Evaluation of predictive formulae for determining metabolic rate during cold water immersion.

Five models predicting shivering thermogenesis on the basis of steady state skin and core temperature were evaluated: Hayward et al., Stolwijk and Hardy,; Nadel et al.,; Timbal et al., and Brown and Brengelmann, using the empirical data derived from a cold water immersion study by Morrison et al. A residual analysis indicated that all models generated substantial errors of prediction. The best overall predictors were expressions suggested by Hayward et al., while the predictive equation of Nadel et al. ranked second. Derivation of personal coefficients significantly improved the prediction of all models and a subsequent modification of the standard models, adding temperature derivative terms, further reduced the magnitude of the error. An analysis of the residuals indicated that peripheral and core temperatures should be weighted according to the characteristics of thermosensitive neural structures in these regions.