气候变化领域集成服务门户

Species responses are influenced by processes operating at multiple scales, yet many conservation studies and management actions are focused on a single scale. Although landscape-level habitat conditions (i.e., habitat amount, fragmentation and landscape quality) are likely to drive the regional persistence of spatially structured populations, patch-level factors (i.e., patch size, isolation, and quality) may also be important. To determine the spatial scales at which habitat factors influence the regional persistence of endangered Ord's kangaroo rats (Dipodomys ordii) in Alberta, Canada, we simulated population dynamics under a range of habitat conditions. Using a spatially-explicit population model, we removed groups of habitat patches based on their characteristics and measured the resulting time to extinction. We used proportional hazards models to rank the influence of landscape and interacting patch-level variables. Landscape quality was the most influential variable followed by patch quality, with both outweighing landscape- and patch-level measures of habitat quantity and fragmentation/proximity. Although habitat conservation and restoration priorities for this population should be in maximizing the overall quality of the landscape, population persistence depends on how this goal is achieved. Patch quality exerted a significant influence on regional persistence, with the removal of low quality road margin patches (sinks) reducing the risk of regional extinction. Strategies for maximizing overall landscape quality that omit patch-level considerations may produce suboptimal or detrimental results for regional population persistence, particularly where complex local population dynamics (e.g., source-sink dynamics) exist. This study contributes to a growing body literature that suggests that the prediction of species responses and future conservation actions may best be assessed with a multi-scale approach that considers habitat quality and that the success of conservation actions may depend on assessing the influences of habitat factors at multiple scales. (C) 2015 Elsevier GmbH. All rights reserved.