A framework for elucidating the temperature dependence of fitness

Climate warming is predicted to cause large-scale extinctions, particularly of ectothermic species. A striking difference between tropical and temperate ectotherms is that tropical species experience a mean habitat temperature that is closer to the temperature at which fitness is maximized (T(opt)) and an upper temperature limit for survival (T(max)) that is closer to T(opt) than do temperate species. Thus, even a small increase in environmental temperature could put tropical ectotherms at high risk of extinction, whereas temperate ectotherms have a wider temperature cushion. Although this pattern is widely observed, the mechanisms that produce it are not well understood. Here we develop a mathematical framework to partition the temperature response of fitness into its components (fecundity, mortality, and development) and test model predictions with data for insects. We find that fitness declines at high temperatures because the temperature responses of fecundity and mortality act in opposite ways: fecundity decreases with temperature when temperatures exceed the optimal range, whereas mortality continues to increase. The proximity of T(opt) to T(max) depends on how the temperature response of development mediates the interaction between fecundity and mortality. When development is highly temperature sensitive, mortality exceeds reproduction only after fecundity has started to decline with temperature, which causes fitness to decline rapidly to zero when temperatures exceed T(opt). The model correctly predicts empirically observed fitness-temperature relationships in insects from different latitudes. It also suggests explanations for the widely reported phenological shifts in many ectotherms and the latitudinal differences in fitness responses.