The leatherback sea turtle (Dermochelys coriacea) is protected by the national laws of Costa Rica (Ley de Conservación de Vida Silvestre No. 7317 in force since 1992) and is classified as critically endangered on the IUCN Red List of Threatened Species (Sarti 2000). Also relevant for Costa Rica is the fact that the species is listed on Appendix I of the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES), which Costa Rica ratified in 1975, prohibiting international trade in parts or products; protected by the Convention on the Conservation of Migratory Species of Wild Animals (or the Bonn Convention), to which both Costa Rica (2007) and Panama (1989), a range state of the Gandoca population, are party; and considered by the Convention on Nature Protection and Wildlife Preservation in the Western hemisphere to be a species for whom protection, “as completely as possible”, is of “special urgency and importance”. In 2000, after more than 3 decades of precedent (as the Western Hemisphere Convention entered into force in Costa Rica in 1967), Costa Rica ratified the Inter-American Convention for the Protection and Conservation of Sea Turtles (IAC), with a stated objective “to promote the protection, conservation and recovery of sea turtle populations and of the habitats on which they depend . . . ”. The national regulatory framework was significantly strengthened in November 2002 with passage by the national legislative assembly of law no. 8325 (Ley de Protección, Conservación y Recuperación de las Poblaciones de Tortugas Marinas), which implements the IAC, designates regulatory agencies, establishes penalties, and so on.

Notwithstanding the various levels of protection conferred on the species by national and international law, leatherbacks are believed to have lost more than 90% of their numbers in recent decades along the Pacific coasts of Central America and Mexico (Eckert and Sarti 1997; Spotila et al. 2000). Declines are attributed primarily to the poaching of eggs and gravid females, as well as the incidental capture of juveniles and adults in gill nets and longlines (Eckert and Sarti 1997; Spotila et al. 2000). Observed declines at some of the hemisphere's most important breeding colonies (e.g., Costa Rica, Mexico), as well as at smaller colonies in Guatemala, El Salvador, and Nicaragua, lent impetus and urgency to the IUCN designation of critically endangered. In contrast, some important nesting colonies outside the Pacific appear to be stable or increasing (Hughes 1996; Chacón 1999; Chevalier and Girondot 2000; Dutton et al. 2000).

Nesting is distributed globally from 40°N to 35°S (Eckert 2001), including the Caribbean shores of Costa Rica, where the highest density nesting grounds are located at Gandoca Beach, Pacuare, and Tortuguero (Berry 1987; Troëng et al. 2004). The species also nests in significant numbers in adjacent northeastern Panama (Meylan et al. 1985; Troëng et al. 2004; Ordoñez et al. 2007).

Nesting at Gandoca Beach (Fig. 1), included in the Gandoca-Manzanillo National Wildlife Refuge (legally designated in 1985), has been systematically monitored since 1990. In Gandoca, the most significant threats to leatherbacks are the illegal collection of eggs for human use and the loss of nesting habitat, primarily because of natural cycles of erosion and the obstruction of nesting habitat by driftwood (Chacón 1999). Poaching along the Caribbean coastline has been documented for 2 decades (see Berry 1987), but, with the exception of efforts undertaken within protected areas, little has been done to reduce poaching pressure and to improve the conservation status of the species.

View Figure

• Download as Powerpoint
• Download Figure

The objectives of the present study were to estimate the number of turtles nesting each year in Gandoca, evaluate the abundance and spatial distribution of nests laid, analyze temporal patterns in the reproductive effort, assess the results of nest conservation actions taken in the Gandoca-Manzanillo National Wildlife Refuge, and formulate policy recommendations for the management and long-term viability of Caribbean leatherback rookies.

METHODS

RESULTS

Data gathered during 15 annual study periods (15 February to 31 July) between 1990 and 2004 indicated that an average of 181.47 females nested each year, composed of 76.8 remigrants (SD = 63.87, range = 7–210) and 105.7 neophytes (SD = 80.21, range = 22–285) (Fig. 2). Because of the degree of movement by individual females between nesting beaches, both within and among years (Table 1), no attempt was made to define internesting or remigration intervals; that is, the true interval between nesting within a reproductive season or the interval (in years) between reproductive seasons.

View Figure

• Download as Powerpoint
• Download Figure

Table 1. Documented intraseasonal movement between nesting beaches for leatherback sea turtles nesting at Gandoca Beach, Costa Rica, 1990–2004.

View Fullscreen

An average of 4 leatherbacks nested per night (SD = 4.38, range = 0–42) for the study period, depositing a total of 8766 nests on Gandoca Beach. An average of 583 nests were laid per year (SD = 303.3, range = 226-1135), with an average density of 75.7 nests/km. The number of nests laid per month ranged from 0 to 408 and peaked during the period 70–90 days from the onset of the nesting season (Fig. 3). The peak corresponded to the months of April and May, which reported 35.2% and 36.1%, respectively, of all nests laid (Table 2). Nest density peaked between markers 15–20, 31–36, 64–76, and 111–116, which was consistent with earlier analyses (Chacón et al. 1996; Chacón 1999). Of the 8766 nests laid, an annual average of 31% (SD = 15.55%) of all nests were deposited on the berm, safely above the high-tide line. In contract, 32% (SD = 16.45%) were sited within the high-tide zone and 37% (SD = 12.10) were sited within the low-tide zone.

View Figure

• Download as Powerpoint
• Download Figure

Table 2. Monthly number of nests laid at Gandoca Beach by leatherback sea turtles during the study period (1990–2004), with the average monthly contribution (%) to the annual reproductive effort.

View Fullscreen

Over the course of the 15-year study period, 2751 gravid females, an estimated 90% of turtles making landfall at Gandoca Beach (during which time the beach was patrolled hourly and all crawls were documented), were uniquely tagged with external metal flipper tags; of these, 529 (19.2% of the tagged population) were also given PIT tags. Of the total, 408 females carried tags applied elsewhere in Central and South America (Table 1).

The average curved carapace length (CCL) was 153.2 cm (n = 2751, SD = 7.39 cm), ranged from 135 to 198 cm. The average curved carapace width (CCW) was 112.0 cm (n = 2751, SD = 5.53 cm). Clutch size averaged 113.3 (n = 5260 nests), composed of an average of 81.2 (SD = 17.88) yolked and 32.1 (SD = 14.24), or 28.4%, yolkless eggs. The average diameter of 3250 yolked eggs, randomly selected during reburial in the hatchery, was 53.2 mm (SD = 0.93 mm); the diameter of yolkless eggs ranged from 4 to 50 mm (mean = 31 mm, n = 2221).

A total of 2254 nests (including 184,828 yolked eggs) were placed in hatcheries, representing an average of 25.4% (SD = 10.60%, range = 0%–37%) of the total number of nests laid per annum. On average, 12.5% (SD = 7.8%, range = 0%–25%) of nests laid per annum were left in situ, unaltered; 12.9% (SD = 8.69%, range = 1%–28%) were left in situ, with field signs camouflaged by beach patrollers; 33.9% (SD = 15.61%, range = 5%–56%) were relocated to lower-risk beach zones, because of erosion, river currents, and/or poaching (Fig. 4). Most nests were lost from high-risk zones near river mouths and also from beach segments most vulnerable to erosion (markers 6–11, 40–61, 67–73, and 80–95).

View Figure

• Download as Powerpoint
• Download Figure

The average incubation period (defined as the number of days from oviposition to the first hatchling emergence at the surface of the sand) in the hatchery was 59.7 days (SD = 9.70 days, range = 47–74 days), with an average emergence success of 42.6% (SD = 35.14%, range = 0%–100%), producing 79,476 hatchlings. Upon emergence, a sample of 2621 hatchlings averaged 59.6-mm straight-line carapace length (SD = 4.46 mm, range = 54–61 mm) and 46.6 g (SD = 6.12 g, range = 39–52 g). Similarly, a sample of 818 nests left in situ, unaltered at the time of deposition and monitored weekly, had an average emergence success of 41.0% (SD = 25.78%, range = 0%–100%).

Preliminary results of temperature readings taken at regular intervals on a daily basis in the hatchery predicted, based on a range of published values for western central Atlantic populations, a female bias in nests incubating in March and April, when the mean temperature in the center of the incubating egg mass was 30.4°C (SD = 1.43°C), and a male bias in nests incubating in June and July, when the mean temperature in the center of the incubating egg mass was 28.2°C (SD = 1.37°C). Temperature profile data and its interpretation will be prepared for later publication.

Incidents of poaching declined logarithmically over the course of the study period, as illustrated by a standard curve fit to the data (Fig. 4). The most serious violations occurred proximal to points of public access to the beach.

DISCUSSION

The annual nesting season at Gandoca Beach extends from early February to late July, peaking in April and May (Rueda et al. 1992; Leslie et al. 1996; Chacón 1999; Suarez 2004). During the 15-year study period, 2751 individual females laid 8766 nests. Average clutch frequency could not be accurately estimated, because the annual reproductive effort is geographically dispersed (Table 1). The average nest density (75.7 nests/km) at Gandoca was higher than that reported by Leslie et al. (1996) for Tortuguero National Park (1990: 15 nests/km; 1991: 36.9 nests/km) on the northern Caribbean coast of Costa Rica, as well as higher than that reported by Suarez (2004) for La Playona, Urabá Gulf, Colombia (1998–1999, 2002–2003: average 69.4 nests/km).

The spatial and temporal distribution of nesting coincides with earlier analyses done in Costa Rica (Gandoca: Chacón et al. 1996; Chacón 1999; Tortuguero: Leslie et al. 1996) and Colombia (Rueda et al. 1992; Suarez 2004). During the study period, nests were roughly equally distributed above (31%) and within (32%) the high-tide zone, and below the low-tide zone (37%). This is a consistent reality in Gandoca (in 1991, e.g., 64% of all nests were laid between the tide lines: Chacón et al. 1996). Eckert (1987) suggests that wide nest dispersal may enhance reproductive success in unpredictable environments; that is, whenever nest survival is not strongly correlated with available environmental information.

The average CCL and CCW (153.5 cm and 112.0 cm, respectively) of actively nesting females were not dissimilar to results previously reported for Caribbean Costa Rica (e.g., Campbell et al. 1996; Chacón et al. 1996; Chacón 1999) and for other regional colonies (e.g., Hirth and Ogren 1987; Chu 1990; Hall 1990; McDonald et al. 1993; Hilterman and Goverse 2004; Suarez 2004) but are noticeably smaller than gravid leatherbacks nesting in the southern hemisphere (Espiritu Santo, Brazil: 159.8 cm, SD = 10.5 cm, range = 139–182 cm) (Thomé et al. 2007). Leatherbacks nesting on the Pacific coast of Costa Rica are smaller, on average, than those nesting on the Caribbean coast, including those nesting at Gandoca (Reina et al. 2002).

Neither clutch size nor the average diameter of yolked eggs differed appreciably from values reported elsewhere in Caribbean Costa Rica (Campbell et al. 1996; Leslie et al. 1996) or global sites summarized by Miller (1997). The average percentage of yolkless eggs per nest (28.4%) is lower than that reported earlier for the population (31.5%: Chacón et al. 1996) but higher than the 26.2% reported from Tortuguero (Campbell et al. 1996) and quite noticeably higher than the 19.6% reported from the southern hemisphere colonies nesting in Espiritu Santo, Brazil (Thomé et al. 2007). Neither hatchling size nor weight differed appreciably from literature values (e.g., Hirth and Ogren 1987; Miller 1997; Pritchard and Mortimer 1999).

Leatherbacks tend to nest in high-energy coastal environments often associated with steep, unobstructed offshore access and chronic cycles of erosion. Nests laid too near the sea are at risk from saltwater inundation, because of the disruption of chemical, gaseous, and thermal conditions optimal for successful embryo development (reviewed by Ackerman 1997). Eggs relocated to hatcheries would have been lost, under natural circumstances, because of the forces of coastal erosion, mainly near the mouths of rivers but also because of the deposition of some one third (37%) of all nests within the low-tide zone. Over the course of the study, 2254 nests (25.7%) were relocated to protected hatcheries after having been laid too near water. Maximizing hatchling production is a key element of the conservation effort at this site, and no effective alternative to the hatchery option has been identified to achieve this goal for otherwise doomed clutches.

The average rate of emergence success did not differ significantly between in situ nests (41.0%) and nests incubated in hatcheries (42.6%). In situ rates of emergence success were low compared with data presented by Hirth and Ogren (1987), Guadamuz (1990), Dutton et al. (1992), and Leslie et al. (1996), but fall within the range (19.8%–54.2%) reported by Bell et al. (2003) for Playa Grande, Costa Rica, between 1990 and 2000. Conversely, results were high compared with those reported by Hilterman and Goverse (2004), who registered a mean hatch rate of 28% in Suriname. Published studies document a wide range of average emergence successes at hatchery sites (for leatherback turtles) throughout the world, ranging, e.g., from 32% to 71.5% over a 19-year period in Malaysia (Siow and Moll 1982; Mortimer 1990). In the wider Caribbean region, hatchery results, including those reported here, generally fall within this range. Reynolds (2000) reported 54.2% emergence success for Playa Grande, on the Pacific coast of Costa Rica.

In the case of both in situ and hatchery incubated nests, 14%–15% of yolked eggs resulted in dead, full-term hatchlings. Reynolds (2000) attributed the premature death of embryos to unacceptably low levels of O₂ and high levels of CO₂ in the nest chamber, and noted that organic matter mixed with the sand could increase CO₂ concentrations and decrease O₂, potentially reducing survival. It is possible that excessive quantities of organic matter deposited by the Sixaola River contribute to a relatively low in situ hatch rate at Gandoca Beach. Erosion cycles and the subsequent loss of preterm nests also reduce overall emergence success.

Nearly 15% (408 of 2751) of the females observed had been previously tagged while nesting outside of Gandoca. All had been tagged elsewhere in Central and South America; there were no observations of turtles tagged in the Eastern Caribbean. Most had been tagged elsewhere along the Caribbean coast of Costa Rica, with some arriving from as far away as Chiriqui Beach, Panama (130 km from Gandoca), and Acandi Beach and Playona, Colombia (600 km from Gandoca). These results support the observations of field workers throughout the region, as well as some quantitative studies (Eckert et al. 1989; Bräutigam and Eckert 2006), that site fidelity in leatherbacks is less focused than in other sea turtle species. Dutton et al. (1999) found virtually no evidence of stock separation, based on maternally inherited mitochondrial deoxyribonucleic acid haplotype frequencies, among the region's major nesting colonies (i.e., French Guiana, Suriname, Caribbean Costa Rica, and Atlantic Florida).

Intraseasonal movement among nesting beaches obscured a definitive assessment of the average internesting interval. Although the calculated modal interval of 8–12 days between nestings by tagged leatherbacks at Gandoca Beach agrees with data presented for other colonies in the region (e.g., National Research Council 1990; Dutton et al. 1992; Boulon et al. 1996; Steyermark et al. 1996; Miller 1997; Chacón 1999; Hilterman and Goverse 2004), it is also the case that 35% (1381 of 3945 documented intervals) exceeded 24 days, suggesting intervening nestings outside the study site. This hypothesis is borne out by records, e.g., in 2004, demonstrating that 25 females bearing tags from Gandoca Beach subsequently nested at Mondonguillo Beach approximately 100 km to the north. The predominance of internesting movement occurs (in both directions) between Gandoca Beach and other beaches, including Playa Negra in Costa Rica and San San, Changuinola, Soropta, Bluff, and Larga beaches in Panama, which are all within 78 km of one another.

Interseasonal remigration was also obscured by the relaxed site fidelity exhibited by the turtles in this population (Table 1). At some Panama beaches, 64% of all foreign-tagged turtles during the 2003 nesting season originated in Gandoca (C. Ordoñez, unpubl. data). And, although we have evidence of long-term tag retention in some individuals (e.g., in 2004, 8% of nesting females carried tags placed in 1994, 23% in 1995, and 11% in 1996), the habit of single tagging (or not tagging at all) by beach patrollers at some nesting beaches at Panama is known to compromise long-term tag retention and, thus, estimates of remigration cycles.

By using estimates published by Boulon et al. (1996) and Miller (1997), indicating an annual clutch frequency of 6.17 or 5.26, respectively, the annual average of 583 nests laid is some 52%–62% of what would be expected if females observed nesting at Gandoca actually laid their full complement of clutches there. Lower than predicted numbers of observed nests reflect localized renesting at neighboring beaches along the coast, both to the north and to the south. Similar complications arise in the evaluation of demographic parameters associated with other mainland nesting colonies in the region, such as in the Guianas (Girondot and Fretey 1996; Reichart et al. 2001; Hilterman and Goverse 2004).

The nesting pattern at Gandoca has high and low years, typical of most sea turtle populations, rendering problematic the determination of a clear trend in the size of the annual breeding cohort. Troëng et al. (2004) concluded that of the 3 “index beaches” on the Caribbean coast of Costa Rica, the most reliable data set for trend analysis is that of Gandoca. Nonparametric regression formulas (Fahrmeir and Lang 2001; Balazs and Chaloupka 2004) produced the most satisfactory results, indicating that the trend at Gandoca is stable (Troëng et al. 2004). Notwithstanding, the number of nests documented in 2004 was less than 25% of the number documented in 2000 and shows a persistent decline in recent years (Fig. 5). Whether this is an actual decline, a reflection of shifting patterns in the spatial distribution of the annual nesting effort, or a natural fluctuation in the breeding cycle of a stable metapopulation cannot be known without greater collaboration in data-sharing among range states and a long-term commitment to the monitoring of beaches known to be used by the “Gandoca population” (Table 1).

View Figure

• Download as Powerpoint
• Download Figure

Based on 15 years of intensive study, it is clear that the most serious threats facing the nesting colony at Gandoca Beach are the continued (though much reduced) illegal taking of eggs, a problem widespread in Costa Rica (Campbell et al. 1996, Leslie et al. 1996, Troëng et al. 2004); shoreline erosion; garbage and mixed debris transported to the coast by river drainages; and egg loss to domestic and wild predators. Although the poaching of adult females is rare, the long-term threat posed to the population by a persistent loss of eggs to poachers, shoreline erosion, and predators should not be underestimated.

Although significantly reduced since 1990 (Fig. 4), poachers still claim an estimated 1%–3% of all clutches laid each year at Gandoca Beach; this is lower than that documented outside the protected area, indicating that an ongoing research presence, the support and involvement of proximal communities, and increased enforcement activity can result in significant conservation achievements, even in remote rural areas.

Seasonal erosion results from strong coastal drift currents that modify or eliminate beach sectors over the course of the annual nesting season, destroying in excess of 10% of all nests laid in some years. In addition, destruction or modification of the nesting beach as a result of upland deforestation, especially along river drainages, typically results in the deposition of a wide variety of debris and garbage on the beach during the rainy season, blocking access to gravid females and fatally trapping an unknown percentage of emergent hatchlings. Similarly, destruction or modification of the nesting beach because of deposition of logs, as well as plastic, metal, and other types of agricultural and domestic garbage, by the littoral current alters patterns of access, prevent gravid females from finding suitable nesting sites, especially near the mouth of the Sixaola River.

A variety of other anthropogenic factors threaten the population to a lesser extent, including beach-sand extraction for construction purposes, drainage of creeks and wetlands, agricultural runoff and discharge, artificial illumination from torches (flashlights) and bonfires, vehicles occasionally driving on the beach, and plastic products and pesticides originating in agricultural plantations and carried by rivers to the sea and later washed ashore. Beachfront illumination is increasing and may become problematic, not only for nesting and hatching sea turtles (see Witherington and Martin 2000) but also for the success of local tourism businesses that rely on the predictable presence of leatherback turtles near developed areas.