Run to the hills: partial migration in large herbivores

Partially migratory populations in which some individuals move to allopatric ranges during one season while others remain on their shared range, offer a unique opportunity to understand which factors shape the realized niche of individuals with plastic movement behaviors. For ungulates, forage and its spatiotemporal variability, risk (predation, humans) and density have been suggested to be the main determinants for the probability, distance and timing of migration. Roe deer (Capreolus caproelus), a small browsing ungulate with a high ecological plasticity and a wide distribution, present an ideal model species to test hypotheses on migration plasticity. The green wave hypothesis predicts that migratory ungulates follow high quality forage, but has not been tested for browsers. I tested for differences in broad-scale vegetation composition and small-scale plant phenology and diet quality (using fecal nitrogen as a proxy) between resident (N=26) and migrant (N=11) roe deer in a diverse alpine study area in Northern Italy. Migrant and resident ranges differed at broader scales, phenology was similar and diet quality was significantly higher for migrants. I conclude that roe deer are selective for forage at different spatiotemporal scales and that the green- wave hypothesis may not be the only explanation for their migratory plasticity. Therefore, I expanded from this single-population feeding niche comparison to testing if roe deer (N=71) switch or follow seasonal niches in five populations spanning wide resource gradients across Europe. I applied a novel integrative approach to classify/describe migration and assessed the determinants for niche differences between seasonal strategies. I found that migrants switched niches between winter and summer and the main niche differences were a function of topography, winter severity, spatiotemporal forage variation and density. Lastly, I scaled up to a between-species comparison of roe and red deer (Cervus elaphus; N >500) to test hypotheses between and within species across a similar latitudinal gradient. Red deer had a much higher probability of migration with pronounced sex differences in contrast to the much less dimorphic roe deer. My results on the determinants driving plasticity in probability, distance and timing of migration, confirm the overall hypothesis that migration behavior is multi-causal and shaped by species’ specific characteristics, including sexual dimorphism, feeding and breeding behavior.