Remotely sensed variables explain microhabitat selection and reveal buffering behaviours against warming in a climate-sensitive bird species

Fine-scale habitat selection modelling can allow a mechanistic understanding of habitat selection processes, enabling better assessments of the effects of climate and habitat changes on biodiversity. Remotely sensed data provide an ever-increasing amount of environmental and climatic variables at high spatio-temporal resolutions, and a unique opportunity to produce fine-scale habitat models particularly useful in challenging environments, such as high-elevation areas. Working at a 10-m spatial resolution, we assessed the value of remotely sensed data for investigating foraging habitat selection (in relation to topography, microclimate, land cover) in nestling-rearing white-winged snowfinch (Montifringilla nivalis), a high-elevation species highly sensitive to climate change. Adult snowfinches foraged at locations with intermediate vegetation cover and higher habitat heterogeneity, also avoiding extremely warm or extremely cold microclimates. Temperature interacted with other environmental drivers in defining habitat selection, highlighting trade-offs between habitat profitability and thermoregulation: snowfinches likely adopted mechanisms of behavioural buffering against physiologically stressful conditions by selecting for cooler, shaded and more snowy foraging grounds at higher temperatures. Our results matched those from previous studies based on accurate field measurements, confirming the species' reliance on climate-sensitive microhabitats (snow patches and low-sward grassland, in heterogeneous patches) and the usefulness of satellite-derived fine-scale modelling. Habitat suitability models built on remotely sensed predictors can provide a cost-effective method for periodic monitoring of species' habitats both at fine grain and over large extents. Fine-scale models also enhance our understanding of the actual drivers of (micro)habitat selection and of possible buffering behaviours against warming, allowing more accurate and robust distribution models, finer predictions of potential future changes and carefully targeted conservation strategies and habitat management.