Sample trace from a wood frog, Rana sylvatica, during freezing and thawing in the respirometer. (A) Reference temperature (grey) and frog temperature (black). (B) Activity (recorded from an infrared activity detector). (C) CO2 production (). b, baseline recordings. Vertical dashed lines indicate: c, beginning of cooling; f, beginning of freezing exotherm; r, beginning of rewarming; and t, increase of temperature, indicating end of thawing.
Schematic of model of energy use for overwintering frogs, based upon the general model of Marshall and Sinclair (Marshall and Sinclair, 2012). Boxes with black borders are the base model. Boxes with blue borders indicate a model that includes metabolic costs of cooling, freezing and thawing. The box with a red border indicates a term present only in some sensitivity analyses. Solid borders indicate values derived from empirical data (present study); dashed borders indicate assumptions of the model. Shaded boxes indicate calculated terms. Asterisks indicate terms that were modified in sensitivity analyses.
Microclimate temperatures representative of Rana sylvatica winter habitat measured by iButton data loggers over the winter of 2008–2009 (1 November–16 April; N=2 data loggers shown). Air temperatures (recorded from an unshielded data logger in the canopy) are included for comparison. Black=ground temperature (no distinction made between the two locations); grey=air temperature. The dashed horizontal line shows the frog freezing temperature of −1.4°C used in our base models.
Breakdown of predicted energy consumption by a 7.1 g Rana sylvatica for two locations over the 2008–2009 winter. It is assumed that metabolism outside of freeze–thaw periods is fuelled by lipid, but that when frogs are frozen they utilise carbohydrate. The pre-freezing increase accounts for an observed increase in metabolism during cooling, and the freeze–thaw value accounts for the observed increase in metabolic rate associated with freezing and thawing processes. See Results for further details. Numbers above bars indicate the estimated number of freeze–thaw cycles for a location.
Example of cumulative overwinter energy use from carbohydrates (grey line) and lipid (dashed line) calculated from microclimate temperatures from Location 2 (black line; see also Tables 2, 3) for a 7.1 g Rana sylvatica in the 2008–2009 winter. Note the long period of snow cover, from 10 December to 17 March, when frogs are predicted to have remained unfrozen. Dates are in day/month/year format.