Effects of Sand Characteristics and Inundation on the Hatching Success of Loggerhead Sea Turtle (Caretta caretta) Clutches on Low-Relief Mangrove Islands in Southwest Florida
ABSTRACT
We determined characteristics of the sand, level and frequency of tidal inundation, and hatching success at loggerhead sea turtle (Caretta caretta) nest sites on 8 low-relief mangrove islands in the northern half of the Ten Thousand Islands in southwest Florida. The sand was generally composed of larger particles and tended to be wetter, more porous, and more saline than on other loggerhead nesting beaches. More than one-third (38.9%) of the nest sites experienced tidal inundation. The mean salinity of inundating water was 26.9‰ (± 9.3, range = 0–40). The water content and salinity of the sand water at nest sites increased with increasing frequency of inundations. The mean particle diameter and total porosity of the sand were negatively related to sand water salinity, perhaps indicating that in the largest-grained, most porous sands, salt was more effectively washed out by rainfall. Hatching success decreased as inundations, sand water content, and sand water salinity increased. However, at nest sites that did not experience inundation, the sand water content and sand water salinity were not related to hatching success. Loggerhead clutches can tolerate a wide range of incubation environments, including a certain amount of inundation. Clutches that are deposited in low-beach areas (close to the water and more prone to inundation) can produce hatchlings. Because the characteristics of loggerhead hatchlings are influenced by the incubation environment, conservation strategies that involve moving all clutches from low-beach areas to high-beach areas may reduce the variety of incubation environments, thus reducing hatchling variability and possibly preventing the expression of characteristics that promote hatchling survival or otherwise increase the reproductive success of females.
When formulating management strategies for sea turtle nesting beaches, it is important to know which characteristics of the sand are most likely to influence hatching success and which general field conditions preclude egg development. Laboratory experiments have shown that hatching success can be affected by unusually high or low temperatures (Yntema and Mrosovsky 1982), limited gas exchange (Ackerman 1980), or unusually wet or dry hydric conditions (McGehee 1990). Some characteristics of the sand such as particle size (Schwartz 1982; Mortimer 1990) and sand water salinity (Bustard and Greenham 1968) have also been shown to affect hatching success in laboratory settings. Except for Mortimer's (1990) work on green turtle (Chelonia mydas) nesting beaches, we are aware of no other studies that have been done to associate sand characteristics with the hatching success of in situ clutches. Inundation of in situ clutches has often been blamed for causing low hatching success in the field (Ragotzkie 1959; Kraemer and Bell 1980; Limpus 1985; Witherington 1986; Milton et al. 1994), but we know of no specific studies that have been conducted to determine the tolerance of sea turtle eggs to various frequencies and levels of inundation. Without more information on the aspects of the incubation environment that are likely to influence hatching success in the field, nesting beach managers cannot appropriately decide when mitigative actions such as nest translocation (Dodd and Seigel 1991) should be undertaken.
In the Ten Thousand Islands of southwest Florida, loggerhead sea turtles (Caretta caretta) regularly nest on the narrow, discontinuous beaches of low-relief mangrove islands (Foley et al. 2000). While conducting nesting surveys on these islands, we noted that there were a wide variety of sand types and various frequencies and degrees of tidal inundation at nest sites. In the present study, we determined characteristics of the sand, level and frequency of tidal inundation, and hatching success at nest sites on these beaches to investigate the tolerance of in situ loggerhead sea turtle clutches to a wide range of naturally occurring incubation environments.
Citation: Chelonian Conservation and Biology 5, 1; 10.2744/1071-8443(2006)5[32:EOSCAI]2.0.CO;2
METHODS
During 1993 and 1994, we determined the characteristics of the sand and the frequency and level of inundation at 90 loggerhead nest sites on 8 keys in the northern half of the Ten Thousand Islands (Fig. 1). Water-content profiles were determined at all nest sites by collecting samples of sand weighing approximately 10 g at the surface and at depths of 5, 10, 20, 30, 40, 50, and 60 cm. We used a spoon to scoop the surface sand sample and a 90-cm tube sampler soil probe to retrieve the subsurface samples. All sand samples were immediately placed into Kapak heat-sealable pouches. These pouches were then flattened, tightly rolled, and held closed with a rubber band. To determine the water content in the sand samples, we weighed each sample before and after drying for 24 hours at 105°C. We determined the percentage of water in each sand sample on a dry-weight basis by the following formula (Gardner 1986):
Sand water salinity was determined at 59 of the nest sites by taking a sample of sand from a depth of 25–30 cm and placing it into a zip-lock plastic bag. In the laboratory, we made a 2:1 saturation extract by adding 100 mL of distilled water to a jar containing 200 g of the sand, and then shaking the jar vigorously by hand for 1 minute, repeated 4 times at 30-minute intervals. We then filtered the suspension using highly retentive filter paper (Whatman Grade 5) and determined the salinity of the filtrate using a portable refractometer (technique modified from Rhoades 1982). Assuming that all the salts in the original sand sample were dissolved in the sand water, we determined the salinity of the water in the 200-g sand sample before saturation (i.e., the sand water salinity) by the following formula:
where Ss‰ was the salinity of the saturation extract, θs was the water content (g H2O/g dry sand) of the saturation extract, and θ was the water content of the sand in the field at a depth of 30 cm (technique modified from Bresler et al. 1982).
The particle-size distribution of the sand was determined at 83 of the nest sites by taking a sample of sand from a depth of 25–30 cm and placing it into a zip-lock plastic bag. In the laboratory, we separated the sand into its various-sized components by wet-sieving 200 g of the sand through the following series of sieves: 2 mm, 1 mm, 0.5 mm, 0.25 mm, 0.125 mm, and 0.063 mm. The sand retained by each sieve was dried for 24 hours at 105°C and weighed. We then calculated the percent of sand by weight for each size class and determined the mean particle diameter (by weight) according to the following formula (Hillel 1980):
where xi was the mean diameter of any particular size range of particles separated by sieving (3 mm was used as the mean diameter of the size range greater than 2 mm) and wi was the weight of the particles in that size range as a fraction of the total dry weight of the sample analyzed.
To determine hydraulic conductivity, sand samples were taken at 49 nest sites. To collect these samples, we first dug by hand to a depth of 25 cm. We then placed a 10.2 × 10.2 cm aluminum cylinder at the bottom of the hole and gently hammered it down completely into the sand. The cylinder was removed by digging around and below it. We carefully cleared away sand from both ends of the cylinder, placed tight-fitting caps over each end, and put the capped cylinder into a zip-lock plastic bag. In the laboratory, we measured the saturated hydraulic conductivity (K) with the constant head method and calculated the value of K (mm/sec) using the following formula (Klute 1965):
where Q was the volume of water (mm3) that passed through the sample in a known time t (seconds), A was the cross-sectional area of the core (mm2), L was the length of the core (mm), and ΔH was the hydraulic head difference (mm).
After determining the saturated hydraulic conductivity, we weighed the saturated core before and after drying for 24 hours at 105°C. The total porosity (St) was assumed to be the percentage of the saturated core's volume that had been occupied by water prior to drying, and we calculated its value by the following formula:
where Vl was the volume (cm3) of water lost from the saturated core by oven-drying (determined by loss of weight), and Vs was the volume of the core sample (cm3). We calculated the air-filled pore space (as a percent of the total pore space) by the following formula:
where St was the total porosity (%), θdw was the water content of the sand in the field at a depth of 30 cm, Mc was the weight of the dry sand in the core sample, and Vs was the volume of the core sample (cm3).
Sand samples that were used to establish water content profiles and to determine sand water salinity were taken once when a nest was first discovered, again 30 days later, and then again 60 days later. Sand samples used to determine particle-size distribution, mean particle diameter, saturated hydraulic conductivity, total porosity, and air-filled pore space were collected after the clutch at the nest site had hatched.
We monitored all nest sites for groundwater inundation at various depths by placing 12.7-mm-diameter chlorinated polyvinyl chloride (CPVC) pipes inside buried 50.8-mm-diameter polyvinyl chloride (PVC) pipes (Fig. 2). One of these devices was placed approximately 0.5 m from each nest cavity in a direction parallel to the shore. Based on measurements we made of 30 nest cavities at the study site during 1992, we chose 15, 30, and 45 cm to represent the depths of the top, middle, and bottom of the clutches, respectively. Groundwater inundation was monitored at depths of 30 and 45 cm in 1993, and at depths of 15, 30, and 45 cm in 1994. The water monitoring devices were checked 1 or 2 times a week throughout the entire incubation period. The salinity of inundating water was measured with a portable refractometer.
Clutches at all the nest sites were protected from predators by self-releasing screens or cages (methodology in Foley et al. 2000). We evaluated the hatching success at 71 of 90 nest sites either 72 hours after the first signs of hatchling emergence or 70 days after the clutch was deposited, whichever occurred first (methodology in Foley et al. 2000). Hatching success evaluations were not conducted at 19 of 90 nest sites because of depredation by raccoons or because eggs were removed prior to hatching for other studies (see Foley 1998).
RESULTS
The combined water-content profile of sand at all nest sites is shown in Fig. 3. All other sand characteristics are given in Table 1.
Thirty-five of the 90 nest sites (38.9%) experienced some degree of groundwater inundation (up to at least 45 cm below the sand's surface). Twenty-five of the nest sites (27.7%) were inundated at least once up to 30 cm below the sand's surface. Of the 61 nest sites monitored in 1994, 10 (16.4%) were completely inundated (up to at least 15 cm below the sand's surface) at least once.
Citation: Chelonian Conservation and Biology 5, 1; 10.2744/1071-8443(2006)5[32:EOSCAI]2.0.CO;2
The overall mean salinity of 277 samples of inundating water was 26.9‰ (± 9.3, range = 0–40). When the inundating water was collected at 2 or more depths within 1 water monitoring device, the water sample collected from the upper layer was usually less saline (and never more saline) than the water sample collected just 15 cm deeper. Some of the water samples collected within the same water monitoring device (and 15 cm apart in depth) varied in salinity by as much as 24‰.
All characteristics of the sand and the inundation frequency at the nest sites were compared to determine their relationships (Table 2). The mean particle diameter, total porosity, air-filled pore space, and saturated hydraulic conductivity of the sand were all positively related. The mean particle diameter and total porosity of the sand were negatively related to the sand water salinity and positively related to the sand water content. As the number of recorded inundations increased, the sand water salinity and sand water content increased.
Citation: Chelonian Conservation and Biology 5, 1; 10.2744/1071-8443(2006)5[32:EOSCAI]2.0.CO;2
The results of the hatching success evaluations are given in Table 3. Most eggs hatched (68.9%) and produced hatchlings that escaped from the nest cavity (68.6%; i.e., emergence success). Of those eggs that did not hatch, most were either not fertilized or stopped development at, or prior to, approximately 27 days of incubation. Of the eggs that were still viable after approximately 27 days, the hatching success was 92.2%. The hatching success was less for nests that were inundated than for nests that were never inundated and hatching success was least for nests that were completely inundated (Table 4). Although some eggs in the evaluated clutches survived up to 2 complete inundations, none survived 3 or more complete inundations.
We also examined the characteristics of the sand at the nest sites in relation to the hatching success there (Table 5). The hatching success was related only to the sand water content and to the sand water salinity. As the sand water content or the sand water salinity increased, the hatching success decreased. However, when only nest sites without any inundation were included in the correlation analyses, the sand water content and the sand water salinity were not related to the hatching success (Table 5).
DISCUSSION
Characteristics of the Sand and Inundation at Nest Sites
The mean water content of the sand at clutch depths of 10 through 50 cm at each of our 90 loggerhead study nest sites ranged from 2.1% to 21.7%. On other loggerhead nesting beaches in Florida, the water content of the sand at clutch depth ranged from 2% to 6% (McGehee 1990; Ackerman 1997; Speakman et al. 1998). On loggerhead nesting beaches in South Carolina (Caldwell 1959) and Israel (Ackerman 1991), the water content of the sand at nest depth ranged from about 2% to 4%.
The mean sand water salinity at our study nest sites (16.5‰) was high for a loggerhead nesting beach. For example, along the east coast of Florida, sand water salinities on loggerhead nesting beaches were never more than 3% that of seawater (about 1‰) (Ackerman 1997). Because there were a few sites with very high sand water salinities in our study, the median sand water salinity (11.7‰) may better describe our typical nest site, yet even that was relatively high. High sand water salinities are not expected in areas where loggerheads nest because the primary source of moisture for the upper layers of beach sand is usually rain (Ackerman 1997). In the absence of rain, the groundwater (at a depth of 1–5 m) may be drawn into the upper layers of sand because of surface drying (Ackerman 1997). But even when beach groundwater is transported upwards, the introduction of salt to areas where loggerhead clutches incubate is not expected because the groundwater is typically either very low in salinity to begin with (De Jong 1979) or may move upward through evaporation and condensation (Olsson-Seffer 1909; De Jong 1979) and presumably lose any salts. Occasional tidal inundations and aerosols may deposit salt on the parts of the beach where loggerheads nest, but frequent rainfalls probably flush these from the upper layers of sand (Ackerman 1997).
We observed frequent rainfalls in the Ten Thousand Islands, but sand water salinities were still relatively high. Presumably, the low relief of these islands accounted for the high sand water salinities. Sand water salinity increases with decreasing distance to the water (i.e., decreasing elevation) (De Jong 1979) and may reach values near that of seawater in areas that are often inundated by tides (Bustard and Greenham 1968). At our nest sites, groundwater was present at a depth of about 0.5–0.75 m, and tidal fluctuations often pushed groundwater up into the level at which loggerhead clutches incubated. The average salinity of the groundwater sampled directly at our nest sites was 26.9‰ and ranged up to 40‰. As the number of inundations at nest sites increased, sand water salinity increased. Evaporation of inundating water from the upper layers of sand may have occasionally driven the salinity of sand water even higher. Rainfall was probably able to flush some of this salt from the sand, but the flushing effectiveness decreased as inundations increased and as the coarseness of the sand decreased (see Table 2).
The sand water salinities at our nest sites were not always high. At 64.4% (38/59) of the nest sites where sand water salinity was determined, at least 1 of the 3 sand samples taken at different times during incubation (day 0, day 30, and day 60) had a sand water salinity of 0‰. Furthermore, at 17% of the nest sites, mean sand water salinities were always less than 5‰, and at 10% of the nest sites, sand water salinities were always 0‰. As with water content of the sand, the sand water salinity varied at our study site from values that were as low as those on other loggerhead nesting beaches to values that were much greater.
The mean diameter of sand particles at our loggerhead nest sites was just over 1 mm, more than double that on other loggerhead nesting beaches (Hughes 1974; Mann 1977; Speakman et al. 1998). The most common sand particle size on loggerhead nesting beaches, including the beaches in the present study, is between 0.25 and 0.5 mm (quartz sand) (Hughes 1974; Mann 1977; Speakman et al. 1998). The frequent occurrence of particles greater than 2 mm in diameter (molluscan-shell gravel) at many of our nest sites was responsible for the relatively large mean particle diameter.
We could not find any published reports of saturated hydraulic conductivity, total porosity, or air-filled pore space of the sand on any other loggerhead nesting beaches. However, because these sand characteristics are all positively related to the particle size of the sand and because the particle size of the sand varied at our site from as small as to much larger than that on other loggerhead nesting beaches, we presume that the saturated hydraulic conductivity, total porosity, and air-filled pore space also varied more than on other loggerhead nesting beaches.
The mean salinity of the water inundating nest sites (26.9‰) was about 10‰ lower than the salinity of the adjacent, offshore water (35‰–40‰), and the salinity of inundating water ranged from that of offshore water to that of freshwater. The groundwater often had a layer of fresher water on top of much more saline water. During inundation, eggs in the bottom of the clutch were typically exposed to more saline water than eggs at the top of the clutch were.
Effects of Sand Characteristics and Inundation on Hatching Success
The loggerhead clutches in the Ten Thousand Islands were exposed to a relatively wide range of sand characteristics and inundation. Nevertheless, the mean hatching success (69%) was similar to that reported from other loggerhead nesting beaches (Dodd 1988).
When all nests in the present study were considered, the hatching success decreased with increasing sand water content. Low hatching success of loggerheads has also been related to high water content of sand in laboratory experiments (McGehee 1990). In our study, however, hatching success was related to water content of the sand only when inundated nests were included in the correlation analysis. Short of inundation, increasing water content of the sand in the present study did not affect the hatching success. Sands with very low moisture content have also been known to decrease the hatching success of sea turtle clutches (McGehee 1990; Mortimer 1990). In the present study, the hatching success of clutches at the driest nest sites was no less than that at the other nest sites.
Hatching success in our study was related to the sand water salinity. As the sand water salinity increased, the hatching success decreased. However, the sand water salinity was also related to nest site inundation, and hatching success was not related to the sand water salinity when inundated nests were excluded from the correlation analysis. There have been reports of decreases in hatching success of green turtle clutches with increases in the sand water salinity. For example, on Ascension Island, as the sand water salinity increased, the hatching success decreased and mortality during pipping increased (Mortimer 1990). In a laboratory experiment, the hatching success decreased as the sand water salinity increased from 0 to 50% that of sea water (about 0‰–17‰) (Bustard and Greenham 1968). In sand with water that was from 75 to 100% the salinity of seawater (about 26‰–35‰), no eggs hatched (Bustard and Greenham 1968). In contrast, some of the loggerhead clutches in our study had a relatively high hatching success despite being exposed to high sand water salinities. For example, at 7 nest sites where the mean sand water salinity was greater than 25‰ (about 70% that of seawater), the average hatching success was 81.2%.
In a laboratory experiment, loggerhead hatching success decreased as the particle size of the incubating sand increased (Schwartz 1982). The highest rate of hatching success was in sand that was between 0.125 and 0.25 mm in diameter. On Ascension Island, the hatching success of green turtle clutches also decreased with increasing sand particle size (Mortimer 1990). At least in the latter case, the relationship between increasing sand particle size and decreasing hatching success may have been related to very dry nest conditions when coarse-grained sands were present, and very low water content in the sand may have been the ultimate cause of lowered hatching success (Ackerman 1997). In the same study, low sand water content may have also explained why there was a decrease in hatching success with an increase in air-filled pore space. Although the present study examined loggerhead hatching success in sands with a wide range of mean particle diameters and amounts of air-filled pore spaces, including sands that were extraordinarily coarse and porous, no effect on hatching success was detected. In our study area, the very moist sand conditions may have prevented lethal amounts of egg-water loss, even in coarse sands with a large amount of air-filled pore space.
About 31% of the eggs in our nests did not hatch. Of the unhatched eggs, 74.9% had no grossly observable signs of embryonic development (they were either unfertilized or died before 27 days of incubation). As on other loggerhead nesting beaches (Caldwell 1959; Hughes 1974; Limpus 1985), eggs with no discernible embryos made up the bulk of unhatched eggs. Assuming that at least 90% of the eggs in each clutch were fertilized (Blanck and Sawyer 1981), at least 63.3% of the fertilized eggs that failed to hatch died during the first 27 days of incubation.
Asphyxiation during inundation was likely the primary cause of embryonic mortality in our study. Embryos most often died during the first half of the incubation period, but we did find later-term embryos that had apparently been killed by inundations. On loggerhead nesting beaches in Australia, early-term loggerhead embryos were thought to be more vulnerable to inundation than later term embryos were (Limpus 1985). In contrast, the early-term embryos (19–22 days old) of red-eared sliders (Trachemys scripta elegans) suffered no significant mortality when exposed to water-saturated substrates for periods up to 48 hours, whereas later-term embryos (39–42 days old) all died when exposed to water-saturated substrates for more than 12 hours (Tucker et al. 1997). In our study, the prevalence of early embryonic death could simply be an artifact of the frequency of inundations at some sites: inundation occurred frequently enough to kill embryos before they reached a later stage of development.
Even though inundation reduced the hatching success in our study, some clutches survived inundation remarkably well. For example, one nest with a hatching success of 70% was completely inundated at least twice, inundated at the middle-clutch level at least 4 other times, and inundated at the bottom-clutch level at least another 4 times. Another nest with a hatching success of 92.8% was inundated at the middle-clutch level at least twice and at the bottom-clutch level at least 1 other time. There were no readily apparent similarities among clutches that tolerated inundation so well: the age of the eggs at inundation varied, as did the characteristics of the sand at the nest sites.
Conservation Implications
To avoid potential egg mortality caused by inundation, nesting-beach management strategies often call for moving sea turtle clutches that are low on the beach to areas of higher elevation (Dutton and Whitmore 1983; Wyneken et al. 1988; Marcovaldi et al. 1997). Although inundations can reduce hatching success, the present study demonstrated that loggerhead clutches could survive inundation. Clutches that were partially inundated many times or completely inundated only once or twice still produced hatchlings.
The incubation environment influences characteristics of turtle embryos such as mobilization of yolk nutrients and patterns of nitrogen excretion (Packard et al. 1984; Gutzke et al. 1987; Janzen et al. 1990) and characteristics of hatchlings such as sex (Bull 1980; Ewert and Nelson 1991), size (Janzen 1993; Spotila et al. 1994), energy reserves (Gutzke et al. 1987; Packard et al. 1988; Janzen et al. 1990), growth (Brooks et al. 1991; Spotila et al. 1994; O'Steen 1998), locomotor performance (Miller et al. 1987; Janzen 1993, 1995), and behavior (Janzen 1993, 1995; O'Steen 1998). In loggerheads, the incubation environment has been found to influence hatchling sex (Yntema and Mrosovsky 1982), size (McGehee 1990; Foley 1998), hydration (Foley 1998), early swimming behavior (Foley 1998), and early growth (Foley 1998). Moving loggerhead clutches from low-beach areas may prevent the eggs from being exposed to unique incubation environments and these unique incubation environments could influence the development of hatchling characteristics that increase survival rates or otherwise increase the reproductive success of females. For example, in the south Florida loggerhead nesting population (as described in Encalada et al. 1998), typical incubation temperatures cause hatchling sex ratios to be strongly biased to females (Mrosovsky and Provancha 1989, 1992; Hanson et al. 1998). Because clutches that are deposited lower on the beach typically experience lower incubation temperatures (Mrosovsky et al. 1984; Baptistotte et al. 1999; Foley et al. 2000), low-beach nests in south Florida may be some of the few that produce male hatchlings. A nesting pattern that includes occasional low-beach nesting (even though low-beach nests may have a lower hatching success than high-beach nests) might lead to an increase in the fitness of females because such a nesting pattern may increase the odds of producing less-common males (Fisher 1930).
To ensure that conservation practices do not negate successful reproductive strategies, the relocation of low-beach nests on a particular beach should be considered only when it is certain that all eggs will be destroyed by inundation if the clutch is not moved. Additionally, because beaches are very dynamic and because loggerhead clutches can tolerate a wide range of incubation environments, including various levels and frequencies of inundation, areas that are assumed to be completely unsuitable for egg incubation should be continually and carefully assessed. The criterion for determining when a nest site is unsuitable should be no hatchling production rather than low hatchling production.
Map of 8 keys (Brush, “B”, Turtle, Gullivan, Whitehorse, Hog, Panther, and Round) in the northern half of the Ten Thousand Islands, Collier County, Florida, where sand characteristics and clutch inundation were monitored at 90 loggerhead nest sites.
Diagram of the device used to detect inundation of loggerhead (Caretta caretta) clutches. A 50.8-mm-diameter polyvinyl chloride (PVC) pipe was buried at the nest site (as indicated) and three 12.7-mm-diameter chlorinated polyvinyl chloride (CPVC) pipes were placed inside the PVC pipe. As groundwater rose to within 60 cm of the sand's surface, water flowed into the PVC pipe through the well points (cuts through the wall of the pipe). If groundwater rose to within 45 cm of the sand's surface, water also rose within the PVC pipe and flowed into the CPVC pipe that had a hole at that depth (A). If groundwater continued to rise, it would eventually flow into the CPVC pipe with a hole at 30 cm (B) below the sand's surface and then into the CPVC pipe with a hole at 15 cm (C) below the sand's surface. Sealed caps at the bottom of the CPVC pipes retained the collected water. The screw cap at the top of the PVC pipe kept rain water out of the device, reduced water loss because of evaporation, and allowed access to the CPVC pipes. The CPVC pipes were checked for collected water by removing the screw cap at the top of the PVC pipe. The salinity of collected water was measured with a portable refractometer.
The combined water-content profile of sand at 90 loggerhead nest sites in the northern half of the Ten Thousand Islands, Florida, where hatching success evaluations were conducted. The sand samples were collected during 1993 and 1994. The error bars represent the standard deviations.
