response. The best-fit model was again chosen using AIC and parsimony. Fixed predictors included in the best-fit models were interpreted as influential predictors of the blood stress parameter in question for the specific species.

Additionally, stress parameters were statistically compared among release condition groups in each species using a one-way analysis of variance (ANOVA), a Welch’s ANOVA, or a Kruskal–Wallis test, depending on whether data were normally distributed and homoscedastic. When ANOVAs or their corresponding tests were significant, a Tukey HSD or Wilcoxon rank sum test was conducted to identify significant pairwise differences. Because each shark was captured at a unique HT, a one-way analysis of co-variance (ANCOVA) was conducted prior to the aforementioned analyses to investigate differences in the stress parameters among release condition groups, while controlling for the covariate HT; however, in all cases but one, the effect of HT was not significant in the model, and in the one case that it was, there were non-overlapping covariate ranges for the stress parameter and HT, making an ANCOVA an inappropriate analysis for these data. All tests were considered significant at α = 0.05.

Results

Blood samples were collected from 86 S. lewini and 85 S. mokarran with mean (±S.E.) FLs of 156 ± 4 and 180 ± 4 cm, respectively. Known HTs for S. lewini ranged from 6 to 382 min and 5 to 538 min for S. mokarran (Fig. 2), with AVM occurring from HTs as low as 83 and 114 min in S. lewini and S. mokarran, respectively (Fig. 3a).

For sharks alive at capture, several blood stress parameters showed relationships with species, HT, water temperature, and shark size (see Supplementary Tables 1–3 for model selection details). Species was maintained as a significant predictor of haematocrit, potassium, magnesium and lactate (Table 1), indicating that for a given HT, water temperature and shark size, haematocrit tended to be lower in S. mokarran than in S. lewini, while lactate, magnesium and potassium overall were higher in S. mokarran than S. lewini (Fig. 2). Whether these differences were because of inherent differences in baseline values of these parameters or a higher susceptibility to stress in S. mokarran compared to S. lewini is unclear. The two species also showed some similarities in patterns of relationships between blood stress parameters and HT, shark size and water temperature. In both species, lactate increased with HT, shark size and water temperature, and glucose increased with water temperature (Table 1, Fig. 2). In addition, in S. lewini pH decreased with HT and water temperature while sodium and chloride both showed positive relationships with HT, and sodium also increased with water temperature (Table 1, Fig. 2). Alternatively, magnesium decreased with water temperature in S. mokarran. No factors significantly influenced haematocrit, potassium or calcium in either species (Table 1, Fig. 2). See Supplementary Tables 1–3 for full model selection results.

The secondary stress parameters glucose, sodium, chloride and magnesium did not vary significantly between release conditions in either species (Fig. 3; Table 2). However, in both S. lewini and S. mokarran, lactate and potassium significantly increased and pH significantly decreased when release condition worsened (Fig. 3; Table 2). Additionally, in S. lewini, calcium significantly increased with worsening release condition (Fig. 3; Table 2), while in S. mokarran, haematocrit significantly decreased with worsening release condition (Fig. 3; Table 2).

Discussion

The results of this study confirmed the previously described susceptibility of S. lewini and S. mokarran to capture stress (Gulak et al., 2015), with a number of blood stress parameters changing character based on hooking duration, shark size and water temperature. Our results confirm that both species are highly susceptible to capture stress, even after relatively short soak times of a few hours or less, and provide insight into how the stress response differs between the species as well as by fishing practices and environmental conditions.

Overall, HT, water temperature and shark size, when significant, showed similar directions of trends relative to each other across blood metrics, indicating that stress generally increases at longer HTs, in warmer water temperatures and in larger animals in both species. Differences in lactate, potassium and haematocrit levels between species also indicated that on average, S. mokarran showed values of these parameters that are characteristic of higher stress than S. lewini for a given capture duration. Whether these interspecific differences are a result of differences in baseline values of these parameters between species or because of a heightened stress response in S. mokarran compared to S. lewini cannot be definitively determined here. Several previous studies have also highlighted the high stress disruption and apparent susceptibility of S. mokarran compared to several sympatric carcharhinid sharks, as S. mokarran tends to exhibit the highest levels of mortality or physiological disruption of commonly caught species (Gallagher et al., 2014; Jerome et al., 2017), including S. lewini (Gulak et al., 2015).

HT was the most common predictor of blood stress indicators and both hammerhead species clearly showed release condition deteriorating as HT increased. These results signal the sensitivity of hammerheads to capture stress even over relatively short durations, including AVM reached at a minimum HT of 83 and 114 min for S. lewini and S. mokarran, respectively. While not as often examined as HT, it is also logical that stress levels tend to increase in warmer water temperatures, as was seen in our results, in particular with glucose increasing with water temperature increasing in both species. As hammerhead sharks are ectotherms, their metabolic rates and muscle performance are governed by environmental water temperatures (Gillooly et al., 2001; Angilletta et al., 2002). Higher water temperatures allow for increased muscle performance, potentially increasing the fight response of sharks on the line, while at the same time increasing oxygen demands through heightened metabolic rates, alongside generally lower oxygenation of the water at warmer temperatures (reviewed in Rummer et al., 2022). Under respiration-limiting scenarios such as longline capture, this could exacerbate the rates of anaerobic respiration and lactic acid production already occurring, while limiting oxygen-dependent recovery as well (reviewed in Skomal and Bernal, 2010). Most previous studies have not examined capture stress responses in elasmobranchs at a large enough range in water temperatures to elucidate any patterns, though similar to the present study, water temperature has also been shown to lead to a higher stress response to capture in Mustelus antarcticus (Guida et al., 2016) and Carcharhinus limbatus (Whitney et al., 2021), though not in C. brachyurus (Dapp et al., 2016). Considering the potential for water temperature to influence physiology and that water temperatures can vary widely even within a single fishery between seasons, this question merits further investigation and could prove useful in determining how susceptibility to mortality via fisheries capture might vary between fisheries and throughout the year.

Lactate and pH

Increases in metabolic stress, as well as overall acidosis (low blood pH), were observed in both S. lewini, and S. mokarran with increasing HT and were also associated with declining release condition. These results indicate that upon capture, these species are likely using burst swimming behaviour to escape, which triggers anaerobic respiration in the muscle and results in a buildup of intracellular metabolic intermediates like lactic acid. Lactic acid dissociates into acidic hydrogen ions and lactate, resulting in increases of blood lactate and pH declines (Black, 1958; Cliff & Thurman, 1984; Wood, 1991; Skomal and Bernal, 2010). Shark baseline lactate levels have been reported as ~ 1.3 mmol L⁻¹ (Spargo, 2001) and baseline pH values ~7.4–8.0 (Spargo, 2001; Mandelman and Skomal, 2009), but with prolonged capture stress, we observed lactate levels up to 32 mmol L⁻¹ and pH values as low as 6.6 in sharks alive at-vessel. Similar increases in lactate with HT have been observed in longline-caught Caribbean reef sharks C. perezi (Brooks et al., 2012), bronze whalers C. brachyurus (Dapp et al., 2016), and several other large coastal species (e.g. Whitney et al., 2021). This potential burst escape behaviour, which is likely contributing to metabolic stress, has been documented in S. mokarran by analyzing data collected from accelerometers attached to fishing gear (Gallagher et al., 2017). Contrasting behaviour has been documented in the gummy shark M. antarcticus, and nurse shark Ginglymostoma cirratum, suggesting these sharks rest on the bottom while still captured, and thus experience no significant increases in lactate (M. antarcticus) or decreases in pH (G. cirratum) with higher HTs (Guida et al., 2016; Bouyoucos et al., 2018). Disparate results have also been observed in smalltooth sawfish Pristis pectinata, which were captured using identical methods to the sharks in this study. Sawfish exhibited a lower stress response for all of the parameters investigated, likely as a result of their fairly calm behaviour post hooking and their low metabolic scope (Prohaska et al., 2018).

Ionic disruption

Ionic disruptions were apparent in both species, with S. lewini experiencing increased sodium and chloride levels with longer HTs, and S. mokarran experiencing reduced magnesium levels as water temperature increased. Increases in potassium were also strongly correlated with worsening release condition in both species. These blood indicators all represent fluctuations in ions that are important in key homeostatic and metabolic pathways. For example, potassium is present in cells as the primary intracellular cation (Hochachka and Somero, 2002), and previous work has also highlighted stress-induced increases in blood potassium and associations with mortality events (Moyes et al., 2006; Mandelman and Skomal, 2009; Frick et al., 2010; Marshall et al., 2012; Dapp et al., 2016; Whitney et al., 2021). Increases in extracellular potassium likely result from repeated cycles of excitation-contraction coupling within the swimming muscles, but they could also be caused by intracellular acidosis and muscle cell damage as a result of extended and exhaustive burst swimming. Disruptions leading to elevated potassium in the blood can affect cardiac functioning and potentially muscle contraction physiology as well (Cliff and Thurman, 1984; Moyes et al., 2006; Skomal and Mandelman, 2012).

Sodium and chloride also play roles in the regulation of homeostasis. Sodium is an important cation in extracellular fluid (Hochachka and Somero, 2002; Skomal and Bernal, 2010), and in combination with concentrated intracellular potassium, is regulated to maintain an appropriate electrochemical gradient across the cell membrane, which is critical for physiological processes such as muscle contraction and nerve conduction (Hochachka and Somero, 2002). Increases in chloride may also be linked to cellular efforts to deal with intracellular acidosis through chloride/bicarbonate exchangers and/or reflective of loss of osmoregulatory control (Perry and Gilmour, 2006; Skomal and Bernal, 2010). Increases in both sodium and chloride levels during capture stress similar to those observed here in S. lewini have also been reported in other shark species (Wells and Davie, 1985; Perry and Gilmour, 2006; Mandelman and Farrington, 2007; Skomal and Bernal, 2010; Marshall et al., 2012; Kneebone et al., 2013).

Conclusions

The results of this study will be useful in informing management on the physiological response of S. lewini and S. mokarran to bottom longline capture, as influenced primarily by HT and water temperature. From our condition assessments at release, 92% of S. mokarran survived initial capture (i.e. were released alive) within a maximum HT of approximately 3.5 h, and 86% of S. lewini survived initial capture with maximum HT of approximately 5.5 h. While the deployment of electronic tags should be used to monitor post-release survivorship in these species over a broad range of HTs, the results of this study suggest that limiting gear soak times could increase overall longline capture survival. Even with the shorter soak times used by this study, both species of hammerheads experienced enough physiological disruption to lead to some AVM. Results of the secondary stress parameter analyses show typical stress responses such as increases in lactate, declines in pH and ionic disruption, particularly with larger individuals and at higher water temperatures. The combined effects of warmer waters and longer soak times, particularly in mature adults, should be considered for future management of both hammerhead species, particularly when considering climate change.

Acknowledgements

We thank the collaborating commercial captains (Dave Campo, Jim Bonnell, Randy Lauser and Luke Hill) and their first mates. We thank the many graduate students and volunteers that assisted in the FSU surveys and many interns and staff at Mote Marine Laboratory. For analysis of ions and metabolites, we thank Diego Bernal (the University of Massachusetts Dartmouth) for the use of the CCX and Charles Innis, Deana Edmunds and Kerry McNally (New England Aquarium) for use of their pHOx benchtop analyser. Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA.

Author Contributions

Research design: B.K.P., H.M., R.D.G. and N.M.W. Funding acquisition: R.D.G and N.M.W. Field sampling: B.K.P., H.M., R.D.G, K.L, B.S.F., A.A., B.A.K. and N.M.W. Laboratory Analysis: B.K.P., H.M., K.L., A.A. and B.A.K. Statistical analysis: B.K.P., H.M. and K.L. Original draft writing: B.K.P., H.M., R.D.G., K.L. and N.M.W. Manuscript revisions: B.K.P., H.M., R.D.G, K.L, B.S.F., J.J.M., A.A., R.E.H., B.A.K. and N.M.W. All authors have reviewed and approved the final manuscript.

Conflict of Interest

The authors declare no conflicts of interest in relation to this study.

Funding

This work was supported by grants from the NOAA Cooperative Research Program awarded to N.M.W. and R.D.G., the NOAA Bycatch Reduction Engineering Program awarded to N.M.W. and the NOAA Office of Protected Resources awarded to R.D.G.

Data Availability

Data used for this article can be accessed by requesting access from the corresponding author.

Supplementary Materials

Supplementary material is available at Conservation Physiology online.

References