Data from: Behavioural thermoregulation in the Australian fur seal (Arctocephalus pusillus doriferus)
Data files
Sep 11, 2025 version files 35.47 KB
-
FG2_summary.csv
679 B
-
FG3_summary_points.csv
636 B
-
FG4_summary.csv
13.39 KB
-
FG5_summary_ANOVA_tukey.csv
15.92 KB
-
README.md
4.85 KB
Abstract
Physiological and behavioural adaptations shape an animal's thermoregulatory capacity. Pinnipeds - true seals (phocids), eared seals (otariids, including fur seals and sea lions), and walruses (Odobenids) - must thermoregulate in both aquatic and terrestrial environments. Fur seals are unique, relying on dense, water-repellent fur and moderate blubber for insulation, while true seals, sea lions, and walruses have lower fur densities and thicker blubber. Fur traps air when wet to reduce heat loss at sea, but its insulating properties when dry on land, along with those of furless flippers, remain unclear for fur seals. This study used behavioural observations and infrared thermography to examine posture and surface temperature of dry, hauled out adult female Australian fur seals (Arctocephalus pusillus doriferus) across a range of air temperatures (Ta). At Ta < 22 °C, huddling minimised heat loss by shielding flippers, and as Ta increased, seals shifted postures - Prone, Curled, Oblique, and Spread - to expose flippers and promote heat dissipation. Generally, flippers had a higher surface temperature than dry fur; however, at 26.5 °C, the fur surface temperature exceeded that of the flipper, with the fur appearing to insulate to retain heat beyond the flippers' dissipation capacity. At Ta > 30 °C, seals entered the water. With rising Ta predicted across this species' range, seals will likely spend more time in the water to overcome heat loss challenges. This shift could increase predation risk, energy demands, and negatively impact maternal investment.
https://doi.org/10.5061/dryad.xwdbrv1qs
Description of the data and file structure
The study was conducted at Kanowna Island (39 ° 10’S, 146 ° 18’E). Weather data and thermographs were taken opportunistically from several locations around Kanowna Island where individuals could be approached with minimal disturbance. Black-bulb ambient temperature was recorded using a digital thermometer with a thermocouple inserted into a matt black plastic ball, all positioned 20 cm above the ground. Hereafter, black-bulb ambient temperature is referred to as air temperature (Ta).
Thermographs were taken using a thermograph camera (AVIO TVS 700, Nippon Avionics, Japan) fitted with a 35 mm lens. Thermographs of seals were from distances of 4 to 17 m, with variation in distance due to seal wariness. Distances were determined with the aid of a laser range finder (Bushnell Yardage Pro Sport 450, Overland Park, KS, USA).
Thermographs were used to assess the posture of an individual and whether the individual was in a huddle. Thermographs were analysed in Goratec Thermography Studio (Version 4.8, Goratec Technology GmbH and Co., Erding, Germany) to gain measures of body surface temperature. Mean temperatures for two body regions consisting of several body parts were calculated. These were the furred parts of the body (trunk, rump, neck, and head) and the unfurred parts of the body (pectoral flippers and tail flippers). Uploaded data has been summarised to directly relate to manuscript sections. NA represents missing data.
Files and variables
File: FG2_summary.csv
Description: This summary data was used to create Figure 2.
Variables
- bb_temp: black-bulb temperature (°C)
- proportion_indiv: percent of individuals (converted from proportion)
- posture: posture demonstrated by individual (prone, curled, oblique, spread)
- n_sample_periods: number of sampling periods
- n_individuals: number of individuals sampled
File: FG3_summary_points.csv
Description: This summary data was used to create Figure 3.
Variables
- bb_temp_1: black-bulb temperature one (°C)
- bb_temp_2: black-bulb temperature two (°C)
- huddle_present_absent: whether huddles were present (1) or absent (0)
File: FG4_summary.csv
Description: This summary data was used to create Figure 4.
Variables
- bb_temp: black-bulb temperature (°C)
- surface_temp: body surface temperature (°C)
- group: age group (adult female [adult] or pups [premoult, postmoult])
- surface: body surfacetemperaturee that body surfactemperaturere (°C) comes from (fur, flipper)
- posture: posture demonstrated by individual (prone, curled, oblique, spread)
File: FG5_summary_ANOVA_tukey.csv
Description: This summary data was used to create Figure 5.
Variables
- bball: black-bulb temperature (°C)
- delta_temp: difference between black-bulb temperature (°C) and body surface temperture (°C)
- posture: posture demonstrated by the individual (prone, curled, oblique, spread)
- resting_type: category of resting per individual (huddled or isolated)
- surface: body surface temperature that body surface temperature (°C) comes from (fur, flipper)
Code/software
Behaviours (four postures and huddling) and Tfur and Tflipper were examined in relation to Ta. Statistical analysis followed the methods of Quinn et al. (2002) using R version 4.3.0 (R Core Team, 2023). Logistic regression (glm) analyses were conducted to assess the relationship between temperature and huddle presence or absence, and McFadden's pseudo-R² was calculated. Linear regression (lm) was used to examine the Tfur and Tflipper against Ta, and Tfur and Tflipper against Ta for each posture. To examine the effect of huddling on surface temperature variation across different postures, an ANOVA was performed on the delta temperature (*i.e., *body surface temperature minus Ta) for both fur and flipper surfaces. Post-hoc Tukey HSD tests were conducted for pairwise comparisons to evaluate the differences between isolated and huddled individuals within each posture. Unless otherwise stated, p-values were considered significant at <0.05.
Quinn, A., Rycraft, J. R., & Schoech, D. (2002). Building a Model to Predict Caseworker and Supervisor Turnover Using a Neural Network and Logistic Regression. Journal of Technology in Human Services,* 19*(4), 65-85
R Core Team (2023). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria.
Access information
Other publicly accessible locations of the data:
- NA
Data was derived from the following sources:
- NA