全球变化知识资源中心

Measures of the seasonal timing of biological events are key to addressing questions about how phenology evolves, modifies species interactions, and mediates biological responses to climate change. Phenology is often characterized in terms of discrete events, such as a date of first flowering or arrival of first migrants. We discuss how phenological events that are typically measured at the population or species level arise from distributions of phenological events across seasons, and from norms of reaction of these phenological events across environments. We argue that individual variation in phenological distributions and reaction norms has important implications for how we should collect, analyze, and interpret phenological information. Regarding phenology as a reaction norm rather than one year's specific values implies that selection acts on the phenologies that an individual expresses over its lifetime. To understand how climate change is likely to influence phenology, we need to consider not only plastic responses along the reaction norm but changes in the reaction norm itself. We show that when individuals vary in their reaction norms, correlations between reaction norm elevation and slope make first events particularly poor estimators of population sensitivity to climate change, and variation in slopes can obscure the pattern of selection on phenology. We also show that knowing the shape of the distribution of phenological events across the season is important for predicting biologically important phenological mismatches with climate change. Last, because phenological events are parts of a continuous developmental process, we suggest that the approach of measuring phenology by recording developmental stages of individuals in a population at a single point in time should be used more widely. We conclude that failure to account for phenological distributions and reaction norms may lead to overinterpretation of metrics based on single events, such as commonly recorded first event dates, and may confound meta-analyses that use a range of metrics. Rather than prescribing a single universal approach to studying phenology, we point out limitations of inferences based on single metrics and encourage work that considers the multivariate nature of phenology and more tightly links data collection and analyses with biological hypotheses.