The temperature of a sea turtle nest during incubation is critical to successful development of hatchlings and the determination of their sex. The temperature of sea turtle nests depends on a complex interaction between physical, chemical, and biological factors, which can vary widely both within and between beaches. For example, environmental factors such as rainfall and air temperature (Bustard and Greenham 1968) and physical factors such as nest depth, sand moisture, and gas exchange (Maloney et al. 1990; Mrosovsky 1994; Broderick et al. 2000) can affect the nest temperature. Metabolism of the egg mass can also raise the nest temperature by 2° to 6°C over the course of incubation (Bustard and Greenham 1968; Bustard 1970; Blanck and Sawyer 1981; Maloney et al. 1990; Godfrey et al. 1997; Broderick et al. 2001; Matsuzawa et al. 2002).

Green turtle eggs develop only between nest temperatures of 25° and 34°C (Bustard and Greenham 1968; Bustard 1970). However, for hatch success rates above 60%, a more optimal range of 27°–32°C is required (Bustard and Greenham 1968). Temperature influences the rate of development of the embryos throughout incubation (Eckert et al. 1988). At higher temperatures, tissue synthesis and tissue maintenance are greater, leading to a higher rate of development and shorter development duration (Booth 1998).

Most studies on sea turtle clutch temperature have been conducted in the laboratory under controlled conditions at constant temperatures. However, in natural nests, temperatures fluctuate both daily and seasonally (Mrosovsky et al. 1984; Bull 1985; Spotila et al. 1987; Georges et al. 1994). During the daily fluctuations of nest temperature, excursions above the daily mean temperature may be more important than excursions below (Bull 1985; Georges et al. 1994). This is because metabolism is faster at high temperatures, which implies that even if the clutch spends half of the day above the mean temperature, more than half of the development for that day will have occurred during this period (Georges et al.1994). A model for a CTE has been proposed, which represents the average temperature over the incubation period, while accounting for the influence of daily temperature fluctuations (Georges et al. 1994). CTE is therefore a better indicator of average temperature of sea turtle nests with high degree of daily temperature fluctuation. However, this generally is only true during the spring and summer months in nesting areas exposed to significant daily and seasonal changes in sand temperature (Booth and Astill 2001). In places closer to the equator—like Malaysia, where the sand temperatures remain relatively constant over the nesting season—daily variation in nest temperature is generally low.

Nest temperature also affects the sex of developing embryos, as sea turtles, like many other reptiles, exhibit temperature-dependent sex determination (TSD) (Mrosovsky and Yntema 1980; Ewert et al. 1994; Georges et al. 1994). It is hypothesized that the nest temperature during the middle third of incubation influences the differentiation of the gonads of the developing embryo through a number of possible mechanisms, including 1) temperature-dependent synthesis and activity of specific enzymes, 2) the presence of heat shock proteins, and 3) temperature sensitive gene expression (Yntema and Mrosovsky 1982; Spotila et al. 1987; Georges et al. 1994; Mrosovsky 1994). The temperature that produces a 50:50 sex ratio is known as the pivotal, or threshold temperature (Mrosovsky 1991; Georges et al. 1994). For green turtles, this is usually between 28° and 31°C but varies between species and populations (Miller and Limpus 1981). In all species of marine turtle, females are predominantly produced at temperatures above the pivotal temperature and males are predominantly produced below it (Mrosovsky and Yntema 1980; Georges et al. 1994; Godfrey et al. 1996).

In hatchery operations, factors such as nest depth and shading can be manipulated to influence nest temperature. Nest depth can influence the nest temperature, with deeper nests being constant, because they are less affected by solar radiation and air temperature (Mrosovsky 1994; Broderick et al. 2000). Shading can also influence the temperature of sea turtle nests by reducing the exposure to direct radiation from the sun. In warm tropical climates like Malaysia, unshaded hatcheries result in nest temperatures near the upper lethal limits for sea turtle development (Limpus 1993; Ibrahim 1994; Ibrahim and Thalathiah 1994; Tiwol and Cabanban 2000). This results in very poor hatch success and a severely female-biased sex ratio of hatchlings produced (Limpus 1993; Tiwol and Cabanban 2000). Shading hatcheries with 70% shade cloth in peninsular Malaysia has reduced nest temperatures to within the optimal range for sea turtle development and produced more desirable hatchling sex ratios (Hamann et al. 2001; Schäuble et al. 2001; van de Merwe et al. 2005).

The present study investigates the influence of hatchery shading and nest depth on the temperature of Chelonia mydas nests relocated to hatcheries in peninsular Malaysia. This study also estimates the influence of metabolic heating of the eggs on nest temperature and how the combination of physical and biological factors contribute to the overall nest temperature over the course of incubation.

METHODS

Study Site

The fieldwork for this study was carried out from 1 July to 1 October 2001 at Ma'Daerah Turtle Sanctuary (4°32′27″N; 103°28′14″E), on the east coast of Malaysia in the state of Terengganu.

Two hatcheries were constructed along the typical plan of Malaysian beach hatcheries. These were wood-framed structures 6 m x 5 m x 2 m and were each designed to hold approximately 40 clutches spaced 90 cm apart. To ensure complete shading of the nests throughout the day, the “cloth hatchery” was enclosed with 70% shade-cloth walls and roof. The “palm hatchery” had 70% shade-cloth walls and locally constructed palm-leaf roof, which effectively represented 100% shading of the hatchery interior throughout the day.

The hatcheries were positioned in the sand above the high-tide mark and immediately below the dune vegetation. To minimize any difference in sand characteristics that might influence nest temperatures, the hatcheries were placed next to each other, the same distance from the high-tide mark.

Egg Collection and Relocation

From 7 July to 6 August, 26 green turtle clutches were relocated into the cloth and palm hatcheries alternatively at depths of 50 and 75 cm. All clutches were handled with care, ensuring not to rotate the eggs, and relocated to the experimental hatcheries within 3 hours of being laid to minimize any movement-induced mortality (Limpus et al. 1979).

Determination of Temperature

Twenty-one DS1921 Thermochron iButtons (Dallas, USA) were activated to record temperature every hour, using the Dallas iButton Viewer software and interface. Each iButton was wrapped in mesh cloth and tied to 15 cm of orange nest tape for protection and ease of retrieval. The iButtons were calibrated in the factory prior to purchase, and are capable of recording and storing temperature readings to the nearest 0.5°C every hour for up to 84 days. This was well within the expected incubation period of 50–60 days. Each unit was a small disc of radius 1 cm and height of 0.5 cm, and was not expected to interfere with either incubation or the movement of hatchlings up the egg chamber.

A single iButton was placed into each of 18 randomly selected clutches at the time of relocation. For each of these clutches, exactly half the eggs were placed into the constructed nest at a depth of 60 cm, followed by the iButton (approximate depth: 45–50 cm) and the remainder of the eggs. During the period that the relocated clutches were incubating, 3 iButtons were placed in the sand adjacent to the relocated clutches at a depth of 60 cm. At the end of incubation, the iButtons were retrieved and nest temperature data were downloaded onto a computer. Mean nest temperature over the entire incubation period was calculated for each nest. Incubation period was defined as the time between clutch relocation and emergence of the first hatchling. Following the experimental period, the iButtons were calibrated against a precision mercury thermometer of known accuracy to ensure that their accuracy was maintained over this period.

Data Analysis

The mean nest temperatures (± SE) for the period of incubation (as defined by the time between nest relocation and emergence of the first hatchlings) were calculated from the hourly temperature recordings of the iButtons using SPSS 11.0 (SPSS Inc, Chicago, IL).

The statistics package Minitab 11 (Minitab Inc, State College, PA) was used to calculate daily mean temperatures, maximum, minimum, and range (max–min) for all retrieved iButtons. From this information, a CTE was calculated for each nest based on the following formula (Georges et al. 1994):

where

M = mean daily temperature

2R = daily range in temperature

T₀ = lower limit of temperature which supports development

T₀ was calculated from the intercept/slope of the regression equation representing the relationship between incubation temperature and the inverse of incubation days (data supplied by the Fisheries Department of Malaysia from a laboratory incubation experiment between 25° and 34°C on C. mydas eggs from a nearby nesting beach in Terengganu). To calculate t′, theoretical values were repeatedly inserted until Equation 2 was balanced.

Forward stepwise regression was performed to determine the variable with the greatest effect on mean nest temperature. Variables used were nest depth, hatchery shading, and number of completely developed eggs. The number of completely developed eggs was defined as the number of eggs that developed to the stage of pipping, and was used to estimate the relative influence of metabolic heating on nest temperature.

To determine the effect of nest depth and hatchery shading on nest temperature prior to the onset of metabolic heating, a two-factorial ANOVA was performed on the mean nest temperature for the first third of incubation, with nest depth and hatchery shading as the factors.

To investigate the influence of hatchery shading and nest depth on the daily temperature range of the nests, a two-factorial ANOVA was performed on the mean daily temperature range with nest depth and hatchery shading as the main factors.

RESULTS

Nest Temperature

The mean daily nest temperature generally increased over the last two-thirds of incubation, followed by a sharp decline before the hatchlings emerged (Fig. 1). This temperature increase was not observed in the temperature data recorded in the nearby sand at 60 cm (Fig. 2).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

The mean temperature of all nests and the mean nest temperature for each of the 4 hatchery/depth treatments are presented in Table 1. The mean nest temperatures (± SE) and constant temperature equivalents (CTE) over the period of incubation were identical to the nearest 0.1°C in all nests. As a result, CTE was not further analyzed.

Table 1. Mean nest temperatures and constant temperature equivalents (CTE) for all nests with iButton temperature loggers.

View Fullscreen

The forward stepwise regression found the number of completely developed eggs (an estimation of metabolic heating) to be the most significant effector of mean nest temperature (Regression: df = 1,12, F = 11.199, p = 0.006). The number of completely developed eggs explained about half of the variation in nest temperatures (R² = 0.43; Fig. 3).

View Figure

• Download as Powerpoint
• Download Figure

There was no effect of nest depth or hatchery shading on the mean nest temperature of the first third of incubation (ANOVA – interaction: df = 1,14, F = 1.366, p = 0.262; depth: df = 1,14, F = 3.619, p = 0.078; shading: df = 1,14, F = 0.274, p = 0.609).

The mean daily range in temperature (± SE) of nests in the cloth hatchery at 50 cm (0.6°C ± 0.03) was larger than nests in the palm hatchery at 50 cm (0.3°C ± 0.04), which were in turn larger than nests at 75 cm in both the palm (0.2°C ± 0.03) and cloth (0.2°C ± 0.03) hatcheries (ANOVA for natural log transformed depth x hatchery interaction: df = 1,14, F = 15.302, p = 0.002; Post Hoc LSD: p < 0.05; Fig. 4).

View Figure

• Download as Powerpoint
• Download Figure

DISCUSSION

Mean nest temperatures between relocation and emergence ranged between 28.0° and 29.8°C. The nest temperatures observed in this study were well within the optimal range for green turtle development (Bustard and Greenham 1968) and near the pivotal temperature of the population (29.4°C, Ibrahim, unpubl. data). The nest temperature profiles for all relocated clutches showed a consistent pattern over the course of incubation. The temperature remained relatively constant over the first two-thirds of incubation followed by a gradual increase peaking towards the end and then falling just before the hatchlings emerge. The increasing temperatures were not observed in temperature data collected from nearby sand at 60 cm depth, which showed no general increase in temperature over the course of the incubation period. This indicates that the increase in nest temperature observed over the latter stages of incubation is not due to external factors (e.g., radiation, rainfall) affecting the sand temperature in the hatcheries, and is presumably due to metabolic heating of the clutch.

Metabolic heating was found to be the dominant factor influencing mean nest temperature over the course of incubation, accounting for a rise in mean daily nest temperature of between 2.0° and 4.4°C. This increase in hatchery nest temperature due to metabolic heating is similar to increases observed in natural C. mydas (Broderick et al. 2001) and Caretta caretta (Matsuzawa et al. 2002) clutches. In this study, metabolic heating was estimated as the number of completely developed eggs in each nest as this gives an indication of both clutch size and hatch success. As each embryo develops, more tissue is synthesized, and heat is generated by the embryo as it maintains tissues and continues growth (Booth 1998). Therefore, each egg that incubated the full term to hatching contributed to the metabolic heating of the nest. Nests with low hatchling numbers demonstrated much smaller increases in nest temperature over the course of incubation compared to nests with high hatchling numbers. For example, nest C12 (cloth, 75 cm) produced 22 hatchlings (compared to a mean of 81 ± 4.1 for the other nests) and showed very little increase in temperature over the last third of incubation.

The dominant influence of metabolic heating on nest temperature indicates that clutch size may be an important predictor of nest temperature. Increased metabolic heating in larger clutches could therefore increase embryo mortality in nesting areas where sand temperatures are already close to the upper tolerable limits for sea turtle egg development. The effect of clutch size on nest temperatures will also influence hatchling sex ratio. Clutches with larger numbers of incubating eggs, and hence greater metabolic heating, will produce warmer nest temperatures and more female hatchlings. It has been estimated that metabolic heating can account for a rise of up to 30% of female hatchlings in C. mydas clutches and should be considered when estimating natural sex ratios (Broderick et al. 2001). In a hatchery, where all nest are incubating under very similar physical and chemical properties, sex ratio is likely to be driven by metabolic heating. Therefore, as clutch size differs between individuals and between subsequent clutches of the same nesting female (Miller 1997), sex ratio in the hatchery could change over the course of the nesting season.

An evolutionary adaptation to reduce nest temperature could be to begin laying nests in cooler months of the year or to reduce the number of eggs in each clutch. A reduction in clutch size would decrease the degree of metabolic heating and allow eggs to incubate successfully within higher sand temperatures. However, with a further 2°C increase in atmospheric temperature predicted over the next 40 years due to global warming (Burgman and Lindenmayer 1998), it is unlikely that marine turtles could adapt these strategies rapidly enough. It is therefore likely that conservation officers will be forced to manage this situation through manipulation of the nest environment. Spraying water on nests has been used to reduce hatchery nest temperatures (Naro-Maciel et al. 1999), and future research could identify the value of splitting clutches into smaller units that will generate less metabolic heat and therefore have lower mean nest temperatures. However, this must be approached with caution, because splitting the clutch further interferes with the natural process of incubation. The effect of metabolic heating could also be used to investigate a theoretical maximum clutch size for a given sand temperature. In hatcheries, where the sand temperature can be manipulated through shading and water spraying, larger clutches may be able to be accommodated in each nest, which would save space and hence costs.

There was no significant effect of nest depth or hatchery shading on the mean nest temperature during the first third of incubation. The nest temperature during this period is determined predominantly by physical and environmental factors (Ackerman et al. 1985; Maloney et al. 1990), because the embryos are at early stages of development and producing negligible metabolic heat. There was, therefore, no extra cooling effect induced by increasing hatchery shading from 70% to 100% or from increasing nest depth from 50 to 75 cm at our study site.

The daily fluctuation of nest temperature can also influence the rate of development and has been the rationale for calculating constant temperature equivalents (Georges et al. 1994). In this study, the significantly larger daily temperature range in nests at 50 compared to nests at 75 cm suggests that these nests, being closer to the surface, were more affected by solar radiation and air temperatures. The mean range in 50-cm nests was also larger in the cloth hatchery, suggesting that the extra shading effect of the palm roofing may have diminished the effect of solar radiation and created a more constant temperature within the palm hatchery at a depth of 50 cm. However, all nest temperature fluctuations were relatively small (less than 0.6°C), leading to the constant temperature equivalents being identical to mean nest temperatures (to the nearest 0.1°C). It is, therefore, not likely that the fluctuations observed in this study were influencing embryonic development. Constant temperature equivalents are probably more useful in more subtropical environments, where greater diurnal temperature fluctuations occur.

Nest temperature studies are a critical part of hatchery management. The results of this study indicated that shading between 70% and 100% and nest depths between 50 and 75 cm were required to maintain nest temperatures within optimal developmental ranges. However, the difference between 70% and 100% shading were minimal. Similarly, nest depths between 50 and 75 cm showed very little variation in nest temperature or daily temperature ranges. More importantly, the effect of the metabolic heating of the eggs was found to be the most significant influence on nest temperature. However, as nest temperature is determined by a complex interaction of physical, chemical, and biological factors, generalizations for other hatchery and in situ nests cannot be drawn. To ensure the correct use of hatcheries elsewhere, this study needs to be repeated in each unique hatchery area. The success of hatcheries depends on being able to correctly manipulate factors that can be controlled, such as nest depth and shading, while taking into account the influence of factors that cannot be easily controlled, such as metabolic heating.