Evidence of site fidelity and deep diving behaviour of scalloped hammerhead shark (Sphyrna lewini) around the Saint Peter and Saint Paul Archipelago, in the equatorial Mid-Atlantic ridge

Additional keywords: oceanic essential habitat, residency, satellite telemetry, swimming speed.

Information on fish behaviour is essential to understanding their life history and devising appropriate conservation measures needed to protect them, as well as their ecosystems (Hayes et al. 2009; Jørgensen et al. 2012). However, tracing the life history of a species can be very complex if the species undertakes great migrations or inhabits a variety of ecosystems (Heupel et al. 2007). This task is particularly difficult for migratory sharks because of the various factors that influence their behaviour, habitat use and distribution, such as age, sex or food availability (Klimley 1987; Hoyos-Padilla et al. 2014; Vandeperre et al. 2014; Bass et al. 2017). As apex predators, some shark species play a key role in maintaining the ecosystem balance as top-down controllers of marine food webs. Therefore, the progressive decline in shark populations in recent decades is having deleterious cascading effects across the entire marine ecosystem (Dulvy et al. 2014; Ruiz et al. 2016). A better understanding of shark migrations, habitat use and essential fish habitats (EFH) has become increasingly crucial to ensure their global conservation.

The scalloped hammerhead shark (Sphyrna lewini Griffith & Smith, 1834) is a cosmopolitan shark that inhabits tropical and warm temperate waters between 40°N and 40°S. It is present in both coastal and oceanic habitats, including island habitats (Compagno 1984). Several studies have shown declines in scalloped hammerhead shark populations, mostly due to the global shark fin trade (Baum and Blanchard 2010; Barreto et al. 2015; Bezerra et al. 2016). High rates of incidental or targeted catches, including the illegal practice of finning, increase the vulnerability of the species. This ultimately results in marked reductions of some populations in a rather short period of time (Stevens et al. 2000; Harrington et al. 2005).

The International Union for the Conservation of Nature’s (IUCN) Red List of Threatened Species classifies the scalloped hammerhead shark as Endangered due to the marked decline in their populations over the past 20 years, due primarily to fishing (Baum et al. 2009). In 2010, the International Commission for the Conservation of Atlantic Tunas (ICCAT) prohibited boarding, landing or trade of any species of the genus Sphyrna (except Sphyrna tiburo) while recommending further studies on their biological and ecological aspects in order to identify nursery areas and potential migration routes for protection (International Commission for the Conservation of Atlantic Tunas 2010). In 2013, S. lewini was also included in Appendix II of the Convention on International Trade in Endangered Species of Wild Fauna and Flora (2013). In Brazil, the species has been classified as Critically Endangered since 2014 (Ministério do Meio Ambiente directive number 445) and its onboarding or selling is prohibited (Teixeira 2014).

The behaviour of migratory sharks in the wild has been investigated using various types of electronic tags and, most notably, pop-up satellite archival tag (PSATs; e.g. Hammerschlag et al. 2011; Musyl et al. 2011; Queiroz et al. 2016). These archival tags collect information on an animal’s vertical and horizontal displacements, helping elucidate key aspects of its spatial ecology, such as the fine-scale habitat use or broad-scale regional migrations. PSATs and ultrasonic telemetry have been used to study the behaviour of scalloped hammerhead shark in the eastern-central Pacific Ocean in the Gulf of California (Jorgensen et al. 2009) and the Malpelo, Cocos and Galapagos islands (Bessudo et al. 2011a; Hoyos-Padilla et al. 2014). According to these studies, adults spend long periods in areas around oceanic islands, moving away from them at night and performing large vertical displacements (dives), presumably to feed, before returning to the island shelves after sunrise. S. lewini seems to have a preference for the thermocline, performing deep and shallow dives when away or near the islands respectively (Klimley 1993; Klimley et al. 1993; Bessudo et al. 2011b; Ketchum et al. 2014a).

Oceanic islands and seamounts are important habitats where pelagic predators concentrate (Morato et al. 2010). The interaction of ocean currents with rough topography often causes local upwelling, contributing to an increase in primary productivity (Hekinian 1982; Hekinian et al. 2000). The Saint Peter and Saint Paul Archipelago (SPSPA) is an isolated group of small oceanic islands that are used by several marine species for protection, reproduction and feeding (Vaske Júnior et al. 2003, 2006; Viana et al. 2013; Dos Santos et al. 2014). In addition, the area hosts a very diverse fish assemblage that includes both resident and migratory species, which use the archipelago along their migratory route (Klimley and Butler 1988; Lessa et al. 1999; Macena and Hazin 2016; Mendonça et al. 2018).

Two species of hammerhead sharks have so far been recorded in the SPSPA: the scalloped hammerhead (Lubbock and Edwards 1981) and, more recently, the smooth hammerhead (Bezerra et al. 2017). Sightings have always been only of solitary individuals, in contrast with the large schools of scalloped hammerheads known to aggregate around other oceanic islands (Klimley and Nelson 1981; Hearn et al. 2010; Bessudo et al. 2011b). The SPSPA lies very close to the equator and is approximately midway between the African and American continents; thus, it is located at a two crossroads for oceanic migratory species such as hammerhead sharks. However, the function and importance of this singular ecosystem to the life history of the hammerheads, as well as the degree of connectivity between this and other oceanic and coastal habitats for Atlantic shark populations, are important questions that remain unanswered. The area is also close to the boundary that regional fisheries management organisations (ICCAT) have long used between the north and south Atlantic stocks for most pelagic predators, including sharks, which means that evaluating species based on local stock is not an effective measure to protect mainly for migratory species.

In this context, the aim of this study was to shed light the vertical and horizontal movements of scalloped hammerhead shark at the Mid-Atlantic Ridge, for the first time in the equatorial Atlantic Ocean.

The SPSPA (Fig. 1), located in the Mid-Atlantic Ridge (0°55′02″N, 29°20′42″W), is included in a vast oceanic marine protected area (MPA), together with Fernando de Noronha Archipelago and Rocas Atoll. Its central position in the Atlantic Ocean, lying 530 nm from Natal, the capital of Rio Grande do Norte state in Brazil, and 985 nm from Guinea-Bissau, in Africa, has long aroused great scientific interest (Mabesoone and Coutinho 1970), particularly after the installation of a research station there in 1998.

Fig. 1.  Location of the Saint Peter and Saint Paul Archipelago (SPSPA) in the equatorial Mid-Atlantic Ridge (left), and the most probable tracks of four scalloped hammerhead sharks (right; HS3, HS5, HS6 and HS7). Circles represent the tagging position, the triangles indicate the pop-up locations and the ‘x’ symbols indicate the SPSPA.

Sharks were captured with a small modified longline consisting of a 500-m main polyamide (PA) monofilament line (3 mm) equipped with 16–25 branch lines (2-mm PA monofilament) with baited circle hooks (size 16/0 or 17/0). The gear was immersed for 4 h during the night after dusk (1900–2300 hours) or before dawn (0000–0400 hours) on both the east and west sides of the archipelago. All sharks were captured and tagged from 1 to 5 nm off the east side.

PSATs (Model MK-10; Wildlife Computers, Redmond, WA, USA) were used to obtain data on the horizontal and vertical movements of six sharks (HS1–3 and HS5–7). A seventh single shark (HS4) was tagged with an Argos satellite-linked smart position and temperature (SPOT) tag (Table 1). HS1 and HS2 were tagged on the boat deck with their eyes covered and a hose running seawater into their mouth to maintain oxygenation and reduce handling stress. These two individuals had their PSATs fixed by a stainless steel dart inserted into the musculature on the base of the first dorsal fin. The remaining five sharks were tagged directly in the water from an inflatable boat. The PSATs were attached using coated monofilament crossed over the dorsal fin ~5 cm from the anterior margin and 12 cm from the base (Hazin et al. 2013). The SPOT tag was attached using the standard method suggested by the manufacturer, fixed parallel to the dorsal fin. All sharks were sexed and measured (total length, TL) to the nearest centimetre. The tagging procedure lasted less than 5 min for PSATs and 11 min for SPOT tags.

Table 1.  Scalloped hammerhead shark tag deployments
Summary of the satellite tag deployments of seven sharks tagged at the Saint Peter and Saint Paul Archipelago. Mk10, a pop-up satellite archival tag or PSAT; SPOT, smart position and temperature; TL, total length

The PSATs collected temperature and depth data with resolutions of 0.05°C and 0.5 m respectively. The tags were programmed to collect and store data every 1 s and send summarised data in 3-h periods across 14 bins. The histogram bins (time at temperature and depth) were defined based on the species’ known vertical habitat preferences (Jorgensen et al. 2009; Bessudo et al. 2011a). The tags were programmed to collect the profiles of depth and temperature (PDT) through the deployment period. PSATs were preprogrammed to pop-up after 70 (n = 3) and 120 (n = 3) days (Table 1). Four tags also had a time series function programmed to record depth and temperature data every 10 min on a 10-day on, 10-day off duty cycle.

All research was conducted with the permission of Chico Mendes Institute for Biodiversity Conservation (number 50119-1), of the Brazilian Ministry of the Environment. The capture of and tagging methods for scalloped hammerhead sharks were approved by the Commission of Ethics on the Usage of Animals of Federal Rural University of Pernambuco (licence number 054/2013, protocol number 23082.022567/2012).

Daily positions (latitude and longitude) of the tagged sharks (HS3, HS5, HS6 and HS7) were estimated by geolocation from light curves using the geolocation tool WC-GPE3 from Wildlife Computers (see https://wildlifecomputers.com/) in order to estimate the most probable track (MPT) of the animal during the tag attachment period. The WC-GPE3 models were calculated with a swimming speed of 2 m s^–1, sea surface temperature (SST) from National Oceanic and Atmospheric Administration (NOAA) high-resolution SST data provided by the NOAA Oceanic and Atmospheric Research Earth System Research Laboratory Physical Sciences Division (http://www.esrl.noaa.gov/psd/, accessed 5 May 2019) and bathymetry from the ETOPO1 1 arc-minute global relief model (https://www.ngdc.noaa.gov/mgg/global/, accessed 5 May 2019). In all, 1005 raw geolocation estimates were obtained for all sharks combined. The daily horizontal swimming speeds were calculated point-to-point (geographic positions) per day or the number of days travelled (km day^–1).

Depth and temperature data from the six PSATs were first analysed separately in day and night periods to evaluate diel patterns in diving behaviour. Then, the two first sharks (HS1 and HS2) with poor data were excluded from the analysis. Because the deployment and pop-up locations, as well as the geolocations, indicated that vertical data had been collected in the equatorial region, where day and night periods have similar durations, twilight periods (dawn and dusk) were not considered. To determine the significance of differences in the time spent at depth and temperature between day (0500–1700 hours) and night (1700–0500 hours), a Kolmogorov–Smirnov (K-S) test was used.

A detailed analysis of the sharks’ vertical behaviour was further conducted on HS3, HS5, HS6 and HS7 using the time series function, producing 288 detailed depth and temperature reconstructions for 24-h periods. The time series depth data were then grouped in four oceanographic depth intervals: a mixed layer (ML; 0–50 m), the thermocline (TC; 50–150 m), the maximum biomass layer (MB; 150–350 m) and a maximum depth layer (MD; >350 m). The TC and MB intervals were defined based on the literature (Travassos et al. 1999; Irigoien et al. 2014). Chi-Square tests were used to determine whether sharks spent significantly more time in any one of these layers.

Individual diving behaviour from time series data was also assessed by visual inspection. The shark was considered to perform a full dive when it left the ML and descended below the TC followed by a return to the ML or TC layer. The mean diving ascent and descent speeds (m s^–1) and their duration (min) were calculated for those dives with complete records, and Student’s t-test was used to compare diving speeds and duration between sharks.

Vertical movements were also assessed using the PDT and relative time spent at depth (TAD) data (Afonso and Hazin 2015). A generalised linear model (GLM) with a Gaussian probability of distribution and an ‘identity’ link function was used to infer the effects of the predictors ‘diel cycle’ (two-level factor comprising 0600–1759 hours as day and 1800–0559 hours as night; GMT-2) and ‘moonlight’ (%; integer) on the response variable ‘maximum depth’ (maxdepth) dives. Moonlight intensity data were obtained from the lunar package in R (ver. 3.6.0, R Foundation for Statistical Computing, Vienna, Austria). A GLM with binomial logistic probability distribution and a logit link function was used to infer TAD intervals, using the TAD bins pooled into five oceanographic water column layers, namely surface (0–1 m), ML (1–50 m), TC (50–150 m), MB (150–300 m) and deep water (>300 m), to assess the effects of the diel cycle and moonlight on the time spent in each depth stratum. The GLMs were conducted using the MASS library in R. All data analysis was performed in R.

Seven scalloped hammerhead sharks were tagged at the SPSPA between October 2010 and May 2014 (Table 1). Five sharks were females measuring between 200 and 260 cm TL. The remaining sharks were two males of 205 and 210 cm TL (Table 1). The PSAT retention time spanned from 5 to 120 days (mean ± s.d., 74 ± 56 days, n = 6), totalling 442 days of tracking. The SPOT tag on HS4 never transmitted.

The deployment and pop-up locations of the three PSATs that remained attached on sharks for up to 4 months were close to each other, indicating that these sharks remained in the surroundings of the SPSPA. This was also confirmed by the most probable tracks whenever data were suitable for calculation of geolocations (i.e. for sharks HS3, HS5, HS6 and HS7; Fig. 1). The mean (±s.d.) swimming speed for all sharks was 17 ± 25 km day^–1 (range 0.10–65.0 km day^–1).

The 3-h binned depth and temperature data showed that the tagged sharks (HS3, HS5, HS6 and HS7) spent most of their time in shallow waters above 150 m, both during the day (98%) and night (87%; Fig. 2). This epipelagic behaviour resulted in the sharks spending 58 and 63% of their time in warm waters >22°C during the day and night respectively (Fig. 2). There was no difference between day and night in TAD (K-S test, D = 0.5021, P = 0.5903) or temperature (D = 0.5043, P = 0.6348). Regardless of this dominant epipelagic behaviour, all sharks frequently dove down to deeper layers, primarily at night, when they eventually descended to depths in excess of 700 m (HS3 728 m, HS5 736 m) and temperatures reached as low as 5.6°C (Fig. 3).

Fig. 2.  Percentage of time spent at depth (left) and temperature (right) at night and during the day by scalloped hammerhead sharks Sphyrna lewini (HS3, HS5, HS6 and HS7) monitored at the Saint Peter and Saint Paul Archipelago. Data are the mean ± s.d. of the number of observations.

Fig. 3.  Profiles of depth and temperature (PDTs) recorded for scalloped hammerhead sharks HS3 (left) and HS5 (right) around the Saint Peter and Saint Paul Archipelago. The PDTs correspond to the data summarised every 3 h, containing 8 depth and 16 temperature readings.

A detailed analysis of the diving behaviour from the time series data showed a clear diel pattern in the vertical movements of scalloped hammerhead sharks monitored at similar seasons in distinct years. During the day, all sharks typically remained in the shallow warm (mean 23.9°C) layer of the ocean. In contrast, most dives below the TC and all dives below a depth of 347 m occurred at night when the sharks experienced substantially colder water (mean 10.4°C, minimum 5.6°C; Fig. 4). All sharks performed several deep dives every night. The single daytime deep dive (below 600 m) was actually recorded at dawn (HS7, 0620 hours; Figs 3, 4). Although the two females (HS3 and HS6) spent more time in the ML (0–50 m) and the two males spent more time in the TC (50–150 m; Fig. 5), this difference was not significant (χ² = 182, P = 0.234). In addition, the relationships between this diel cycle and moonlight intensity inferred by the logistic regressions on time spent in different oceanographic layers were only significant for TC during the night period (see Table S1, available as Supplementary material to this paper). The prediction for the effects of the same covariates on maximum depth attained by the scalloped hammerhead sharks revealed a deep diving preference for the night period (Table S1; Fig. 4).

Fig. 4.  Detailed vertical behaviour of four scalloped hammerhead sharks (HS3, HS5, HS6 and HS7) from 10-min resolution time series. Black lines represent depth, grey dots represent temperature.

Fig. 5.  Time spent by four tagged scalloped hammerhead sharks in the mixed layer (ML; 0–50 m), thermocline layer (TC; 50–150 m), maximum biomass layer (MB; 150–350 m) and maximum depth layer (MD; >350 m).

Visual inspection of the diving records indicated that deep dives >500 m typically occurred interspersed with returns to the ML before another deep dive (Fig. 4). These deep dives lasted 10–40 min, with the sharks travelling at a mean (±s.d.) speed of 0.80 ± 0.36 m s^–1 (range 0.65–0.92 m s^–1). Descent rates were always than ascent rates faster (mean 0.50 v. 0.32 m s^–1 respectively), but this difference was not significant (Student’s t-test, P > 0.05).

The use of satellite tags to study the fine- and broad-scale behaviour of scalloped hammerhead sharks is unprecedented in the equatorial Atlantic Ocean. Despite the reduced number of animals tagged, which reflects both the difficulty in working in a remote place such as the SPSPA and the current rarity of this species, the present study is the first to provide this type of information for this endangered species in this region of the globe.

In fact, PSAT technology was only recently used on this species and so even the tag attachment methods need to be improved to enhance the success of tag retention and shark survival (Jorgensen et al. 2009; Bessudo et al. 2011a). In the present study, the short period of tag retention for sharks HS1 and HS2 was similar to those observed by Bessudo et al. (2011a) in the eastern Pacific, as well as that observed in other shark species (Meyer et al. 2010; Hammerschlag et al. 2011; Afonso and Hazin 2015; Campana et al. 2016). These two premature releases may be related to several reasons, but premature death resulting from the stress of the capture and on-board tagging procedure (Musyl et al. 2011) appears to be the most probable cause, as opposed to all other sharks (HS3, HS5, HS6, HS7) that were tagged directly in the water and swam freely after the tagging process.

The position data from the geolocations together with the pop-up locations near the SPSPA indicate that the three scalloped hammerhead sharks monitored up to for 120 days (Table 1) did not swim great distances away from the archipelago, and may suggest that adult scalloped hammerhead sharks remain in the area for at least 4 months. At a finer horizontal scale, previous studies have shown that adult scalloped hammerhead sharks at oceanic archipelagos perform diel movements between their inshore daytime resting or cleaning aggregation areas and their offshore night-time feeding areas (Hearn et al. 2010; Bessudo et al. 2011b; Ketchum et al. 2014b). The PSAT geolocations in the present study prevented us from a finer-scale horizontal analysis and there may be an associated spatial error of the geopositions at the lower latitudes. The common low horizontal speed and very clear reverse diel vertical behaviour (night-time deep diving, daytime shallow depths), similar to that reported in other studies (Bessudo et al. 2011a; Hoyos-Padilla et al. 2014), most possibly indicates that sharks around the SPSPA also perform such diel movements where sharks disperse and return to the archipelago. In the Galapagos Marine Reserve, migration between the islands has been associated with abiotic factors, such as ocean currents and seasons, while also varying because of differences in individual behaviour (Ketchum et al. 2014a, 2014b). In the present study, sharks did not migrate towards the continent or to the nearest archipelago (Fernando de Noronha), and instead moved towards the south-east quadrant, with marked individual differences as well (Fig. 1).

The most likely reason for such a diel migration pattern seen or inferred in this and other studies is searching for prey for feeding at night in the mesopelagic vicinity of the seamounts. This behaviour may be related to the high productivity of a remote insular ecosystem compared with the surrounding oligotrophic waters (Klimley 1987; Hearn et al. 2010). Diving to greater depths is most probably associated with searching for prey due to the species capacity to forage in the mesopelagic zone and its preference for oceanic cephalopods, such as Chiroteuthis sp. and Vampyroteuthis infernalis, which are known to occur at depths of 300 m (Smale and Cliff 1998; Vaske Júnior et al. 2009). Local small-scale upwelling and turbulence resulting from the interaction between the seamount relief and marine currents promote the vertical transport of nutrients that may be retained close to the seamount by eddies, contributing significantly to an increase in food availability and the resulting aggregation of marine organisms (Worm et al. 2003; White et al. 2007). Depths below 350 m have been considered the range of high biomass in the SPSPA surroundings and may thus possibly explain the deep dives of the species into that layer due to intraspecific and interspecific competition and overlapping feeding niches (Cortés 1999; Heupel et al. 2007; Vaske Júnior et al. 2009; Heithaus et al. 2013; Bornatowski et al. 2014; Irigoien et al. 2014). Diving to deep waters for trophic reasons has also been suggested for other shark species (Bonfil et al. 2010; Cartamil et al. 2011; Howey-Jordan et al. 2013; Afonso and Hazin 2015), although this hypothesis remains to be tested for the scalloped hammerhead shark.

The return of sharks at dawn also indicates they may be using the seamount as a reference point (Klimley 1993; Dagorn et al. 2000; Fréon and Dagorn 2000; Bezerra 2017). The use of geomagnetic fields as a geographical reference by scalloped hammerhead sharks was well described by Klimley (1993) in Baja California, indicating directional movements probably affected by local island topography. The SPSPA and its surrounding oceanic area may also be an essential habitat for these sharks, which may use this archipelago and associated seamounts as a daily refuge. Although we do not know the longer-term residency (or migrations) of these animals, it is conceivable that individuals may temporarily embark on large-scale migrations, for example to pup in coastal nurseries (Bonfil et al. 2005; Vandeperre et al. 2014) or to explore other seamounts and islands in the region, eventually coming back to the SPSPA. It is also possible that the SPSPA serves not only as a feeding ground, but also as a shelter from predation and a mating area (Klimley and Butler 1988; Heupel and Hueter 2002; Vaske Júnior et al. 2003, 2006; Viana et al. 2013; Dos Santos et al. 2014).

The vertical movements indicate that the scalloped hammerhead shark is able to use a wide range of temperatures and depths between the epipelagic and deeper mesopelagic zones. However, a notable preference for epipelagic waters was clear, together with a marked diel pattern, with deeper dives occurring mostly at night. Hoffmayer et al. (2013) reported multiple deep dives to the mesopelagic zone occurring at night, possibly to explore deep-water prey. Incursions of sharks to depths between 100 and 300 m at Wolf Island, Galapagos, also occurred during the night, coinciding with the time in which the scalloped hammerhead sharks swam away from the island (Hearn et al. 2010).

The fact that adult scalloped hammerhead shark dive to the mesopelagic zone mostly at night (e.g. Bessudo et al. 2011a; present study) contrasts with the pattern commonly described for other shark species, which remain in deep waters during the day and move to shallow waters searching for food at night (Cartamil et al. 2010, 2011; Queiroz et al. 2012). This inverse vertical behaviour of scalloped hammerhead shark could be an adaptive response to a potential reduced competition for prey in deeper waters (Jorgensen et al. 2009; Bessudo et al. 2011a), especially from other sharks that may also prey on hammerhead sharks.

In this study, after performing a deep dive the shark would return to the TC to warm up before undertaking another deep dive. This frequent daily diving between surface and greater depths, called a ‘yo-yo’ diving pattern (Klimley et al. 2002; Bessudo et al. 2011a; Nakamura et al. 2011; Afonso and Hazin 2015), likely results from the sharks’ physiological need to thermoregulate in warmer shallow waters so they can maintain key capacities, such as searching for prey and navigation, while diving in deeper, colder waters (Weihs 1973; Bonfil et al. 2010; Campana et al. 2011; Thums et al. 2013). This behaviour is also common for other shark species, such as Prionace glauca, Alopias vulpinus, Alopias superciliosus, Isurus oxyrinchus and Carcharhinus longimanus (Carey and Scharold 1990; Nakano et al. 2003; Abascal et al. 2011; Cartamil et al. 2011; Howey-Jordan et al. 2013). In particular for hammerhead sharks, the yo-yo movements may also be associated with their greater perception of the Earth’s magnetic field, useful for guiding migratory movements (Klimley et al. 2002). Furthermore, the speed of descent was greater than that of ascent (fast down, slow up), suggesting that sharks may be feeding during their return to shallow waters, a pattern also reported by Hoffmayer et al. (2013) for scalloped hammerhead sharks. Alternatively, this behaviour may simply reflect their lowered metabolism because of the time spent in colder waters (Thums et al. 2013).

Despite the small sample size, the shallower distribution of females in relation to males found in this study may also have a physiological reason. Judging from their size, all tagged females were likely adults and could therefore be using the warmer waters of the ML to save energy or accelerate the ovulation and fertilisation processes, as hypothesised for other shark species (Hazin et al. 2000; Hight and Lowe 2007; Nosal et al. 2014).

In summary, we observed a horizontal and vertical behaviour of adult scalloped hammerhead shark at the SPSPA consistent with that described for other oceanic islands in the eastern Pacific. In order to elucidate the specific habitat use of scalloped hammerhead sharks around the SPSPA and the Mid-Atlantic ridge, more telemetric studies are clearly needed, preferably combining techniques that can detect long-term residency at the SPSPA (e.g. acoustic telemetry) with large-scale migrations and fine-scale vertical behaviour (PSATs with longer deployment times). This study showed that sharks remained near the SPSPA for at least 4 months. The SPSPA, together with the Rocas Atoll Biological Reserve and Fernando de Noronha Archipelago, are included in the MPA of Fernando de Noronha. Currently, the SPSPA has been going through changes since the Brazilian government converted a part of archipelago for restricted use. The use of the SPSPA by scalloped hammerhead sharks highlights the importance of adopting conservation measures for the species in this insular ecosystem. The great importance of these insular oceanic ecosystems to several other marine species, which depend on them for their life cycles and ultimate survival, should also be emphasised.

The authors declare that they have no conflicts of interest.

The authors thank Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) and Fundação de Amparo à Ciência e Tecnologia de Pernambuco (FACEPE) for an international scholarship to N. P. A. Bezerra and for postdoctoral fellowships to B. C. L. Macena and N. P. A. Bezerra. The authors also thank the Brazilian Navy through the Secretaria da Comissão Interministerial para os Recursos do Mar (SECIRM) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) for logistic and financial support. This research was funded by CNPq under grant number 482557/2011-7.