Microclimate affects the distribution of grassland birds, but not forest birds, in an Alpine environment

Predictions derived from species distribution models (SDMs) are strongly influenced by the spatial scale at which species and environmental data (e.g. climate) are gathered. SDMs of mountain birds usually build on large-scale temperature estimates. However, the topographic complexity of mountain areas could create microclimatic refuges which may alter species distributions at small spatial scales. To assess whether fine-scale data (temperature and/or topography) improve model performance when predicting species occurrence, we collected data on presence-absence of bird species, habitat and fine-scale temperature at survey points along an elevational gradient in the Alps (NW Italy). Large-scale temperature data, and both large- and fine-scale topography data, were extracted from online databases for each point. We compared species models (fine-scale vs large-scale) using an information-theoretic approach. Models including fine-scale temperature estimates performed better than corresponding large-scale models for all open habitat species, whereas most forest/ecotone species showed no difference between the two scales. Grassland birds such as Northern Wheatear Oenanthe oenanthe and Water Pipit Anthus spinoletta were positively associated with warmer microclimates. These results suggest that alpine grassland species are potentially more resistant to the impact of climate change than previously predicted, but that indirect effects of climate change such as habitat shifts (forest- and shrub encroachment at high elevations) pose a major threat. Therefore, active management of alpine grassland is needed to maintain open areas and to prevent potential habitat loss and fragmentation. SDMs based solely on large-scale temperatures for open habitat species in the Alps should be re-assessed.