Plasticity in migratory behaviour of large herbivores with contrasting species-specific traits

Ungulates show high plasticity in migration across and within species providing an ideal setting to enhance our understanding of the determinants of seasonal migration. Also, understanding seasonal ungulate movements is needed for management. The forage maturation hypothesis (FMH) predicts that plant green-up determines migration tendencies in ungulates. However, the relative costs and benefits to migration vary among species and sexes due to different foraging strategies, social organization or physiological constraints. Consequently, relationships between plant green-up and migration propensity, distance and timing may vary between species and sexes within species. Aiming to shed some light onto migration plasticity along the migration-residency continuum, we used GPS radio-collar data from >500 individuals in 10 roe deer (Capreolus capreolus) and 11 red deer (Cervus elaphus) populations across Europe to test for differences in migration behavior (e.g. migration distance, timing, probability) in these contrasting ungulate species. We used several methods, including ecological niche modeling, time-to event models and generalized linear models testing for effects of intrinsic factors (sex) and extrinsic conditions (e.g. topography, seasonality, canopy closure, plant productivity/ NDVI) on parameters describing migration plasticity. We found large variation in migration behavior in both roe deer and red deer. On average red deer were three times more migratory than roe deer (56% vs. 18% respectively). Red deer males were twice as migratory as females (82% vs. 38% respectively) in contrast to the much less sexually dimorphic roe deer, which had similar probability to migrate. Also timing of migration differed between species and sexes. Roe deer of both sexes migrated first in spring and roe deer males were the last to leave summer ranges in autumn. We found general support that variation in plant phenology affects migration probability in both species, but for red deer plant productivity (high absolute NDVI) determined the onset of migration in spring, while for roe deer the rate of plant green-up (high delta NDVI) was important. In autumn low NDVI values, suggesting reduced plant productivity, correlated with return migrations in both ungulate species, but the relationship was stronger for red deer. We also found switching rates between 1% and 23%, confirming that migration is facultative in both species. Our results confirm that migration behavior is shaped by species’ specific traits, which can mediate the effects of plant green-up on migration. The degree to which climatic and landscape changes may affect future migrations needs to be evaluated further, but our results suggest a stronger disconnect between plant green-up and migration for roe deer than for red deer. Also, our results suggest that across the populations we sampled, functional management units could be smaller for roe deer than for red deer based on annual movement distances. Ungulate management may benefit from sex-specific harvest timing and migration dates, given that males migrated later than females, especially in roe deer. Lastly, impairment of migration (e.g. under human landscape alteration) could contribute to population declines.