Coping with heat: behavioural and physiological responses of savanna elephants in their natural habitat

Abstract Most of southern Africa's elephants inhabit environments where environmental temperatures exceed body temperature, but we do not know how elephants respond to such environments. We evaluated the relationships between apparent thermoregulatory behaviour and environmental, skin and core temperatures for tame savanna elephants (Loxodonta africana) that were free-ranging in the hot parts of the day, in their natural environment. Environmental temperature dictated elephant behaviour within a day, with potential consequences for fine-scale habitat selection, space use and foraging. At black globe temperatures of ~30°C, elephants adjusted their behaviour to reduce environmental heat load and increase heat dissipation (e.g. shade use, wetting behaviour). Resting, walking and feeding were also influenced by environmental temperature. By relying on behavioural and autonomic adjustments, the elephants maintained homeothermy, even at environmental temperatures exceeding 40°C. Elephants clearly have the capacity to deal with extreme heat, at least in environments with adequate resources of forage, water and shade. Future conservation actions should provide for the thermoregulatory, resource and spatial needs of elephants.

S3: Summary of methods used to measure the parameters needed to record skin temperature on the elephants in the present study.
We recorded skin temperature using an infrared thermal camera (FLIR T640, FLIR Systems Inc., Portland, Oregon, USA), which was mounted on a monopod. The camera had a 25˚ lens, with a 7.5-14 µm spectral range and was factory (FLIR Systems Inc., Portland, Oregon, USA) calibrated to record temperature with an accuracy of ± 1.0˚C. The camera detects infrared radiation emitted from the surface of an object, which is converted into a temperature value. These values are depicted in the form of an infrared thermal image (thermograph).
The radiation that is detected by the camera originates from three different sources; radiation from the target object, reflected radiation from the object's surroundings and radiation from the surrounding air. Therefore, the camera output can be described as follows: where ߝ is the emissivity of the target object, ߬ is the transmittance of the atmosphere, ܹ is the amount of radiation emitted from the target object, ܹ is the amount of radiation emitted from the object's surroundings and ܹ ௧ is the amount of radiation present in the surrounding air. To ensure accurate skin temperature recordings we supplied the camera with the following parameters: emissivity=0.98, distance=10 m (unless specified otherwise), air temperature, relative humidity and reflected temperature.

Emmisivity
The amount of radiation emitted from an object comes from two major sources; radiation from the object itself (emissivity, ߝ) and radiation from the surrounding environment that is reflected off of the object (reflectivity, ߩ). This is described as follows: To calculate the emissivity of elephant skin, we placed an object of known emissivity (black scotch tape, ߝ=0.95) onto a piece of elephant skin. Both the skin and tape were cooled to the same temperature (5˚C). The emissivity of the camera was adjusted to 0.95 and the surface temperature of the black tape was recorded. We then adjusted the emissivity of the camera until the temperature of the elephant skin was identical to the temperature of the tape.
The emissivity at which the skin temperature and the tape temperature were identical represented the emissivity of elephant skin. This value was 0.98, which is similar to the emissivity of human skin. Therefore, 98% of the radiation emitted from elephant skin is radiation from the skin itself and 2% is from reflected radiation. An emissivity of 0.98 was used for elephant skin throughout the study.

Distance
The amount of radiation emitted from the air surrounding the target object is dependent on the volume of air between the object and the observer (Wolfe and Zissis 1989). Therefore, we supplied the camera with the distance between the camera and the focal elephant. We consistently maintained a distance of 10 m to avoid sampling error. Where it was not possible to maintain this distance, we visually estimated the distance and supplied this to the camera.

Air temperature and relative humidity
The amount of radiation emitted from the air surrounding the target object is also dependent on the temperature and water content of the air (Wolfe and Zissis 1989). Therefore, we supplied the camera with ambient temperature and relative humidity that we recorded using a portable psychrometer (ExTech ® HD500, Townsend West, Nashua, U.S.A).

Reflected temperature
To account for reflected radiation from the environment, we calculated the reflected temperature by setting emissivity to 1.00 and distance to 0 m, before obtaining a thermal image of a diffuse reflector. The diffuse reflector comprised of a wrinkled sheet of aluminium foil placed over a spherical polystyrene ball (250 mm in diameter). Aluminium foil has an emissivity of 0.04. Therefore 99.6% of radiation from the surrounding environment is reflected from the aluminium foil. By setting ߝ to 1.00, we assumed that all radiation from the surrounding environment was represented by the average temperature of the diffuse reflector.
We then substituted this average temperature into the camera parameters before obtaining a thermograph of the focal elephant. This ensured that skin temperature measurements excluded all reflected radiation from the surrounding environment.

Reference:
Wolfe WL, Zissis GJ (1989) The infrared handbook. Department of the Navy, Washington, Note: K=number of parameters in model; LogLik= log likelihood; AIC= Akaike's information criterion; ∆ AIC i = Difference in AIC between the model and best fitting model; w i =Akaike weight; R 2 =adjusted coefficient of determination. Parameters: mini-globe = black mini-globe temperature; time = time of day; state = whether elephant is wet or dry.