Quantification of climatic niches in birds: adding the temporal dimension

Quantification of the climatic niche from geographic occurrences is an increasingly important tool for studying species’ relationships to their environment, for example to predict responses to climate change. However, as the geographic distributions of birds are seasonally dynamic, they pose a challenge to carrying out comparable and appropriate quantification of climatic niches. In this review, we first assess how relevant seasonal dynamics are across birds as a whole by compiling a database of migratory behaviour for 10 443 bird species. Second, we examine how studies have quantified climatic niches of birds. Finally, using Australia as a case study, we investigate how well existing distribution datasets represent temporal dynamics by comparing seasonal patterns of species richness obtained from point‐occurrence data with those from range maps and assess the consequences for niche quantification. We provide a consistent classification of migratory behaviour across all birds, and find that a huge variety exists between and within species that should be considered when quantifying climatic niches. Despite this, our review of the literature revealed that seasonal dynamics have often not been accounted for. For future studies, we provide a framework for selecting appropriate occurrence data depending on migratory behaviour and data availability. Our comparison of seasonal species richness patterns obtained from extent‐of‐occurrence range maps and point‐occurrence data suggests that range maps are less able to detect temporal dynamics of bird distributions than point‐occurrence data. We conclude that seasonally explicit range maps combined with climatic data for the corresponding time period can be used to adequately quantify climatic niches for resident birds, but are not adequate to quantify the climatic niches of migratory and nomadic species. Therefore, consistent quantification of climatic niches across all birds requires temporally explicit occurrence points. As such, increasing the availability of occurrence data and methods correcting biases should be a priority.