Insufficient considerations of seasonality, data selection and validation lead to biased species–climate relationships in mountain birds

Linking organism distribution to climate is key to understanding factors determining species occurrence and evaluating the potential impacts of ongoing climate change. A common analytical tool to assess the link between species and climate is represented by ecological niche modelling and by the tightly related species distribution models (SDMs). Those approaches have been widely used to explain key features related to species occurrence in relation to current climatic and other environmental variables, or as a response to past colonization and extinction events (Pons et al. 2021), as well as to predict the future distribution of target species (Scridel et al. 2021). Moreover, the link between climatic changes and niche evolution, i.e. niche changes (or conservatism) over time and especially across phylogenies remains complex and fascinating. In the context of change, the rise of community science (‘citizen science’) data provides an invaluable opportunity to expand our knowledge on species distribution. While much attention has been given to aspects related to the statistical approaches and algorithm settings adopted in distribution or niche modelling (Engler et al. 2017, Ortega-Huerta and Rivera 2017, Kozma et al. 2018), less attention has been paid to validate very fundamental features which can greatly impact models performance, such as the quality of data for a given species, its spatial and temporal congruence with the climatic and environmental data used, and the risks associated with the inclusion of records of occasional vagrant individuals observed outside of their usual environment (but see Engler et al. 2017, Eyres et al 2017, Andrew and Fox 2020, Taheri et al. 2020).

Keywords: distribution, ecological niche, Montifringilla nivalis, mountains, seasonality, snowfinch