Non-invasive methods for obtaining occupancy probabilities and density estimates of Interior Alaska's mesocarnivore populations

Abstract

Mesocarnivore species worldwide have been shown to be significant drivers of ecological communities. Changes in their abundance and distributions are known to cause cascading effects throughout ecosystems, and changes to the landscape and climate will likely lead to shifts in mesocarnivore population sizes and distributions. However, the current status of these species in some of the world's most susceptible landscapes is not known. I assessed the impacts of abiotic factors on the distributional patterns and abundance of boreal mesocarnivores and evaluated methods commonly used to estimate density and occupancy. I conducted non-invasive winter surveys of coyotes (Canis latrans), red foxes (Vulpes vulpes), lynx (Lynx canadensis), wolverines (Gulo gulo), and marten (Martes americana) in the interior of Alaska. Overall, mesocarnivore occupancy was most strongly influenced by snow depth and snow compaction as well as habitat type. Canid species used areas with shallow and compact snow while mustelid species used deeper and fluffier snow conditions most often, and lynx used areas with shallow and fluffy snow. Forested habitat types were used most commonly across all mesocarnivores. Prey abundance and the presence of human activity were less influential to mesocarnivore occupancy patterns than snow conditions and habitat, suggesting that a changing boreal climate may have a strong, direct influence on the distribution of these mesocarnivores. Estimating current population status of these species is particularly important in areas that are most susceptible to change, and I used two occupancy-modeling methods and a spatially explicit capture-recapture density estimator to assess coyote and red fox populations. Occupancy and density are two distinct parameters, however, the simplicity of occupancy (both in terms of sampling and modeling) makes its use as a proxy for density an appealing possibility. I found that occupancy and density estimates were not consistent and led to significantly different inference about coyote and red fox populations. Coyotes and red fox occupancy probabilities were similar to each other (range: 0.34-0.48), but red fox density was nearly four times greater than coyote density. While both methods produced precise parameter estimates, top-ranking occupancy and density models were different. I suggest that managers use caution when using occupancy as a proxy for density. Occupancy is best used to address questions related to spatial use, while density should be used to assess population size. Together, these findings provide valuable information about the current status of a previously unstudied mesocarnivore community and provide managers with useful insight into study design and management actions that should be taken to best protect this guild.