Lethal wolf control elicits change in moose habitat selection in unexpected ways

Abstract

Moose (Alces alces) and woodland caribou (Ranger tarandus caribou) are the 2 large prey species for wolves (Canis lupus) in the Nearctic boreal forest in North America. Caribou have declined, with widespread anthropogenic disturbance as the ultimate cause and wolf predation as the proximal cause. To conserve caribou the government of Alberta, Canada initiated a wolf control program to reduce predation rates on caribou populations and contribute to caribou population recovery. Predators play an important role in shaping the structure and function of ecosystems through top-down forces. We hypothesized that the strongest factors influencing moose occurrences would reflect changes in predation risk before and after the onset of wolf control. We weighed evidence for competing hypothesis by deploying cameras across a highly industrialized landscape in Alberta for 3 years after wolf control (2017–2020), capitalizing on 3 years of existing data before the onset of wolf control (2011–2014). We created generalized linear models representing competing hypotheses about moose response to natural and anthropogenic landscape features before and after wolf control, examining support for each in an information-theoretic framework. Prior to wolf control, the model containing landscape features providing security cover was best-supported, but this was scale-dependent. After wolf control, the model containing landscape features that offer increased forage opportunities was best-supported. Unexpectedly, the direction of effect was often opposite to predictions, with moose avoiding some features thought to provide security and forage. We demonstrate that lethal predator control affects the spatial distribution of its primary prey species but in ways we do not fully comprehend, highlighting the need for a better understanding of community dynamics following wolf control.