作者: Luca Börger , Thomas D Nudds
DOI: 10.1890/12-1683.1
关键词:
摘要: Understanding the effects of landscape change and environmental variability on ecological processes is important for evaluating resource management policies, such as emulation natural forest disturbances. We analyzed time series detection/nondetection data using hierarchical models in a Bayesian multi-model inference framework to decompose dynamics species distributions into responses variability, spatial variation habitat conditions, population interspecific interactions, while correcting observation errors sampling regimes. modeled distribution 14 waterbird (broadly defined, including wetland riparian species) from two different breeding bird surveys collected Boreal Shield ecozone within Ontario, Canada. Temporal occupancy (2000-2006) was primarily driven by climatic variability. Only showed evidence consistent temporal trends distribution: Ring-necked Duck (Aythya collaris) decreased, Red-winged Blackbird (Agelaius phoeniceus) increased. The had good predictive ability independent over (1997-1999). Spatial strongly related specific land cover types disturbance: Fire harvesting influenced more than did roads, settlements, or mines. Bioclimatic heterogeneity indices geographic coordinates exerted negligible influence most distributions. Estimated suitability spatially (Hudson Bay Lowlands ecozone). Additionally, we detected interactions. Species fire were similar 13 species; thus, forest-harvesting practices Ontario generally appeared emulate waterbirds timescales 10-20 years. Extrapolating all 84 Shield, however, suggested that up 30 may instead have altered (short-term) due forestry practices. Hence, disturbances are critical components ecology boreal which aim approximate them succeed allowing maintenance associated species, but improved monitoring modeling large-scale will be necessary resolve existing uncertainties, especially less-common species.