Abstract 201 – Large-scale estimates of Arctic grizzly bear abundance using spatial capture recapture methods.

John Boulanger, Murray Efford, Malik Awan, Kim Poole

One of the challenges of estimating grizzly bears is obtaining estimates of abundance for large
regional areas given the low densities and wide-ranging movements of bears. Traditionally,
practitioners have extrapolated estimates from smaller grid-based sampling areas to larger
regions, which can create uncertainty in how well the smaller area represents the region. This
challenge is very apparent in the Arctic where grizzly bears have large ranges and occur at low
densities, making it problematic to obtain regional estimates given that all sampling areas
require helicopter access. In this study, we developed a sub-grid cluster sampling method to
estimate grizzly bear abundance across a 156,500 km2 regional area (approximately the size of
Illinois) in the Kitikmeot Region, Nunavut, using DNA-based spatially explicit capture recapture
methods. In the initial phase of the study optimization methods were used to estimate sub-
grid dimension, numbers of sub-grids, and number of sampling sessions required to equal the
precision of previous grid-based estimates (2008-09) in a smaller area (54,200 km2) near
Kugluktuk. Estimates from the initial study in 2021 demonstrated that estimates of
comparable precision could be derived using subgrids with savings in effort employed. The
subgrid approach has now been expanded to the larger survey area allowing estimates of
abundance for the entire regional area (156,500 km2) with sampling occurring in 2022 and 2023. The results of our study illustrate the use of optimization methods as a means to
evaluate and design larger-scale surveys. We discuss strengths and weaknesses of this
approach with suggestions on how it may be applied to other areas.