Site affinity of whitespotted eagle rays Aetobatus narinari assessed using photographic identification

Introduction

Anthropogenic threats such as overfishing, target and incidental fisheries, habitat loss and climate change can negatively affect populations of elasmobranchs (sharks, skates and rays) both locally and globally, which can ultimately lead to extinction (Dulvy et al., 2014). Five of the seven most threatened families of elasmobranchs are batoids (superorder Batoidea: skates, stingrays, sawfishes, wedgefishes and guitarfishes; Dulvy et al., 2014). There are over 630 species of batoids, of which 112 are considered threatened with extinction and 243 are considered data deficient by the International Union for the Conservation of Nature (www.iucn.org; Last et al., 2016). Many management and conservation actions for rays will require an understanding of their behaviour and movements (Flowers et al., 2016).

Movements of animals can genetically structure populations through a behaviour known as philopatry (Chapman et al., 2015; Flowers et al., 2016), i.e. when individuals frequently return to or stay in their home ranges, birthplaces or other defined locations (Mayr, 1963; Speed et al., 2010; Chapman et al., 2015; Flowers et al., 2016). Relevant to this study are three specific types of philopatry. The first is residency, when individuals remain in one area for at least 12 months (Chapman et al., 2015; Flowers et al., 2016). This movement pattern has been documented in many species of sharks (Chapman et al., 2015), but there has been limited published evidence for residency in batoids (Flowers et al., 2016). Longer study durations, however, would probably show many seasonally resident batoid species to be yearly residents (Flowers et al., 2016). The second form of philopatry relevant to this study is site fidelity, where an individual will return to a specific site after making a long-distance movement (Chapman et al., 2015; Flowers et al., 2016). This behaviour is commonly exhibited by manta rays, Mobula birostris (Walbaum 1792) (Dewar et al., 2008) and Mobula alfredi (Krefft 1868) (Germanov & Marshall, 2014; Jaine et al., 2014; Braun et al., 2015). Couturier et al. (2011) introduced a new term, site affinity, to include the first two types of philopatry. Mobile animals are regularly deemed to be displaying site affinity when the evidence does not clearly show the difference between site fidelity and residency (Couturier et al., 2011; Flowers et al., 2016).

Various methods can be used to determine philopatric behaviour. One that is common and relatively inexpensive for elasmobranchs is photographic identification (photo-ID; Marshall & Pierce, 2012). The whitespotted eagle ray Aetobatus narinari (Euphrasen 1790) has intricate white dorsal markings that facilitate the identification of individuals (Corcoran & Gruber, 1999; Speed et al., 2007; Marshall & Pierce, 2012; González-Ramos et al., 2016). The taxonomic classification of this species is in flux (Richards et al., 2009; White et al., 2010; Naylor et al., 2012; Last et al., 2016) and the globally distributed species complex is currently listed as near threatened by the IUCN (Kyne et al., 2006). Although its population trend is decreasing (Kyne et al., 2006), A. narinari potentially occupies an important trophic position as an intermediate predator (Schluessel et al., 2010a; Ajemian et al., 2012; Newby et al., 2014). Movement data for A. narinari are lacking, but it is a semi-pelagic animal that is capable of long-distance movements, which is supported by both population genetic and tagging studies that suggest a seasonal north to south migration (Bassos-Hull et al., 2014; Sellas et al., 2015). In contrast to these findings, individuals in Bermuda may be exhibiting residency (Ajemian et al., 2012; Ajemian & Powers, 2014).

Three previous studies have used variations of A. narinari dorsal surface to identify individuals within populations (Corcoran & Gruber, 1999; Bassos-Hull et al., 2014; González-Ramos et al., 2016). The first study examined the essential photograph requirements for this species in the Bahamas, which included the pelvic fins for sex determination and the dorsal side of the pectoral fins for visual comparison of individuals (Corcoran & Gruber, 1999). The interactive individual identification system (I³S; www.reijns.com/i3s) was used by the second study to visually confirm recaptures of tagged individuals, but the performance and matching ability of the programme were not examined (Bassos-Hull et al., 2014). The third study validated I³S Spot for use with A. narinari (González-Ramos et al·2016); however, the species studied was likely the Pacific eagle ray Aetobatus laticeps (Gill 1865) rather than A. narinari (Naylor et al., 2012; Last et al., 2016). This could have implications for the use of the software tool with A. narinari, which tends to have more complex markings than individuals of Aetobatus laticeps (K. I. Flowers, pers. obs.). González-Ramos et al. (2016) discussed a limitation to I³S Spot as only being able to digitize 30 landmarks (or spots), yet it is possible to manually increase the number of spots for annotation in the programme′s metadata (den Hartog & Reijns, 2014a). Furthermore, these authors appear to have treated the top ranked photograph match as a confirmed individual match (González-Ramos et al., 2016), when in fact this should not be assumed (den Hartog & Reijns, 2014a). Even though these parameters were overlooked, 83·1–88·2% of images were correctly identified to their individual, depending on which region of the dorsal surface was used in the annotation process, validating I³S Spot as a photo-ID assistance tool for aetobatid species (González-Ramos et al., 2016). Photographs were collected from specimens brought onto a boat for the Bassos-Hull et al. (2014) and González-Ramos et al. (2016) studies, but there is a need to develop inexpensive and less invasive methodologies for movement studies. Assessing the usefulness of I³S Spot with in-water photographs of free-swimming A. narinari could open new research possibilities.

The Turks and Caicos Islands consist of over 40 islands and cays along the Lucayan Archipelago in the north-west Atlantic Ocean. The islands contain and are surrounded by mangrove and coral-reef habitats, separated by sand flats to the south (Caicos Bank; an area of c. 4015 km²) and a deep channel to the east (Turks Islands Passage). Aetobatus narinari is commonly observed along the east Caicos Bank in the Turks and Caicos Islands, providing an opportunity to study philopatry of this species. According to roving-diver surveys, where over 200 surveys were conducted by Reef Environmental Education Foundation (REEF, 2017; www.reef.org) volunteers at any one location, the sighting frequency of A. narinari in the tropical western Atlantic Ocean is the highest in South Caicos (19·9%; www.reef.org), suggesting that the east Caicos Bank is a key area for the life history of this species. Further studies are required to understand the underlying reasons for their abundance in and frequenting of this area. This study used in-water photographs of individual A. narinari in combination with the I³S Spot 4.0 (www.reijns.com/i3s/download/I3S_download.html; den Hartog & Reijns, 2014b) to discern individual ray resightings in South Caicos, Turks and Caicos Islands, over the course of approximately 6 years to determine if there was any evidence of philopatry among these individuals. Knowledge of A. narinari movements will help designate the scale of protection required for this species.

Materials and methods

Field data collection

Photographs and videos of A. narinari were collected opportunistically by in-water observers from the School for Field Studies Center for Marine Resource Management (SFS-CMRS; www.fieldstudies.org) around South Caicos Island, Turks and Caicos Islands from 16 February 2009 to 12 August 2015, excluding 2010 and the month of January in all years due to logistical limitations. The primary study sites (Fig. 1) were Plandon Cay Cut (PCC; 21° 34·467′ N; 71° 29·676′ W), Dove Cay (DC; 21° 29·102′ N; 71° 31·700′ W), Shark Alley (SA; 21° 29·006′ N; 71° 32·055′ W) and the south end of Long Cay (SELC; 21° 27·189′ N; 71° 34·328′ W).

These sites were selected due to prior anecdotal observations of aetobatids in these areas, as well as proximity to the field station. At the beginning of the study in 2009, photographs and videos were collected by snorkellers and scuba divers from February to November at locations between the northernmost and southernmost points of the islands (PCC and SELC). In 2011, snorkel surveys to collect photographs were conducted in November at DC, SA and SELC, while roving-boat surveys took place on the sand flats in between. When rays were sighted, snorkellers would enter the water to obtain photographs. From 2012 to 2014, photographs and videos were collected by snorkel surveys in the months of March, April, October, November and December from the primary study sites PCC, DC, SA and SELC. In 2015, photographs and videos were collected by snorkel sessions from May to August, also at the four primary study sites. Session durations ranged from 4 to 150 min and were often dependent on environmental conditions, particularly at PCC where tidal currents occur. In 2015, individual sex was determined in the field whenever possible. The presence of claspers indicated male and the absence of claspers indicated female. When there was not a clear view of the ray′s ventral pelvic region underwater, sex was recorded as unknown. An image of a hand gesture was taken in between individual sightings to represent male, female or unknown, following Marshall & Pierce (2012). For all other years, sex was determined to be unknown, unless claspers were visible in the image. Because of this, a χ²-test was conducted on data from 2015 only to test the hypothesis that the sex ratio across all primary study sites was not significantly different from 1:1. In order to determine if a ray was mature (adult-sized), size was visually estimated in-water. If greater than approximately 127 cm disc width (W_D; Tagliafico et al., 2012; Bassos-Hull et al., 2014), it was presumed that the ray was adult-sized.

In 2013, an outreach programme was established to get the dive community of Providenciales, the main tourism island in the Turks and Caicos, to participate in the Turks and Caicos eagle ray project. Divers and snorkellers uploaded photographs and associated information to the project′s website (www.fieldstudies.org/eagle-ray-project).

Photographic identification (photo-id) of individuals

Video screenshots were taken of each individual ray after the field session. Each field session was defined as an encounter. By-eye comparisons of all photographs and video stills were conducted after every encounter in order to choose the best quality image of each individual sighted. Taking into consideration that the rays are usually swimming away from the photographer and that the pelvic fins were already determined to be reasonable for photo-ID (Corcoran & Gruber, 1999), three reference points delineating the area of the pelvic fin to be studied using I³S Spot were selected: (1) where the pelvic fin meets the pectoral fin, (2) the posterior lateral tip of the pelvic fin and (3) the posterior medial tip of the pelvic fin [Fig. 2(a)]. Programme metadata were edited prior to uploading the images into I³S Spot. Reference points were named: (1) pectoral, (2) pelvic and (3) tail [Fig. 2(a)]. Left and right pelvic fin images were required to be uploaded and annotated separately in I³S Spot, therefore one metadata element was added, fin side, with a custom drop-down menu for left or right. Whenever both fin sides were visible in one image, the image was duplicated in order to annotate both left and right pelvic fins separately. Under expert settings, the minimum number of spots was changed to five (as some of the rays have few spots within the reference points) and the maximum number of spots was changed to 60 (as some of the rays have many spots within the reference points). This procedure follows the I³S Spot user manual (den Hartog & Reijns, 2014a). All markings present within the three reference points were annotated, regardless of their shape [Fig. 2(b)]. Since the programme only allows for circular and oval annotations, thin ovals or a combination of thin ovals were used to properly annotate markings that were linear or straight-edged. Images were sometimes blurry or pixelated, therefore only markings that were clearly visible were annotated. Images were not included in this study when all markings of the pelvic fin were not identifiable, the fin was flexed in a manner that hid portions of the pelvic fin or the animal was too far away in the image. All of these factors were influenced by the proximity to the ray, the amount of movement in the photograph or video and in the case of snorkelling, the breath-holding and free-diving abilities of the snorkeller. In order to maintain consistency and reduce human error, all images were entered into the software and annotated by a single operator. The image listed in the top rank was not necessarily a direct match, but rather gave the operator substantial evidence to compare the potential match by-eye for confirmation that the two images were of the same individual (den Hartog & Reijns, 2014a). This greatly reduced the amount of time spent cross-analysing images manually. For each image that was uploaded into the program, the top 50 ranked potential matches were examined by-eye. A second author validated all matching images. Results are reported using the baseline database, which only includes individuals with images for both right and left pelvic fins.

This research was conducted under the scientific and research licence (number D2-001) granted by the Turks and Caicos Islands Department of Environment and Coastal Resources (DECR; www.gov.tc/dema).

Results

Resightings

Within the database, A. narinari were observed year-round at South Caicos except for the months of January where no effort occurred. Anecdotal sightings of the rays occurred in January 2013 (K. I. Flowers, pers. obs.). During the course of the study period, rays were observed in coastal habitats, including seagrass beds, coral reefs, channels with patch reefs, patch reefs dominated by octocorals and on sand flats. They were rarely sighted in mangrove habitats. There was a tendency to see adult-sized rays [>127 cm W_D (Tagliafico et al., 2012; Bassos-Hull et al., 2014)], with only several smaller individuals (<100 cm W_D) sighted in shallow water near SELC and around a patch of mangroves near 21° 30·686′ N; 71° 32·662′ W over the study period (A. C. Henderson, pers. obs.). A total of 662 images were uploaded into I³S Spot 4.0. There were 256 image verified individuals of A. narinari from the entire database (including individuals with only left or right pelvic fin images). Of the 256, 38 individuals had only right-side images and 53 individuals had only left-side images. One hundred sixty five individuals formed the baseline database (those with both left and right pelvic fin images), from which 26 sightings occurred at PCC, 37 sightings occurred at DC, 129 sightings occurred at SA, 62 sightings occurred at SELC, one sighting occurred at Highland House (21° 29·754′ N; 71° 29·634′ W), one sighting occurred at the Spanish Chain dive site (21° 28·926 N; 71° 31·686 W) and one sighting occurred at Leeward Reef, Providenciales (21° 50·324′ N; 72° 9·7972′ W). Of the 165 individual rays, 19 were female (F), 44 were male (M) and 102 were of undetermined sex (U) (Table S1, Supporting information). In 2015, using baseline individuals only, 19 individuals were F, 10 were M and one was U. Using these data only, sex ratio was not found to differ from parity (χ² = 2·8, P > 0·05). The discovery curve of new individuals over time did not reach an asymptote during the study period (Fig. 3). Outreach to the dive community of Providenciales yielded four photographs that were uploaded to the Turks and Caicos eagle ray project website. Only one photograph was included in the baseline database from Leeward Reef, Providenciales because the remaining three photographs did not contain both fin sides in the field of view, one of the requirements of the baseline database. Of the 165 rays in the baseline database, 56 were resighted at least once (33·9%; Fig. 4 and Table S1, Supporting information), where the maximum number of days between the first and last sighting was 1640 (median 165, interquartile range, IQR = 698; Fig. 5). There were few apparent changes to the shape and size of dorsal markings of the individual ray that was sighted 1640 days apart (Fig. 5 and Table S1, Supporting information).

Discussion

The majority of A. narinari sightings were of adult-sized individuals, suggesting that the rays may separate by life stage around South Caicos. Catch data from both targeted and by-catch A. narinari fisheries in Mexico and Brazil suggest that the rays segregate by size (Yokota & Lessa, 2006; Cuevas-Zimbrón et al., 2011). Data from the Southern Gulf of Mexico show that juvenile rays account for a significantly higher proportion of the catch at a site only 8–15 km offshore, compared with the site 30–50 km offshore (Cuevas-Zimbrón et al., 2011). In north-east Brazil, juvenile and neonate rays were caught by gillnets in depths <10 m (Yokota & Lessa, 2006). These studies corroborate accounts of smaller rays in the shallows around South Caicos.

Some species of sharks frequently segregate by sex (Springer, 1940; Klimley, 1987; Stevens & McLoughlin, 1991; Sims et al., 2001), while the pelagic stingray Pteroplatytrygon violacea (Bonaparte 1832) is also known to segregate by sex (Wilson & Beckett, 1970). Previous studies that examined targeted A. narinari fishery catch data in Mexico and Venezuela also reported differing sex ratios (Cuevas-Zimbrón et al., 2011; Tagliafico et al., 2012), with two fishing sites in Mexico dominated by the opposite sex (Cuevas-Zimbrón et al., 2011) and one fishing site in Venezuela dominated by females (Tagliafico et al., 2012). In south-west Florida, scientific boat-based surveys demonstrated overall catch was biased towards males, but on a month-to-month basis the sex ratio only differed significantly during one of the six study years (Bassos-Hull et al., 2014). Some inshore species of rays, however, do not routinely show a sex ratio that is significantly different from 1:1 (Snelson et al., 1988; Ismen, 2003). Schluessel et al. (2010b) reported a 1:1 sex ratio in Australia and Taiwan for Aetobatus narinari, a species now considered to be Aetobatus ocellatus (Kuhl 1823) in those regions (White et al., 2010; Naylor et al., 2012; Last et al., 2016). This was similar to the findings of this study where there was a 1:1 sex ratio of A. narinari across all sites in 2015. Sample size was low, however, and the proportion of unsexed individuals was high and may be female biased. Images would require a ventral view to confirm the lack of claspers, which is extremely difficult to accomplish while snorkelling. Collectively, results from this study and aforementioned studies suggest that sex composition in A. narinari varies by species and location (Schluessel et al., 2010b; Cuevas-Zimbrón et al., 2011; Naylor et al., 2012; Tagliafico et al., 2012; Bassos-Hull et al., 2014).

The degree of change in dorsal markings of A. narinari has only been studied on small sample sizes over short time periods (Corcoran & Gruber, 1999; Bassos-Hull et al., 2014; González-Ramos et al., 2016). During this study, the longest time between sightings was about 4 years (1640 days), in which few apparent changes were observed in the dorsal surface markings of the individual (Fig. 5). At the initial sighting of this individual, it was estimated to be adult size. More research needs to be done to explore the possibility of dorsal marking changes during different life stages (González-Ramos et al., 2016), ideally in aquaria (Marshall & Pierce, 2012). The lack of an asymptotic relationship on the discovery curve suggests that the entire population of A. narinari in South Caicos or using the area around South Caicos was not sampled in the study period. Nevertheless, there was a high proportion of individual ray resightings, which indicates residency, site fidelity or a combination of both. Since the data cannot definitively distinguish between residency and site fidelity without additional sampling effort, results concluded that the rays around South Caicos were exhibiting site affinity. This finding is consistent with the recapture data from Bermuda and off southwest Florida that suggested the rays were resident, displaying site fidelity or both within the study area (Ajemian et al., 2012; Bassos-Hull et al., 2014). It is possible that the South Caicos A. narinari exhibited residency, since they were sighted throughout the entire sampling period and some individuals were sighted across multiple seasons of multiple years. Photo-ID often underestimates residency as the movements of the animals are not monitored outside of the sampling period and data are entirely dependent on sampling effort (Delaney et al., 2012). It cannot be ruled out, however, that South Caicos individuals make large-scale movements and frequently return to the study sites. Population genetic studies and survey data in the Gulf of Mexico suggest a possible seasonal north to south migration of A. narinari in response to temperature fluctuations (Bassos-Hull et al., 2014; Sellas et al., 2015). Long-term telemetry, long-term satellite tracking or population genetic studies would help determine the actual type of philopatry exhibited by A. narinari in South Caicos (Flowers et al., 2016).

Although the rays are capable of moving between sites, the majority of them were resighted multiple times at their original capture site. This suggests that individual A. narinari have specific home ranges around South Caicos. Home ranging in sharks is a common behaviour, particularly in juvenile lemon sharks Negaprion brevirostris (Poey 1868) (Morrissey & Gruber, 1993; Yeiser et al., 2008) and adult grey reef sharks Carcharhinus amblyrhynchos (Bleeker 1865) (McKibben & Nelson, 1986). It has also been documented in two benthic stingrays, the brown stingray Bathytoshia lata (Garman 1880) (Cartamil et al., 2003) and the round stingray Urobatis halleri (Cooper 1863) (Vaudo & Lowe, 2006), as well as one semi-pelagic ray with a larger home range, the cownose ray Rhinoptera bonasus (Mitchell 1815) (Collins et al., 2008). Future telemetry studies could elucidate the home range size and stability of A. narinari around South Caicos.

South Caicos is an island that has been attempting to transition from a queen conch Lobatus gigas and Caribbean spiny lobster Panulirus argus fisheries-dominated industry to a tourism industry (Béné & Tewfik, 2003; Zuidema et al., 2011) for a number of years. Although fishers in South Caicos do not target A. narinari, they have been known to harpoon the rays in the past (A. C. Henderson, pers. comm.). This species could be an attractive incentive for tourists (divers and snorkellers) to visit the island, suggesting that the rays potentially have a high ecotourism value. When properly managed, ecotourism that is focused on elasmobranchs can supply a nation with millions of dollars annually and enhance diver experiences (Anderson & Waheed, 2001; Gallagher & Hammerschlag, 2011; O′Malley et al., 2013; Haas et al., 2017). Granting this species protection in the Turks and Caicos Islands could be done via marine protected area (MPA) design, a ray sanctuary [where there is a no-take (incidental or targeted) policy for all species of rays] or full species protection, as is the case in the state of Florida (Florida Fish and Wildlife Conservation Commission, 2016). Protection in this area would be considered proactive: a measure to prevent fishery development in the future that would safeguard the ray′s tourism value. Understanding that these rays are spending part or all of their time in South Caicos waters will help inform local policy makers on the potential benefits of the various forms of protection.

Photo-ID is a useful tool to inform resource managers about the movements of rays that have distinctive markings. I³S Spot effectively guided researchers to make individual matches of A. narinari. Since A. narinari around South Caicos were displaying site affinity, at least some management action could be on a local scale. If the South Caicos A. narinari are resident, local management could include well-designed protected areas or granting the species nationwide protection. If the rays are exhibiting site fidelity and moving across jurisdictions, trans-boundary management would be necessary to effectively conserve this population. Overall, this study contributes to a small but growing body of evidence that large, semi-pelagic rays may be more localized in their movements than expected from their swimming abilities (Flowers et al., 2016).

The authors gratefully acknowledge the logistical and financial support provided by The School for Field Studies. We are indebted to the staff and undergraduate students of The School for Field Studies Center for Marine Resource Studies who have helped with data collection. The authors would also like to thank the I³S team, J. den Hartog and R. Reijns, for their communication with us throughout this process. Acknowledgement is extended to members of the Stony Brook University community including R. Cerrato, A. Fields and L. Thorne.