Meeting Abstract
Thermal performance curves (TPCs) are widely used to predict mean and variation in performance for fluctuating weather and climate conditions. TPCs are typically estimated by measuring performance at a series of constant temperatures, but these curves sometimes yield poor predictions in fluctuating conditions. We develop a new statistical method for estimating TPCs in fluctuating conditions. The method assumes that there is an underlying TPC function, and uses maximum likelihood to estimate the parameters of this function from data on mean performance of individuals and the temperatures experienced by those individuals. We apply the model to data for mean growth rate of Manduca sexta larvae from two sets of experiments, in deterministically or stochastically fluctuating thermal regimes. The estimated TPCs are similar to those estimated from data at constant temperatures over short (4-24h) time scales. Surprisingly, TPCs can be estimated from data based on only 1-2 thermal regimes if they include a wide range of temperatures; and stochastic fluctuations can provide more information than deterministic fluctuations for estimating TPCs. The method is applicable to both experimental and observational (including field) data.
