Clark's nutcracker (Nucifraga columbiana) conifer seed resource use in Yellowstone National Park: A hierarchical distance sampling approach

Thomas McLaren and Diana F. Tomback, Integrative Biology, University of Colorado Denver, Denver, CO, Walter Wehtje, Ricketts Conservation Foundation, Little Jackson Hole, WY, Lauren Walker, Yellowstone Center For Resources, National Park Service, Yellowstone National Park, WY, Douglas W. Smith, Yellowstone Center for Resources, National Park Service, Yellowstone National Park, WY, Nels Grevstad, Department of Mathematical and Computer Sciences, Metropolitan State University, Denver, CO

Background/Question/Methods
Clark’s nutcrackers (Nucifraga columbiana) harvest and cache seeds from many western conifer species, contributing to tree regeneration. Multiple seed resources within a region are required to sustain nutcracker populations but especially seeds of whitebark pine (Pinus albicaulis), a preferred but declining species. Nutcrackers are primary seed dispersers for whitebark pine, and loss of whitebark pine may trigger a downward spiral in nutcracker populations and whitebark pine regeneration. We examined how nutcrackers use forest habitat within Yellowstone National Park to determine all potential seed resources. In 2019 and 2020, we established eleven transects in five different forest community types with five point-count stations along each transect, and used distance sampling methods, relative cone abundance indices, and behavioral observations to determine habitat and seed resource use. To determine which forest types in the park are used by nutcrackers within a season and how variation in cone production influences their use within and across years, we applied a hierarchical approach to habitat modeling using distance observations to estimate detectability within each habitat type in a regression modeling framework. Additionally, with distance analysis we were able to assess the effect of habitat covariates on detection probability.

Results/Conclusions

Preliminary results are as follows: A negative binomial regression model of cone count data indicated that cone crops were generally larger at our study sites in 2020 compared to 2019. Using an information criteria approach to assess model parsimony, we found that cone crop index appeared to be positively correlated with nutcracker abundance. Additionally, forest type appeared as a cofactor in the detection process, indicating that observed nutcracker count data may be influenced by the available species and conifer density at our study sites. Notably, we observed high variation in limber pine site visitation, related to cone availability. Results from our regression models indicate that nutcracker abundance during seed harvesting season is correlated with cone crop index, which shows significant interannual variation. In addition to producing a habitat use model that integrates detectability based on distance sampling analysis, we are performing a power analysis with the study design to determine its ability to detect a decline in the nutcracker population.