Statistical population reconstruction to evaluate grizzly bear trends in British Columbia, Canada

Abstract Grizzly bear (Ursus arctos) populations are costly to monitor by traditional survey methods. In British Columbia, Canada, hunter kill data are available and provide relatively inexpensive information that possibly can be used to estimate trends in hunted populations. We applied statistical population reconstruction (SPR) using Program PopRecon 2.0 to evaluate trends in abundance of ≥3-year-old male grizzly bears for 3 large areas in British Columbia. Model inputs included annual estimates of age-at-kill and hunter effort, combined with auxiliary information on population abundance in 2012, and a non-hunting survival rate. Modeled abundance in all 3 areas was sensitive to the auxiliary abundance estimate but less so for the auxiliary survival estimate or the length of the time series. Relative trends in abundance appeared to be primarily affected by kill and effort data and were less affected by the auxiliary data. The gradual increase in abundance within the Temperate Mountains area from 1985 to 2004 followed by an apparent decline was consistent with other independent studies and supported the premise that grizzly bear numbers were recovering from a population low until between 2000 and 2005. Our results suggest that the grizzly bear population in the Boreal–Sub-boreal area was also recovering during this period. Our analysis demonstrates the potential utility of SPR for monitoring grizzly bear population trends, but results from the Coastal area also highlight the importance of sufficient hunter-kill and -effort data, in addition to quality auxiliary data, to detect population change. Future enhancements in Program PopRecon may help improve the performance and utility of SPR for grizzly bears in British Columbia.