摘要 : Context. Climate and land use are among the most important drivers of global biodiversity change, and they may be operating at different spatial scales. The effects of cross-scale interactions (CSIs) between these drivers on avian... 展开 Context. Climate and land use are among the most important drivers of global biodiversity change, and they may be operating at different spatial scales. The effects of cross-scale interactions (CSIs) between these drivers on avian abundance are poorly understood. Aims. Our primary objective was to assess whether the abundances of eight forest bird species in the eastern United States were significantly associated with CSIs involving four subregional climate variables (breeding- and pre-breeding-season temperature and precipitation) and two landscape variables (percentage exurban cover and forest patch size). Methods. For North American Breeding Bird Survey routes in six U.S. Environmental Protection Agency Level II ecoregions, we measured subregional climate variables within species maximum natal dispersal distances, and we measured landscape variables within species median natal dispersal distances. Using Akaike's information criterion and negative-binomial regression, we compared the fits of 21 a priori competing models separately for each of the eight species, and separately for percentage exurban cover and forest patch size (8 x 2 = 16 model sets). Key results. Total abundances during 2009-2013 of all eight species were associated with CSIs, which were informative in nine of the 16 best-supported models. The informative CSIs in the best-supported models involved all four subregional climate and both landscape variables. These results were evident after we accounted analytically for various methodological and environmental covariates, including within-scale interactions, that may otherwise have obscured the effects of CSIs. In some models, CSIs were more influential than were the associated additive effects, similar within-scale interactions, or other environmental variables, whereas in other models they were not. The associations between species abundances and CSIs were species-specific. Conclusions. CSIs among global drivers of change may be common, and failure to identify CSI effects may result in misleading bird-landscape models. Implications. Understanding how CSIs modify the effects of variables at different spatial scales may be crucial for effective broad-scale management of declining species. Conservation attention to species that are presently common but declining in abundance may help avoid extirpation in parts of their geographic ranges. 收起