Predator responses to non-stationary rodent cycles

Abstract: Regular fluctuations in population size, cycles, are common in small mammals and have important effects on predator populations and life histories. In this thesis, I identify long-term patterns and processes in two specialist predators, the arctic fox Vulpes lagopus and the rough-legged buzzard Buteo lagopus, in relation to their prey (lemmings and voles) and in the case of the arctic fox also to a dominant competitor, the red fox Vulpes vulpes. The results demonstrate that the specialist predators as expected were limited by food supply, which was revealed by combining long-term monitoring with a pseudo-experimental approach. But dampening of cycles has led to long-term declines in the predator populations. Cycles in red and arctic foxes disappeared in the 1970s and 1980s. A return to cyclic dynamics was found locally in the red fox in the boreal zone, but monitoring in the mountain region identified a widespread return of cycles in both lemmings and voles in the early 2000s. This increase in natural food was sufficient to halt the decline in the arctic fox population, but a large-scale field experiment revealed that only supplementary fed subpopulations increased in size. Competition with the red fox further had a negative impact on the arctic fox population. A theoretical model suggested that this asymmetric competition is context dependent and most severe if red fox numbers are independent of rodent density due to access to alternative food. The number of breeding rough-legged buzzards was determined primarily by rodent abundance, but has been nearly halved since the 1970s and was decoupled from rodents in the most recent years. Reproductive output is currently also lower due to smaller clutch sizes. The functional responses of the rough-legged buzzard were complex and differed between lemmings and voles. Rodent cycles are clearly essential for maintaining biodiversity, but spatiotemporal patterns and effects on ecological communities are increasingly variable.