Study Site

San José Island is located in northwest Costa Rica (10°51′24.58″N, 85°54′45.07″W) and is part of the Murciélago Archipelago (Fig. 1), which is protected within the Área de Conservación Guanacaste (ACG), under the authority of the Ministry of the Environment. As mentioned above, the beach is largely free of anthropogenic threats because of the park's protection and the remoteness of the site. The waters surrounding the archipelago are also protected by ACG, and activities such as fishing or extraction of marine life are not permitted, although enforcement is not robust. The nesting area is divided into 3 adjacent sections by rocky outcrops. The lengths of these sections from north to south are 80, 125, and 125 m. The edge of the vegetation of each section is populated mainly by saltgrass (Distichlis spicata), with a scarce presence of trees. The beach zone adjacent to the vegetation line is a sand dune, which is strongly modified by the wind.

Data Collection

Trained field assistants (FAs) conducted daily surveys to determine nesting activity from October to late May over 4 nesting seasons: 2012–2013, 2013–2014, 2014–2015, and 2015–2016. Nightly patrols started at 1900 hrs and ended at 0400 hrs. Additionally, FAs monitored the beach every morning to count nesting tracks to confirm that we observed all turtles that visited the beach the previous night.

Every turtle encountered was inspected for tags. Turtles without tags were tagged with Inconel tags (model 681, National Band and Tag Company, Kentucky). Tagging procedure followed these general guidelines: 1) females were allowed to complete the egg laying process and tagging occurred as soon as nest covering began (except in those cases when females were tagged while returning to the ocean); 2) the scale was disinfected with a solution of vanodine prior to tag application; and 3) each turtle was tagged with one Inconel tag on the second proximal scale of each front flipper. After females had completed the egg laying process and were tagged, we measured the curved carapace length (CCL) notch to tip and curved carapace width (CCW) in centimeters with a flexible measuring tape (Bolten 1999). We determined clutch size at the time of exhumation.

We calculated the observed internesting period (OIP) from the number of days elapsed between 2 observed oviposition events by an individual female, as identified by her unique tag number. We excluded data where the interval was two times greater than the mode during the season, as this increases the likelihood that a turtle laid a clutch that was not observed (Reina et al. 2002). We calculated the estimated clutch frequency (ECF) for each turtle by dividing the number of days between her first and last observed nesting events by the mean OIP of the season and added 1 to account for the first nest (Santidrián Tomillo et al. 2014).

We marked nests with two 8-cm sections of 0.50-inch-diameter PVC tubes joined by 75-cm-long monofilament thread. One of the tubes was placed inside the nest chamber during oviposition, whereas the other was left outside the nest on the surface of the beach. The latter tube was labeled with the nest number and the oviposition date. We tied a plastic bottle to the tube at the surface to facilitate finding the clutch later, especially after high tides. To estimate clutch density, we divided the mean number of clutches per season by the length of the beach (km), following the methodology of Godley et al. (2001).

We excavated the contents of nests to determine hatching success 2 d after the last observed hatchling emergence or 55 d after oviposition. At excavation, we counted the number of shells, where more than 50% of a broken shell was counted as an egg, and classified unhatched eggs into 4 developmental categories (Santidrián Tomillo et al. 2009). We estimated hatching success using the formula H = S/(S + U), where S is number of hatched eggshells and U is number of unhatched eggs.

Statistical Analysis

We used R Software v. 3.3.1 to conduct statistical analyses. We used 1-way analysis of variance (ANOVA) to compare differences between seasons in OIP and hatching success because the data were normally distributed. Statistical significance was assumed at p ≤ 0.05.

Importance of San José Island as Nesting Site for EP Green Turtles

Our results showed that San José Island hosts the largest nesting aggregation of EP green turtles in Central America in terms of absolute numbers and appears to have the second highest clutch density of any nesting beach in the entire EP region. Within the EP, the green turtle population of Colola is considered the largest in the region and is increasing because of the near-complete cessation of illegal hunting at foraging grounds and reduction of egg poaching at nesting beaches (Alvarado-Díaz et al. 2001; Delgado Trejo and Alvarado Díaz 2012).

In Costa Rica, 4 nesting sites were previously identified along the Pacific Coast: Naranjo (Cornelius 1976; Drake et al. 2003), Cabuyal (Santidrián Tomillo et al. 2014), Matapalo (Urena Lopez 2014), and Nombre de Jesús (Blanco et al. 2012a, 2013b). All of these sites registered fewer than 300 clutches per year, which constitutes much lower levels than those of San José Island. Cabuyal Beach, where there has been a consistent monitoring effort since 2011–2012 (Santidrián Tomillo et al. 2014), accounts for ∼ 25% of the mean number of turtles identified at San José Island in a given year.

Despite the presence of park rangers since 1996, there are no historical or anecdotal records of sea turtle nesting activity at San José Island. The best recollection of nesting activity by park rangers is the nesting of 3 or 4 turtles per night. Before park rangers were present, the island was used as a refuge by local fishers who poached turtle eggs. Thus, protection of this beach over the past 20 yrs, and possibly low poaching levels before then, may have contributed to the currently high nesting abundance.

Nesting Ecology

The nesting season coincides with that of leatherback turtles (Dermochelys coriacea) in the EP (October–February) (Reina et al. 2002; Sarti Martínez et al. 2007). The nesting season at San José Island mainly occurs between November and February, and the peak season between January and February, with January being the peak month when all 4 seasons were taken into account. However, there was some variability in the nesting peak among years by up to 2 mo. For example, November and December registered the most nesting in 2014–2015 and 2012–2013, respectively. This temporal nesting pattern was previously documented at Cabuyal, where the nesting peak varied by 1 mo between years (Santidrián Tomillo et al. 2014). The nesting peak in the area coincides with the end of the rainy season and with the beginning of the north winds, which induce a significant decrease in sea surface temperature as a result of upwelling in the area (Jímenez 2001). Perhaps the number of nesting turtles is regulated in some fashion by local environmental conditions, such as winds and sea surface temperature.

Nesting phenology of green turtles differed somewhat among the 3 main nesting beaches in the EP. In general, peak nesting months in Galápagos rookeries occur after those of rookeries in the northern hemisphere. In addition, the Galápagos green turtle peak nesting months overlapped with peak rainy months, unlike the rookeries in Costa Rica and Mexico. These patterns suggest that the nesting phenology of EP green turtles is likely influenced by local climatic conditions, leading the season to begin earlier in the north and later in the south.

Some (11.7%) of the San José Island green turtles returned to nest after 2 seasons and 9.2% after 3 seasons, but it is too soon to know what the dominant remigration intervals are for this population. At Colola Beach, which has maintained a tagging program for over 2 decades, the great majority of the turtles had a remigration interval of 3 or 4 yrs (Delgado-Trejo 2016); in the Galápagos Archipelago, the majority of turtles had a remigration interval of 4 or 5 yrs (Zárate 2009). A longer record is required to elucidate the actual situation at San José Island because remigration rates may vary according to climate changes such as the El Niño Southern Oscillation (ENSO), the warm phase of which has been linked to a reduction in primary productivity and an increase in the remigration interval (Limpus and Nicholls 1988; Wallace and Saba 2009).

As expected, because of geographic proximity and the observed migration of green turtles from Galápagos to Costa Rican foraging areas (Seminoff and Shanker, 2008), female size (CCL) at San José Island (85.4 cm) seemed similar to that of Cabuyal (86.2 cm) and the Galápagos Archipelago (85.0 cm) and to those of Colola Beach (83.1 cm). Similarly, clutch size at Colola (∼ 67 eggs), Cabuyal and San José (both 76 eggs), and the Galápagos (73 eggs) (Zárate 2002, 2009; Santidrián Tomillo et al. 2014), was similar. Thus, female and clutch sizes fall well within the sizes of EP green turtles that are generally smaller than those of other populations in the world, such as the green turtles that nest in the Caribbean, where green turtles can exceed 104 cm (Reyes and Troëng 2002) and lay over 100 eggs per clutch (Fowler 1979).

OIP fluctuated between 12 and 14 d, being slightly shorter than the 15.4 d recorded at Cabuyal (Santidrián Tomillo et al. 2014). Because OIP is affected by sea surface temperature (Hays et al. 2001), differences in mean OIP between Cabuyal and San José may reflect differences between nesting seasons rather than differences between sites. Likewise, fluctuations in OIP have also been observed among seasons in Colola and the Galápagos Archipelago, where OIP oscillated between 13 and 16 d (Zárate 2002, 2004, 2009; Alvarado-Díaz et al. 2003). Strong northeast winds are typical in the months of December and January in Costa Rica; the winds normally cause generalized upwelling in the region, leading to a significant reduction in sea surface temperatures (Jímenez 2001). This low temperature regime likely influences OIP in the later months, increasing its length because of an expected reduction in metabolic rate. In fact, longer OIPs were registered later in the season in Cabuyal (Santidrián Tomillo et al. 2014).

Estimated clutch frequency was high and similar to that reported for Cabuyal (4.4 vs. 4.3 clutches; Santidrián Tomillo et al. 2014) and higher than that exhibited by the turtles of Colola (3.1 clutches; Alvarado-Díaz et al. 2003). This difference may be explained by the high monitoring effort that we maintained on this beach and the high nest-site fidelity that turtles exhibited. Because of the small size of the nesting area (350 m), the great majority of nesting females were seen, and only 2% of the females were found to nest on other beaches, although it is likely that some turtles nested on beaches that were not monitored actively.

Hatching success and clutch survival were high at San José Island, likely leading to a higher production of hatchlings than those at other sites. The anthropogenic pressures that reduce sea turtle reproductive output along the continental coast such as egg poaching (Santidrián Tomillo et al. 2014), urban development (Urena Lopez 2014), and predation by jaguars, coyotes, and raccoons (Drake et al. 2003; Alvarado et al. 2016) are nonexistent at San José, increasing clutch survival. Additionally, hatching success was high and consistent with that reported at other green turtle rookeries, such as Colola (Delgado Trejo and Alvarado Díaz 2012), Tortuguero (Fowler 1979), Cyprus (Broderick and Godfrey 1996), and Taiwan (Cheng et al. 2009) (all greater than 80%), and higher than that of the Galápagos, which was < 50% because of predation by beetles and feral pigs (Zárate et al. 2013).

There was some destruction of clutches by other females, but the rate of destroyed clutches was lower than expected (0.8%) for a 350-m beach. Because turtles showed some preference to nest near the vegetation or under trees, the available nesting space was greater than that provided by just the beach area. This preference was reported at Cabuyal (Santidrián Tomillo et al. 2014), where clutches under the vegetation represented a higher proportion than those located at the edge of the vegetation. Another factor that reduced the probability of clutch destruction by a turtle was the presence of a sand dune with a surface of approximately 1500 m². This dune has an elevation of more than 7 m at its highest point relative to sea level, and turtles were seen nesting up and down the dune. The presence of the dune increased the surface area of the beach.

We estimated that 282,335 hatchlings were produced at San José Island over 4 yrs, which was proportional to the number of clutches laid during the study period. Production was lowest during the 2015–2016 season, when we recorded the lowest number of nesting turtles and lowest hatching success. This low hatching success coincided with a warm El Niño event. Such El Niño events are thought to reduce precipitation and increase air temperature in the area, which increases embryonic and hatchling mortality in sea turtle clutches (Santidrián Tomillo et al. 2012). However, for reasons not understood, the effect of El Niño on egg mortality was not as pronounced in green turtles as in the leatherback (Santidrián Tomillo et al. 2012) and olive ridley (Lepidochelys olivacea) turtles (Valverde et al. 2010).

As mentioned above, San José Island seems to be the main nesting site for EP green turtles in Central America, but there are several nearby beaches along the continental coast of Costa Rica where EP green turtles also nest. If we add the numbers of turtles nesting at Coloradas, Nancite, Naranjo, Cabuyal, Matapalo, Nombre de Jesús, and the beaches of difficult access of the Papagayo Peninsula, at least 500 females could nest every year in northwest Costa Rica. This number implies that this region is at least as important for EP green turtles as the beaches of Mexico and the Galápagos. The high nesting and high hatchling productivity at San José Island suggest that passive management (i.e., minimal or no interference with the natural reproduction of the species in the area) is appropriate for the conservation of green turtles here because there are not significant anthropogenic threats or natural predators at this location. However, the impact of incidental fishing and degradation of foraging habitats should also be evaluated, which may threaten San José Island and other nesting populations in the area.