Improving inferences on wildlife responses to oak-centered forest management with models that account for imperfect detection.

摘要

Oak-hickory (Quercus-Carya) species are important mast producers that can regulate wildlife populations. Specifically, mast is a critical energy source during winter months and positively influences small mammal abundance. Wildlife also can affect oak reproductive success through seed and seedling predation and dispersal. Current oak-focused forest management strategies attempt to reverse declines in oak regeneration. An important corollary of this work is to develop a greater understanding of wildlife responses to these management strategies. In an attempt to understand these responses, I characterized small mammal populations and communities at different oak-managed successional sites, both pre- and post-harvest. I used methodologies that explicitly acknowledged the limitations inherent in the sampling of cryptic species by incorporating detection probability into the estimates of occupancy, abundance, species richness, colonization rates, and extinction rates for small mammals. I also incorporated detection probability into the indicator value method to more accurately select indicator species. This method, which I demonstrate with song birds in oak-managed stands, can be used when selecting wildlife species to monitor for responses to forest management. I found dramatic differences in occupancy and abundance estimates for small mammals relative to naive estimates, especially for rarer species. Small mammal assemblages at my study sites responded to stand structure and microsite characteristics, with an Eastern chipmunk (Tamias striatus) dominated community at early seral stages developing into a more diverse, white-footed mouse ( Peromyscus leucopus) dominated community at later seral stages. I also found that imperfect detection influenced the selection of indicator species, with the greatest influence occurring when species counts and probability of detection are low or when probability of detection strongly varies with site-association. Failure to account for imperfect detection could result in erroneous estimates of occupancy, abundance, and incorrect species-site associations. Monitoring plans or management decisions made based on methods that do not account for imperfect detection could have negative repercussions for conservation.