Cutaneous heat flux models do not reliably predict metabolic rates of marine mammals.
Heat flux models have been used to predict metabolic rates of marine mammals, generally by estimating conductive heat transfer through their blubber layer. Recently, Kvadsheim et al. (1997) found that such models tend to overestimate metabolic rates, and that such errors probably result from the asymmetrical distribution of blubber. This problem may be avoided if reliable estimates of heat flux through the skin of the animals are obtained by using models that combine calculations of conductive heat flux through the skin and fur, and convective heat flux from the surface of the animal to the environment. We evaluated this approach based on simultaneous measurements of metabolic rates and of input parameters necessary for heat flux calculations, as obtained from four harp seals (Phoca groenlandica) resting in cold water. Heat flux estimates were made using two free convection models (double-flat-plate and cylindrical geometry) and one forced convection model (single-flat-plate geometry). We found that heat flux estimates generally underestimated metabolic rates, on average by 26-58%, and that small variations in input parameters caused large variations in these estimates. We conclude that cutaneous heat flux models are too inaccurate and sensitive to small errors in input parameters to provide reliable estimates of metabolic rates of marine mammals.