Through the taste buds of a large herbivore: foodscape modeling contributes to an understanding of forage selection processes.

How large herbivores track resource quantity and quality through time has formed the core of an abundance of literature on migratory populations in recent decades. Yet, relating foraging processes and habitat selection patterns in resident populations, where spatial heterogeneity of food resources is fine‐grained and/or where the portion of edible plants (i.e. the foodscape) is low, is challenging. We addressed this issue in a mountain population of chamois Rupicapra rupicapra, an intermediate feeder, whose individuals do not migrate. We relied on a rare combination of data on habitat use of 50 GPS‐collared females and data on the quantity (biomass) and quality (phenology) of edible resources in their landscape, derived from field sampling of vegetation, remote sensing and diet (DNA barcoding). The foodscape of the chamois was composed of a low proportion of the available biomass (<18%), including relatively high‐quality plants, with low spatial covariation between plant phenology and biomass. Chamois avoided areas with a low edible biomass (where the intake rate may be too low) and focused on areas with plants at approximately the flowering stage, whatever the average plant phenological stage available. Due to this constant preference for flowering plants, home range selection ratios therefore shifted during the summer from a selection of more advanced plants in June to their avoidance in August. When the phenology scores of all plants available, rather than edible plants only, were considered, areas with relatively more advanced plants were selected all summer long. This exemplifies that, when traits from edible plants are different from those of all plants available, it is crucial to consider the actual foodscape to decipher forage and habitat selection processes. By integrating species‐specific dimensions of resources in habitat selection studies, we believe understanding of the foraging processes will be improved. [ABSTRACT FROM AUTHOR]