Heat flux models have been used to predict metabolic rates of marine mammal
s, generally by estimating conductive heat transfer through their blubber l
ayer. Recently, Kvadsheim et al. (1997) found that such models tend to over
estimate metabolic rates, and that such errors probably result from the asy
mmetrical distribution of blubber. This problem may be avoided if reliable
estimates of heat flux through the skin of the animals are obtained by usin
g 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 en
vironment. We evaluated this approach based on simultaneous measurements of
metabolic rates and of input parameters necessary for heat flux calculatio
ns, as obtained from four harp seals (Phoca groenlandica) resting in cold w
ater. Heat flux estimates were made using two free convection models (doubl
e-flat-plate and cylindrical geometry) and one forced convection model (sin
gle-flat-plate geometry). We found that heat flux estimates generally under
estimated 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. (C) 2000 Academic Press.