Abstract
Understanding causal mechanisms of decline for species at risk is critical for effective conservation. Caribou (Rangifer tarandus) face threats from habitat loss and degradation due to human activities, and many caribou populations across Canada have experienced dramatic declines in recent decades. Disturbance‐mediated apparent competition (DMAC) has been implicated in many of these declines, but its generality has been questioned, particularly for low‐productivity caribou ranges. The DMAC hypothesis leads to the following predictions: 1) a vegetation productivity pulse after disturbance, 2) primary ungulate prey attraction to disturbed areas, 3) predator attraction to primary prey and disturbance, and 4) increased caribou predation risk due to overlapping habitat use with primary prey and predators. We tested these predictions for the declining Itcha‐Ilgachuz caribou population, located in the low‐ productivity Chilcotin Plateau region of west‐central British Columbia, Canada. We used a remotely sensed productivity index to examine vegetation recovery patterns after disturbance and used camera traps and Bayesian mixed effects negative binomial regression models to estimate the responses of primary prey, predator, and caribou relative abundance to landscape disturbances <40 years old, interacting species, and other habitat features. We identified a productivity pulse in harvested and burnt forest patches, but overall productivity was lower than in other caribou ranges where DMAC occurs. Primary prey, moose (Alces alces) and mule deer (Odocoileus hemionus), showed strong positive responses to burnt areas and weak positive responses to harvested forest. For predators, wolves (Canis lupus), black bears (Ursus americanus), and grizzly bears (Ursus arctos) were positively associated with primary prey species, while coyotes (Canis latrans) and Canada lynx (Lynx canadensis) were more strongly associated with snowshoe hare (Lepus americanus), and wolverines (Gulo gulo) were not associated with any focal prey species. Wolves, grizzly bears, and wolverines were not associated with habitat disturbance, but black bears, coyotes, and lynx responded positively to burned and harvested areas. Caribou did not have reduced relative abundance in harvested forests or burns, potentially increasing their overlap with predators. Overall, we found mixed support for DMAC for the Itcha‐Ilgachuz caribou population, with stronger evidence for a pathway mediated by disturbance from forest fire, rather than forest harvest. We recommend further research and action on wildfire management for the recovery of this population, including monitoring population trends of caribou and interacting species in response to habitat management. Our results emphasize the context‐dependency of mechanisms of decline for caribou and underscore the need for population‐specific knowledge to effectively conserve threatened species.