ESTIMATING GRIZZLY BEAR POPULATION SIZE USING CAMERA SIGHTINGS

Abstract

Estimating population size of grizzly bears (Ursus arctos horribilis) is hampered by certain characteristics (Harris 1986, Interagency Grizzly Bear Comm. [IGBC] 1987, Miller et al. 1987, Miller 1990a): grizzly bears are secretive, aggressive, and difficult to observe; populations often inhabit mountainous, densely vegetated, remote habitats; individual bears often have different probabilities of capture or observation that violate the assumption of equal catchability; age and gender are difficult to determine without handling; sampling opportunities are limited because bears spend 6-7 months in dens; and sample sizes are typically small. These factors impose logistical and financial constraints on researchers obtaining point estimates and confidence intervals (CI's) for the population. The most reliable estimator for grizzly bears is the Petersen capture-recapture design (Miller et al. 1987); it accommodates small samples and requires only a single sighting period after the initial marking period (Seber 1982). The few published estimates of grizzly bear populations have wide CO's (Harris 1986, IGBC 1987, Miller et al. 1987). Model biases are more serious than small-sample bias