Variation in top‐down control of avian reproductive success across a fragmentation gradient

Abstract

Understanding mechanisms that inhibit population persistence in fragmentation‐sensitive species is a key to conservation and management. It has become clear that species vary in their sensitivity to habitat fragmentation caused by urbanization and other land uses. We studied variation in avian reproductive success in coastal sage scrub habitat across a fragmentation gradient (interior of large habitat blocks (>2300 ha)→edge of large habitat blocks→large habitat fragments (37–80 ha)→small habitat fragments (5–17 ha)) in urban southwestern California. We focused on four species of birds: the shrub‐nesting wrentit (Chamaea fasciata) and California towhee (Pipilo crissalis) and the ground‐nesting spotted towhee (P. maculatus) and rufous‐crowned sparrow (Aimophila ruficeps). Snake abundance was the best predictor of nest failure in the ground‐nesting species, whereas abundance of avian predators (Corvidae, Mimidae) was associated with nest failure of the wrentit and a combination of snake and avian predator abundance was correlated with nest failure of the California towhee. Avian nest predator abundance increased across the fragmentation gradient; by contrast, snake abundance decreased across this gradient. As a result, top‐down control of reproductive success in populations of the spotted towhee and rufous‐crowned sparrow was largely absent in habitat fragments. These two species enjoyed much higher nest success on fragments yet, paradoxically, both species are sensitive to fragmentation whereas the shrub‐nesting species are not. Our findings suggest that another process – perhaps differential survivorship or dispersal – yields reduced abundance of ground‐nesting species in habitat fragments. Studies of top‐down control and trophic cascades in fragmented landscapes should reflect that the strength of top‐down forces across a landscape varies with characteristics of predator and prey and their autecological responses to ecological gradients.