Identifying individual polar bears at safe distances: A test with captive animals

Area of Science:

• Wildlife biology
• Animal behavior
• Conservation science

Background:

• Individual identification is vital for population and behavioral studies.
• Automated photo processing offers speed but requires high-quality, close-range images.
• Existing methods like whisker spot analysis for polar bears (Ursus maritimus) have limitations due to photographic requirements.

Purpose of the Study:

• To evaluate the accuracy of a physiognomic identification method for polar bears using photographs taken at greater distances.
• To compare the effectiveness of this method with automated whisker spot analysis.
• To assess the impact of observer training on identification accuracy.

Main Methods:

• Photographs of 15 zoo-housed polar bears were taken from distances up to 400m.
• Ten observers, some with photogrammetric training, identified individuals across multiple trials.
• Accuracy and error rates were compared between observers and with automated whisker spot analysis.

Main Results:

• The physiognomic method achieved observer repeatability of 0.68.
• Trained observers had significantly lower error rates (0.13) compared to untrained observers (0.72).
• Physiognomic method error rates were lower than automated whisker spot analysis (0.26-0.58).

Conclusions:

• Physiognomic identification from distant photographs is a viable alternative for polar bear studies.
• Photogrammetric training enhances the accuracy of this identification method.
• Combining multiple identification techniques may optimize data collection in wildlife studies under varying conditions.