Mark-recapture validates the use of photo-identification for the widely distributed blue-spotted ribbontail ray, Taeniura lymma

doi:10.21203/rs.3.rs-2952109/v1

Download PDF

Article

Mark-recapture validates the use of photo-identification for the widely distributed blue-spotted ribbontail ray, Taeniura lymma

https://doi.org/10.21203/rs.3.rs-2952109/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 29 Jul, 2024

Read the published version in Scientific Reports →

You are reading this latest preprint version

The ability to identify individual animals can provide valuable insights into the behaviour, life history, survivorship, and demographics of wild populations. Photo-identification (photo-ID) uses unique natural markings to identify individuals and can be effective for scalable and non-invasive research on marine fauna. The successful application of photo-ID requires that chosen distinguishing markings are unique to individuals and persist over time. In this study, we validate the use of dorsal spot patterns for identifying individual blue-spotted ribbontail rays (Taeniura lymma) in conjunction with traditional tagging methods. Spot patterns were unique among T. lymma with 90.3% of individuals correctly identified using I³S photo-matching software from images taken up to 496 days apart. In comparison, traditional physical tagging methods showed a tag loss rate of 27% and a maximum tag retention period of only 356 days. Our findings demonstrate the effectiveness of photo-ID as a tool to monitor populations and better understand the ecology of the blue-spotted ribbontail ray without the need for physical tagging. The validation of photo-ID for this widespread species is important as it enables behavioural and demographic changes to be easily tracked in relation to coastal threats such as human development and habitat degradation.

The ability to accurately identify individual animals is critical for effective species management. The ability to differentiate among individuals of the same species enables researchers to investigate many aspects of a species’ biology, including social behaviour, reproductive patterns, population size, residency, site fidelity, connectivity, and habitat preferences^1–4. Traditionally, populations have been monitored using mark-recapture techniques, where individuals are physically marked with distinguishing tags or catalogued using natural physical identifiers. To reduce the invasiveness of monitoring techniques, photo-identification (photo-ID) has frequently been employed as an alternative to more intrusive methods of identifying individual animals⁵.

The use of photo-ID has grown in recent years, especially among marine vertebrates, where pigmentation patterns, scale patterns, scarring, or other markings are used as distinguishing features^6–10. Successfully identifying individuals of a given species with photo-ID assumes that such markings are unique and persist through time. Several studies have demonstrated that batoids in particular are suitable candidates for photo-ID and that photo-matching software can provide statistical validation for these methodological assumptions (e.g., reef manta rays Mobula alfredi¹¹, spotted eagle rays Aetobatus narinari¹², and smalleye stingrays Megatrygon microps¹³). However, spot patterns may change in size, shape, or quantity throughout an individual animal’s life which may hinder re-identification over time. For example, some batoid lineages appear to exhibit spot patterns that may not remain stable through time¹⁴, and the utility of photo-ID must be evaluated on a species-by-species basis⁵.

The bluespotted ribbontail ray Taeniura lymma (Forsskål, 1775) is a demersal dasyatid stingray that is widely distributed throughout the Indo-West Pacific Ocean, inhabiting coral reefs, mangrove habitats, and lagoonal systems¹⁵. This species is characterised by its conspicuous blue spots across the dorsal surface and two blue stripes along the dorsolateral edges of the tail (Supplementary Fig. 1). The reproductive strategy of T. lymma is viviparous¹⁶, with sexually mature individuals giving birth to live young that already exhibit a distinct spot pattern similar to that of adult individuals. Global declines in shark populations and the consequent reduction in predation has led to speculation about a potential increase in T. lymma populations via mesopredator release^17,18. However, in order to comprehensively assess the status of T. lymma, it is necessary to gather more information on population trends and life history characteristics^17,19. The ability to easily identify individuals of this species through photo-ID would greatly facilitate the collection of such information. This is particularly pertinent to regions undergoing rapid coastal development, such as the Red Sea, where planned developmental activities overlap with key habitats for T. lymma²⁰. Monitoring regional populations T. lymma through photo-ID can aid in understanding the species' behavioural and population-level responses to habitat degradation as well as provide evidence-based management insights for small-scale fisheries which harvest the species^21,22. To systematically identify T. lymma individuals using photo-ID, the persistence of uniquely identifiable spot patterns needs to be first validated through a standardised approach. Although extremely short-term individual identifications of T. lymma using spot patterns have been implemented (i.e., < 60 minutes^17,20), there has been no investigation into the reliability and longevity of these markings over longer time scales relevant to population monitoring.

This study investigates the performance of photo-ID methodology to differentiate T. lymma individuals in conjunction with mark-recapture tagging. We assess the accuracy of using unique dorsal spot patterns of T. lymma for individual identification and examine the longevity of these spot patterns to evaluate the use of photo-ID for long-term monitoring. Determining the distinctiveness and temporal stability of dorsal spot patterns in T. lymma is a crucial prerequisite for enabling low-cost, scalable, and non-invasive monitoring of this widespread, coastal species.

Tagging and recapture events

In total, there were 218 T. lymma capture events from Thuwal (n = 185) and the Farasan Banks (n = 33), Saudi Arabia (Fig. 1). From all capture events, 161 individuals were identified: 80 females, 56 males, and 25 individuals for which sex was undetermined. Average disc width in females was 22.9cm and 22cm in males (Supplementary Fig. 2). Of the 161 identified rays, 29 (18%) were physically recaptured at least once within the study period. The mean recapture interval was 164 days (± 30 SE).

Among recaptured stingrays that had previously been tagged (n = 15) the maximum observed period of tag retention was 356 days (Table 1). Several individuals were recaptured after shedding their original physical tag and were subsequently identified through photo-ID. Analysis of the data showed that 27% of recaptured T. lymma had shed their physical tags as early as 109 days post-release (Table 1), as indicated by the use of photo-ID. Notably, no visible scarring was observed at the tag attachment site after a minimum period of three months.

Table 1

Tag retention periods for recaptured *Taeniura lymma* individuals from Thuwal, Saudi Arabia. Specimen identification number (Specimen ID), sex, date of initial and final capture events, Maximum Recapture Interval (MRI; days between first and last capture event), observed tag retention (number of days tag was physically observed on the individual), and the presence (Y) or absence (N) of the tag on the final recapture event.
Specimen ID	Sex	Initial capture	Last recapture	MRI (days)	Tag retention (days)	Tag present
BS159	F	15-May-19	16-May-19	1	1	Y
BS145	F	04-Aug-19	08-Aug-19	4	4	Y
BS025	M	25-Apr-19	13-May-19	18	18	Y
BS132	F	13-Mar-19	09-May-19	57	57	Y
BS036	M	18-Mar-19	01-Jun-19	75	75	Y
BS137	F	15-May-19	05-Aug-19	82	82	Y
BS008	M	17-Apr-19	04-Aug-19	109	< 109	N
BS015	M	13-Mar-19	05-Aug-19	145	145	Y
BS126	F	03-Mar-19	04-Aug-19	154	< 154	N
BS236	F	01-Sep-20	24-May-21	265	264	Y
BS114	M	02-Sep-20	04-Jul-21	305	305	Y
BS107	M	31-Aug-20	04-Aug-21	338	< 226	N
BS103	M	28-Aug-20	04-Aug-21	341	336	Y
BS118	M	30-Aug-20	24-Aug-21	359	356	Y
BS135	F	18-Apr-19	26-Aug-20	496	< 496	N

Photo-ID validation

Dorsal spot patterns of T. lymma were found to be unique among the individuals examined in this study. No notable differences were observed between male and female spot patterns. The ROC analysis showed that individual spot patterns of T. lymma were sufficiently different to successfully discriminate between individuals using the I³S software. With an AUC value of 0.98, the I³S software demonstrated a high level of accuracy in its ability to successfully identify T. lymma individuals using dorsal spot patterns (Fig. 2). Photo-ID using I³S had an optimum match score (i.e., threshold score) of 8.39, below which there was a 2.96% false-positive detection rate and true-positive rate of 90.32%.

Multiple physical features for each individual (such as the number, shape, and relative position of spots) indicate that the overall dorsal spot pattern of T. lymma remains stable over time. There was a notable exception with the individual with the longest recapture interval (496 days) which displayed observable changes in the number of spots. For example, a relatively small spot (< 3 mm diameter) appearing on the cranium and a spot (< 5 mm) on the outer pectoral disc disappearing (Fig. 3). These changes were not observed in other individuals that were recaptured and rephotographed (Supplementary Fig. 3). The match score between photographs of the longest recapture interval identified using I³S was 1.03, which was below the threshold score (Fig. 2A). The I³S mean match score for all recaptured individuals was 3.77 (± 0.5 SE). A significant positive correlation was found between disc width and mean spot diameter in T. lymma (r = 0.81, p < 0.001).

The current study establishes that photo-ID can be used effectively to identify T. lymma individuals using dorsal spot patterns. Our results reveal a high degree of individual variation in T. lymma spot patterns (Supplementary Fig. 4). I³S photo-matching software correctly identified 161 different T. lymma individuals and proved to be an effective method to rapidly recognise individual spot patterns among T. lymma individuals. The physical recapture of individuals demonstrated that the overall pattern of larger spots persists and allows for accurate identification for at least 496 days. Given the longevity of spot patterns and the high probability of tag loss, we suggest photo-ID as a preferable, non-invasive alternative to traditional tagging methods for individual identifications of T. lymma.

The implementation of computer-aided identification with I³S confirmed that the spot patterns of T. lymma were unique and distinguishable, and validated its use for photo-ID of individuals. With a true-positive detection rate exceeding 90%, the results of our study affirm the reliability of I³S as a photo-matching tool. These findings are consistent with previous studies which also used I³S to identify individual teleost fish such as Paracirrhites forsteri²³ and Stereolepis gigas¹⁰, and more commonly elasmobranchs such as A. narinari, Carcharodon carcharias, and Carcharias taurus^12,24,25. Among these studies focusing on elasmobranchs, only C. taurus achieved a higher true-positive detection rate of 92% exceeding that established here, highlighting the suitability of T. lymma for photo-ID based identification. Our study also establishes a threshold score of 8.39 for the identification of T. lymma individuals, where match scores below this value could be considered true-positive matches (Fig. 2). This threshold score is also comparable to those previously established for other species of the family Dasyatidae, such as M. microps (10.06¹³).

Assessing the long-term applicability of photo-ID requires knowledge of the degree of change in T. lymma dorsal markings over time. Of the 29 individuals that were recaptured, only one individual had an observable change in the spot pattern (Fig. 3; adult female). However, this change was only observed in smaller spots (< 5 mm diameter) located on the outer edge of the disc and on the cranium, both of which had not been selected for the I³S fingerprint file. Despite this, the larger spot patterns in the centre of the disc remained unchanged, allowing for successful re-identification of the individual for up to 16 months using I³S. Specifically, the image of the re-captured individual had the lowest match score to the image from the initial capture event, and retained its position as the most likely match within the database. This emphasises the importance of considering the stability of individual spots, as well as the role of larger and more central spots in maintaining the overall stability of the pattern for long-term identification. While the positive correlation between spot size and disc width suggests that photo-ID may be more reliable for adult individuals of T. lymma, it is unclear whether spot patterns change with age or developmental stage. For example, some species, such as the reticulate whipray, Himantura uarnak, appears to gain new spots with age instead of having its existing spots grow in size¹⁴. To fully understand population dynamics and ontogenetic distribution shifts, it will be essential to confirm the stability of spot patterns over longer durations (i.e., > 500 days). While this study was conducted on wild populations of T. lymma, it would be valuable to independently validate these findings in a captive setting, such as those in aquaria. Based on these observations, it is advisable to use only spots larger than 5mm in diameter and those located in the centre of the disc to ensure matching accuracy and minimise errors due to potential spot pattern changes over time. Due to the limited tag retention times observed in the current study, alternative marking methods (such as, microchips or Passive Integrated Transponder (PIT) tags^26,27) should also be considered in future studies to assess spot-stability in T. lymma over greater time scales.

While substantial scars or deformities can serve as secondary features for identification (e.g., flapper skate, Dipturus intermedius²⁶), the high healing capacity of T. lymma means that scar tissue cannot be relied upon as a defining feature for photo-ID. Anecdotally, in the current study recaptured individuals did not have any visible scarring at the tag attachment site three months after losing their tag. It is also important to note that injuries to the dorsal surface may not result in long-term scarring but may still have the capacity to alter boundaries of individual spots (see Supplementary Fig. 5). Though injuries may temporarily alter spot boundaries, they do not affect the relative position of spots or their pattern, and thus will not impact the placement of points in the fingerprint files and subsequent I³S matches, as the software relies solely on the central position of the spots rather than their shape.

Photo-ID offers several advantages over conventional tagging methods for T. lymma. Firstly, the physical risks associated with collecting stingrays for tagging, including the potential harm from their barbed tails, pose a significant challenge to both researchers and the animals. Additionally, the physical attachment of the tags can restrict movement and cause capture/handling stress, potentially impacting the survival or behaviour of the individual^28,29. In contrast, photo-ID will not always require stingrays to be captured, thus avoiding the risks and impacts associated with traditional tagging methods when identifying individuals. It is still important to note, however, that the successful application of photo-ID in the wild is dependent on several factors, such as the clarity of the water, the behaviour of the animals towards observers, and the accessibility of the population in terms of depth and habitat³⁰. Nevertheless, photo-ID is likely to be more cost-effective than conventional tagging methods, particularly when the cost of tagging equipment and the labour required to attach and monitor the tags are considered³¹. Additionally, since tags can be shed, photo-ID is more effective at identifying individuals over longer durations as the dorsal spot pattern of T. lymma appears to be stable over the course of the study. However, future research is needed to evaluate the factors that may affect dorsal spot pattern stability in T. lymma and to determine which parts of the body future photo-ID efforts should focus on to ensure maximum pattern stability. Overall, photo-ID offers a safer and more cost-effective alternative to traditional tagging methods, making it a more advantageous method for individual identification of T. lymma, especially in areas where the collection of organisms is prohibited. Lastly, photo-ID can provide longer monitoring time scales while avoiding the negative impacts associated with invasive methods, making it a suitable option for long-term research on this species.

The current study highlights the potential for computer-aided photo-ID to improve the efficiency of long-term monitoring of T. lymma populations. Future applications of photo-ID may not require physical capture as photographs of T. lymma can be easily taken alongside common research methods, such as dive surveys^10,28,29 or aerial surveys over shallow-water habitats²⁰ (Supplementary Fig. 1). With increasing capabilities in machine learning, the process of identifying individual animals from photographs could likely be fully automated, making large-scale photo-ID projects even more feasible. Collaborative efforts between researchers and local stakeholders can also expand the spatial coverage of data collection and complement non-invasive monitoring programmes to increase knowledge on the biology and conservation of this species³¹. Recreational scuba divers or fishermen can also help collect photo identification data on stingrays by participating in community science programs or by contributing their own photographs to ongoing research projects, similar to MantaMatcher³¹ and the Common skate photo-ID database for Scotland³². Overall, by expanding photo-identification databases in both size and duration, the performance of photo-matching software will improve, as there will be increased availability of matches to draw upon^24,33.

The present study has validated the accuracy and reliability of using natural dorsal spot patterns to identify T. lymma individuals. Our data suggests that T. lymma can be reliably recognised using photographs over at least 496 days. Considering tag retention rates and the potential impacts from tag application, the use of photo-ID can provide more reliable information with minimal impacts to individuals within a population. Long-term studies involving T. lymma can further validate the use of photo-ID to monitor this species over longer periods and could potentially be expanded to include other species with similar pigmentation patterns (e.g., bluespotted maskray, Neotrygon kuhlii; and Oceania fantail ray, Taeniura lessoni)^17,34. With the validation of this method for T. lymma, long-term monitoring using photo-ID can provide new insights about the population trends, behaviour, and biology of T. lymma to better inform their management.

This study complied with the Kingdom of Saudi Arabia's animal welfare laws, guidelines and policies as approved by the King Abdullah University of Science and Technology (KAUST) Institutional Animal Care and Use Committee (IACUC) under approval #18IACUC14.

Collections and tagging

Taeniura lymma were collected at two locations in the Saudi Arabian Red Sea where the species is locally abundant^20,35: Thuwal (22.3245, 39.0894) and the Farasan Banks (19.6207, 40.6067) (Fig. 1). The collection site in Thuwal lies along a fringing coastal reef system backed by a shallow, sandy lagoon and mangrove forest in the central Red Sea. Collections in the Farasan Banks were conducted off small islands (Dubarah, Thara, and Sirryan), which also contained both mangrove and coral reef systems, in the southern Red Sea.

Hand-seine nets (1.5 x 15m) were used to capture T. lymma individuals in shallow water environments (< 1 m deep) and dip nets were used in deeper environments (> 1 m deep) while freediving. Once captured, T. lymma were transferred into a dip net if captured by hand-seine, and the barb(s) were immediately secured using fabric cloth. Disc width (DW; mm), sex, identifying features (e.g., injury), and capture location were documented for each individual that was collected. Individuals were then briefly removed from the water and placed on a flat surface (< 30 seconds), where 2 to 3 photographs were taken perpendicular to the dorsal surface using an Olympus TG waterproof camera, positioned above a metric measurement strip which was visible in every photo. All T. lymma individuals > 20 cm DW collected in Thuwal were fitted with polyolefin dart-style identification tags (Floy Tag Inc., Seattle, USA). Tags displayed unique identification numbers and were implanted into pectoral musculature away from internal organs. Each T. lymma individual was promptly released at their capture site following morphological measurements and photographs.

Photo-Identification

Photographs were retrospectively examined to select the image with the clearest spot patterns per individual, which was used as the final photo-ID shot for subsequent analyses. Only one image per individual was added to the main photo-ID database, while any additional images taken during the same capture event or upon recapture and identification from tags were retained in the main stingray database (Supplementary Fig. 6). To assess the accuracy of manual spot-pattern identification in identifying known individuals of T. lymma, a validation test-set consisting of 20 individuals that were recaptured, re-photographed, and already present in the main photo-ID database was created. These individuals were considered “known” due to identification from their physical tags, allowing the applicability of manual spot-pattern identification to be assessed. Each image from the validation test-set was visually compared against every image in the main photo-ID database in an iterative pairwise comparison process. Visual comparison of spot shape, their relative positions, and quantity were used as identifiable features during this process. The initial pairwise comparisons were verified by two independent researchers who conducted randomised blind trials on the validation test-set against every image in the main photo-ID database, consistent with methods used in previous photo-ID studies ^13,26. Once the applicability of individual identification was confirmed, this manual validation process was conducted for all images in the photo-ID database to check for duplicate or misidentified individuals in the database to account for possible recapture events of untagged rays (i.e., individuals < 20 cm DW) or tag losses.

The photo-matching software, I³S Classic (Version 4.0.2³⁶), was used to analyse a digital library of T. lymma photographs. Fingerprint files were created in I³S for each individual by defining an identification area based upon three fixed reference points on the dorsal surface: tip of rostrum, left pectoral insertion and right pectoral insertion (Supplementary Fig. 7). These points of reference were chosen as there is no clear apex to the pectoral disc, and this area is less prone to flex with pectoral movement ¹³. The centre of 30 dorsal spots were user-selected in the software interface within the identification area to create a fingerprint file for each individual. To avoid selecting very small spots, the selection process began with all conspicuous spots on the cranium and then continued with those radiating from the centre of the cranium that were anterior of the eye stalks (Supplementary Fig. 7).

To evaluate the performance of I³S in matching photos of T. lymma individual, we tested rates of true-positive and false-positive identifications. In this instance, a true-positive can be confirmed when two images are of the same T. lymma individual, such as an image taken from a recapture event identified from the physical tag or a consecutive image of the same individual from the same capture event. In contrast, a false-positive would occur when the algorithm incorrectly matches two images of different individuals, such as erroneously pairing an image taken in Thuwal with one taken in the Farasan Banks. Given that the home-range of T. lymma is assumed to be relatively small³⁷, T. lymma captured in Thuwal and those from the Farasan Banks were considered to be different individuals as these locations are approximately > 350 km apart, making it highly unlikely for T. lymma to travel such distances. We developed two test-sets to evaluate the performance of the algorithm: the database test-set and the error test-set. The database test-set was compiled using one image of each T. lymma individual captured in Thuwal directly from the main database (n = 128). The error test-set consisted of a subset of T. lymma individuals that were included in the database test-set, but with distinct images that were not used in the database test-set. This included photographs from recapture events (n = 31) and consecutive photos (n = 30) taken of the same individual to serve as examples of true-positives. In addition, the error test-set contained images of T. lymma individuals that were not included in the training database, such as those captured in the Farasan Banks (n = 15), which serve as examples of false-positives. The database test-set was used as the training dataset for the algorithm, with the error test-set used to evaluate the algorithm's ability to distinguish true-positives from false-positives.

Fingerprint files for each individual in the accuracy test-set were queried against the photo-ID training database. I³S numerically ranks images within the database against the queried image based on how closely the fingerprint files match in terms of relative distance between the user-selected spots. The resulting match score of each image describes how closely the spot patterns match the query image, with a higher match score indicating a closer match. Once matched, a thorough manual comparison was conducted to assess ‘true-positives’ (matched images of T. lymma known to be the same individual) or ‘false-positives’ (known to be different).

The efficacy of I³S software to successfully match T. lymma individuals was assessed using a receiver operating characteristic (ROC) curve from the ROCR packaging³⁸ of R version 4.0.3³⁹. The ROC analysis compared the rate of ‘true-positives’ against ‘false-positives’. A cut-off value was identified as the point at which the rate of ‘true-positives’ was highest and ‘false-positive’ was lowest using Youden’s index. This cut-off value indicates the optimal threshold that should be considered when discriminating T. lymma individuals. The proportion of true-positive and false-positive identification rates were determined by analysing the matches from I³S that were below the aforementioned cut-off value. The true-positive rate represents the percentage of correct matches, while the false-positive rate represents the percentage of matches that were incorrect. Overall accuracy was assessed using the area under curve (AUC), where values close to one indicate higher accuracy of I³S software to correctly match images⁴⁰. To determine the probability of I³S returning a ‘true-positive’ within the first five matched results against the query image, we used the I³S “Elaborate Evaluation” tool⁴¹.

Spot pattern stability was assessed by the ability of I³S to correctly match photographs of recaptured rays between the time of first, any subsequent, and final capture events, and was reported as the mean match score of all recaptured individuals. To account for potential changes through time, the relationship between spot pattern and disc width was explored using adapted methods and parameters outlined by Borsa et al.^14,42 for a subset of 64 randomly selected individuals. The mean dorsal spot diameter was calculated in ImageJ⁴³ using the known disc width of the individual to relatively measure the diameter of spots that were on the pectoral transect line in millimetres (i.e., mid-scapular point to the extremity of the right pectoral fin; Supplementary Fig. 8). The Pearson correlation method was used to measure a linear dependence between mean spot diameter and disc width and to plot a linear regression curve. A significance level of α = 0.05 was considered for all analyses. Normality and heteroscedasticity testing was conducted using Shapiro-Wilk and Levene’s tests, respectively.

Acknowledgments

The authors are grateful to the members of the Reef Ecology Lab, King Abdullah University of Science and Technology, foremost to Charlotte Young, Lyndsey Tanabe, and Alexander Kattan for their assistance in the field, and to Ute Langer for their preparation of the included map.

Author contributions

Conceptualization, A.M.; Data curation, A.M, C.W., W.R. and A.K.; Funding acquisition, M.B.; Methodology, A.M., N.B.; Supervision, M.B.; Validation, C.W., W.R., A.K., N.B.; Writing—original draft, A.M.; Writing—review and editing, A.M., C.W., W.R., A.K., N.B., M.B. All authors have read and agreed to the submitted version of the manuscript.

Data Availability

All data generated during this study are available via GitHub: https://github.com/AshlieJMcivor/Taeniura_PhotoID_analysis. Photographs of all specimens are available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.

Thompson, P. M. & Wheeler, H. Photo-ID-based estimates of reproductive patterns in female harbor seals. Mar. Mammal Sci. 24, 138–146 (2008).
Araujo, G. et al. Using minimally invasive techniques to determine green sea turtle Chelonia mydas life-history parameters. J. Exp. Mar. Bio. Ecol. 483, 25–30 (2016).
Schofield, G. et al. Long-term photo-id and satellite tracking reveal sex-biased survival linked to movements in an endangered species. Ecology 101, (2020).
Dinis, A. et al. Large-scale movements of common bottlenose dolphins in the Atlantic: Dolphins with an international courtyard. PeerJ 9, (2021).
Marshall, A. D. & Pierce, S. J. The use and abuse of photographic identification in sharks and rays. J. Fish Biol. 80, 1361–1379 (2012).
Carpentier, A. S., Jean, C., Barret, M., Chassagneux, A. & Ciccione, S. Stability of facial scale patterns on green sea turtles Chelonia mydas over time: A validation for the use of a photo-identification method. J. Exp. Mar. Bio. Ecol. 476, 15–21 (2016).
Dunbar, S. G. et al. HotSpotter: Using a computer-driven photo-id application to identify sea turtles. J. Exp. Mar. Bio. Ecol. 535, (2021).
Landeo-Yauri, S. S., Ramos, E. A., Castelblanco-Martínez, D. N., Niño-Torres, C. A. & Searle, L. Using small drones to photo-identify Antillean manatees: A novel method for monitoring an endangered marine mammal in the Caribbean Sea. Endanger. Species Res. 41, 79–90 (2020).
Thornycroft, R. E. & Booth, A. J. Computer-aided identification of coelacanths, Latimeria chalumnae, using scale patterns. Mar. Biol. Res. 8, 300–306 (2012).
Love, M. S., Seeto, K., Jainese, C. & Nishimoto, M. M. Spots on Sides of Giant Sea Bass (Stereolepis gigas Ayres, 1859) are Likely Unique to Each Individual. Bull. South. Calif. Acad. Sci. 117, 77–81 (2018).
Kitchen-Wheeler, A. M. Visual identification of individual manta ray (Manta alfredi) in the Maldives Islands, Western Indian Ocean. Mar. Biol. Res. 6, 351–363 (2010).
González-Ramos, M. S., Santos-Moreno, A., Rosas-Alquicira, E. F. & Fuentes-Mascorro, G. Validation of photo-identification as a mark–recapture method in the spotted eagle ray Aetobatus narinari. J. Fish Biol. 90, 1021–1030 (2017).
Keeping, J. A. et al. Computer‐aided photo‐identification of a rare stingray, Megatrygon microps. J. Fish Biol. 96, 815–819 (2020).
Borsa, P. et al. Neotype designation and re-description of Forsskål’s reticulate whipray Himantura uarnak. Mar. Biodivers. 51, 1–10 (2021).
Last, P. R. et al. Rays of the World. vol. 6 (CSIRO publishing, 2016).
Pereira, N., Batista, H. & Baylina, N. Ultrasound assessment of pregnant ribbontail stingrays, Taeniura lymma (Forsskål, 1775). in The Elasmobranch Husbandry Manual II: Recent Advances in the Care of Sharks, Rays and their Relatives (eds. Smith, M. et al.) 325–330 (Ohio Biological Survey, 2017).
Sherman, C., Heupel, M., Moore, S., Chin, A. & Simpfendorfer, C. When sharks are away, rays will play: effects of top predator removal in coral reef ecosystems. Mar. Ecol. Prog. Ser. 641, 145–157 (2020).
Simpfendorfer, C. A. et al. Widespread diversity deficits of coral reef sharks and rays. Science. Accepted, (2023).
Sherman, C. S. et al. Taeniura lymma. The IUCN Red List of Threatened Species 2021: e.T116850766A116851089. The IUCN Red List of Threatened Species 2021: e.T116850766A116851089 https://dx.doi.org/10.2305/IUCN.UK.2021-1.RLTS.T116850766A116851089.en (2021).
McIvor, A. J., Spaet, J. L. Y., Williams, C. T. & Berumen, M. L. Unoccupied aerial video (UAV) surveys as alternatives to BRUV surveys for monitoring elasmobranch species in coastal waters. ICES J. Mar. Sci. 79, 1604–1613 (2022).
Madduppa, H., Ayuningtyas, R. U., Subhan, B., Arafat, D. & Prehadi, P. Exploited but unevaluated: DNA barcoding reveals skates and stingrays (Chordata, Chondrichthyes) species landed in the Indonesian fish market. Ilmu Kelaut. Maret 21, 29–36 (2016).
Asut, H., Herman, H., Junianto, J. & Dewanti, L. P. Analysis of stingray catches which landed in fish landing site of Labuan Bajo, West Manggarai Regency of East Nusa Tenggara. World News Nat. Sci. 24, (2019).
McInnes, M. G. et al. A new model study species: high accuracy of discrimination between individual freckled hawkfish (Paracirrhites forsteri) using natural markings. J. Fish Biol. 96, 831–834 (2020).
Van Tienhoven, A. M., Den Hartog, J. E., Reijns, R. A. & Peddemors, V. M. A computer-aided program for pattern-matching of natural marks on the spotted raggedtooth shark Carcharias taurus. J. Appl. Ecol. 44, 273–280 (2007).
Andreotti, S. et al. Semi-automated software for dorsal fin photographic identification of marine species: application to Carcharodon carcharias. Mar. Biodivers. 48, 1655–1660 (2018).
Benjamins, S. et al. Evaluating the potential of photo-identification as a monitoring tool for flapper skate (Dipturus intermedius). Aquat. Conserv. Mar. Freshw. Ecosyst. 28, 1360–1373 (2018).
Rigby, C. et al. Rapid Assessment Toolkit for Sharks and Rays. (WWF International and CSTFA, James Cook University, 2019).
Cerutti-Pereyra, F. et al. Observations of spotted eagle rays (Aetobatus narinari) in the Mexican Caribbean using photo-ID. Environ. Biol. Fishes 101, 237–244 (2018).
Boggio-Pasqua, A., Flam, A. L. & Marshall, A. D. Spotting the ‘“small eyes”’: Using photo-ID methodology to study a wild population of smalleye stingrays (Megatrygon microps) in southern Mozambique. PeerJ e7110 (2019) doi:10.7717/peerj.7110.
Pierce, S., Holmberg, J., Kock, A. & Marshall, A. Photographic identification of sharks. in Shark research: emerging technologies and applications for the field and laboratory (ed. Carrier, J. C.) 220–234 (CRC Press, 2018).
Marshall, A. & Holmberg, J. MantaMatcher Photo-identification Library. https://www.mantamatcher.org/ (2020).
NatureScot & SAMS. Common skate photo-identification database for Scotland. COSPIDS Photo-identification Library https://skatespotter.sams.ac.uk/ (2017).
Calmanovici, B., Waayers, D., Reisser, J., Clifton, J. & Proietti, M. I3S Pattern as a mark-recapture tool to identify captured and free-swimming sea turtles: An assessment. Mar. Ecol. Prog. Ser. 589, 263–268 (2018).
Last, P. R., White, W. T. & Naylor, G. Three new stingrays (Myliobatiformes: Dasyatidae) from the Indo-West Pacific. Zootaxa 4147, 377–402 (2016).
Bonfil, R. Consultancy on elasmobranch identification and stock assessment in the Red Sea and Gulf of Aden. (2003).
Hartog, J. E. & Reijns, R. A. Interactive Individual Identification Software (I3S). Version 4.2.0. http://www.reijns.com/i3s (2014).
Udyawer, V. et al. Scaling of activity space in marine organisms across latitudinal gradients. Am. Nat. (2022) doi:10.1086/723405.
Sing, T., Sander, O., Beerenwinkel, N. & Lengauer, T. ROCR: Visualizing classifier performance in R. Bioinformatics 21, 3940–3941 (2005).
R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing vol. 2 https://www.R-project.org (2020).
Hanley, J. A. & McNeil, B. J. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 143, 29–36 (1982).
Hartog, J.D. & Reijns, R. A. I3S Manual: Interactive Individual Identification System. (2014).
Borsa, P. et al. Himantura tutul sp. nov. (Myliobatoidei: Dasyatidae), a new ocellated whipray from the tropical Indo-West Pacific, described from its cytochrome-oxidase i gene sequence. Comptes Rendus - Biol. 336, 82–92 (2013).
Schneider, C. A., Rasband, W. S. & Eliceiri, K. W. NIH Image to ImageJ: 25 years of image analysis. Nat. Methods 9, 671–675 (2012).

No competing interests reported.

TaeniuraPhotoIDSuppMat.docx

Download PDF

Journal Publication

published 29 Jul, 2024

Read the published version in Scientific Reports →

Reviewers agreed at journal
04 Sep, 2023
Reviews received at journal
07 Aug, 2023
Reviewers agreed at journal
28 Jul, 2023
Reviewers invited by journal
05 Jul, 2023
Editor assigned by journal
27 Jun, 2023
Editor invited by journal
22 May, 2023
Submission checks completed at journal
22 May, 2023
First submitted to journal
18 May, 2023

You are reading this latest preprint version

Mark-recapture validates the use of photo-identification for the widely distributed blue-spotted ribbontail ray, Taeniura lymma

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Tagging and recapture events

Photo-ID validation

Discussion

Conclusion

Methods

Collections and tagging

Photo-Identification

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1