Mon, Aug 02, 2021:On Demand
Background/Question/Methods
Logistical and analytical constraints of traditional analyses of habitat selection have required the assumption that purpose is homogeneous across contexts. Such an assumption can result in underestimation of the strength of selection or failure to observe selection, rendering subsequent strategies to manage populations ineffective. In the Prairie Potholes Region of Manitoba, raccoons (Procyon lotor) are an important nest predator of waterfowl, but it is unclear how raccoons use this landscape during the waterfowl nesting season. Our objective was to use high-resolution GPS telemetry to differentiate among behaviors by raccoons, evaluate behavior-specific habitat selection during the waterfowl nesting season, and describe the characteristics of wetlands used for different purposes. We used hidden Markov models (HMM) to fit 4-state models to the movement trajectories of collared animals and classified observations into discrete behaviors. We then fit behavior-specific Random Forest resource selection functions to evaluate the relative importance of proximity to landscape features on habitat selection for resting, foraging, slow travel, and directed travel behaviors.
Results/Conclusions We collected ~ 35,000 locations from 22 animals during the 2018 and 2019 nesting seasons, amounting to 475 animal-night’s worth of movement data. Proximity to a wetland was the most important variable contributing to selection for the resting and slow travel states, whereas proximity to a human-use site was the most important variable contributing to selection for foraging animals. The HMM predicted increasing probability of switching from the resting and foraging state to a directed travelling state with increasing distance from a wetland. Even though half of our observations were classified as resting, raccoons used far fewer wetlands for resting than for slow travel behavior. This finding suggests the buffers around wetlands serve a variety of functions, including as movement corridors in an otherwise inhospitable matrix. Although wetlands used for foraging and resting were larger than those used for travel, proximity of wetlands to other landscape features was consistent across uses. These results illuminate the complexity of habitat selection by an important predator in this patchy landscape. This insight can contribute to the development of conservation strategies that both accommodate the existence of the predator and support a robust population of breeding waterfowl.
Results/Conclusions We collected ~ 35,000 locations from 22 animals during the 2018 and 2019 nesting seasons, amounting to 475 animal-night’s worth of movement data. Proximity to a wetland was the most important variable contributing to selection for the resting and slow travel states, whereas proximity to a human-use site was the most important variable contributing to selection for foraging animals. The HMM predicted increasing probability of switching from the resting and foraging state to a directed travelling state with increasing distance from a wetland. Even though half of our observations were classified as resting, raccoons used far fewer wetlands for resting than for slow travel behavior. This finding suggests the buffers around wetlands serve a variety of functions, including as movement corridors in an otherwise inhospitable matrix. Although wetlands used for foraging and resting were larger than those used for travel, proximity of wetlands to other landscape features was consistent across uses. These results illuminate the complexity of habitat selection by an important predator in this patchy landscape. This insight can contribute to the development of conservation strategies that both accommodate the existence of the predator and support a robust population of breeding waterfowl.