The hawksbill turtle, Eretmochelys imbricata, is distributed in tropical and subtropical areas of the Atlantic, Pacific, and Indian oceans and is currently listed globally as Critically Endangered by the International Union for Conservation of Nature (IUCN 2018). The Caribbean is a critical region for hawksbill turtles that includes nesting beaches, coral reef ecosystems, fringing and patch reef habitats, and associated seagrass habitats that are important environments for mating, development, and foraging for this species (Meylan et al. 2011; Campbell 2014). Hawksbill turtles are thought to be major predators in coral reef ecosystems, with sponges constituting the main component of their diet (Meylan 1988; Anderes and Uchida 1994; León and Díez 1999; Rincón-Díaz et al. 2011). This species plays a key role in maintaining the structure, ecology, and evolution of coral reefs in the region (León and Bjorndal 2002; Bjorndal and Jackson 2003).

Similar to other sea turtle species, hawksbill hatch-lings enter an initial pelagic phase, where they spend several years in the high seas until they recruit to neritic foraging habitats as small juveniles. Juvenile hawksbills reside in these areas for extended periods of time (van Dam and Díez 1998; León and Díez 1999; Blumenthal et al. 2009b), eventually recruiting into adult reproductive populations (Velez-Zuazo et al. 2008; Hawkes et al. 2014). Knowledge about juvenile hawksbill aggregations in the region indicates that size distribution, somatic growth rates, habitat use, and feeding habits are highly variable among sites and are habitat dependent (León and Díez 1999; Díez and van Dam 2002; Blumenthal et al. 2009b; Bjorndal and Bolten 2010; Krueger et al. 2011; Hawkes et al. 2014; Bjorndal et al. 2016). This highlights the importance of conducting research into demographic parameters for different populations and foraging grounds.

In the Caribbean, hawksbill turtle foraging aggregations are mixed stocks, composed of individuals originating from multiple, genetically distinct nesting subpopulations from throughout the region (Moncada et al. 1998; Bowen et al. 2007; Mortimer et al. 2007; Blumenthal et al. 2009a; Browne et al. 2010). The waters of the Cuban archipelago constitute one of the most important foraging grounds for hawksbills in the region, as haplotypes of this aggregation contribute most heavily to all Caribbean foraging populations (0%–44.7%) including that in Cuba (43.4%–72.0%; Bowen et al. 2007). It is estimated that up to 55% of foraging hawksbills in Cuba originate from other Caribbean countries (Bahamas, Dominican Republic, Puerto Rico, and the US Virgin Islands; Bowen et al. 2007).

The extensive shallow fringing reefs of the Jardines de la Reina Archipelago (JdlRA) provide important foraging habitats for both juvenile and adult hawksbill turtles and constitute the core area of distribution for the species in the Cuban archipelago (Moncada et al. 1999, 2010). Nonetheless, scarce information is available for hawksbills assembled in this foraging and developmental region. Here, we analyze data resulting from an intensive tagging and monitoring program of this species carried out between 1992 and 2009. Data are presented on hawksbill spatial distribution, site fidelity, individual size composition, and somatic growth rates for this Cuban population.

METHODS

Study Site. — The JdlRA is the largest contiguous marine reserve in the Caribbean, and its neritic habitats are among the best preserved in the region, including a variety of coral reef, sea grass, and mangrove ecosystems (Appeldoorn and Lindeman 2003). The reserve is located ∼ 50 km off the southeastern coast of Cuba (20°86′732″N, 79°03′969″W); it extends over ∼ 360 km and includes ∼ 661 low-lying islands, or cays (Fig. 1). The majority of the cays have beaches, interior lagoons, and abundant coral reefs, forming part of the Labyrinth Cays of Doce Leguas inside the JdlRA. The reserve harbors well-developed fringing reefs, mainly in the southern region, that offer important foraging habitat for hawksbill turtles. In addition, the beaches of the cays are important for hawksbill nesting (Moncada et al. 2010).

View Figure

• Download as Powerpoint
• Download Figure

Surveys for juvenile hawksbills were carried out at 6 external cays of JdlRA (Cayos Cinco Balas, Cayo Grande, Cayo Caballones, Cayo Anclitas, Cayo Cachiboca, and Cayo Boca Seca, distributed within of Labyrinth Cays of Doce Leguas region) and 3 internal cays (Cayo Rabihorcado, Cayos Paloma, and Cayo Cuervo) located north of the Labyrinth Cays of Doce Leguas (Fig. 1). These sites were selected based on previous knowledge obtained through surveys of fishermen and from available data on captures during legal turtle fisheries before the activity was banned in the region.

Data Collection. — In-water hawksbill surveys at potential foraging sites on the reef systems were conducted from 1992 to 2009. Between 1992 and 1995, exploratory in-water hawksbill surveys were conducted at each of the 9 external cay study sites (Fig. 1). Starting in 1996, Cayo Grande and Cayo Caballones were selected as the primary study areas due to better accessibility to perform the work and the higher abundance of juveniles detected there. Surveys were typically conducted 2–5 times per month during all months of the year, although most surveys were carried out between October and March. The survey team generally consisted of 2–4 divers, and turtles were captured using 2 methodologies: 1) hand capture through free diving or with self-contained underwater breathing apparatus (SCUBA), and 2) using entanglement nets made of monofilament (180 × 2 m, 90-mm knot-to-knot mesh size). Nets were deployed during daylight hours within the vicinity of reefs and parallel to the coastline (5–500 m); water depths at netting sites were ≤ 2.5-m depth and nets were monitored at 30-min intervals or more frequently as splashes or other signs of potential capture dictated. Nets were always deployed at the same locations at the 2 sites.

Once captured, turtles were tagged on the trailing edge of the front flipper using numbered monel or titanium tags. Biometric measurements were taken, including curved carapace length (CCL_n–t; from anterior notch to posterior-most tip) and curved carapace width (CCW), using a tape measure to the nearest 0.1 cm. After processing, turtles were released as close as possible to the site of original capture; capture locations were recorded for all turtles.

Data Analysis. — For size analysis, the size range (CCL) of turtles encountered was separated into 5-cm size classes for each study site. Turtles < 65 cm CCL were considered immature (Witzell 1983), as were other turtles of larger size that did not exhibit secondary sexual characteristics (see Fig. 2). Subadults, although often not yet displaying secondary sexual characteristics, will already show some distinctive characteristics (e.g., the carapace is smooth except for overlapping scutes, which show growth bands).

View Figure

• Download as Powerpoint
• Download Figure

Local movements were determined from the distance traveled, calculated as a straight line between tagging and recapture sites, using Global Positioning System (GPS) MapSource software v.3.02 (1999, Garmin, Olathe, KS) and the elapsed time between captures. Somatic growth rates were estimated as the change in CCL between captures divided by the elapsed time (Mariscal Loza 2008). To avoid measurement errors in growth increments due to short intervals, we carried out growth rate estimations only for turtles with recapture intervals ≥ 60 d (Hawkes et al. 2014). Somatic growth data are summarized for each 10-cm size class.

The comparison of sizes among the study areas was carried out via analysis of variance (ANOVA), given that the data complied with the premises of normality and homogeneity of variance. All statistical analyses were conducted with R.3.4.3 (R Development Core Team 2019) with a confidence level of p ≤ 0.05.

RESULTS

From 1992 to 2009, we captured 496 hawksbill turtles (juvenile and subadults) at JdlRA, which were measured, flipper tagged, and released soon after capture (Fig. 1). Most of the turtles (73%) were captured using nets, whereas the remaining captures (27%) were carried out via free diving or SCUBA. Most turtles (n = 454) were not seen again; 42 turtles were recaptured between 1 (n = 27), and 4 (n = 1) times. Although not included in this study, green turtles (Chelonia mydas) were occasionally captured during our field efforts.

Size Distribution. — Among all hawksbills studied, size at initial capture ranged 19.0–68.0 cm CCL, with the smallest turtle (19 cm CCL) captured at Cayo Cinco Balas. Small juveniles with CCLs of 26–30 cm were the most frequent size class of turtles encountered across all study sites (Fig. 2A). Mean ± standard devision (SD) CCL for hawksbills captured in the external cays was 35.6 ± 9.6 cm (n = 487), whereas that for hawksbills captured at the internal cays was 59.6 ± 7.7 cm (n = 9; Fig. 2B). In the internal cays, only large juveniles (> 40 cm CCL) were encountered, with turtles larger than 60 cm CCL accounting for 66.7% of all captures at these sites (Fig. 2B).

Recaptures and Movement. — Over the duration of the study, 42 juvenile and subadult hawksbills were recaptured 1–4 times: 27 (64.3%) turtles were recaptured once, 10 (23.8%) were recaptured twice, 4 (9.5%) were recaptured 3 times, and 1 (2.4%) was recaptured 4 times. Local recaptures occurred at time intervals (i.e., time-at-large) ranging from 15 d to 8.4 yrs (mean ± SD = 559.4 ± 533.5 d). Time elapsed between the first and last capture of individual turtles had a median of 1.01 yrs, with a median period between sequential captures of 0.92 yrs (see Fig. 3 for time-at-large summaries).

View Figure

• Download as Powerpoint
• Download Figure

Most of the recaptured hawksbills showed limited movements away from their first capture site. Thirty-one hawksbills—all within the Doce Leguas Cays—remained in practically the same location as initial capture (range = 0–1 km); 4 individuals were 3 km from their site of original capture, 2 were from 3 to 6 km away, and 1 was recaptured between 6 and 9 km distance from site of initial capture. One hawksbill moved out of the Doce Leguas Cays (125 km) but into the JdlRA. There was no correlation between distance traveled and turtle CCL at initial capture (Spearman's r = 0.09, p > 0.05). Four of the tagged hawksbills were recaptured outside of Cuba: 2 were recaptured in Nicaragua (1000 km), 1 in Colombia (2500 km; see Moncada et al. 2012), and 1 was observed nesting in Barbados (Moncada et al. 2020).

Growth and Morphometrics. — Among juvenile turtles that were recaptured, the smallest size at initial capture was 24.0 cm CCL and the largest size at recapture was 55.0 cm CCL (Fig. 4A). In addition, 1 subadult turtle was recaptured, measuring 61.0 cm at initial capture and 77.0 cm at final capture.

View Figure

• Download as Powerpoint
• Download Figure

Overall, juvenile hawksbill turtles grew at an average of 7.4 cm yr^–1 (mean ± 4.0 SD, range 0.4–22.3; Fig. 4B). The greatest rate of growth (22.3 cm yr^–1) was extrapolated for 1 turtle measuring 24.0 cm CCL at initial capture, which was recaptured after 229 days at liberty and had gained 14.2 cm in carapace length. The greatest recorded change in carapace length between captures was observed for a hawksbill that was recaptured after being at large for 1050 days. This turtle's initial and final CCLs were 33.1 cm and 55.0 cm, respectively, which amounted to a 21.9-cm difference between initial and final capture, and was equivalent to an average growth rate of 7.61 cm yr^–1.

Somatic growth rates varied significantly with size at initial carapace (generalized linear mixed models [GLMM]: X1 = 10.9, p < 0.05), with smaller turtles growing faster than larger turtles (Fig. 4). For example, hawksbills between 20.0 and 29.9 cm CCL at initial capture grew on average at 8.2 cm yr^–1, whereas turtles larger than 40.0 cm CCL grew at 4.6 cm yr^–1 (Fig. 4).

DISCUSSION

This article provides the first description of a significant foraging aggregation of critically endangered hawksbill turtles within Cuban waters. Development habitats for hawksbills include sites where juveniles establish a home range within which they feed, rest, and obtain appropriate shelter (Meylan and Meylan 1997; Cuevas et al. 2007). The results of this long-term study around the JdlRA demonstrate that the archipelago constitutes an important area of recruitment and development for juvenile and subadult hawksbill turtles, respectively.

Immature hawksbills were found in all sampled cays, from Cinco Balas to Boca Seca, and the inshore cays of the archipelago (Rabihorcado, Palomas, and Cuervo; Fig. 1). Differences in size class size distribution were observed among study sites within the JdlRA. Recorded CCL for hawksbills captured at the inshore cays indicates that all but 1 turtle were 20 cm CCL or more, with the single smaller individual measuring 19 cm CCL. This size threshold is consistent with the size at recruitment from pelagic zones into benthic foraging habitats for hawksbills elsewhere in the Caribbean and beyond (Boulon 1994; Musick and Limpus 1997). The data also indicate that the demographic structure of hawksbills in the JdlRA is consistent with other foraging sites reported for the wider Caribbean region, such the Yucatan Peninsula, Mexico (Maldonado and Garduño 1999), Mona Island, Puerto Rico (van Dam and Díez 1998), Parque Jaragua, Dominican Republic (León and Díez 1999), and Bermuda (Meylan et al. 2011). Indeed, as has been described for these foraging grounds, the hawksbill size class distribution in the external cays of the JdlRA shows that most turtles are in the small juvenile size classes (peaking in the 25–30-cm range) and relatively fewer are larger, subadult turtles. This could indicate that, upon reaching larger sizes, hawkbills transition out of the JdlRA and move on to other regional foraging grounds, as has been suggested for a small juvenile-dominated population in Florida (Wood et al. 2013).

In contrast to the generally smaller size of hawksbills assembled at the external cays, individuals captured at the inshore cays of the JdlRA are typically larger subadults. The disparity in sizes of turtles at the external vs. inshore cays suggests a size, or life-stage–based, difference in habitat preference of individual hawksbills in the JdlRA, as has been observed for other sea turtle species and in other areas (Seminoff et al. 2002; Moncada et al. 2006). However, it is important to point out that, although in this study the largest individuals are located at the inshore cays, no mature turtles were encountered in these areas, which perhaps suggests yet a different habitat preference among adult hawksbills in the region.

Somatic growth rates among hawksbills in the JdlRA reefs were comparable to other Caribbean sites, although there appears to be substantial variability in somatic growth rates throughout the region for the species. Growth rates in the present study are similar to those observed in the Dominican Republic (León and Díez 1999), but higher than those reported for the Cayman Islands (Blumenthal et al. 2009b) and lower than rates observed in Anegada, British Virgin Islands (Hawkes et al. 2014). The significant differences in juvenile growth rates reported throughout the Caribbean region suggest that there are major variations in environmental quality among hawksbill foraging sites (even among some areas which are relatively nearby, within a single country). However, the mechanisms that influence somatic growth of hawksbills are still poorly understood. Díez and van Dam (2002) speculated that differences in food abundance and quality are responsible for the different growth rates they found at Mona and Monito islands, Puerto Rico. Water temperature and turtle population density can also be expected to influence an animal's growth rate. A regional study in the West Atlantic, where growth rates were similar for numerous hawksbill subpopulations, concluded that hawksbill somatic growth rates are likely driven by region-wide environmental forces (Bjorndal et al. 2016). However, recent hawksbill data from Avens et al. (2021) for turtles stranded along the United States coastline indicated that somatic growth varied significantly relative to size, age, and stranding location.

In the present study, short recapture intervals (60 d) were used when calculating annual somatic growth rates. Because somatic growth may vary seasonally within a year, we acknowledge that it is important to be cautious when using short-term recapture intervals to estimate annual growth rates. Bjorndal et al. (2016) used recapture interval for hawksbills of as low as 146 d, but only after verifying that doing so did not significantly influence the results of their growth models. As a result, the somatic growth rates reported here should be considered preliminary; our future growth studies will evaluate the efficacy of using short-term recapture intervals for measuring individual growth.

The low spatial displacement of individuals from their original capture sites observed for recaptured juveniles in the external cays could be related to their small size and the foraging habits of this life stage. Immature hawksbills are thought to show high site fidelity within their developmental habitats, maintaining a small home range for extended periods, with individuals selecting different microhabitats depending on their body size or ontogenetic stage (van Dam and Díez 1997; Cuevas et al. 2007; Blumenthal et al. 2009b). Hawksbills live in close association with coral reefs, and the home range extent of juveniles is likely heavily influenced by habitat quality (Scales et al. 2011). The JdlRA coral reefs are characterized as sandy or hard-bottom habitat, harboring sponges such as Chondrilla caribensis and Spirastrella coccinea, which are known to be preferred prey of hawksbills (Meylan 1988; Anderes and Uchida 1994; León and Bjorndal 2002; Berube et al. 2012). High site fidelity to small areas for long periods has also been described in other foraging areas for the species (León and Díez 1999; Blumenthal et al. 2009b; Witt et al. 2010; Chevis et al. 2017).

Flipper tagging and satellite tracking studies show that juvenile hawksbills can also migrate long distances and undertake transoceanic crossings to access distant foraging grounds (Whiting and Koch 2006; Grossman et al. 2007; Rainer et al. 2017). In this study, 4 individuals left Cuban waters: 2 were recaptured in Nicaragua and 1 in Colombia, with all 3 having been tagged in Cuba about 4 yrs prior to their long-distance recapture (see Moncada et al. 2012 for details). The remaining turtle was recaptured nesting in Barbados 14 yrs after having been tagged in the JdlRA (Moncada et al. 2020), providing further evidence that at least some juveniles use the archipelago as a benthic developmental habitat before traveling to elsewhere in the Caribbean.

The results of this work provide further knowledge about the spatial distribution and size-class composition of hawksbill turtles in the JdlRA. Our data also highlight the importance of the marine habitats of this region for recruitment, development, and feeding of juvenile and subadult hawksbills. Clearly, the reefs of the JdlRA constitute an appropriate area for this species in the Caribbean. Hence, the JdlRA should continue as an important reference site for research and conservation of Caribbean hawksbill turtles.