| Summary: | The assessment of disturbance effects on wildlife and resulting mitigation efforts are founded on edge-effect theory. According to the classical view, the abundance of animals affected by human disturbance should increase monotonically with distance from disturbed areas to reach a maximum at remote locations. Here we show that distance-dependent movement taxis can skew abundance distributions toward disturbed areas. We develop an advection-diffusion model based on basic movement behavior commonly observed in animal populations and parameterize the model from observations on radio-collared caribou in a boreal ecosystem. The model predicts maximum abundance at 3.7 km from cutovers and roads. Consistently, aerial surveys conducted over 161,920 km2 showed that the relative probability of caribou occurrence displays nonmonotonic changes with the distance to anthropogenic features, with a peak occurring at 4.5 km away from these features. This aggregation near disturbed areas thus provides the predators of this top-down-controlled, threatened herbivore species with specific locations to concentrate their search. The edge-effect theory developed here thus predicts that human activities should alter animal distribution and food web properties differently than anticipated from the current paradigm. Consideration of such nonmonotonic response to habitat edges may become essential to successful wildlife conservation. Average Vector DirectionData used to evaluate changes in the orientation of caribou with respect to the nearest road or cut block as a function of distance from these anthropogenic features (Fig. B1 Online Appendix B).AvgVectorDirection.xlsxData for GAM analysisData used to model the relative probability of occurrence of caribou track networks in winter as a function of the distance from the nearest anthropogenic disturbance (Table 1). Landscape features (distance from the nearest cut block or road, various land cover types) are provided at observed track network locations and at random locations.GAM.xlsHome ...
|
|---|