Drivers and Consequences of Partial Migration in the Red Kite (Milvus milvus)

Understanding the drivers and consequences of behavioural plasticity provides insight into how animals can respond to environmental change. Migratory behaviour is a prime example, where migration can vary in space and time, between and within individuals and can even vary in whether it occurs or not. Such behavioural plasticity may make migrant species more robust to rapid climate change and anthropogenic influences. This doctoral dissertation examines the phenomenon of partial migration, where both seasonal migrants and year-round residents coexist within the same population. It is assumed that use of the migrant versus resident strategy is a trade-off between survival and reproduction on the individual level. Residents are typically expected to gain reproductive benefits from prolonged presence on the breeding grounds, while migrants are predicted to gain survival benefits through avoidance of harsh overwintering conditions in the breeding area. My study system was a population of GPS-loggered Swiss red kites (Milvus milvus), in which migration strategy can vary within individuals, and incidence of the resident strategy is increasing. In order to explain this increase, I first examined how variation in individual condition (i.e., innate characteristics or current state) and environmental factors influenced migratory decisions throughout ontogeny. I then empirically quantified differences in reproductive output and survival between migration strategies.

In Chapter 2, I focused on the individual level drivers of migratory decisions during ontogeny. I modelled the ontogeny of migration strategy from birth to early adulthood, and how this is modulated by sex and body size. Individuals almost always migrated in their first year, and increasingly transitioned to residence with advancing age. This transition was largely directional, as the switch from residence back to migration was rarely observed. The age at which the transition to residence was made varied with sex and size, such that males transition to residence earlier than females, with the exception of very large females, which transition earliest. These results highlight that migratory decisions are not only facultative responses to the environment, but likely reflect important life-history decisions.

In Chapter 3, I examined the environmental drivers of migration strategy and fall migration phenology, modelling the effects of ambient temperature and food availability. I used a supplementary feeding experiment in fall and winter to manipulate food availability in territorial adults with the purpose of disentangling effects of temperature from food availability, which can be affected by temperature. Warmer than average winters were found to delay the departure of migration. Increased food availability decreased the overall probability to migrate, with a particularly strong effect in females. A tendency for temperature to affect control, but not supplementarily fed individuals suggests that food availability may be the primary pathway through which temperature acts on migratory behaviour. This chapter demonstrates that climate warming may result in decreased migratoriness, both in terms of propensity and later fall migration phenology. Furthermore, it suggests that additional sources of anthropogenic food, such as the garden feeding of raptors in Switzerland, may be another important pathway through which food availability can reduce migratoriness in our population.

In Chapter 4, I examined the effect of migratory decisions on reproductive output. I found annual variation and an elevation-dependent effect of migration strategy on reproduction. The reproductive advantage between strategies varied between years, where migrants intermittently exhibited an advantage. Territory elevation had a positive effect on migrant reproductive output in terms of fledgling number, while residents exhibited no such elevation-dependence. This discrepancy resulted in a resident reproductive advantage at low elevations, corresponding to a large part of the Swiss red kite distribution range and encompassing areas with the highest breeding density. Therefore, these results indicate that the resident reproductive advantage is likely a main driver of the increasing incidence of the resident strategy within this population.

In Chapter 5, I examined the effect of migratory decisions on survival. Monthly survival rates in migrants and residents varied with age, sex and season (breeding versus non-breeding). In residents in the younger age classes there is a general disadvantage in both the non-breeding and breeding seasons demonstrating direct and carry-over effects of winter conditions. These effects were particularly strong in young females. Survival differences between strategies disappeared with age, but adult male migrants had lower survival in the breeding season that adult male residents. Cumulative survival rates indicated that low survival in females than transitioned to residence young, as well as in males that remained migrant into adulthood exhibited particularly low survivorship across time. These results suggest that low survival in young individuals, and particularly in females, likely contribute to sex-dependent patterns in the ontogeny of migratory decisions observed in Chapter 2, while low survival in adult male migrants likely contributes to the spread of the resident strategy at the population level.

In Chapter 6, we examined how breeding red kites reacted to predator (eagle owl, Bubo bubo) exposure in the vicinity of their nest, based on brood-specific and environmental parameters using a supplementary feeding experiment. Food conditions determined the intensity of the anti-predator response. Under experimentally enhanced food conditions, red kite parents exhibited a reduced anti-predator response in young broods and an increased response in old broods, and this was further modulated by weather and brood size. These results suggest that, under favourable food conditions, brood value is the main driver of the intensity of antipredator responses, while under poor food conditions, brood vulnerability becomes more important.

In summary, I could confirm that migratory behaviour is highly plastic in red kites, enabling individuals to respond immediately to changes in their environment. Ontological and sexdependent differences in both the drivers and consequences of migration strategy suggest that differential migratory behaviour is crucial for success in different life phases and contexts. Furthermore, spatio-temporal variation in fitness emphasized the importance of considering large spatial and long temporal scales when estimating fitness parameters and drawing population-level conclusions based on single studies. Ultimately, the increase in the resident strategy observed in this partially migrant population is likely driven primarily by reproductive benefits across a large geographic scale, combined with condition-dependent survival benefits in certain demographic subgroups.

Keywords: condition-dependence, environmental change, food availability, life-history decision, migratory plasticity, migration strategy, migration threshold, ontogeny, raptor, reproduction, survival