A flexible model for thermal performance curves

Abstract

Temperature responses of many biological traits-including population growth, survival, and development-are described by thermal performance curves (TPCs) with phenomenological models like the Briere function or mechanistic models related to chemical kinetics. Existing TPC models are either simple but inflexible in shape or flexible yet difficult to interpret in biological terms. Here we present flexTPC, a model that is parameterized exclusively in terms of biologically interpretable quantities: the thermal minimum, optimum, and maximum, the peak trait value, and thermal breadth. FlexTPC can describe unimodal temperature responses of any skewness and thermal breadth, enabling direct comparisons across populations, traits, or taxa with a single model. We apply flexTPC to various microbial and entomological datasets, compare results with the widely used Briere model, and find that flexTPC often has better predictive performance. The interpretability of flexTPC makes it ideal for modeling how thermal responses change with ecological stressors or evolve over time.