You are what you eat: Examining the effects of provisioning tourism on shark diets

Biological Conservation 224 (2018) 300–308

Contents lists available at ScienceDirect

Biological Conservation journal homepage: www.elsevier.com/locate/biocon

You are what you eat: Examining the effects of provisioning tourism on shark diets

T

⁎

Kátya G. Abrantesa, , Juerg M. Brunnschweilerb, Adam Barnetta a b

College of Science and Engineering, James Cook University, Townsville, QLD 4811, Australia Independent Researcher, Gladbachstrasse 60, 8044 Zurich, Switzerland

A R T I C LE I N FO

A B S T R A C T

Keywords: Bayesian mixing models Conservation Ecotourism management Shark Stable isotope analysis Wildlife provisioning

Wildlife tourism is a growing industry, with significant conservation and socio-economic benefits. Concerns have however been raised about the possible impacts of this industry on the long-term behaviour, health and fitness of the animal species tourists come to see (the target species), particularly when those species are regularly fed to improve the tourism experience. Information on the contribution of food rewards to the diet of the target species at feeding sites is critical to assess the dependency on handouts and to identify possible health/fitness problems that might be associated, if handouts become the main part of animals' diets. Here, we use stable isotopes (δ13C and δ15N) to evaluate the importance of handouts for a marine predator, the bull shark (Carcharhinus leucas), at a feeding site (Fiji) where shark feeds occur 5 days/week and sharks (up to 75 individuals/dive) are fed ~200 kg of tuna heads/day. There was no evidence of incorporation of food provided, even for individuals that regularly consume food rewards. Results, when combined with those from previous studies on bull shark movements and feeding rates at our study site, show that current levels of provisioning likely have no long-term impacts on bull shark diet or behaviour. This study also demonstrates the applicability of stable isotope analysis to assess and monitor the contribution of food rewards to wildlife, and highlights the benefits of using multi-sources of information to gain a holistic understanding of the effects of provisioning predators.

1. Introduction Wildlife tourism is a growing industry that often involves feeding (provisioning) wildlife to increase the chances of viewing animals up close. Despite that provisioning tourism can have significant conservation, economic and social benefits (Cisneros-Montemayor et al., 2013; Gallagher and Hammerschlag, 2011), concerns are being raised about the impacts this expanding industry might have on the long-term behaviour, health (i.e. functional and metabolic efficiency) and fitness (i.e. reproductive success) of the target animals, with several studies asking if the negative impacts might outweigh the positive (e.g. Burgin and Hardiman, 2015; Shannon et al., 2017; Trave et al., 2017). Although tourism-related behavioural changes have been documented for many species (e.g. of the 48 provisioning studies in the marine environment reviewed by Trave et al. (2017), behavioural changes were evident in 89.5% of the cases), little information is available on tourism effects on health and fitness. This issue has been repeatedly pointed out in recent wildlife tourism reviews for both the terrestrial (e.g. Newsome et al., 2015; Penteriani et al., 2017) and aquatic (e.g. Brena et al., 2015; Burgin and Hardiman, 2015; Trave et al., 2017) environments.

⁎

Corresponding author. E-mail address: [email protected] (K.G. Abrantes).

https://doi.org/10.1016/j.biocon.2018.05.021 Received 20 July 2017; Received in revised form 23 May 2018; Accepted 28 May 2018 0006-3207/ © 2018 Elsevier Ltd. All rights reserved.

To better understand if wildlife provisioning might have a negative impact on the animal species tourists come to see (i.e. the target species), information on the basic factors that may affect the health and fitness of those animals is crucial. This includes information on parameters such as individual consumption rates at feeding events, energy content of provisions, daily energetic requirements, and the contribution of these food rewards to the overall diet of the target species. This information is critical to assess the level of dependency on food handouts and any possible problems that might be associated, if food rewards become the main part of the animal's diets. Predators regulate ecological communities by controlling prey populations through both direct predation and by influencing prey behaviour (risk effects) (Beschta and Ripple, 2009; Estes et al., 2011; Wirsing and Ripple, 2011). Artificially providing food can lead to direct effects such as altered predator abundance, diet, life-history, social behaviour and spatial use, as well as to indirect effects such as increasing competition between co-occurring species, increased predation pressure on native prey, and prey switching (see Newsome et al. (2015) and Burgin and Hardiman (2015) for examples). Food provisioning therefore has the potential to alter trophic structure and dynamics of

Biological Conservation 224 (2018) 300–308

K.G. Abrantes et al.

both aquatic and terrestrial systems (Brena et al., 2015; Burgin and Hardiman, 2015; Newsome et al., 2015). However, its effects greatly vary depending on species- and case-specific settings (Brena et al., 2015; Newsome et al., 2015; Penteriani et al., 2017). Diving with sharks is becoming increasingly attractive to recreational divers worldwide. Many dive operators regularly feed sharks to ensure that animals attend the dive sites, so that shark-tourist interactions can be guaranteed. Although some studies have detected effects of feeds/attracts on shark community composition (e.g. Brunnschweiler et al., 2014; Clarke et al., 2013; Meyer et al., 2009), shark behaviour (e.g. Clua et al., 2010; Fitzpatrick et al., 2011) and metabolic rate (Barnett et al., 2016), it is still not known if those changes have any negative long-term consequences for the sharks' health and fitness (Brena et al., 2015; Burgin and Hardiman, 2015; Gallagher et al., 2015). In the present study, stable isotope analysis of carbon (δ13C) and nitrogen (δ15N) are used to evaluate the importance of food subsidies for a marine predator, the bull shark (Carcharhinus leucas, Carcharhinidae), the main species at the Shark Reef Marine Reserve, a tourism feeding site in Fiji (Brunnschweiler et al., 2014). According to a recent study, the consumption of 2.3 tuna heads per week can meet the energetic requirements of a ~2.8 m bull shark individual, suggesting that bull sharks could meet their energy requirements exclusively from provisioning (Brunnschweiler et al., 2017). This stresses the importance of determining the contribution of bait for these animals' diets, so that potential impacts on their health and fitness and on the overall food web can be assessed. Such information can be used to inform the tourism operators of “safe” feeding levels, therefore contributing to setting best practices for the appropriate management of the industry. Whitetip reef sharks (Triaenodon obesus, Carcharhinidae) were also sampled as this species has relatively small home ranges and, unlike bull sharks which move in and out of the study area (Brunnschweiler and Barnett, 2013), whitetip reef sharks are typically resident in the reefs where they occur (Barnett et al., 2012; Speed et al., 2012). Since whitetip reef sharks are also an important part of the shark dive in Fiji (Brunnschweiler et al., 2014), it is likely that this species would more easily reflect any significant incorporation of bait provided. Stable isotope analysis is useful for this study because δ13C and δ15N can differ among primary sources of nutrition, and because δ13C and δ15N change in a predictable way as they are passed on from food source to consumer. δ13C is particularly useful as a source indicator in coastal systems as different primary producers (e.g. plankton, algae, seagrass) typically have different δ13C values, and because δ13C values change little as material is passed on from food to consumer (0–1‰; DeNiro and Epstein, 1978; McCutchan et al., 2003). δ15N is generally used as a trophic level indicator (Post, 2002) because of the higher δ15N trophic fractionation (2–3‰; McCutchan et al., 2003; Minagawa and Wada, 1984). δ15N will be particularly important in the present study because sharks at our study site are fed tuna heads/frames. Since tunas are typically high trophic level pelagic species (Olson et al., 2010), they have δ13C values in the lower part of the marine δ13C spectrum, and δ15N values much higher than those of local reef prey, meaning that it will be possible to identify and quantify any contribution of tuna provided to shark nutrition. Given that tuna heads provided can fuel the energetic needs of bull sharks (Brunnschweiler et al., 2017), we expect that tuna will make up a significant component of the diet of at least some bull sharks individuals. Note however that the study from Brunnschweiler et al. (2017) was conducted in 2008, and that the number of sharks attending the dive site in 2008 was lower than in 2015, when the present study was done (see Fig. 1). This means that in 2008 bait was shared among a smaller number of bull shark individuals and therefore that bait contribution could have been higher than in 2015, the year the present study was conducted.

Fig. 1. Monthly mean ( ± SD) number of bull sharks sighted per dive for 2008 (when feeding rate data used in Brunnschweiler et al. (2017) was collected) and 2015 (when stable isotope data from the present study was collected).

2. Methods 2.1. Study site and study species This study was conducted at the Shark Reef Marine Reserve, a multispecies shark diving site on the southern coast of Viti Levu, Fiji (Brunnschweiler et al., 2010). For a detailed description of the dive protocol and for information on species composition and relative abundances of sharks at this provisioning site, see Brunnschweiler and Baensch (2011) and Brunnschweiler et al. (2014). Currently, shark feeds occur 5 days/week in two feeding dives/day, where ~100 kg of tuna (Thunnus spp. (Scombridae)) heads (each ~2.4 ± 0.9 kg ( ± SD)) are fed on the first and again on second dive of the day. Several shark species are attracted to the dive site but the bull shark is the most abundant species and the main attraction of this shark dive (Brunnschweiler et al., 2014). In 2015 (when stable isotope samples for the present study were collected), up to 75 individual bull sharks could be seen in each dive. Bull sharks inhabit coastal and tropical reef habitats (Brunnschweiler et al., 2010; Carlson et al., 2010; Daly et al., 2014; Espinoza et al., 2016), and feed mostly on fish including teleosts and also on elasmobranchs such as batoids and smaller sharks (Cliff and Dudley, 1991; Olin et al., 2013; Trystram et al., 2016). The bull sharks of the Shark Reef Marine Reserve range in size from ~1.8 m to ~3.5 m (subadults and adults; average size ~2.8 m) and include both males and females. Although bull sharks can be seen throughout the year, there are seasonal cycles in abundance: more individuals are present between January and September (mean ± SD: 39.0 ± 11.5 individuals/dive; mode: 40 individuals/dive for 2015; unpubl. data) and less between October and December (~16.6 ± 6.5 individuals/dive, mode: 15 individuals/dive for 2015) (Fig. 1), as animals move out of the area for weeks to months at the end of the calendar year (likely for reproductive purposes), typically returning in the beginning of the following year (Brunnschweiler et al., 2014; Brunnschweiler and Baensch, 2011; Brunnschweiler and Barnett, 2013). Tracking studies show that different bull shark individuals have different degrees of site fidelity to the feeding site, with some individuals being present almost year round while others spend longer periods of time away (Brunnschweiler and Barnett, 2013). These regular movements out of the study area, coupled with the relative slow muscle turnover rate of large sharks (Logan and Lutcavage, 2010; MacNeil et al., 2006) can limit our ability to quantify a possible incorporation of bait based on stable isotope analysis. Therefore, the importance of bait for whitetip reef sharks, a resident shark species that is also an important focus of the shark dive at the 301

Biological Conservation 224 (2018) 300–308

K.G. Abrantes et al.

2.3. Data analysis

Shark Reef Marine Reserve (Brunnschweiler et al., 2014), was also quantified. On diving days, the second part of the first dive (~20 min) is dedicated to feeding whitetip reef sharks (average 15 individuals), blacktip reef sharks (Carcharhinus melanopterus; ~15 individuals) and grey reef sharks (Carcharhinus amblyrhynchos; up to ~15 individuals, with numbers depending on season), when these smaller species are fed ~12 kg tuna heads (~4 kg/species) (M. Neumann pers. comm). Unlike bull sharks, whitetip reef sharks typically have limited home ranges and strong site fidelity (Barnett et al., 2012; Speed et al., 2012) and therefore the individuals sampled are likely to have been exposed to shark feeds since regular feeding started in 2000. Therefore, if bait is important for whitetip reef sharks, this should be easily identifiable and quantifiable using stable isotope analysis.

Because urea was not removed, and urea content has been shown to influence elasmobranch δ15N values (Churchill et al., 2015; Hussey et al., 2012; Kim and Koch, 2011), the relationship between C:N ratios (which are related to urea content) and bull shark δ15N values was tested using linear regression analysis. This was not done for whitetip reef sharks as these had a narrow range in C:N ratios, which were close to 3 (2.9–3.3), the value of pure protein, meaning that urea content was unlikely to influence shark δ15N. Classification and regression tree analyses (CART) were run to determine if there was an effect of sex, size (medium, large, very large) and month or year of sampling on bull shark δ13C and δ15N. CARTs were not run for whitetip reef sharks because only females were sampled, and within a short period (May–June 2015), so no temporal differences would be expected. For details on the CART analysis methodology and applicability in ecological studies see De'ath and Fabricius (2000) and Urban (2002). Since bull shark numbers decrease towards the end of the calendar year, when animals move out of the sampling area for reproductive purposes (Brunnschweiler et al., 2014; Brunnschweiler and Baensch, 2011; Brunnschweiler and Barnett, 2013), the relationships between sampling day (day of the year, from 1 to 365) and δ13C/δ15N were also analysed using circular statistics (Oriana 4 software). Firstly, Rao's Spacing Test was used to test for uniformity in the temporal distribution of samples collected, and then circular-linear correlation coefficients (range from 0 to 1) were calculated to test for a correlation between day of the year and bull shark δ13C/δ15N (Mardia and Jupp, 2000). Here, the explanatory variable was day of the year (2014 and 2015 combined), and the response variables were δ13C and δ15N (analysed separately). Linear regression analysis was also used to determine if there is a significant relationship between encounter rates (a proxy for the time spent in the area) and bull shark δ13C/δ15N values, that could be related to differences in reliance on tuna heads provided. Encounter rates were calculated as the proportion of dives each individual was observed at the feeding site by a trained diver over the 2–3.5 year period prior to stable isotope sampling (January 2012 to June 2015), and varies between 0 and 1 (Brunnschweiler and Barnett, 2013). Data for 15 individuals for which there was both stable isotope data and encounter rate data between January 2012 and June 2015 were used. Bayesian mixing models were then used to quantify the contribution of bait and fish prey to bull sharks and to whitetip reef sharks using the package SIMMR (Stable Isotope Mixing Model in R v.3; Parnell et al., 2013). Details of these models can be found in Parnell et al. (2010, 2013). Fish prey were firstly grouped into habitat and trophic categories: demersal low trophic level teleosts (herbivorous, detritivorous and corallivorous species combined a priori, as these groups had similar stable isotope composition), planktivorous teleosts (only considered as potential prey for whitetip reef sharks), macrobenthic carnivores, demersal piscivores and pelagic piscivores (Table A1). These multi-species groups were considered as potential prey in the Bayesian mixing models. For bull sharks, whitetip reef sharks were also considered as potential prey, as bull sharks are known to also prey on elasmobranchs (Cliff and Dudley, 1991; Olin et al., 2013; Trystram et al., 2016). The stable isotope composition of whitetip reef sharks was therefore considered as indicative of the stable isotope composition of local elasmobranch prey. Trophic discrimination factors (TDFs) of +0.8 ± 0.2‰ for δ13C and + 2.4 ± 0.2‰ for δ15N (Δδ13C and Δδ15N respectively) were used, as most appropriate for non-lipid/urea extracted samples of large sharks (Hussey et al., 2010). The TDF values reported by Hussey et al. (2010) were also found to be the most appropriate for bull sharks based on the analysis of the standard ellipse area overlap between sharks and their known prey (identified from stomach contents studies) in South Africa (Olin et al., 2013). However, because of the uncertainty in

2.2. Sample collection and processing Muscle samples from free-swimming bull sharks (March 2014–June 2015) and whitetip reef sharks (May–June 2015) were sampled using a small speargun fitted with a 5 mm Ø biopsy punch, which was fired into the back of the shark, collecting a ~1 cm deep biopsy. Sex and size information was recorded for each individual (size classes: bull sharks medium:- ~1.8–2.4 m; large: ~2.5–3.0 m; very large: > ~3.0 m; whitetip reef sharks – small: ~0.6–0.8 m; medium: ~0.9–1.2 cm; large: > ~1.2 m). These size classes were selected because they encompass the size range observed at the SRMS and because these size classes were easy to distinguish visually. A large proportion (> 150 individuals) of bull sharks that attend this shark dive have been named based on external markings (Brunnschweiler and Baensch, 2011) so that individuals can be identified, reducing the probability that the same shark was sampled twice. Shark tissues were stored frozen until processing. Muscle samples from bait (tissue removed from tuna heads with a scalpel) and from representative species of potential prey groups were also analysed. For potential prey, samples were removed from the flank of fish bought from local markets and roadside stalls in the coastal communities adjacent to the dive site (4–6 km away) in July 2015. Zooplankton and seagrass were also collected from the area to give an indication of the δ13C and δ15N values at the base of the local food web. Zooplankton was collected with a 250 μm plankton net towed behind the boat until sufficient material was collected. Debris was then removed by filtering the collected material through a 500 μm mesh sieve, and the mixture was placed into a 200 ml jar for ~8 h to allow zooplankton guts to clear. The top part of the mixture was then passed through a 250 μm mesh to collect zooplankton (mostly copepods). Seagrass was collected by hand picking green leaves. For seagrasses, each replicate was composed by 6–10 blades from different individuals combined. Prey, zooplankton and seagrass samples were dried (60 °C, 48 h) in the field on the day of collection. In the laboratory, shark muscle tissue was removed from the adjacent dermis with a scalpel and dried to a constant weight at 60 °C. All dried samples were then homogenised into a fine powder with a mortar and pestle, weighed and encapsulated into tin capsules. Furthermore, to remove carbonates for δ13C analysis (Jaschinski et al., 2008), a second set of zooplankton samples were weighed in silver capsules and acidified by adding 5% HCl drop-by-drop onto the weighed samples until no further bubbling was evident. Samples were analysed at the Davis UC Davis Stable Isotope Facility (USA) with a PDZ Europa ANCA-GSL elemental analyser interfaced to a PDZ Europa 20–20 isotope ratio mass spectrometer. Results are expressed as per mil (‰) deviations from standards, as defined by: δ13C, δ15N = [(Rsample / Rreference) − 1] × 103, where R = 13C/12C for carbon and 15N/14N for nitrogen, and had a precision of ± 0.1‰ ( ± SD) for both δ13C and δ15N, calculated from standards.

302

Biological Conservation 224 (2018) 300–308

K.G. Abrantes et al.

elasmobranch Δδ13C and Δδ15N (e.g. Caut et al., 2013; Olin et al., 2013), a sensitivity analysis was done in which Bayesian mixing models were also run using TDFs and respective SDs from 12 other studies, seven of which based on elasmobranchs, to determine if this would lead to different results regarding the importance of bait provided to bull and whitetip reef shark diets. Moreover, because urea content affects elasmobranch δ15N values, models using shark δ15N values corrected by +0.6‰ (corresponding to the mean shift in δ15N with urea/lipid removal identified for bull sharks (Hussey et al., 2012)), and TDFs of 1.3 ± 0.5‰ for Δδ13C and 2.8 ± 0.3‰ for Δδ15N (i.e. the estimated Δδ13C/Δδ15N values based on lipid/urea extracted shark muscle tissue and non-treated prey samples (Hussey et al., 2012)) were also computed. When potential prey groups had similar stable isotope composition and strongly negatively correlated source proportions, the contribution of those prey groups was combined a posteriori to lead to more precise results regarding the importance of the various prey types (Phillips et al., 2014; Phillips et al., 2005). For bull sharks, this means that the contributions of macrobenthic carnivores and demersal piscivores were combined into a “demersal carnivore” category; and for whitetip reef sharks, low trophic level teleosts (herbivorous, detritivorous and corallivorous species), macrobenthic carnivores and demersal piscivores were combined into a “demersal teleost” category. Because whitetip reef shark formed two groups well separated in both δ13C (−14.2 ± 0.4‰ vs. −11.3 ± 0.1‰ ( ± SD); with individual sharks separated by > 2‰) and δ15N (12.1 ± 0.3‰ vs. 10.3 ± 0.4‰; with individuals separated by > 2‰), mixing models were run for each group separately. For bull sharks, separate models were also run for the 10 individuals with stable isotope composition closest to bait.

Fig. 2. Stable isotope composition of individual bull shark (grey circles) and whitetip reef shark (black squares), along with possible prey (white symbols; mean ± SD) and primary producers (black circles; mean ± SD). Producer stable isotope values are included to provide information on the baseline δ13C and δ15N of the local food web. For producers, numbers in brackets correspond to the number of samples analysed, with the exception of coral, for which values are based on 123 samples from eight sites taken from a review by Heikoop et al. (2000). For prey, sample sizes can be seen in Table A1. Prey: ○ - herbivorous and corallivorous fish species; ⊙ - detritivorous fish; △ - planktivorous fish; ▽ - carnivorous fish (macrobenthic carnivores); □ - demersal piscivores; ◊ - pelagic piscivores.

between −14.7‰ and −14.2‰, and δ15N between 11.5‰ and 12.3‰, while two individuals (a juvenile and a large individual) had higher δ13C and slightly lower δ15N (Table 1, Fig. 2). Bait had very high δ15N values, much higher than all potential prey and > 2‰ higher than all sharks analysed (Fig. 2), suggesting that bait is not important for the diet of any individual. Presence data at each dive (January 2012–July 2015) was available for 15 bull shark individuals for which stable isotope samples were also collected. Encounter rates varied between 0.01 and 0.30 (n = 527 dives). There were no significant relationships between encounter rates and δ13C or between encounter rates and δ15N (p > 0.05) (Fig. 4). Bayesian mixing model results suggest bull sharks relied on a range of teleost prey, including demersal and pelagic species (Fig. 5a). When the model was run based on all bull shark individuals, there was no evidence of contribution of bait as the 95% credibility interval (CI) of bait contribution was narrow, had a low maximum and included 0% (0–19%, Fig. 5a) (Benstead et al., 2006). Results on bait contribution were similar when models were run using TDF values from all various available studies, with lower limits of the 95% CIs ≤ 1%, median potential contribution < 10% and low maximum values (Table A2). Even when only the group of 10 bull shark individuals with stable isotope composition closest to bait were considered, mixing model results did not provide evidence of bait contribution to bull shark diets, as the lower limit of the 95% CI was still 0% (95% CI: 0–15%), and results suggest that those sharks fed mostly on local pelagic piscivores such as mackerels (59–85%; Fig. 5a, Table A2). With one exception, models run using TDFs from other studies led to similar results (Table A2). This included the model run based on bull shark δ15N values corrected for the effect of lipid/urea extraction (+0.6‰), and Δδ13C/Δδ15N values based on lipid/urea extracted shark muscle tissue and non-treated prey samples (Hussey et al., 2012). Only in one case (Galván et al., 2016) did the mixing model results lead to a significant importance of bait (9–31%; Table A2). There was also no evidence that whitetip reef sharks (a proxy for elasmobranch prey) were important prey for bull shark diet (Fig. 5a; Table A2). Whitetip reef sharks also relied mostly on local prey, with mixing model results indicating that the whitetip reef shark group with lower

3. Results A total of 58 bull shark and nine whitetip reef shark samples were analysed (Table 1). Bull sharks had a wide range in both δ13C and δ15N values, and individuals were well spread within those ranges (Figs. 2, 3). Bull shark's C:N ratios varied between 2.7 and 3.3 (mean ± SD: 2.9 ± 0.2), and whitetip reef shark's between 2.9 and 3.3 (3.1 ± 0.1) (Table 1). There was no relationship between C:N ratios and δ15N for bull sharks (linear regression analysis, p > 0.05; Fig. A1), likely because of the relatively low range in C:N values (0.7). CART analyses did not find a significant effect of sex, size, month, or year of collection on bull shark δ13C or δ15N. Moreover, there was no evidence of directionality of stable isotope values (Rao's Spacing Test U = 66.567, p > 0.99 for both δ13C and δ15N), meaning that collected δ13C/δ15N values were distributed uniformly throughout the calendar year. Circular-linear correlation did not find a relationship between date of collection and δ13C (r = 0.206, p = 0.100) or δ15N (r = 0.063, p = 0.807) (Fig. 3). Whitetip reef sharks were less variable in δ13C and δ15N, and seemed to form two groups: most (7 out of 9) individuals had δ13C Table 1 Details of bull sharks and whitetip reef sharks sampled, including sex, size class, stable isotope values (in ‰), C:N ratios (mean ± SD) and sample size. δ13C

δ15N

C:N ratios

n

Bull sharks Females Medium Large Very large Males Medium Large

−13.1 ± 1.1 −13.6 ± 1.4 −14.3 −12.2 ± 0.8 −15.1

11.7 ± 0.6 12.0 ± 0.9 12.5 11.3 ± 0.3 13.0

2.8 ± 0.2 2.9 ± 0.2 3.1 2.7 ± 0.0 3.0

15 37 1 4 1

Whitetip reef sharks Females Small (juvenile) Medium Large

−12.5 ± 1.5 −13.9 −13.9 ± 1.3

11.0 ± 1.5 11.9 11.9 ± 0.7

3.2 ± 0.1 3.2 3.1 ± 0.2

2 1 6

Species

Size

303

Biological Conservation 224 (2018) 300–308

K.G. Abrantes et al.

Fig. 4. Encounter rates (for the period between January 2012 and June 2015) and δ13C (top) and δ15N (bottom) values of 15 sharks. No significant relationship was present (linear regression analysis; p > 0.05).

13 models were the lower bounds of the 95% CIs of bait contribution higher than 5% (Table A3), meaning that it is possible that the importance of bait is in reality lower than that estimated using the TDFs of Hussey et al. (2010). 4. Discussion There was no evidence of bait incorporation by bull sharks collected at the Shark Reef Marine Reserve, with results showing that, as in southern Mozambique (Daly et al., 2013) and Reunion Islands (Trystram et al., 2016), bull sharks rely mostly on demersal and pelagic teleosts. The variability in bull shark stable isotope composition also indicates that this species has a broad trophic niche, a result previously reported in stomach content- (Cliff and Dudley, 1991; Olin et al., 2013; Trystram et al., 2016) and stable isotope-based studies (Daly et al., 2013; Matich et al., 2011; Trystram et al., 2016). Indeed, bull sharks sampled in Florida (USA) (Matich et al., 2011), Reunion Island (Trystram et al., 2016) and Mozambique (Daly et al., 2013) also showed wide ranges in δ13C, similar in magnitude to that found in the present study, and in two cases (Florida and Mozambique) the ranges in δ15N were also similar. This variability is explained by the fact that, although bull sharks are considered to be generalists at the population level, populations are composed by several specialist individuals, i.e. different bull shark individuals specialise on different prey (Matich et al., 2011; Trystram et al., 2016). Despite this variability and the number of individuals sampled, there was no evidence that food rewards are a significant component of bull shark diets. The lack of bait incorporation was somewhat surprising, particularly given the frequency of shark feeds (5 days/week, with 2 dives/ day) and bull shark encounter and feeding rates at the Shark Reef Marine Reserve (Brunnschweiler and Barnett, 2013; Brunnschweiler et al., 2017). Indeed, Brunnschweiler et al. (2017) estimated that one tuna head could meet the energy requirements of a 200 kg (corresponding to ~2.8 m (Branstetter and Stiles, 1987)) bull shark for 3.1 days (2.3 heads/week). In that study, it was estimated that an average ~2.6 tuna heads/week were consumed by each of the 10 individuals that most regularly fed on bait (focal individuals), meaning

Fig. 3. Circular plots of the distribution of bull shark δ13C (top) and δ15N (bottom) values sampled throughout the calendar year, showing the lack of a relationship between date of collection and δ13C or δ15N. White circles – males; black circles – females. The circular means (line to the center of the plot) and circular standard deviation (just outside the circular-linear plot) are also presented.

δ13C values relied on a combination of demersal (35–51%) and pelagic teleost prey (31–53%), while the group with higher δ13C fed primarily on demersal teleost prey (61–96%; Fig. 5b). However, unlike with bull sharks, bait also had some contribution to the whitetip reef shark group with lower δ13C values (8–22%; Fig. 5b). This group was composed by 7 out of the 9 individuals, meaning that bait could contribute to longterm diet for ~78% of the population. Note that the group with higher δ13C was composed by only two individuals (one small and one large), so it could be argued that it is not representative of a different group. However, this corresponds to 22% of all individuals analysed, so it is possible that it represents a group with distinct dietary preferences. Models using TDFs from other studies led to a similar general pattern of relative importance of natural prey (Table A3). However, for the group of whitetip reef sharks with lower δ13C, only for three out of the other 304

Biological Conservation 224 (2018) 300–308

K.G. Abrantes et al.

Fig. 5. Bayesian mixing model results on the contribution of the different prey to (a) bull shark and (b) whitetip reef shark diets. For bull sharks, white boxes are results for all individuals sampled, and grey boxes are results for the 10 individuals with stable isotope composition closest to bait. For whitetip reef sharks, white boxes are the results for the group with lower δ13C, and grey boxes are the results for the shark group with higher δ13C values. Boxes are the 50% credibility intervals (CI), lines within the boxes are the medians and whiskers are the 95% CI.

and for longer periods of time (weeks to months) at the end of the calendar year before returning to in the beginning of the following year (Brunnschweiler and Barnett, 2013). Therefore, the stable isotope composition of their muscle tissue reflects the averaged diet from different areas. Since it takes over a year for the stable isotope composition of large elasmobranchs to represent a new diet (Logan and Lutcavage, 2010; MacNeil et al., 2006; Vander Zanden et al., 2015), it is possible that the time spent at the Shark Reef Marine Reserve was not long enough to allow for bull shark muscle to reflect the incorporation of bait. However, some bull shark individuals that can be identified based on external markings have been regular visitors to the Shark Reef Marine Reserve since 2003 and regularly feed when visiting the provisioning site, and electronic tracking data shows that many individuals remain in the area for extended periods of time (Brunnschweiler and Barnett, 2013). Although muscle turnover rates of elasmobranchs are slow (MacNeil et al., 2006; Logan and Lutcavage, 2010), if tuna was important we would expect higher shark δ15N, particularly since there was a large difference in δ15N between bait and natural prey. This was however not observed, and all shark individuals had much lower (> 2‰ difference) δ15N than bait (see Fig. 2). Differences in stable isotope composition that result from differences in diet or habitat use can be detected in muscle of large elasmobranchs, even if diet/habitat differs for only part of the year. For example, Abrantes and Barnett (2011) found differences in both δ13C and δ15N in large (up to 3 m) sevengill sharks (Notorynchus cepedianus) in Tasmania. Those differences were related to intraspecific differences in movement patterns, and were detectable despite the regular movement of sevengill sharks in and out of the study area, and despite that the stable isotope composition of sharks was not in equilibrium with that of the local coastal prey (Abrantes and Barnett, 2011). The use of tissues with faster turnover rates such as blood or plasma would likely give more precise short-term dietary information (e.g. Kinney et al., 2011; Matich et al., 2010; Matich et al., 2011). However, such tissue is not available for the sharks sampled in the present study since it requires the capture and handling of animals, which is not allowed in the Shark Reef Marine Reserve. It is however important to note that although no significant incorporation of bait was found for muscle tissue, bull sharks could be drawing on (reliable) provisioning to fill gaps or boost resources at key stages of their life cycle such as during egg production (Hobson, 2006; Tanaka et al., 2016). If this is the case, the importance of provisioning could be minor in muscle tissue, but could have an important effect on reproduction. Although there was no evidence of bait incorporation by bull sharks, results suggest some (8–22%) importance of bait for some whitetip reef shark individuals. This could be because, unlike bull sharks, whitetip

that it would be energetically possible for those bull shark individuals to rely significantly on bait provided. However, it is important to note that data for the feeding rate part of the Brunnschweiler et al. (2017) study was collected in 2008 (31 dives)/2009 (five dives), when only between 2 and an estimated 20 sharks were present on each dive (Brunnschweiler unpubl. data). From those, 10 individuals (focal individuals; corresponding to a high proportion of the total sharks attending the dive) were present on ≥10 of the 36 sampling dives and had feeding rates that could lead to a major reliance on food provided (Brunnschweiler et al., 2017). Since 2008, the number of sharks attending the shark dive and shark feeds increased (M. Neumann pers. comm, Brunnschweiler unpubl. data; Fig. 1), meaning that the food provided is now shared among a higher number of bull shark individuals (note that feeding duration and the amount of food provided remained constant). Indeed, for the same period of the year when the feeding rate data of Brunnschweiler et al. (2017) was collected (February–March), much higher bull shark numbers were sighted per dive in 2015 (43.3 ± 10.1 individuals) than in 2008 (10.6 ± 5.4 individuals), a pattern that was present for all months of the year (Fig. 1). While a maximum of 30 sharks were sighted per dive in 2008, in 2015 it was common to have 50–75 sharks attend a dive (M. Neumann pers. comm, Brunnschweiler unpubl. data). Currently, > 150 of the bull sharks that attend the Shark Reef Marine Reserve dive site have been identified and named, meaning that the total number of individuals, including un-named sharks, is likely higher. Therefore, different results could have been found if sampling was conducted in previous years. Indeed, stable isotope-based information on the importance of food subsidies will always be site- and time-specific, and results from one study cannot be considered as representative of the situation in other site/time settings. Note also that, as in other shark feed operations (e.g. Fitzpatrick et al., 2011, Maljković and Côté, 2011; and also in other predator provisioning sites, e.g. bears (Ursus spp.) (Penteriani et al., 2017)), (1) encounter rates vary widely among individuals, (2) different individuals consume different amounts of food, and (3) not all individuals feed when present (Brunnschweiler and Barnett, 2013, Brunnschweiler et al., 2017). The focal individuals could be bolder and dominant, so their feeding patterns might not be representative of the overall Shark Reef Marine Reserve bull shark population (Brunnschweiler et al., 2017). It could be argued that the slow turnover rates of shark muscle (Logan and Lutcavage, 2010; MacNeil et al., 2006) would make it unlikely that any incorporation of bait could be detected, particularly given the movement patterns of bull sharks in the area. Bull sharks intermittently move out of the area for a few days throughout the year, 305

Biological Conservation 224 (2018) 300–308

K.G. Abrantes et al.

wildlife tourism (e.g. Penteriani et al., 2017).

reef sharks have relatively small home ranges and are often resident in the reefs where they occur (Barnett et al., 2012; Speed et al., 2012). However, even for whitetip reef sharks, the importance of bait was low, despite that feeding occurs 5 times/week. Note that, in each dive, an average of 15 whitetip reef shark individuals are fed ~4 kg tuna heads/ day (12 kg shared with ~15 black tip reef sharks and up to ~15 grey reef sharks), which corresponds to an average 266 g/dive day for each whitetip reef shark individual. Overall, results suggest that at current levels, the impacts of provisioning tourism on shark diets at the Shark Reef Marine Reserve are limited. To our knowledge, only two other studies used biochemical tracers to identify the importance of food provisioning to elasmobranchs at feeding sites. In one study, Maljković and Côté (2011) investigated the importance of bait to Caribbean reef sharks Carcharhinus perezi in the Bahamas, and found that only a small proportion of individuals regularly consumed bait, and those had δ15N 0.5‰ higher than conspecifics that did not regularly consume bait or that did not have access to bait (control group). Higher δ15N was explained by a higher bait consumption, as bait was mostly groupers i.e. high trophic level species. However, groupers are also part of the natural diet of Caribbean reef sharks (Maljković and Côté, 2011), meaning that the stable isotope composition of bait (which was not measured) was not likely to be very different to that of natural prey, limiting the suitability of stable isotope analysis to identify incorporation of bait in that particular case. Indeed, δ13C of fed and non-fed Caribbean reef sharks were similar. The difference in δ15N between groups could have been caused by intraspecific differences in behaviour, particularly as the fed group was composed by individuals with larger mean size than the remaining of the population (Maljković and Côté, 2011). Size-related dominant behaviour of larger individuals at feeding sites has been reported for other elasmobranch feeding sites (e.g. Newsome et al., 2004; Semeniuk and Rothley, 2008). Larger sharks are more likely to be dominant and more aggressive and therefore to dominate the shark feed and to be “bolder”, stronger, and more able to prey on larger prey, which would have higher δ15N due to their likely higher trophic level. Moreover, each of Maljković and Côté's (2011) fed sharks consumed only ~5% (average) of the bait provided, and residency time in the feeding area was only ~50%, meaning that bait consumed was far from sufficient to fulfil their energetic requirements (Maljković and Côté, 2011), and could not lead to the measured δ15N shift of +0.5‰. Therefore, it is likely that, as for bull sharks in the present study, bait provided was also not important for the Caribbean reef sharks at the Bahamas provisioning site. In another study, Semeniuk et al. (2007) used fatty acid analysis to identify the importance of artificially provided squid to southern stingrays (Hypanus americanus) at the Grand Cayman Island (Caribbean), and found that fed stingrays had fatty acid profiles that indicate a significant, almost exclusive, incorporation of squid. Feeding and consequent increased stingray densities at that site also led to a range of negative behavioural and physiological effects (Corcoran et al., 2013; Semeniuk et al., 2009; Semeniuk and Rothley, 2008; Semeniuk et al., 2007). The differences between the southern stingray and the Caribbean reef shark studies are striking, but not unexpected. Stingray City receives over a million tourists per year and is one of the world's most intensive and most popular provisioning sites, representing an extreme situation due to the intensive year-round provisioning of 90–100 stingrays (Vaudo et al., 2017) by tourists not only diving/snorkeling but also waddling in the water. In the Caribbean reef shark case, however, despite that sharks are fed daily (Maljković and Côté, 2011), much less food is provided than at Stingray City (Corcoran et al., 2013) and shark densities and residency at the feeding site are also much lower (Corcoran et al., 2013; Maljković and Côté, 2011). Results from those two studies and from our present study therefore suggest that the impacts of provisioning activities depend largely on the species and the relative amount of food provided, both significant considerations in managing and/or developing best practices for any species involved in

4.1. Caveats to take into account in stable isotope results interpretation A number of factors should be considered when interpreting the stable isotope data from the present study. Firstly, it could be argued that the relatively small number of local prey analysed (27 individuals; Table A1) could be insufficient to represent well the local food web and available prey. However, δ13C and δ15N of the different prey species agreed well with their trophic ecology and with the stable isotope values of local plankton, seagrass and coral. This suggests that, despite the limited number of individual samples analysed, prey stable isotope values represent well the local food web and the local prey available to bull and whitetip reef sharks. Another aspect that could affect stable isotope results interpretation is related to the bull sharks' movement ecology. Indeed, because bull sharks regularly move out of the study area (Brunnschweiler et al., 2014; Brunnschweiler and Barnett, 2013), it is possible that bull sharks also feed on natural prey with stable isotope composition different to that of prey sampled in the present study and that, if those prey were considered in the mixing models, mixing model results would suggest a significant incorporation of bait. However, information on bull shark habitat use and on the δ15N isoscape of the region suggest that this is not likely (see Appendix 1 for details). The uncertainty in elasmobranch Δδ15N (e.g. Caut et al., 2013; Olin et al., 2013) could also affect mixing model results (Bond and Diamond, 2011). This uncertainty was addressed by running Bayesian mixing models based on Δδ15N (and Δδ13C) values from 13 different studies, in a sensitivity analysis. The validity of our conclusions was confirmed as, in general, the different models led to similar results of bait contribution (Tables A2, A3). Finally, despite recommended (Carlisle et al., 2017; Hussey et al., 2012; Kim and Koch, 2011; Shipley et al., 2017), urea was not removed before δ15N analysis. However, in the present study, mixing models based on Δδ15N values derived from three studies on elasmobranchs where urea was not extracted were also computed, and those led to similar results of limited bait incorporation (see Tables A2, A3). These issues and respective implications are discussed in depth in Appendix 1. 4.2. Conclusions Predator-focused wildlife tourism has the potential to contribute significantly to the conservation of aquatic and terrestrial predators (Macdonald et al., 2017). However, information on the effects of tourism activities on the long-term behaviour, health and fitness of target animals is required to 1) assist in best practice/least impact management, and 2) to adequately assess conservation benefits. For example, if stakeholders understand the effects of wildlife provisioning on the target animals, operational methods can be modified if effects with long-term behaviour, health or fitness implications are apparent (Penteriani et al., 2017; Trave et al., 2017). Results from the present study, particularly when combined with information from previous studies at this feeding site (Brunnschweiler et al., 2010, 2014; Brunnschweiler and Baensch, 2011; Brunnschweiler and Barnett, 2013; Brunnschweiler et al., 2017), suggest that current levels of provisioning lead to no detrimental long-term impacts on the behaviour or diet (and probably health) of sharks at the Shark Reef Marine Reserve site. Since the threshold number of bull sharks relative to quantity of bait provided where a contribution of bait in the diet is expected is not known, we suggest that (1) the amount of bait provided and the frequency of feeds should not be increased at this site, and (2) if the number of bull sharks attending the dive site decreases, the amount of bait provided should be reduced and a follow-up study should be conducted to assess potential changes in bait importance to bull shark diets. The suite of studies available for this site is likely the most comprehensive body of work to date comparing the impacts of tourism on a predator species' behaviour and fitness with conservation and economic 306

Biological Conservation 224 (2018) 300–308

K.G. Abrantes et al.

benefits. The various studies address a range of potential issues (e.g. residency, habitat use, consumption of bait, energetic value of bait and diet) and provide a good example of how site-specific information, obtained from multiple methods, can contribute for an effective, evidence-based, management of the predator tourism industry. Similar multi-methods approaches can be applied to other aquatic and terrestrial wildlife tourism species/operations, to move beyond our current understanding of tourism-driven behavioural and ecological changes (Newsome et al., 2015; Trave et al., 2017), and provide a more holistic understanding of the effects of provisioning on the health and fitness of the target species. For example, the two recent reviews addressing provisioning predators in terrestrial systems mainly discuss behavioural and ecological changes, and only touch on physiological changes (Newsome et al., 2015; Penteriani et al., 2017), and there is a lack of studies that quantify the energetic value of food rewards and the relative assimilation of human-provided food by predators. The use of methods such as stable isotopes and analysis of the energetic value of food rewards could contribute to an improved understanding of the effects of provisioning on the long-term health and fitness at many of the study locations reviewed. Similarly, for marine systems, several of the studies reviewed by Trave et al. (2017) could benefit from including or following-up with studies incorporating information on the energetic value and assimilation of supplemental feeding. Note also that a recent review of 115 studies on the effects of provisioning on wildlife heath found that, for tourism-related feeding operations, there were negative effects and likely altered health in 85% of the cases (Murray et al., 2016). Most impacts were however related to increased pathogen transmission and prevalence. It is therefore clear that studies addressing the assimilation of food provisions in relation to the natural diet are needed, and will greatly contribute to the field of wildlife provisioning research. The present study shows that stable isotope analysis can be particularly useful for this purpose. Indeed, since the stable isotope composition of animal tissues provides quantitative time-integrated information on assimilated diets (and not just the ingested material), this time- and cost-effective method should be used (1) to identify and quantify the incorporation of artificial food by the target species and (2) to monitor impacts of feeding operations through time. Moreover, because different animal tissues have different turnover rates, reflecting diets at different time scales, the use of multiple tissues can be used to identify temporal/ontogenetic variations in diets (e.g. Matich et al., 2010, 2011) and in importance of supplementary foods. We therefore suggest that this method should be more widely used in wildlife provisioning studies.

Brunnschweiler, J., Queiroz, N., Sims, D., 2010. Oceans apart? Short-term movements and behaviour of adult bull sharks Carcharhinus leucas in Atlantic and Pacific Oceans determined from pop-off satellite archival tagging. J. Fish Biol. 77, 1343–1358. Brunnschweiler, J.M., Abrantes, K.G., Barnett, A., 2014. Long-term changes in species composition and relative abundances of sharks at a provisioning site. PLoS One 9, e86682. Brunnschweiler, J.M., Barnett, A., Payne, N.L., 2017. Hand feeding can periodically fuel a major portion of bull shark energy requirements at a provisioning site in Fiji. Anim. Conserv. http://dx.doi.org/10.1111/acv.12370. Burgin, S., Hardiman, N., 2015. Effects of non-consumptive wildlife-oriented tourism on marine species and prospects for their sustainable management. J. Environ. Manag. 151, 210–220. Carlisle, A.B., Litvin, S.Y., Madigan, D.J., Lyons, K., Bigman, J.S., Ibarra, M., Bizzarro, J.J., 2017. Interactive effects of urea and lipid content confound stable isotope analysis in elasmobranch fishes. Can. J. Fish. Aquat. Sci. 1–10. Carlson, J., Ribera, M., Conrath, C., Heupel, M., Burgess, G., 2010. Habitat use and movement patterns of bull sharks Carcharhinus leucas determined using pop-up satellite archival tags. J. Fish Biol. 77, 661–675. Caut, S., Jowers, M.J., Michel, L., Lepoint, G., Fisk, A., 2013. Diet-and tissue-specific incorporation of isotopes in the shark Scyliorhinus stellaris, a North Sea mesopredator. Mar. Ecol. Prog. Ser. 492, 185–198. Churchill, D.A., Heithaus, M.R., Grubbs, R.D., 2015. Effects of lipid and urea extraction on δ 15 N values of deep-sea sharks and hagfish: can mathematical correction factors be generated? Deep-Sea Res. II Top. Stud. Oceanogr. 115, 103–108. Cisneros-Montemayor, A.M., Barnes-Mauthe, M., Al-Abdulrazzak, D., Navarro-Holm, E., Sumaila, U.R., 2013. Global economic value of shark ecotourism: implications for conservation. Oryx 47, 381–388. Clarke, C., Lea, J., Ormond, R., 2013. Changing relative abundance and behaviour of silky and grey reef sharks baited over 12 years on a Red Sea reef. Mar. Freshw. Res. 64, 909–919. Cliff, G., Dudley, S., 1991. Sharks caught in the protective gill nets off Natal, South Africa. 4. The bull shark Carcharhinus leucas Valenciennes. S. Afr. J. Mar. Sci. 10, 253–270. Clua, E., Buray, N., Legendre, P., Mourier, J., Planes, S., 2010. Behavioural response of sicklefin lemon sharks Negaprion acutidens to underwater feeding for ecotourism purposes. Mar. Ecol. Prog. Ser. 414, 257–266. Corcoran, M.J., Wetherbee, B.M., Shivji, M.S., Potenski, M.D., Chapman, D.D., Harvey, G.M., 2013. Supplemental feeding for ecotourism reverses diel activity and alters movement patterns and spatial distribution of the southern stingray, Dasyatis americana. PLoS One 8, e59235. Daly, R., Froneman, P.W., Smale, M.J., 2013. Comparative feeding ecology of bull sharks (Carcharhinus leucas) in the coastal waters of the southwest Indian Ocean inferred from stable isotope analysis. PLoS One 8, e78229. Daly, R., Smale, M.J., Cowley, P.D., Froneman, P.W., 2014. Residency patterns and migration dynamics of adult bull sharks (Carcharhinus leucas) on the east coast of southern Africa. PLoS One 9, e109357. De'ath, G., Fabricius, K.E., 2000. Classification and regression trees: a powerful yet simple technique for ecological data analysis. Ecology 81, 3178–3192. DeNiro, M.J., Epstein, S., 1978. Influence of diet in the distribution of carbon isotopes in animals. Geochim. Cosmochim. Acta 42, 495–506. Espinoza, M., Heupel, M.R., Tobin, A.J., Simpfendorfer, C.A., 2016. Evidence of partial migration in a large coastal predator: opportunistic foraging and reproduction as key drivers? PLoS One 11, e0147608. Estes, J.A., Terborgh, J., Brashares, J.S., Power, M.E., Berger, J., Bond, W.J., Carpenter, S.R., Essington, T.E., Holt, R.D., Jackson, J.B., 2011. Trophic downgrading of planet Earth. Science 333, 301–306. Fitzpatrick, R., Abrantes, K.G., Seymour, J., Barnett, A., 2011. Variation in depth of whitetip reef sharks: does provisioning ecotourism change their behaviour? Coral Reefs 30, 569–577. Gallagher, A.J., Hammerschlag, N., 2011. Global shark currency: the distribution, frequency, and economic value of shark ecotourism. Curr. Issue Tour. 14, 797–812. Gallagher, A.J., Vianna, G.M., Papastamatiou, Y.P., Macdonald, C., Guttridge, T.L., Hammerschlag, N., 2015. Biological effects, conservation potential, and research priorities of shark diving tourism. Biol. Conserv. 184, 365–379. Galván, D., Jañez, J., Irigoyen, A., 2016. Estimating tissue-specific discrimination factors and turnover rates of stable isotopes of nitrogen and carbon in the smallnose fanskate Sympterygia bonapartii (Rajidae). J. Fish Biol. 89, 1258–1270. Heikoop, J., Dunn, J., Risk, M., Tomascik, T., Schwarcz, H., Sandeman, I., Sammarco, P., 2000. δ15N and δ13C of coral tissue show significant inter-reef variation. Coral Reefs 19, 189–193. Hobson, K.A., 2006. Using stable isotopes to quantitatively track endogenous and exogenous nutrient allocations to eggs of birds that travel to breed. ARDEAWAGENINGEN 94, 359. Hussey, N.E., Brush, J., McCarthy, I., Fisk, A.T., 2010. δ15N and δ13C diet-tissue discrimination factors for large sharks under semi-controlled conditions. Comp. Biochem. Physiol. A 155, 445–453. Hussey, N.E., Olin, J.A., Kinney, M.J., McMeans, B.C., Fisk, A.T., 2012. Lipid extraction effects on stable isotope values (δ13C and δ15N) of elasmobranch muscle tissue. J. Exp. Mar. Biol. Ecol. 434–435, 7–15. Jaschinski, S., Hansen, T., Sommer, U., 2008. Effects of acidification in multiple stable isotope analyses. Limnol. Oceanogr. Methods 6, 12–15. Kim, S.L., Koch, P.L., 2011. Methods to collect, preserve, and prepare elasmobranch tissues for stable isotope analysis. Environ. Biol. Fish. http://dx.doi.org/10.1007/ s10641-011-9860-9. Kinney, M.J., Hussey, N.E., Fisk, A.T., Tobin, A.J., Simpfendorfer, C.A., 2011. Communal or competitive? Stable isotope analysis provides evidence of resource partitioning within a communal shark nursery. Mar. Ecol. Prog. Ser. 439, 263–276.

References Abrantes, K.G., Barnett, A., 2011. Intrapopulation variations in diet and habitat use in a marine apex predator, the broadnose sevengill shark Notorynchus cepedianus. Mar. Ecol. Prog. Ser. 442, 133–148. Barnett, A., Abrantes, K.G., Seymour, J., Fitzpatrick, R., 2012. Residency and spatial use by reef sharks of an isolated seamount and its implications for conservation. PLoS One 7, e36574. Barnett, A., Payne, N.L., Semmens, J.M., Fitzpatrick, R., 2016. Ecotourism increases the field metabolic rate of whitetip reef sharks. Biol. Conserv. 199, 132–136. Benstead, J.P., March, J.G., Fry, B., Ewel, K.C., Pringle, C.M., 2006. Testing IsoSource: stable isotope analysis of a tropical fishery with diverse organic matter sources. Ecology 87, 326–333. Beschta, R.L., Ripple, W.J., 2009. Large predators and trophic cascades in terrestrial ecosystems of the western United States. Biol. Conserv. 142, 2401–2414. Bond, A.L., Diamond, A.W., 2011. Recent Bayesian stable-isotope mixing models are highly sensitive to variation in discrimination factors. Ecol. Appl. 21, 1017–1023. Branstetter, S., Stiles, R., 1987. Age and growth estimates of the bull shark, Carcharhinus leucas, from the northern Gulf of Mexico. Environ. Biol. Fish 20, 169–181. Brena, P.F., Mourier, J., Planes, S., Clua, E., 2015. Shark and ray provisioning: functional insights into behavioral, ecological and physiological responses across multiple scales. Mar. Ecol. Prog. Ser. 538, 273–283. Brunnschweiler, J.M., Baensch, H., 2011. Seasonal and long-term changes in relative abundance of bull sharks from a tourist shark feeding site in Fiji. PLoS One 6, e16597. Brunnschweiler, J.M., Barnett, A., 2013. Opportunistic visitors: long-term behavioural response of bull sharks to food provisioning in Fiji. PLoS One 8, e58522.

307

Biological Conservation 224 (2018) 300–308

K.G. Abrantes et al. Logan, J.M., Lutcavage, M.E., 2010. Stable isotope dynamics in elasmobranch fishes. Hydrobiologia 644, 231–244. Macdonald, C., Gallagher, A.J., Barnett, A., Brunnschweiler, J., Shiffman, D.S., Hammerschlag, N., 2017. Conservation potential of apex predator tourism. Biol. Conserv. 215, 132–141. MacNeil, M.A., Drouillard, K.G., Fisk, A.T., 2006. Variable uptake and elimination of stable nitrogen isotopes between tissues in fish. Can. J. Fish. Aquat. Sci. 63, 345–353. Maljković, A., Côté, I.M., 2011. Effects of tourism-related provisioning on the trophic signatures and movement patterns of an apex predator, the Caribbean reef shark. Biol. Conserv. 144, 859–865. Mardia, K.V., Jupp, P.E., 2000. Statistics of Directional Data, 2nd edition. John Wiley & Sons, Chicester (429pp). Matich, P., Heithaus, M.R., Layman, C.A., 2010. Size-based variation in intertissue comparisons of stable carbon and nitrogen isotopic signatures of bull sharks (Carcharhinus leucas) and tiger sharks (Galeocerdo cuvier). Can. J. Fish. Aquat. Sci. 67, 877–885. Matich, P., Heithaus, M.R., Layman, C.A., 2011. Contrasting patterns of individual specialization and trophic coupling in two marine apex predators. J. Anim. Ecol. 80, 294–305. McCutchan, J.H., Lewis Jr., W.M., Kendall, C., McGrath, C.C., 2003. Variation in trophic shift for stable isotope ratios of carbon, nitrogen and sulfur. Oikos 102, 378–390. Meyer, C.G., Dale, J.J., Papastamatiou, Y.P., Whitney, N.M., Holland, K.N., 2009. Seasonal cycles and long-term trends in abundance and species composition of sharks associated with cage diving ecotourism activities in Hawaii. Environ. Conserv. 36, 104–111. Minagawa, M., Wada, E., 1984. Stepwise enrichment of 15N along food chains: further evidence and the relation between δ15N and animal age. Geochim. Cosmochim. Acta 48, 1135–1140. Murray, M.H., Becker, D.J., Hall, R.J., Hernandez, S.M., 2016. Wildlife health and supplemental feeding: a review and management recommendations. Biol. Conserv. 204, 163–174. Newsome, D., Lewis, A., Moncrieff, D., 2004. Impacts and risks associated with developing, but unsupervised, stingray tourism at Hamelin Bay, Western Australia. Int. J. Tour. Res. 6, 305. Newsome, T.M., Dellinger, J.A., Pavey, C.R., Ripple, W.J., Shores, C.R., Wirsing, A.J., Dickman, C.R., 2015. The ecological effects of providing resource subsidies to predators. Glob. Ecol. Biogeogr. 24, 1–11. Olin, J.A., Hussey, N.E., Grgicak-Mannion, A., Fritts, M.W., Wintner, S.P., Fisk, A.T., 2013. Variable δ15N diet-tissue discrimination factors among sharks: implications for trophic position, diet and food web models. PLoS One 8, e77567. Olson, R.J., Popp, B.N., Graham, B.S., López-Ibarra, G.A., Galván-Magaña, F., LennertCody, C.E., Bocanegra-Castillo, N., Wallsgrove, N.J., Gier, E., Alatorre-Ramírez, V., Ballance, L.T., Fry, B., 2010. Food-web inferences of stable isotope spatial patterns in copepods and yellowfin tuna in the pelagic eastern Pacific Ocean. Prog. Oceanogr. 86, 124–138. Parnell, A.C., Inger, R., Bearhop, S., Jackson, A.L., 2010. Source partitioning using stable isotopes: coping with too much variation. PLoS One 5, e9672. Parnell, A.C., Phillips, D.L., Bearhop, S., Semmens, B.X., Ward, E.J., Moore, J.W., Jackson, A.L., Grey, J., Kelly, D.J., Inger, R., 2013. Bayesian stable isotope mixing models. Environmetrics 24, 387–399. Penteriani, V., López-Bao, J.V., Bettega, C., Dalerum, F., del Mar Delgado, M., Jerina, K.,

Kojola, I., Krofel, M., Ordiz, A., 2017. Consequences of brown bear viewing tourism: a review. Biol. Conserv. 206, 169–180. Phillips, D.L., Newsome, S.D., Gregg, J.W., 2005. Combining sources in stable isotope mixing models: alternative methods. Oecologia 144, 520–527. Phillips, D.L., Inger, R., Bearhop, S., Jackson, A.L., Moore, J.W., Parnell, A.C., Semmens, B.X., Ward, E.J., 2014. Best practices for use of stable isotope mixing models in food web studies. Can. J. Zool. 92, 823–835. Post, D.M., 2002. Using stable isotopes to estimate trophic position: models, methods and assumptions. Ecology 83, 703–718. Semeniuk, C.A., Rothley, K.D., 2008. Costs of group-living for a normally solitary forager: effects of provisioning tourism on southern stingrays Dasyatis americana. Mar. Ecol. Prog. Ser. 357, 271–282. Semeniuk, C.A., Speers-Roesch, B., Rothley, K.D., 2007. Using fatty-acid profile analysis as an ecologic indicator in the management of tourist impacts on marine wildlife: a case of stingray-feeding in the Caribbean. Environ. Manag. 40, 665–677. Semeniuk, C.A., Bourgeon, S., Smith, S.L., Rothley, K.D., 2009. Hematological differences between stingrays at tourist and non-visited sites suggest physiological costs of wildlife tourism. Biol. Conserv. 142, 1818–1829. Shannon, G., Larson, C.L., Reed, S.E., Crooks, K.R., Angeloni, L.M., 2017. Ecological consequences of ecotourism for wildlife populations and communities. In: Ecotourism's Promise and Peril. Springer, pp. 29–46. Shipley, O.N., Olin, J.A., Polunin, N.V., Sweeting, C.J., Newman, S.P., Brooks, E.J., Barker, S., Witt, M.J., Talwar, B., Hussey, N.E., 2017. Polar compounds preclude mathematical lipid correction of carbon stable isotopes in deep-water sharks. J. Exp. Mar. Biol. Ecol. 494, 69–74. Speed, C.W., Meekan, M.G., Field, I.C., McMahon, C.R., Abrantes, K., Bradshaw, C.J.A., 2012. Trophic ecology of reef sharks determined using stable isotopes and telemetry. Coral Reefs 31, 357–367. Tanaka, H., Yoneda, M., Kitano, H., Kawamura, K., Imanaga, Y., Matsuyama, M., Okamura, K., Ohshimo, S., 2016. Stable isotope evidence for income resource allocation to egg production in the Japanese anchovy Engraulis japonicus. Mar. Biol. 163, 28. Trave, C., Brunnschweiler, J., Sheaves, M., Diedrich, A., Barnett, A., 2017. Are we killing them with kindness? Evaluation of sustainable marine wildlife tourism. Biol. Conserv. 209, 211–222. Trystram, C., Rogers, K.M., Soria, M., Jaquemet, S., 2016. Feeding patterns of two sympatric shark predators in coastal ecosystems of an oceanic island. Can. J. Fish. Aquat. Sci. 74, 216–227. Urban, D., 2002. Classification and regression trees. In: McCune, B., Grace, J. (Eds.), Analysis of Ecological Communities. MjM Sofware Design, Gleneden Beach, pp. 222–232. Vander Zanden, M.J., Clayton, M.K., Moody, E.K., Solomon, C.T., Weidel, B.C., 2015. Stable isotope turnover and half-life in animal tissues: a literature synthesis. PLoS One 10, e0116182. Vaudo, J.J., Wetherbee, B.M., Harvey, G.C.M., Harvey, J.C., Prebble, A.J.F., Corcoran, M.J., Potenski, M.D., Bruni, K.A., Leaf, R.T., Henningsen, A.D., Collie, J.S., Shivji, M.S., 2017. Characterisation and monitoring of one of the world's most valuable ecotourism animals, the southern stingray at Stingray City, Grand Cayman. Mar. Freshw. Res. 69, 144–154. Wirsing, A.J., Ripple, W.J., 2011. A comparison of shark and wolf research reveals similar behavioral responses by prey. Front. Ecol. Environ. 9, 335–341.

308