MSc 2005

Dr Andrew Trites

A multiple regression model for predicting the energy requirements of marine mammals.

Marine ecosystem management has prompted the need for better understanding of the impact of marine mammals in the oceans. Using the wealth of information available in the literature, a rule based multiple regression model was developed to estimate the energy requirements of all 124 marine mammal species. This meta-analysis modeling framework provided a simple means for estimating the energetics (metabolism or consumption) of marine mammals under varying conditions, as a function of easily obtained or estimated physiological and environmental variables, including morphology, developmental stage, growth, sex, reproductive status, health, activity, postabsorptive state, thermoneutral condition, and season.

Based on different combinations of input variables, a set of empirical equations was developed. The empirical equations provide an objective predictive tool for estimating the energy requirements of data deficient marine mammal species. Extensive model validation indicated that all models were robust to their statistical assumptions, including phylogenetic independence, and captured a substantial amount of the observed heterogeneity in energy requirements (up to 82% residual variance).

Equations also synthesize evidence of a uniform pattern of energy use, from consumption to expenditure, and provide quantitative rough estimates of the components of the bioenergetic framework for all marine mammal species. Results suggest that body mass is a better predictor of energy requirements than body length, although length may be used in circumstances when mass cannot be estimated or measured.

Of the parameters considered, model predictions were most sensitive to uncertainty in morphology, developmental stage, activity, and growth. By including flexibility in prediction and uncertainty in estimates, results extend the simple allometric scaling relationships with mass alone (e.g., Kleiber’s Equation), and refine estimates of marine mammal energy requirements currently available. Results serve as a useful starting point from which complex analyses can proceed, and provide a basis against which other models can be compared. The method provides an objective means for researchers and resource managers to select an equation most appropriate for their predictive needs, even for data-deficient species, given different levels of available input information. The empirical models are useful tools for parameterizing ecosystem models and can be used to help address ecological questions and issues pertaining to conservation and resource management.


A multiple regression model for predicting the energy requirements of marine mammals. Hunter, A.M.J. 2005. MSc Thesis, University of British Columbia, Vancouver. 300 pages (PDF)

Food consumption by sea lions: existing data and techniques. Winship, A.J., A.M.J. Hunter, D.A.S. Rosen, and A.W. Trites. 2006. In A.W. Trites, S. Atkinson, D.P. DeMaster, L.W. Fritz, T.S. Gelatt, L.D. Rea and K. Wynne (eds), Sea Lions of the World. Alaska Sea Grant College Program, University of Alaska, Fairbanks. pp. 177-191. (PDF)

An annotated bibliography of scientific literature (1751-2000) pertaining to Steller sea lions (Eumetopias jubatus) in Alaska. Hunter, A.M.J. and A.W. Trites. 2001. Fisheries Centre Research Reports Vol 9 pp. 45 (PDF)

Estimates of basal metabolic and feeding rates for marine mammals from measurements of maximum body length. Hunter, A.M.J., A.W. Trites and D. Pauly. 2000. In C.L.K. Baer (ed.), Proceedings of the Third Comparative NutritionSociety Symposium. Pacific Grove, California, August 4-9, 2000. 3:103-106. (PDF)