Reproductive responses of an apex predator to changing climatic conditions in a variable forest environment

Abstract

Apex predators are ideal subjects for evaluating the effects of changing climatic conditions on the productivity of forested landscapes, because the quality of their breeding habitat depends primarily on the availability of resources at lower trophic levels. Identifying the environmental factors that influence the reproductive output of apex predators can, therefore, enhance our understanding of the ecological relationships that provide the foundation for effective forest management strategies in a variable environment. To identify the determinants of breeding-habitat quality for an apex predator in a forest food web, I investigated the relationships between site-specific environmental attributes and the reproductive probabilities of northern goshawks (Accipiter gentilis) on the Kaibab Plateau, Arizona during 1999-2004. I used dynamic multistate site occupancy models to quantify annual breeding probabilities (eggs laid) and successful reproduction probabilities (≥1 young fledged) relative to temporal and spatial variation in climatic conditions (precipitation and temperature), vegetation attributes (forest composition, structure, and productivity), and prey resources (abundances of 5 mammal and bird species). Climatic conditions during the study period varied extensively, and included extreme drought in 2003 and record-high precipitation in 2004. There was also substantial variation in the amount and distribution of 4 forest cover types among 102 goshawk territories within the 1,285-km² study area. The abundance of most prey species, especially mammals, also varied considerably among years, cover types, and goshawk territories. I identified the environmental components that best explained spatiotemporal variation in goshawk reproductive parameters using an information-theoretic approach to evaluate the relative weight of evidence for each model. Out of 62 potential models representing how the distribution of territories with breeding goshawks varied among years and across the landscape, the model with the most evidence indicated that breeding attempts depended on the interaction between the current-year conditions and the site-specific attributes of territories, whereas fledgling production only varied among years. The best-supported model containing variables for climatic conditions included the effects of average annual precipitation and ambient temperature prior to egg-laying on breeding probabilities, and cumulative precipitation during the first 3 weeks post-hatching on successful reproduction probabilities. The best-supported model incorporating vegetation attributes included year-specific effects of forest cover type on breeding probabilities, and conifer cone production on successful reproduction probabilities. Of the prey models I considered, the best-supported model included the effects of the abundance of prey species with distinct ecological niches on both breeding and successful reproduction probabilities. Overall, a habitat model representing food resource availability was the most parsimonious explanation of variation in goshawk reproduction, because it incorporated the effects of temporal variation in climatic conditions and spatial variation in vegetation attributes on the abundance and distribution of prey species with different functional traits. I found that greater heterogeneity in habitat attributes mitigated the effects of changing climatic conditions on reproductive probabilities, because more diverse prey communities increased the abundance of food resources for goshawks during and following drought. Based on the environmental factors that had the greatest influence on reproductive output of an apex predator, I identified management actions designed to enhance the diversity of niches at a landscape scale and reduce the effects of climatic extremes on the productivity of forest food webs.