Identifying the Critical Climatic Time Window That Affects Trait Expression

Abstract

Identifying the critical time window during which climatic drivers affect the expression of phenological, behavioral, and demographic traits is crucial for predicting the impact of climate change on trait and population dynamics. Two widely used associative methods exist to identify critical climatic periods: sliding-window models and recursive operators in which the memory of past weather fades over time. Both approaches have different strong points, which we combine here into a single method. Our method uses flexible functions to differentially weight past weather, which can reflect competing hypotheses about time lags and the relative importance of recent and past weather for trait expression. Using a 22-year data set, we illustrate that the climatic window identified by our new method explains more of the phenological variation in a sexually selected trait than existing approaches. Our new method thus helps to better identify the critical time window and the causes of trait response to environmental variability.