Nesting Distribution and Hatching Success of the Leatherback, Dermochelys coriacea, in Relation to Human Pressures at Playa Parguito, Margarita Island, Venezuela
ABSTRACT
Spatial use of nesting habitat by Dermochelys coriacea is analyzed on a beach with intense tourism use on Margarita Island. There are no previously published data in Venezuela about the use of nesting habitat by leatherback sea turtles and the effects of human activities on nest location and hatching success. Data were collected during nightly intensive beach patrols on Playa Parguito beach (1.6-km long) from March through August 2001. Thirty-one females nested on the beach, and we found 74 nests. The average distance between pairs of nests from the same female was 498 m, although modal distance was between 200 and 300 m. The average distance between randomly selected pairs of nests from any female was 451 m. No special preference for a certain part of the beach for each female was found. Spatial analysis indicated nesting aggregates toward the most suitable portions of the beach: those with fewer risk factors (light pollution, concentration of beach furniture and umbrellas). Our results show nesting behavior change, causing decreased nesting in areas where the best hatching rates occurs. Hatching success was significantly higher in the southern part of the beach, but the larger number of clutches were laid in the northern area, where hatching success was the lowest. Hatching rate for in situ clutches was 47.18%, which was significantly higher than that obtained for nests moved into a hatchery (33.1%) or for nests relocated to a safer part of the beach (27.8%).
Female leatherback sea turtles (Dermochelys coriacea) in Venezuela face threats on nesting beaches both directly and from nest poaching. However, there are few projects providing nest inventory baselines, and conservation plans for the nesting areas of this species in Venezuela are scarce. The most important nesting areas for the leatherback in Venezuela are in the eastern part of the country, along the north coast of Paria Peninsula and Margarita Island (Guada 2000; Guada and Solé 2000; GTTM 2001; GTTM-NE 2001).
On Margarita Island, the main nesting beaches for leatherbacks are El Humo, Playa El Agua, Playa Parguito, and Guacuco (GTTM 2001; Hernandez et al. 2005). All these beaches are located in developed areas with a high human population density, and they sustain intense tourism use. These impacts strongly degrade the habitat and increase nest poaching, despite efforts by government enforcement agencies and conservation groups.
An important aspect to be considered in the management and conservation of marine turtles is the condition of the reproductive habitat and its influence on a population's reproductive success. The geomorphology of nesting beaches is naturally dynamic; weather and oceanic factors influence erosion and accretion processes. The global action of these factors also modifies coastal profiles seasonally through sediment transport (Holmes 1971; Meadows and Campbell 1978). The effect of these seasonal processes on sea turtle nesting habitat and reproductive potential can be aggravated by human activities affecting beach dynamics.
Human alteration of nesting beaches can affect hatching success by altering a site's environmental conditions, e.g., sand quality, but may also modify reproductive potential by making some sites suboptimal. Any significant change in the hatching success of naturally incubated clutches may be an indicator of a problem affecting the nesting beach (Miller 2000).
This study focuses on the nesting habitat of Playa Parguito, its characteristics, the use of this beach by the leatherback, and how these affect the reproductive success under different management conditions. Assessment of nesting characteristics, estimated numbers of nesting females, nesting periodicity, frequency, and success were analyzed in Hernandez et al. (2005). On beaches where clutches are managed, the management technique affects beach reproductive potential. At Playa Parguito, nesting density and human disturbance varied along the beach and allowed comparing zones with low- and high-intensity disturbance. These characteristics were the main reason for selecting Playa Parguito as the study area.
METHODS
Playa Parguito is located on the northeastern coast of Margarita Island (Fig. 1). This 1600-m-long beach is bordered to the south by a rocky formation (Cabo Blanco Point) and to the north by Playa El Agua Beach. This area has been established as a “tourist interest zone” within the Venezuelan Land Use Plan. This means that tourist activities are a priority.
Citation: Chelonian Conservation and Biology 6, 1; 10.2744/1071-8443(2007)6[79:NDAHSO]2.0.CO;2
To evaluate spatial use by sea turtles in relation to human pressures, the beach was divided into nine 200-m sections beginning at the southern end. In each section, a profile line was established with control stakes, which were surveyed to a permanent benchmark. Beach sections allowed analysis and comparison among factors affecting nesting process and their distribution along the beach (Schroeder and Murphy 2000).
Between March and September 2001, nightly surveys were conducted from 2100 to 0500 hours along the entire beach. When a nesting leatherback was found, it was tagged (on the rear left flipper, after oviposition) with Iconel metal tags (National Tag and Band Company No. 681).
Egg clutches were managed in 3 different ways. If the turtle was sighted before oviposition and the area was considered adequate for clutch development, eggs were counted while laying and left in situ. If the place selected by the turtle was considered extremely risky for the nest, i.e., too near high-tide level or a heavily used tourist area, eggs were relocated. Two different methods of egg collecting were used according to the stage at which the turtle was found nesting. If the turtle was found before nesting activity began, eggs were collected directly from the female with a clean plastic bag. If the turtle was found after nesting had already occurred, eggs were collected from the nest in the sand. In both cases, latex gloves were used to avoid egg contamination, and transport to another place on the beach or to the hatchery was done as soon as possible, always within 6 hours after nesting. For clutch size determination, only normal size and shape yolked eggs were counted by using a mechanical counter. In situ nest locations were marked by triangulation by using plastic flags in nearby vegetation and distances between flags and nests. In addition, nest location was also marked by using a GPS receiver Garmin 45™. The nest position in relation to the highest spring tide line (HSTL) was classified in 3 zones: a, near HSTL; b, open dry area of the beach; and c, near the dune. All clutches laid in zone a were relocated.
All nests were excavated 2 days after emergence of the first hatchlings or 65 days after oviposition (after Troëng et al. 2000). We recorded the number of emerged hatchlings, as well as live and dead hatchlings that remained in the nest. Eggs were counted from empty shells, unhatched eggs without embryos and with embryos, predated eggs, and yolkless eggs. Hatching success was calculated according to Miller (2000), where hatching percentage is the number of shells in the nest divided by the total number of eggs in the nest (shells plus unhatched, undeveloped, and depredated eggs) multiplied by 100.
Nesting site fidelity was analyzed by comparing distances between randomly selected nests (in pairs) and distances between (not necessarily consecutive) pairs of nests dug by the same female. The tendency to nest in the same vertical position was tested by using a contingency table of individual females and beach zones in relation to HSTL.
Morphodynamic assessment of Playa Parguito was made through the measurement of beach profiles and collection of sediment samples. A profile for each beach sector was measured by using the Emery (1961) method with a 2-m leveling bar, a popular and simple method widely used to follow beach-profile changes (Hill et al. 2002; Krause 2004). A total of 5 profile measurements for each of the 9 beach sections were made between October 2000 and September 2001. The record from each transect consisted of a series of horizontal distances, vertical distances, and appropriate annotations. Data were transformed to mean sea level by using the nearest sea-level gauge. A bathymetric map of the ca. 0.5-km near-shore area was created by using a boat equipped with a GPSMAP 235 Garmin™ echo sounder. A time series of beach profiles at the same location were evaluated in a variety of ways. Graphically, the profile represented the sand volume in a given time. A time series of beach heights at fixed locations revealed processes of sedimentation or erosion responding to climatic seasons or events. The evaluation of total changes provided quantitative information on the magnitude of the transported sand volumes (Hill et al. 2002; Krause 2004).
Sediment samples for grain-size analysis were taken 3 times between April and September 2001. Three sediment samples were taken along each profile, one from the surf area, the second from the swash area, and the last one from the dry area seaward of the dune. To describe the sediment texture, oven-dried samples were treated with sodium hexametaphosphate as dispersant (Shepard and Moore 1955; Boltovskoy 1965; Llano 1987) and were separated into particle size fractions with a Tyler column by using graded mesh-size sieves described as phi intervals (Llano 1987; Mortimer 1990). The calculation of statistics (mean grain size, sorting coefficient, and grain-size distribution kurtosis) was performed according to the momentum method (Folk and Ward 1957; Brazeiro 2001). Sediments were classified according to the Shepard (1954) triangle.
Spatial and nonspatial data were integrated in a database and thematic maps for each factor (beach zone, nesting density, beach characteristics, infrastructure, and sediment characteristics) were integrated by using Map-info™ 6.0 and Vertical Mapper™ software. In this way a collection of vector maps (points, lines, and polygons) represented complete geographic information of this area.
Data were analyzed by using Statgraphic™ Plus 5.1 software. Differences in hatching success between nest management methods, nesting months, and beach sectors were analysis of variance (ANOVA) tested, after variance check for homogeneity. The post-hoc test used to discriminate among the means was Fisher's least significant difference (LSD) procedure. Regression analysis was used to test relation between nesting month and beach section.
RESULTS AND DISCUSSION
During the period between 16 March and 16 September 2001, 24 leatherback females were tagged, and 73 nests were found and marked. The mean number of eggs per nest was 80.6 (n = 73, SD = 16.4, range = 42–119). We found an additional 9 nests upon their emergence, making a total of 82 successful nests for a total of 6600 eggs (95% confidence limits = 7134–6232), as estimated from the excavated nests. Mean hatching success was 38.3% (n = 44, SD = 21.3, range = 0–78.8), and the mean incubation duration to emergence was 58.3 days (n = 42, SD = 1.9 days, range = 54–62 days), yielding a potential recruitment of 2528 hatchlings (95% confidence limits = 2732–2387) during the season.
The hatching success and incubation period length for the 3 different nest management methods are shown in Table 1. A significant difference (F(1, 41 df) = 4.97, p = 0.031) (ANOVA) was found between the hatching success of the in situ nests and transported clutches (both hatchery and translocated taken as a whole). Hatching rates of leatherback turtles under different conditions worldwide show great variability. Whitmore and Dutton (1985) found a 68.7% hatching success in a hatchery in Suriname. On the Pacific coast of Mexico, Alvarado and Figueroa (1989) found rates of 69.9% and Fuenmayor (2000) documented 60% hatching success. On the Caribbean coast of Costa Rica, Chacón et al. (1996) indicated a hatching success of 55.1% in their hatchery. On the northeastern mainland coast of Venezuela, Martínez (2001) found a hatching success of 54.3% for their hatchery. Eckert and Eckert (1990), in Sandy Point, St. Croix, obtained a hatching success of 53.7% in nests translocated to a hatchery. These results show that the hatchery and translocated nests hatching success at Playa Parguito may be considered low. It must be noted that in this study, hatchery management and some nest transportation was done by personnel of another project, and, therefore, conditions were not entirely standardized.
Meanwhile, the 47.2% hatching success for in situ nests found at Parguito lies between those found in other studies in the Caribbean. In Tortuguero, Troëng et al. (2000) reported a hatching success of 24.8% for in situ nests. Hilterman and Goverse (2004) found an average hatching success of 28.0% in Suriname. Eckert and Eckert (1990) found 64.1% for their study of in situ nests. This study did not find significant differences in hatching success, but differences in the incubation period were observed. The last observation indicates that nest conditions in hatcheries are not the same as those observed in in situ nests, but, if a hatchery is managed properly, hatching success could be similar or even higher than in situ nests, because these often show low hatching rates. Low hatching success of in situ nests for this species has been attributed to high nesting density in flood-prone areas (Troëng et al. 2000) and nest oxygen supply depletion because of clutch bulk (Ackerman 1980). However, Wallace et al. (2004) noted that hypoxia apparently does not cause the low hatching success of leatherback clutches, but they maintained the hypothesis that the large metabolic mass of leatherback clutches and limits to gas flux imposed by the sand create a situation in which leatherback embryos collectively affect their own environment.
Hatching success changed significantly (ANOVA, F(3, 17 df) = 4.17, p = 0.022) between nesting months in the nests translocated to the hatchery, with July having significantly higher hatching success than May and June (p < 0.05, LSD) (Fig. 2a). Similarly, in situ nests experienced higher hatching success in July, and a significant difference (p < 0.05, LSD) between April and June–July (ANOVA, F(3, 13 df) = 3.68, p = 0.0407) (Fig. 2b) was found. Temporal differences in hatching success rates between months are common in sea turtle nesting beaches and may be attributed to seasonal environmental changes, such as the onset of the rainy period and temperature variability.
Citation: Chelonian Conservation and Biology 6, 1; 10.2744/1071-8443(2007)6[79:NDAHSO]2.0.CO;2
Nesting-site fidelity was analyzed by comparing distances between randomly selected pairs of nests and distances between the nests dug by the same female. The mean distance between 504 randomly selected pairs of nests was 451.5 m (SD = 377.3 m, range = 6.4–1363 m). Because we observed 73 nests from 24 identified females, we were able to measure the distance between 68 pairs of nests (consecutive and nonconsecutive) of the same female for the 15 turtles that nested more than once (range, 2–6) (Hernandez et al. 2005). The mean distance was 498.6 m (SD = 307.8 m, range = 78.9–1161.8 m). This mean distance was 31% of the total length of the beach, quite similar to the measurement obtained by Eckert and Eckert (1990) of 594.9 m, or 25% of the total length of the Sandy Point Beach (St. Croix).
Nordmoe et al. (2004) rejected an independence model along the coastal axis for leatherbacks nesting at Playa Grande, Costa Rica, but along the ocean-to-vegetation axis, consecutive nest placements were independent. On Parguito Beach, we tested vertical fidelity by using contingency tables analysis to determine if the 2 classification factors are related. The lamba value obtained was 0.5625, giving the previous vertical zone site selection a predictive value for the next nest of 56%. Of 15 turtles that nested more than once only 5 switched from zone b (open area) to c (near dunes) or vice versa.
These results are consistent for the species, which differs from green turtles by using nesting habitat along an entire beach or even several adjacent beaches in a wide coastline section. It appears that leatherbacks have adopted a regional rather than a local optimum for nest placement patterns, possibly resulting from their weak beach fidelity and the frequent erosion and destruction of their nesting beaches (Kamel and Mrosovsky 2004).
Spatial variability for in situ hatching success is shown in Fig. 3 for the 5 beach sectors where in situ nests were maintained. The highest hatching success (78.8%) was observed in section 2 and the lowest (28.3%) in section 6. The difference was statistically significant (ANOVA F (4, 12 df) = 5.02 p = 0.0131). LSD post-hoc test found significant differences (p < 0.05) between sections 6 and 5 and sections 2 and 3 (Fig. 3), indicating that the south zone of the beach (sections 2 and 3) offers more suitable conditions for clutch development. However, the beach section most frequently used during the nesting season was section 6, with 21.88% of the nesting events, as can be seen in Fig. 4. The fact that sections 4, 5, and 6, which accounts for over half of the nests laid during the season, are the beach sectors with lower hatching success has consequences for the reproductive potential of leatherback turtles in Playa Parguito.
Citation: Chelonian Conservation and Biology 6, 1; 10.2744/1071-8443(2007)6[79:NDAHSO]2.0.CO;2
Citation: Chelonian Conservation and Biology 6, 1; 10.2744/1071-8443(2007)6[79:NDAHSO]2.0.CO;2
Spatial analysis using thematic maps built with Mapinfo™ allowed to overlap nest location, topographic, bathymetric, and human stressors location maps (Fig. 5). As seen in Fig. 5, bathymetry of the study area is very irregular, with deep contours almost parallel to the coast line in the southern half of the beach. Isobaths get closer to the beach indicating strong slopes in the northern half. It is noted in Fig. 5 that most nests were located in the middle part of Playa Parguito, where the −6-m-deep contour gets closer to the beach, and the submarine slope is abrupt. In the southern and northern sections where the −6-m isobath is farther away from the coast and beach slope is smooth, nest concentration is lower, indicating a trend to nest in beach areas with a stepper approach, as was noted by Pritchard and Trebbau (1984).
Citation: Chelonian Conservation and Biology 6, 1; 10.2744/1071-8443(2007)6[79:NDAHSO]2.0.CO;2
All sediment samples were classified as sand according to Shepard (1954) with small gravel and pebble percentages. The mean grain diameter (Mz) along the sampling period in Playa Parguito was 1.76 phi (n = 81; SD = 0.29 phi; range, 2.7–0.8 phi), indicating a constantly high mean kinetic energy. The mean value of the sorting coefficient (Folk and Ward 1957; Brazeiro 2001), for the whole beach was of 0.72 phi, with a range between 0.34 phi in section 4 and 1.02 phi in section 2 (n = 81; SD = 0.11). Beach sectors with higher nesting density (6) and with higher hatching success (2) have moderately sorted sands.
The evolution of the sand volume for each beach profile along the study period is shown in Fig. 6. The section with highest variability was 9, in the northern border of the beach; its sediment volume changed from a minimum of 1 193 m3 in February, at the starting of the nesting season, to a maximum of 6 301 m3 in June. Seasonal variability of sand volume in a profile is an indicator of the accretion and erosion processes in that beach sector. As seen in Fig. 6, beach sectors 1 to 6, where 84% of the nesting occurred, showed a smaller seasonal variability and thus lesser risk of nest loss because of beach erosion. This relative homogeneity in erosion processes along the beach may explain why no shift in nest distribution as the season progressed was observed as in other beaches (Lum 2005). The relation between nesting month and beach section used was absent (R2 = 0.000002%, p = 0.999).
Citation: Chelonian Conservation and Biology 6, 1; 10.2744/1071-8443(2007)6[79:NDAHSO]2.0.CO;2
All beach profiles confirmed their utility as a storm beach damage prediction tool (Benavente et al. 2002), showing the beginning of erosive processes in September, at the end of the nesting season, giving way to the lower sediment volumes in October, when North Atlantic winter storms began affecting the northern and eastern coasts of Margarita Island. For this time of the year very few, if any, nest remained on the beach.
The homogeneity of the sand found at Playa Parguito, with a mean of 1.76 phi, if compared with other beaches, indicates that the leatherback turtle may use a great variety of sediments to nest but that nesting females prefer medium-sized grains sands. Ramírez and Torres (1995) conducted a study on sediments of the main beaches used by the leatherback turtle in the Pacific coast of Mexico and found that the most important beaches have a Mz between 1.50 and 1.92 phi; other beaches, less important as nesting sites have a Mz of 0.83 phi, classifying sediments as coarse sand and others with a Mz of 2.83 and 2.17 phi as fine sand. For other species, the preferred sand varies greatly (Mortimer 1990). In Ascension Island, Stancyk and Ross (1978) found a Mz between 0 and 1 phi in several nesting beaches of Chelonia mydas. In Aves Island, an important nesting site for C. mydas in the Caribbean Sea, Buitrago and Ziegler (2004) found homogeneous sediments that measured approximately 1 phi. Preference for moderately sorted sands has been also reported by Ramírez and Torres (1995). Buitrago and Ziegler (2004) also found a C. mydas preference for less well sorted sands in Aves Island.
Other aspects of the beach sedimentology may be important for site selection and hatching success. Mortimer (1990) observed a significant relation between the physical characteristics and the chemical composition of sediment, with respect to the site selection and nest survival in the nests of C. mydas, in different beaches around the world. Reyes et al. (2001) found a significant relation between some physical characteristics of the sediment, such as humidity and temperature, and nest site selection for C. mydas in Tortuguero, Costa Rica.
Spatially related anthropogenic stressors present in Playa Parguito were restaurant lights, shower structures, umbrellas, and beach chairs among other tourist services in the southern part and buildings lights and beach armor in the northern part. This allowed division of the beach into 3 main sectors according to human disturbance type and strength (Fig. 5). Sectors 1 to 3 in the south, accounted for 30% of the beach length, and 7 and 8 in the north side with 19% of the beach length were considered highly human-disturbed areas, while, in sectors 4 to 6, human pressures were much reduced. Sector 9 was a marginal area, not intensively used both for humans and turtles, in the northern tip of the beach. Other human stressors, such as vehicular traffic, harassment of females attempting to nest, and egg poaching, were randomly distributed along the beach, so they were not considered in the spatial analysis. It is noteworthy that all the tourist units are unlawful, because they are partially or in full, within the legal protected zone of the beach (80 m behind the high-tide line).
Sectors 1 to 3 in the southern part of the beach, the area showing the higher hatching success, were not the preferred nesting area may be because of physical characteristics of the beach as the steeper slope in other areas like sectors 4 to 6 but may also be because of the higher degree of human disturbances in the southern sector. The intense tourist use of the area has developed recently, in the last 20 years, but there are no prior records of nesting activity to compare.
The fact that the zone offering more suitable conditions for the hatching success in the southern part is strongly affected by anthropogenic activities, driving the females to nest in the less optimum areas, affects the reproductive potential of the leatherback turtle in Playa Parguito and offers an opportunity for environmental authorities to take novel management measures in favor of this species critically endangered.
Playa Parguito Beach, Margarita Island, eastern Venezuela.
Seasonal variability in hatching success for leatherback (D. coriacea) clutches relocated to hatchery (2a) where July (month 7) and April (4) have significantly higher (ANOVA, F(3, 17 df) = 4.17, p = 0.022) hatching rates than during June (6) and May (5). In situ incubated nests (2b) at Playa Parguito, where June (6) and July (7) had a significant higher (ANOVA, F(3, 13 df) = 3.68, p = 0.0407) hatching rate than April (4). (* = mean; bars = 95% confidence interval).
Beach section variability in hatching success for leatherback (D. coriacea) in situ incubated nests on Playa Parguito, Margarita Island, during the 2001 season. Results for the 5 beach sections where in situ nests were maintained. Hatching success differed significantly (ANOVA F(4, 12 df) = 5.02, p = 0.0131). LSD post-hoc test found significant differences (p < 0.05) between sections 5–6 and sections 2–3. (* = mean; bars = 95% confidence interval). Sections 1–9 are 200-m-wide beach sectors from south to north.
Spatial distribution of leatherback (D. coriacea) nests on Playa Parguito, Margarita Island, during the 2001 season. Sections 1–9 are 200-m-wide beach sectors from south to north.
Study region map showing Playa Parguito coastal zone, nest locations (crosses), tourism infrastructure (patterned blocks), and main topographic and bathymetric characteristics.
Seasonal variations of sediment volume along 1-m-wide beach profiles during the study period on Playa Parguito, Margarita Island. The x-axis labels P-1 to P-9 are the 200-m-wide beach sectors from south to north.
