Predicting grizzly bear density in western North America

Area of Science:

• Ecology and Conservation Biology
• Wildlife Management
• Population Dynamics

Background:

• Grizzly bear (Ursus arctos) conservation is hampered by controversial and often subjective density estimates used for management.
• Existing objective methods for predicting grizzly bear density are limited by their poor generalizability across different study areas.
• Accurate grizzly bear density data is crucial for effective conservation planning and determining allowable human-caused mortality.

Purpose of the Study:

• To develop objective, generalizable regression models linking grizzly bear density to ecosystem productivity and mortality.
• To predict current grizzly bear population sizes across Canada using these novel models.
• To provide alternative population estimates for conservation status assessment and mortality limit setting.

Main Methods:

• Regression models were constructed using 90 interior and 17 coastal North American grizzly bear density estimates.
• Models incorporated measures of ecosystem productivity (e.g., terrestrial vegetation, salmon availability) and mortality factors (e.g., human use, topographic ruggedness).
• Coastal models included tree cover, salmon proportion in diet, and topographic ruggedness; interior models used terrestrial productivity, vegetation cover, human use, and ruggedness indices.

Main Results:

• The best coastal model showed negative relationships with tree cover and positive relationships with salmon proportion and topographic ruggedness.
• The best interior model integrated terrestrial productivity, vegetation cover, human use, and topographic ruggedness.
• Model predictions suggest fewer grizzly bears in British Columbia but more bears in Canada compared to the latest status review.

Conclusions:

• The developed regression models provide objective and generalizable tools for estimating grizzly bear density across diverse ecosystems.
• These predictions can inform conservation status assessments and aid in setting sustainable human-caused mortality limits.
• While useful for static population estimates, the models do not assess population trends over time.