An inconvenient truth about temperature–time data from thermocouples

Abstract

Plant ecologists have long been interested in the effects of fire on vegetation. Thermocouples have been in their proverbial toolbox for decades, despite temperature not being a direct product or measure of wildland fire behaviour or fire effects. To better represent the cumulative impact of high-temperature exposure on organisms, ecologists often use temperature–time curves from thermocouples to calculate residence time—the duration of heat exposure above a threshold temperature—which can be used to calculate another popular metric, degree $\cdot$seconds. A systematic literature review of 105 published papers shows that residence time, especially, and degree $\cdot$ seconds are common metrics derived from raw temperature–time data. While several errors in thermocouple readings have been previously identified and addressed—responsiveness to heating, discrepancy between thermocouple temperature and actual temperature of the medium surrounding the thermocouple—this paper highlights a previously unconsidered source of error that must be reconciled for metrics like residence time to be biologically valid: the disproportionately long time it takes for thermocouples to cool once heat input is complete. Using an array of thermocouples in a fume hood over a Bunsen burner before and after the flame is extinguished, this paper shows that after being exposed to flame, 30-gauge K-type thermocouples require 80–100 s to register ambient temperatures despite taking only about 5 s to respond to heating. The review indicates ecologists give no consideration for this disproportionately slow cooling response. These findings indicate that residence time (and therefore degree $\cdot$ seconds) have been over-estimated in the fire ecology literature. The proposed solution is to simply truncate temperature–time curves at the point temperature begins to decline, which indicates a shift from the biologically relevant effect of heat input to the biologically irrelevant, physical properties (heat diffusivity) of the thermocouple itself. Conceptual models present these biologically relevant portions of the temperature–time curve and identify parts of the biologically relevant curve that might be useful in quantifying components of flammability.