Background: Intraguild predation (IGP) is widespread but it is often neglected that guilds commonly include many layers of dominance within. This could obscure the effects of IGP making unclear whether the intermediate or the bottom mesopredator will bear higher costs from the emergence of a new top predator.
Methodology/principal findings: In one of the most extensive datasets of avian IGP, we analyse the impact of recolonization of a superpredator, the eagle owl Bubo bubo on breeding success, territorial dynamics and population densities of two mesopredators, the northern goshawk Accipiter gentilis and its IG prey, the common buzzard Buteo buteo. The data covers more than two decades and encompass three adjacent plots. Eagle owls only recolonized the central plot during the second decade, thereby providing a natural experiment. Both species showed a decrease in standardized reproductive success and an increase in brood failure within 1.5 km of the superpredator. During the second decade, territory dynamics of goshawks was significantly higher in the central plot compared to both other plots. No such pattern existed in buzzards. Goshawk density in the second decade decreased in the central plot, while it increased in both other plots. Buzzard density in the second decade rapidly increased in the north, remained unchanged in the south and increased moderately in the center in a probable case of mesopredator release.
Conclusions/significance: Our study finds support for top-down control on the intermediate mesopredator and both top-down and bottom-up control of the bottom mesopredator. In the face of considerable costs of IGP, both species probably compete to breed in predator-free refugia, which get mostly occupied by the dominant raptor. Therefore for mesopredators the outcome of IGP might depend directly on the number of dominance levels which supersede them.