Core and Body Surface Temperatures of Nesting Leatherback Turtles (Dermochelys Coriacea)

Journal of Thermal Biology 51 (2015) 15–22

Contents lists available at ScienceDirect

Journal of Thermal Biology

journal homepage: www.elsevier.com/locate/jtherbio

Core and body surface temperatures of nesting leatherback turtles (Dermochelys coriacea)

Thomas J. Burns n, Dominic J. McCafferty, Malcolm W. Kennedy n

Institute of Biodiversity, Animal Health and Comparative Medicine, College of Medical, Veterinary and Life Sciences, Graham Kerr Building, University of Glasgow, Glasgow G12 8QQ, Scotland, UK article info abstract

Article history: Leatherback turtles (Dermochelys coriacea) are the largest species of marine turtle and the fourth most Received 27 November 2014 massive extant reptile. In temperate waters they maintain body temperatures higher than surrounding Received in revised form seawater through a combination of insulation, physiological, and behavioural adaptations. Nesting in- 27 February 2015 volves physical activity in addition to contact with warm sand and air, potentially presenting thermal Accepted 1 March 2015 challenges in the absence of the cooling effect of water, and data are lacking with which to understand Available online 3 March 2015 their nesting thermal biology. Using non-contact methods (thermal imaging and infrared thermometry) Keywords: to avoid any stress-related effects, we investigated core and surface temperature during nesting. The Thermal biology mean7SE core temperature was 31.470.05 °C (newly emerged eggs) and was not correlated with en- Thermography vironmental conditions on the nesting beach. Core temperature of leatherbacks was greater than that of Non-invasive techniques hawksbill turtles (Eretmochelys imbricata) nesting at a nearby colony, 30.070.13 °C. Body surface tem- Core temperature Dermochelys coriacea peratures of leatherbacks showed regional variation, the lateral and dorsal regions of the head were Eretmochelys imbricata warmest while the carapace was the coolest surface. Surface temperature increased during the early nesting phases, then levelled off or decreased during later phases with the rates of change varying between body regions. Body region, behavioural phase of nesting and air temperature were found to be the best predictors of surface temperature. Regional variation in surface temperature were likely due to alterations in blood supply, and temporal changes in local muscular activity of ﬂippers during the different phases of nesting. Heat exchange from the upper surface of the turtle was dominated by radiative heat loss from all body regions and small convective heat gains to the carapace and front ﬂippers. & 2015 Elsevier Ltd. All rights reserved.

1. Introduction Casey et al., 2014). Thermoregulation is aided by counter-current heat exchangers within the front and rear ﬂippers (Greer et al., Leatherback turtles (Dermochelys coriacea) are the largest spe- 1973), a vascular plexus lining the trachea to reduce respiratory cies of marine turtle, with adult females having a mean mass of heat loss, analogous to that of nasal turbinates found in birds and around 400 kg (Georges and Fossette, 2006). As adults they inhabit mammals (Davenport et al., 2009a), extensive adipose tissues in a broad range of water temperatures, migrating between high la- the head and neck, and major blood vessels buried deep within titude, prey-rich temperate waters and the tropics or subtropics, the insulated neck (Davenport et al., 2009b). where beaches provide the conditions for laying and development In addition to these physiological adaptations for controlling of eggs. Leatherbacks exhibit a range of physiological and beha- heat loss, behaviour plays a key role in temperature control. vioural adaptations to cope with different environmental tem- Swimming (and the consequent metabolic heating) maintains a peratures, allowing them to remain active in temperate waters and high body to water temperature differential (Bostrom and Jones, prevent overheating in the tropics. Their thermoregulatory strat- 2007) and turtles change ﬂipper stroke rate in response to dif- egy likely involves aspects of both gigantothermy due to their ferent water temperatures (Bostrom et al., 2010). Furthermore, large body mass, extensive fatty insulating tissue layers and leatherbacks may dive to cooler waters to lose heat in tropical adaptable blood circulation system (Paladino et al., 1990), and waters (Southwood et al., 2005; Bostrom and Jones, 2007) and, endothermy due to internal heat production (Bostrom et al., 2010; conversely, in temperate seas may bring prey items to the surface to warm them before ingestion (James and Mrosovsky, 2004).

n Nesting is the only time when female leatherbacks are known to Corresponding authors. E-mail addresses: [email protected] (T.J. Burns), return to land, the process may last over two hours, and requires, [email protected] (M.W. Kennedy). with the exception of egg laying, extensive use of the ﬂippers http://dx.doi.org/10.1016/j.jtherbio.2015.03.001 0306-4565/& 2015 Elsevier Ltd. All rights reserved. 16 T.J. Burns et al. / Journal of Thermal Biology 51 (2015) 15–22

Table 1 Published values for leatherback body temperature shown alongside values derived from this study.

o Mean TB7SE (range) C Method of measurement Location (turtle activity) n Source

31.4 70.05 (30.6–32.5) Egg surface temperature (corrected) Trinidad (nesting-laying) 65 This study 29.170.23 (27.3–30.6) Sub-carapace Costa Rica (inter-nesting) 3 Southwood et al. (2005) 30.270.35 (27.1–33.2) Gastrointestinal Costa Rica (inter-nesting) 4 Southwood et al. (2005) 28.370.07 (28.1–28.7) Gastrointestinal USVI (inter-nesting) 8 Casey et al. (2010) 31.1n (30.5–33.5) Egg temperature Mexico (nesting) 10 Mrosovsky (1980) 30.6 (29.8–31.4) Egg temperature French Guiana /Suriname (nesting) 24 Mrosovsky, Pritchard (1971) 3370.45 (32–34) Not speciﬁed Costa Rica (nesting-restrained post laying) 5 Lutcavage et al. (1992) 31.6 70.3 Inserted 25–30 cm into body through drilled hole Costa Rica (nesting-restrained post sand scattering) 10 Paladino et al. (1996) 30.870.2 Inserted 25–30 cm into body through drilled hole Costa Rica (nesting-laying) 3 Paladino et al. (1996) 31.4 70.4 Inserted 25–30 cm into body through drilled hole Costa Rica (nesting-‘exercising’)10Paladino et al. (1996) 31.5 Gastrointestinal (juvenile- 37 kg) Laboratory (in water tank) 1 Bostrom et al. (2010) 31.8 Gastrointestinal (juvenile- 16 kg) Laboratory (in water tank) 1 Bostrom et al. (2010) 26.470.23 (25.4–27.3) Gastrointestinal Northwest Atlantic (foraging) 7 Casey et al. (2014) 24.3370.94 (21.6–25.8) Cloacal Nova Scotia (restrained) 4 James and Mrosovsky (2004)

n Represents a median value, values from this study are shown in bold.

(Eckert et al., 2012). Flipper movement has been suggested as the the nest hole; 4-laying; 5-refilling the nest hole; 6-sand scattering primary source of heat production at sea (Bostrom and Jones, (also termed ‘camouflaging’ and ‘disguise’ elsewhere in the lit- 2007; Bostrom et al., 2010) and it is likely that flipper activity erature); 7-return to the sea (for a full description of nesting during nesting will result in substantial heat production. Biophy- phases see Table S1). sical modelling suggests that core temperature of leatherbacks increases throughout the nesting process and rises in core tem- 2.2. Body core temperature perature during nesting are predicted as a result of future rising temperatures in their breeding range (Dudley and Porter, 2014). The surface temperature of freshly laid eggs was used as a non- However, measurements on land of core temperature in this contact proxy of core body temperature. Measurements were ta- species are, surprisingly, relatively scarce (Table 1). þ The aims of this study were, first, to estimate the core tem- ken using a Fluke 62 Max portable infrared thermometer ¼ – m ¼ ° perature of nesting leatherbacks, second, to examine spatial and (spectral range 8 14 m, accuracy 1 C or 1%, thermal ¼ ° temporal variation in surface temperature through all phases of sensitivity 0.1 C) set to an emissivity of 0.98, which is within the nesting behaviour, and, third, to estimate the relationships be- range of previously used values for emissivity of living tissues tween both core and surface body temperatures and the en- (McCafferty et al., 2013; Rowe et al., 2013; Mortola, 2013). Infrared vironmental conditions on the nesting beaches. An important and thermometers were calibrated against a thermocouple, which had novel aspect of this study was that all measurements were made itself been calibrated against a mercury thermometer of 0.1 °C with minimal or no disturbance using non-contact thermometry sensitivity. The recorder lay in the sand with one arm extended and thermal imaging. For comparison we also estimated the core into the nest hole (some sand was removed from one side of the temperatures of ectothermic nesting hawksbill turtles (Eretmo- nest hole to allow better access) and observed the eggs as they chelys imbricata). were laid. Measurements were taken only on eggs that were freshly laid (within one or two seconds of emergence) and had no sand attached to the side which was being measured. As mea- 2. Materials and methods surements of surface temperature were taken so soon after the eggs emerged from the female the effects of variation in cooling 2.1. Study area and discrimination of nesting phases rates between eggs will be negligible. Measurements were also taken for nesting hawksbill turtles (to provide a smaller ec- Fieldwork on leatherback turtles was carried out at Fishing tothermic comparator) at Hermitage Bay, Tobago (approximately Pond Beach (approximately 10.58° N, 61.02° W) on the East coast 11.31° N, 60.57° W) using the same methods. of Trinidad, West Indies. Fishing Pond is one of three protected To correct for potential post-laying cooling and/or inaccuracy beaches on the island which receive high densities of nesting fe- due to imprecise emissivity setting, a correction factor for egg male leatherbacks during the nesting season which runs from surface temperature was estimated and applied to all measure- March to the end of August, peak nesting occurring during April ments of both study species. The internal temperatures of up to and May (Bacon, 1970). Beach visits were made on a regular basis seven of the small shelled albumen gobs (Sotherland et al., 2003), from mid-June through to mid-August 2013 and 2014 during the often referred to as ‘yolkless eggs’ (Eckert et al., 2012), produced – hours 8 pm 1 am (local time). The beach was regularly patrolled by 20 leatherbacks were measured using a K-type thermocouple by rangers but rarely by tourist groups, so observations could be (calibrated against a mercury thermometer 70.1 °C) connected to made with minimal disturbance to the nesting females. a digital thermometer (RS 206-3750); leatherbacks are a vulner- Discrimination of the different behavioural phases of the able species (Wallace et al., 2013) so only these sterile ‘eggs’ were nesting process were modified from those in previous publications used for this purpose. The linear regression of mean surface (Hendrickson, 1958; Carr and Ogren, 1959, 1960; Pritchard, 1971). against the mean internal temperature was Seven behavioural phases were defined in this study, with the prospecting and body pitting phases being combined for practical Internal egg temperature purposes: 1-approach (movement from surf to upper beach); 2 – =×(0.8262 egg surface temperature) + 5.9378 prospecting/body-pitting (selection of a nest site once in nesting 2 area and preparing the nesting site for excavation); 3-excavation of (r ¼0.62, F1,18 ¼29.8, p¼ o0.01), (see Fig. S1) T.J. Burns et al. / Journal of Thermal Biology 51 (2015) 15–22 17

Fig. 1. Example of thermal images collected. From left to right: top, full body proﬁle and from posterior; middle, front and rear ﬂippers; bottom, head lateral and dorsal. False colour palette setting is ‘ironbow’ and emissivity (ε)¼0.98 (turtle surface). At this setting the sand temperature (ε¼0.76) is approximately 0.7–2.0 °C warmer than shown. The full colour images are available in the online version of this paper.

2.3. Thermal imaging and body surface temperature each body region using polygons fitted around the outline of the dorsal surface of the front flipper, the dorsal surface of the head All images were collected using a Fluke Ti20 thermal camera and the lateral surface of the head. Rear flipper mean temperature (spectral range¼7.5–14 mm, accuracy¼2 °C or 2%, thermal was estimated by a straight line running through the proximal two sensitivity¼0.1 °C). The thermal camera was calibrated against a thirds of the centre of the dorsal surface of the rear flipper and thermocouple, which had been calibrated against a mercury carapace temperature was estimated from the mean of ten point thermometer 70.1 °C. A series of six images were taken once and measurements from sand free areas. In cases where multiple on occasion twice (excavation and camouflaging phase only) for images had been taken for a body region during a phase the each of the behavioural phases of the nesting process (Fig. 1). average of the two extracted values was used. Images were taken from only one side of a turtle, with the same side being used consistently throughout nesting unless access 2.4. Environmental measurements limitations made it necessary to image the opposite side, the side used varied between individuals (for a description of images taken Multiple recordings were taken for each phase of the nesting see Table S2). Variation in the side of the turtle which was ther- process for each of the following (unless stated otherwise). Wind mally imaged could potentially affect recorded surface tempera- speed was measured at the highest point of the carapace using an tures due to asymmetrical arrangement of major organs such as Explorer Skywatch 2 handheld anemometer (accuracy¼ 73%, the liver, stomach and intestines. However, we expect that any sensitivity¼0.1 ms1, range¼0–42 ms1). Substrate temperature influence of this asymmetry will be dampened due to insulation of at a depth of 6, 8, 10, 12, 14 and 16 cm depth below the sand the organs from the skin surface; Davenport et al. (1990) reported surface was recorded using a probe which consisted of several K the major organs to be within a continuous blubber capsule en- type thermocouples glued to the side of a plastic rod, read by a closed within the thick oily skin, with the points of penetration of digital thermometer (RS 206-3744). It was placed into the sand the capsule by the limbs and tail having a thick cuff of blubber. adjacent to the turtle at equal beach height. A mean value of Emissivity was set at 0.98 (as explained in Section 2.2). substrate temperature across the depths was calculated. Air tem- Images were analysed using InsideIR™ image analysis software perature and relative humidity were measured at ground level version 4.0 (http://www.fluke.com/fluke/uken/support/software/ using a UTI Hygrometer (accuracy¼ 71 °C and 5% relative hu- ti-update.htm). Mean temperature values were extracted from midity (r.h.), resolution¼0.1 °C and 1% r.h., range¼0–49.9 °C and 18 T.J. Burns et al. / Journal of Thermal Biology 51 (2015) 15–22

20–95% r.h.). Cloud cover was estimated for the area of sky directly 35 above the nesting female using a quadrat (16 squares of 8 8cm2) Leatherback held at arm’s length above an observer lying on the sand adjacent Hawksbill to the turtle (recorded three times during the nesting process). 30 Sea surface temperature (SST) was estimated from daytime satellite derived data (MODIS aqua Level 2, http://oceancolor.gsfc. 25 nasa.gov/). Data was gathered for every available day during ﬁ eldwork periods in 2013 and 2014 from seven to 20 point mea- y c surements in areas within 30 km from the coastline of both Tri- n 20 e nidad and Tobago. Within the Matura Bay and Galera Point area in u q e

Trinidad, and in areas east of Roxborough and Castara Bay, on the r south and north coast of Tobago, respectively. Areas sampled in F 15 Trinidad have been shown to be regularly used by inter-nesting female leatherbacks (Eckert, 2006). Data availability was limited 10 by satellite coverage of the area and quality, only data from the highest quality retrieval category were extracted for analysis. 5 2.5. Statistical analysis

All statistical procedures were carried out using R software 0 (http://www.r-project.org), unless otherwise stated. Spearman 28 29 30 31 32 33 Rank Correlations were used to examine relationships between Egg temperature (oC) body core temperature and the environmental conditions individual turtles experienced during the period leading up to and Fig. 2. Distribution of egg temperatures, used as an estimation of core temperature, of leatherback and hawksbill turtles in Trinidad and Tobago, respectively, estimated including egg laying. A Bonferroni Correction was applied to ac- from infrared thermometry of freshly laid eggs and calibrated as described in count for the use of multiple testing and to account for ties in the Section 2 (see also Fig. S1). Curves show fitted normal distributions. A mean of test data, a sample size of 10,000 was simulated under the null 14.1 75.19 and 17.978.87 eggs, during a total of 65 and 29 nestings were sampled hypothesis of no correlation and the p-value was defined as the for leatherbacks and hawksbills, respectively. probability that the observed value is greater than or equal to the null distribution. Power analysis (pwr package) was also carried of eggs (see Section 2 and Fig. S1). The mean core body tem- out to assess the power of this dataset has to test significance, perature (Tb) of our nesting leatherbacks was thereby estimated at using what may be considered conventional significance and 31.4 70.05 °C(7SE). For comparison with a species not con- power levels (Stefano, 2003). sidered to be an endothermic/gigantothermic mesotherm like Linear mixed effects models (lme4 package) were used to ex- leatherbacks, we carried out similar measurements on hawksbill amine factors affecting body surface temperature. Behavioural turtles from a geographically nearby colony. Estimated hawksbill phase of nesting and region of body surface were categorical fixed core temperatures were significantly lower, 30.070.13 °C (GLM, effects, the environmental conditions during each phase of nest- F 1, 92 ¼229.4,po0.001) (Fig. 2). Moreover, Tb was greater during ing: air temperature, relative humidity, sand temperature and 2013 compared to 2014 for both species, (GLM, F 1, 91 ¼21.7, wind speed-were continuous fixed effects and individual was a po0.001) with the differential between years depended on spe- random effect. Cloud cover was not included as a factor because cies (GLM, interaction F 1, 90 ¼11.3, p¼0.001), hawksbills showing a sampling was insufficient across all phases of nesting. Models greater differential than leatherbacks (Fig. S2). Leatherbacks also were compared using second order Akaike Information Criterion clearly exhibit a narrower distribution of temperatures than values (AICc), AICc differences and Akaike weights. hawksbills (Fig. 2). Tb was not correlated with curved carapace lengths (CCL) of either species (leatherbacks, r2¼0.03, n¼53, 2 F1, 51 ¼1.66, p40.05; hawksbills, r ¼0.04, n¼29, F1,27¼1.19, 3. Results p40.05 ).

There was no significant correlation between Tb of the lea- 3.1. Environmental conditions therbacks and beach environmental conditions- air temperature, relative humidity, substrate temperature, wind speed and cloud Air and sand temperatures were found to decrease and relative cover (Spearmann Rank pZ0.01 in all cases, n¼14 individuals). humidity increase during the nesting process (Fig. S3). Mean Power analysis, however, indicated that the sample size was in- (7SE) wind speed was 3.570.13 ms1 and mean cloud cover was sufficient for detection of significance in factors having a relatively 2 3870.05%, with no discernible pattern of change. Mean sea sur- weak effect on Tb (r ¼0.78, when n¼14, significance level¼0.01, face temperature was significantly higher during 2013 than 2014 power¼0.8); observed correlation coefficients suggest a sample

(GLM, F1, 72 ¼17.29, po0.001) (Fig. S4). There was no signiﬁcant size of 21 to 59 individuals is necessary to detect a signiﬁcant difference in SST between locations for either year (GLM, correlation.

F1, 70 ¼0.12, p¼0.73). 3.3. Surface temperature 3.2. Core temperature Turtles emerging from the sea should brieﬂy cool by evapora- In order to obtain an estimate of core temperatures of the tion before exposure to sources of heating (day-warmed sand, air, population under study at nesting, we measured surface tem- physical exercise) and loss of cooling by water. Compensatory peratures of between ﬁve and 25 eggs laid by each of 65 leather- mechanisms for heat loss would be detectable at the body surface, back turtles using infrared thermometers. We assumed an emis- and may be regional. Thermographic images show regional dif- sivity of 0.98 for this and then applied a correction factor based ferences in surface temperatures (Fig. 1), and sequential mea- upon parallel thermocouple readings from the centres of a sample surements taken throughout several nestings showed a consistent T.J. Burns et al. / Journal of Thermal Biology 51 (2015) 15–22 19

28.5 4. Discussion

4.1. Core temperatures

In this study core temperature was estimated from the surface 28.0 temperature of freshly laid eggs using infrared thermometry.

C) Measurements were made within one or two seconds of laying o ( and several eggs were measured for each turtle. To our knowledge e r

u no other study has used this non-invasive technique to estimate t a

r core temperature, or reported sample sizes of this magnitude.

e 27.5

p Mean core temperature fell within the range of previously pub- m

e lished values recorded during nesting, using various techniques t

e (Table 2). Values were well below that of the estimated critical c a

f thermal maximum temperature for leatherbacks of 40 °C(Spotila r u 27.0 et al., 1997). The greater core temperature of leatherbacks com- S pared with hawksbills may be related to the large difference in body size between the species, leading hawksbill body temperatures to more closely follow environmental temperatures. Mean core temperatures were within the range of tempera- 26.5 tures reported for freely swimming individuals in tropical waters (Table 1), showing that, at least up until the point of laying, nesting leatherbacks do not show core temperatures elevated beyond the Carapace Front flipper Rear flipper Head dorsal Head lateral range reached while in tropical seas. However, this does not ne- cessarily mean that leatherbacks are at no risk of overheating Fig. 3. Surface temperature of different body regions of nesting leatherback turtles during nesting. Thermoregulatory mechanisms are used in tropical (n¼17 individuals). Mean (7SE) of all measurements taken throughout all the nestings reported, of turtles emerging from the sea between approximately 8 pm waters (Southwood et al., 2005; Bostrom and Jones, 2007), sug- and midnight. gesting that leatherbacks are required to avoid significant rises in core temperature. Data on core temperature changes during nesting is currently lacking, but a previous biophysical model rank order in which the head was consistently hottest, followed by predicted that core temperatures will continue to rise throughout the flippers and the carapace (Fig. 3). Detailed breakdowns of the nesting process (Dudley and Porter 2014). If this were so it may these data (Fig. 4) show that different surface regions increased in be expected that body surface temperature would also rise due to fl temperatures at different rates, the rear flippers showing the lar- changes in peripheral blood ow. This was not observed in this gest and most rapid increase up to the laying phase. study, as surface temperatures tended to level out, and even de- Body region, behavioural phase of nesting, and air temperature crease in the later phases, even for areas such as the lateral region of the head, where reduced insulation (Davenport et al., 2009b) (T ) were found to be the best predictors of surface temperature a and proximity to the core may lead surface temperature to more (linear mixed effect models, Table 2). There was a positive re- closely track core temperature (Fig. 4). The relatively constant lationship between leatherback surface temperature and air tem- surface temperature during laying could be explained by a de- fi perature. The 95% con dence intervals for these factors: body re- creased metabolic rate and inactivity of the flippers reducing – gion (maximum range across all levels; 0.0586 1.2661); beha- metabolic and muscular heat production during this phase (Pala- vioural phase of nesting (0.1828–2.1878); and air temperature dino et al., 1996). (0.1196–0.5247), exclude zero indicating a significant influence It is also noteworthy that there were significant differences in upon Ts. It is worth noting that there is only a small difference in fit the core temperatures of turtles between years of study, with between this model and one that also includes humidity; con- temperatures being lower during 2014 for both species (Fig. S2). fidence intervals for the model including humidity also exclude Whether this co-variation is the result of chance or has an en- zero, suggesting a significant effect (Table 2). vironmental cause requires further years of observation, but one Surface temperature (7SE) was greatest on lateral and dorsal potential explanation for this is that the sea temperature regimes regions of the head (28.070.09 and 27.870.1 °C, respectively), to which individuals were exposed varied between years. Satellite- derived sea surface temperatures during our fieldwork periods the carapace had the lowest temperature (26.870.1 °C) and the were also found to be significantly lower during 2014 for both temperature of the rear flipper was greater than front flipper study sites (Fig. S4), seemingly consistent with this hypothesis. 7 ° 7 ° (27.6 0.08 C and 27.0 0.07 C, respectively) (Figs. 1 and 3). However, due to the limited data available for SST during our field Surface temperatures increased during the early nesting phases, observation periods, and broad spatial scales of satellite-derived remained relatively constant in mid-phases and then decreased or data, there is insufficient information as yet to infer a casual re- remained constant following egg laying, with rates of change dif- lationship, though this clearly merits further investigation. fering between body regions (Fig. 4A and C). The temperature differential between body surface and air 4.2. Regional variation in temperature temperature (Ts – Ta) was negative in the early phases, then as Ts increased and air temperature decreased, the differential was close Body region, behavioural phase of nesting, and air temperature to zero and positive for head and rear flipper regions during the were found to be the best predictors of surface temperature of ° later stages of nesting (Fig. 4B). leatherbacks. Surface temperature varied by up to 1.2 C, with the head regions, in particular the lateral region, having the greatest surface temperature. This was followed in descending order by the dorsal region of the head, rear flippers, front flippers and carapace (Fig. 3). That the carapace had the lowest surface temperature was 20 T.J. Burns et al. / Journal of Thermal Biology 51 (2015) 15–22

Fig. 4. Regional differences in body surface temperatures of leatherback turtles during nesting. (A) Surface temperature of leatherback turtle body regions during behavioural phases of the nesting process. (B) Temperature differential between body surfaces and air temperature during nesting. (C) Change in leatherback turtle body surface temperature from values at start of prospecting/body pitting phase. Mean timings of phase transitions are indicated by the dotted lines and are: approach (A), prospecting/ body pitting (P/B), excavation (E), laying (L), reﬁlling (R), sand scattering (SS) and return to sea (S). All temperature values are expressed as mean7SE. Data for which sample size was less than four individuals for a phase have been removed. All individuals emerged from the sea between approximately 8 pm and midnight.

Table 2 Results of linear mixed models that account for variation in surface temperature of nesting leatherback turtles (n¼17 individuals). Model including body region, behavioural phase of nesting, air temperature and individual as a random effect was found to be the best ﬁt. df¼degrees of freedom, AICc¼second order Akaike information criterion and

ΔAICc ¼AICc differences.

Model df AICc ΔAICc Akaike weight

Body surface temperature Body regionþbehavioural phase þair temperatureþhumidityþsand temperatureþwind speedþ(1|individual) 17 644.4585 7.7338 0.0095 Body regionþbehavioural phase þair temperatureþhumidityþsand temperatureþ(1|individual) 16 640.4831 3.7584 0.0695 Body regionþbehavioural phase þair temperatureþhumidityþwind speedþ(1|individual) 16 640.4861 3.7614 0.0694 Body regionþbehavioural phase þair temperatureþhumidityþ(1|individual) 15 637.653 0.9283 0.286 Body regionþbehavioural phase þair temperatureþ(1|individual) 14 636.7247 0 0.455 Body regionþbehavioural phaseþ(1|individual) 13 639.5521 2.8274 0.1107 not surprising given its thick, poorly vascularised, insulating et al., 2009b). Measurements of the lateral view of the head in- blubber lining (Davenport et al., 1990) that receives little blood cluded both the jaw and eye and most likely explain the small flow during nesting (Penick, 1996). Substantial adipose tissues are temperature difference (0.2 °C) with the dorsal head region. also found in the head and neck, but the area around the jaw However, in general, the head was the warmest region of the body, muscles are poorly insulated, and eyes are exposed (Davenport reflecting a general lack of insulation compared to other body T.J. Burns et al. / Journal of Thermal Biology 51 (2015) 15–22 21 parts. Blood flow to the skin surface increases during nesting, with In contrast, surface body temperature was positively correlated the skin appearing to flush red and the pineal spot, located on the with air temperature suggesting that the temperature differential dorsal region of the head, turns pink compared to its usual white with air was driving heat exchange from these upper body sur- appearance at sea (Davenport et al., 2014). Blood flow has been faces, rather than wind speed. As both the air and sand tempera- reported to be ten times greater when the skin appears pink ture changed during nesting, the time after sunset of emergence compared to when pale (Spotila et al., 1997). In this study, the may be an important factor determining a leatherback’s heat ex- pineal spot was 0.3–0.4 °C higher in temperature than the sur- change with the environment (Fig. S3). A basic heat transfer model rounding skin surface of the dorsal region of the head (T. J. Burns, of the exposed upper surfaces of a nesting leatherback was de- per obs.). This may be due to bone thickness and lack of insulation veloped to estimate convective and radiative heat exchange during immediately dorsal to the pineal gland (Davenport et al., 2014). the laying phase of nesting (Table S3). The model used mean va- Furthermore, the blood supply to the head lacks the counter cur- lues of body surface temperature and beach environmental con- rent heat exchangers found in the flippers (Greer et al., 1973). ditions, as well as estimates of the surface area of different body Previously, the head region has not been included in analyses of regions. Total surface area was estimated to be 5.34 m2 where heat exchange (Bostrom et al., 2010) but our findings suggest that body, front flippers, rear flippers and head represented 67.6, 19.9, heat loss from the head of nesting leatherbacks merits 8.6 and 3.7% of total surface area, respectively. For simplicity upper consideration. surfaces of the body were estimated as 50% of total body surface. The likely reason for higher surface temperatures of the flippers The total heat loss from the upper surface of the animal was fl is increased blood ow under warm ambient conditions, as has estimated to equal 164 W was and heat loss was greatest from the fl been shown in the front ippers of green (Chelonia mydas) and carapace (62.9%), followed by front flippers (21.1%), rear flippers loggerhead (Caretta caretta) turtles (Hochscheid et al., 2002). In (11.4%) and head (4.6%). Heat loss was principally by radiation fl juvenile leatherbacks 30% of total heat loss was from the ippers 180 W while there was an overall convective heat gain of 16 W which accounted for 27% of total body surface area (Bostrom et al., driven by the fact that the surface temperature of carapace and fl 2010). Furthermore as their study measured heat ux in only the front flippers were below air temperature. fl fi fl front ipper, our ndings, that the rear ipper is warmer than the An additional consideration when examining heat exchange in front, suggest this effect may be even greater. nesting leatherbacks is the effect of the covering of sand which fl Variation in the temperature between fore and rear ippers often gathers on the body surface, particularly the carapace, dur- may be caused by differing levels of muscular heat produced ing nesting. Sand throwing has been associated with thermo- fl during nesting. Other than during egg laying, the rear ippers are regulation in pinnipeds (Lewis and Campagna, 1998), whether the fl actively used for all phases of the nesting process. The front ip- sand layer on nesting leatherbacks increases evaporative cooling pers are relatively inactive during the excavation and refilling or provides insulation is not known. phases, when they instead perform a supporting role for the pi- voting of the body mass from side to side. During nesting, rear 4.4. Conclusion flippers are active for approximately 89% of the overall nesting period, compared with 46% for the front flippers (T. J. Burns and M. This study found that core temperatures of nesting leatherback W. Kennedy, unpublished data). Additionally the proportion of turtles were similar to core temperatures measured in tropical time which different body surfaces spend in contact with sand seas, suggesting that the energy expenditure involved in nesting versus air during the nesting process will likely influence heat activity and the thermal conditions encountered are not severe transfer, as heat flux between the body surface and these two enough to elevate core temperatures beyond the range of freely mediums is likely to be considerably different. However, quanti- swimming individuals. Leatherback body surfaces initially in- fying these proportions to estimate the influence on heat transfer would require careful calculation. creased in temperature from their arrival on the beach before le- A limitation of the thermal imaging methods used to measure velling off or decreasing in later nesting phases. Regional variation surface temperature in this study is that areas of the body surface in surface temperature was related to heat production from the fl not exposed to the air were not measurable, including the ventral ippers during use and the degree of insulation in carapace and surfaces of the flippers and head, and the plastron. Furthermore, head. It appears that time ashore leads to relatively small increases these surfaces will spend most of, or, in the case of the plastron, all in surface temperature, despite substantial energy expenditure fl of the nesting process in contact with sand rather than air, which while removed from the cooling in uence of water. will vary in temperature with depth. It was therefore not possible to estimate heat transfer across ventral surfaces. Heat flux rates across the plastron of a juvenile leatherback under aquatic la- Acknowledgements boratory conditions were reported to be very similar to values for the carapace across a range of temperature regimes (Bostrom We would like to thank the Wildlife Section of the Trinidad et al., 2010) and Davenport et al. (1990) noted that the blubber of Government for allowing us access to Fishing Pond Beach, and the plastron, which was of similar thickness to that of the car- Turtle Village Trust and all the members of the tagging and pa- apace, was poorly vascularised. trolling team at Fishing Pond Village. We are particularly grateful to Sookraj Persad for accompanying us in the field throughout and 4.3. Environmental conditions for passing on his considerable experience on leatherback turtle nesting. We would also like to thank Hannah Davidson and other Heat exchange between any mass and the environment will members of the Glasgow University expeditions to Trinidad 2013 occur through convection, radiation, conduction, and evaporation. and 2014 for assistance in the field, and the Glasgow University No significant correlation between core temperature and any en- expeditions to Tobago 2013 and 2014 for data on hawksbill turtles. vironmental variable was found, a point that merits larger sample Paul Johnson provided invaluable advice and assistance with sta- sizes to resolve. The correlation between sand temperature and Tb tistical modelling. Alan Cooper provided advice on the physics of was closest to significance (Spearman Rank Correlation, r2 ¼0.67, heat transfer. Thanks to Margaret Reilly (Hunterian Museum, p¼0.01), this would suggest that the most important factor in- University of Glasgow) for access to museum specimen for mea- fluencing core temperature may have been conduction (Fig. S5). surement. Furthermore we would like to thank our funders, field 22 T.J. Burns et al. / Journal of Thermal Biology 51 (2015) 15–22 research was supported by the Carnegie Trust for the Universities Department of Interior, Fish and Wildlife Service, Biological Technical Pub- of Scotland, T. J. B. was funded by the IBAHCM Summer Student- lication BTP-R4015-2012, Washington, D.C, pp. 33–34. Eckert, S.A., 2006. High-use oceanic areas for Atlantic leatherback sea turtles ship Scheme and the thermal imaging camera for the project was (Dermochelys coriacea) as identified using satellite telemetered location and kindly donated by Douglas Neil. dive information. Marine Biol. 149, 1257–1267. Funding sources had no involvement in study design; in the Greer, A., Lazell, J., Wright, R., 1973. Anatomical evidence for a counter-current heat exchanger in the leatherback turtle (Dermochelys coriacea). Nature 244, 181. collection, analysis and interpretation of data; in the writing of the Georges, J.V., Fossette, S., 2006. Estimating body mass in leatherback turtles Der- report; and in the decision to submit the article for publication. mochelys coriacea. Marine Ecol. Progress Ser. 318, 255–262. Hendrickson, J. 1958. The green sea turtle, chelonia mydas (linn.) In malaya and sarawak. In: Proceedings of the Zoological Society of London 130, 455–535. Hochscheid, S., Bentivegna, F., Speakman, J., 2002. Regional blood flow in sea tur- Appendix A. Supplementary material tles: implications for heat exchange in an aquatic ectotherm. Physiol. Biochem. Zool. 75, 66–76. Supplementary data associated with this article can be found in James, M., Mrosovsky, N., 2004. Body temperatures of leatherback turtles (Der- mochelys coriacea) in temperate waters off Nova Scotia, Canada. Can. J. Zool. 82, the online version at http://dx.doi.org/10.1016/j.jtherbio.2015.03. 1302–1306. 001. Lewis, M.N., Campagna, C., 1998. Flipping sand in elephant seals. Aquat. Mammals 24.3, 85–90. Lutcavage, M.E., Bushnell, P.G., JoNEs, D.R., 1992. Oxygen stores and aerobic metabolism in the leatherback sea turtle. Can. J. Zool. 70, 348–351. References McCafferty, D., Gilbert, C., Thierry, A.-M., Currie, J., Maho, Y., Ancel, A., 2013. Em- peror penguin body surfaces cool below air temperature. Biol. Lett. 9, 20121192. Mortola, J., 2013. Thermographic analysis of body surface temperature of mammals. Bacon, P.R., 1970. Studies on the Leatherback turtle, Dermochelys coriacea (L.), in Zoolog. Sci. 30 (2), 118–124. – Trinidad, West Indies. Biol. Conserv. 2 (1), 213 217. Mrosovsky, N., 1980. Thermal biology of sea turtles. Am. Zool. 20 (3), 531–547. Bostrom, B., Jones, D., 2007. Exercise warms adult leatherback turtles. Comp. Bio- Mrosovsky, N., Pritchard, P.C.H., 1971. Body temperatures of Dermochelys coriacea – chem. Physiol. Part A: Mol. Integr. Physiol. 147, 323 331. and other sea turtles. Copeia 4, 624–631. Bostrom, B., Jones, T., Hastings, M., Jones, D., 2010. Behaviour and physiology: the Paladino, F., O’Connor, M., Spotila, J., 1990. Metabolism of leatherback turtles, gi- thermal strategy of leatherback turtles. PLoS One 5 (11), e13925. gantothermy, and thermoregulation of dinosaurs. Nature 344, 858–860. Carr, A., Ogren, L.H., 1959. The ecology and migrations of sea turtles, 3. Dermo- Paladino, F.V., Spotila, J.R., O'Connor, M.P., Gatten, R.E., 1996. Respiratory physiology – chelys in Costa Rica. Am. Mus. Novit. 1958, 1 30. of adult leatherback turtles (Dermochelys coriacea) while nesting on land. Carr, A., Ogren, L., 1960. The ecology and migrations of sea turtles, 4. The green Chelonian Conserv. Biol. 2 (2), 223–229. – turtle in the Caribbean sea. Bull. Mus. Nat. Hist. 121, 1 48. Penick, D.N., 1996. Thermoregulatory Physiology of Leatherback (Dermochelys Casey, J., Garner, J., Garner, S., Williard, A., 2010. Diel foraging behavior of gravid coriacea) Green Sea Turtles (Chelonia mydas). Drexel University, Philadelphia, leatherback sea turtles in deep waters of the Caribbean Sea. J. Exp. Biol. 213, PA, USA, PhD dissertation. – 3961 3971. Pritchard, P.C.H., 1971. The leatherback or leathery turtle Dermochelys coriacea. Casey, J., James, M., Williard, A., 2014. Behavioral and metabolic contributions to IUCN Monograph 1, 1–39. thermoregulation in freely swimming leatherback turtles at high latitudes. J. Rowe, M., Bakken, G., Ratliff, J., Langman, V., 2013. Heat storage in Asian elephants – Exp. Biol. 217, 2331 2337. during submaximal exercise: behavioral regulation of thermoregulatory con- Davenport, J., Fraher, J., Fitzgerald, E., McLaughlin, P., Doyle, T., Harman, L., Cuffe, T., straints on activity in endothermic gigantotherms. J. Exp. Biol. 216, 1774–1785. Dockery, P., 2009a. Ontogenetic changes in tracheal structure facilitate deep Sotherland P.R., Reina, R.D., Bouchard, S., Wallace, B.P., Franks, B.F., Spotila, J.R., dives and cold water foraging in adult leatherback sea turtles. J. Exp. Biol. 212, 2003. Egg mass, egg composition, clutch mass, and hatchling mass of leather- – 3440 3447. back turtles (Dermochelys coriacea) nesting at Parque Nacional Las Baulas, Davenport, J., Fraher, J., Fitzgerald, E., McLaughlin, P., Doyle, T., Harman, L., Cuffe, T., Costa Rica. Pp. 31 in J.A. Seminoff, comp. In: Proceedings ofthe 22nd Annual 2009b. Fat head: an analysis of head and neck insulation in the leatherback Symposium on Sea Turtle Biology andConservation. NOAA Technical Memor- – turtle (Dermochelys coriacea). J. Exp. Biol. 212, 2753 2759. andum NMFSSEFSC-503, Miami. Davenport, J., Jones, T., Work, T., Balazs, G., 2014. Pink spot, white spot: the pineal Southwood, A., Andrews, R., Paladino, F., Jones, D., 2005. Effects of diving and skylight of the leatherback turtle (Dermochelys coriacea Vandelli 1761) skull swimming behavior on body temperatures of pacific leatherback turtles in and its possible role in the phenology of feeding migrations. J. Exp. Marine Biol. tropical seas. Physiol. Biochem. Zool. 78, 285–297. – Ecol. 461, 1 6. Spotila, J.R., O'Connor, M.J., Paladino, F.V., 1997. The Thermal Biology of Sea Turtles Davenport, J., Holland, D.L., East, J., 1990. Thermal and biochemical characteristics of In: Lutz, P.L., Musick, J.A. (Eds.), The biology of Sea Turtles, Vol. 1. CRC Press, Inc., the lipids of the leatherback turtle Dermochelys coriacea- evidence of en- Boca Raton, USA, pp. 297–314. – dothermy. J. Marine Biol. Assoc. United Kingdom 70, 33 41. Stefano, D.J., 2003. How much power is enough? Against the development of an Dudley, P., Porter, W., 2014. Using empirical and mechanistic models to assess arbitrary convention for statistical power calculations. Funct. Ecol. 17, 707–709. global warming threats to leatherback sea turtles. Marine Ecol. Progress Ser. Wallace, B.P., Tiwari, M., Girondot, M., 2013. Dermochelys coriacea. The IUCN Red – 501, 265 278. List of Threatened Species. Version 2014.3. 〈http://www.iucnredlist.org〉. Eckert, K.L., Wallace, B.P., Frazier, J.G., Eckert, S.A., Pritchard, P.C.H., 2012. Synopsis of the Biological Data on the Leatherback Sea Turtle (Dermochelys coriacea). U.S.

All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and read them immediately.