Search

Background: To assess population persistence of species living in heterogeneous landscapes, the effects of habitat on reproduction and survival have to be investigated., Methodology/principal Findings: We used a matrix population model to estimate habitat-specific population growth rates for a population of northern wheatears Oenanthe oenanthe breeding in farmland consisting of a mosaic of distinct habitat (land use) types. Based on extensive long-term data on reproduction and survival, habitats characterised by tall field layers (spring- and autumn-sown crop fields, ungrazed grasslands) displayed negative stochastic population growth rates (log lambda(s): -0.332, -0.429, -0.168, respectively), that were markedly lower than growth rates of habitats characterised by permanently short field layers (pastures grazed by cattle or horses, and farmyards, log lambda(s): -0.056, +0.081, -0.059). Although habitats differed with respect to reproductive performance, differences in habitat-specific population growth were largely due to differences in adult and first-year survival rates, as shown by a life table response experiment (LTRE)., Conclusions/significance: Our results show that estimation of survival rates is important for realistic assessments of habitat quality. Results also indicate that grazed grasslands and farmyards may act as source habitats, whereas crop fields and ungrazed grasslands with tall field layers may act as sink habitats. We suggest that the strong decline of northern wheatears in Swedish farmland may be linked to the corresponding observed loss of high quality breeding habitat, i.e. grazed semi-natural grasslands.