Using GPS-accelerometry to track energy expenditure in a breeding seabird

Background/Question/Methods

Energy is a common currency for any living organism, and understanding how animals use energy is critical to understanding their ecology and consequently how they might adapt to changes in their environment. Reproduction is a costly event for many animals, especially birds which must deliver food to a central place. Measuring energy expenditure over the length of the breeding season can provide valuable insight on physiological limitations, for example; by highlighting periods of high demand and identifying when physiological overload might occur. Doubly-labelled water (DLW) has been widely used to assess energy expenditure in wild animals, but requires multiple recaptures at fixed intervals and only provides a single, average value for energy expenditure. Alternatively, animal-borne devices such as GPS-accelerometers offer the opportunity to measure energy expenditure at different time scales. Although advancement of technology has made tracking behaviour more accessible, the analysis of accelerometry data can pose a barrier when classifying behaviour. Yet, using broad behavioural categories when classifying behaviour has proven to be relatively easy and accessible, as few predictor variables can yield accurate time-activity budgets. Here, we developed a novel DLW-accelerometry calibration to measure energy expenditure in breeding black-legged kittiwakes (Rissa tridactyla).

Results/Conclusions

We found that kittiwakes varied their time-activity budgets throughout the breeding season by increasing time spent in flight and decreasing their time spent on water in chick-rearing compared to both incubation (p = 0.03 and p = 0.009 respectively) and pre-laying (p = 0.03 and p = 0.04 respectively). Time-activity budgets were comparable across pre-laying and incubation, yet kittiwakes significantly increased their energy expenditure in incubation (p = 0.02), averaging 752 ± 36 kJ d-1 in incubation and 623 ± 33 kJ d-1 in pre-laying. We obtained activity-specific metabolic rates for each behaviour, and found that energy expenditure measured using our calibration versus measured using DLW did not vary (p = 0.19, R2 = 0.55, slope = 0.80). The calibration produced will facilitate estimation of daily energy expenditure using only GPS-accelerometers, reducing the impact on animals, researcher effort, and expense relative to traditional DLW methods. Using broad behavioural categories, time-activity budgets can be easily obtained from free-ranging individuals to accurately track energy expenditure in response to behavioural and environmental changes.