Maternal Denning. Worldwide, most available denning information has been derived from visual surveys. Except in Alaska and Canada, few animals have been followed to dens by radiotelemetry. Outside Alaska, most radio-collared polar bears have not been resighted on emergence to confirm denning outcome. Still fewer have been monitored long enough to assess patterns of fidelity to den sites or habitat types. Hence, data that might provide an understanding of the significant differences in denning patterns worldwide are not available. Why polar bears in different regions choose to den where they do is largely unknown.

The absence of comparative data among different geographic regions minimizes our understanding of the influence of different sea-ice, climate, and biotic conditions on the chronology and geographic distribution of denning. Polar bears in dens are more vulnerable to anthropogenic as well as natural disturbances than at other times in their life cycle. Therefore, denning data that are comparable over broad geographic regions are essential to interpretation of proposed increases in human uses of the Arctic.

Estimating Numbers. Worldwide, 20 populations of polar bears currently are recognized (see "Movements" Fig. 27.8). The total combined number of bears in those populations is thought to be between 21,500 and 25,000 (Lunn et al. 2002, Table 1). Population estimates vary in quality from educated guesses in some of the populations to rigorous values complete with confidence intervals in others. Because most of the populations are hunted and many are subject to other potential human perturbations, better estimates of abundance are needed. Optimistic population estimates in the 1950s and 1960s resulted in excessive harvest and declines in Beaufort Sea polar bear numbers (Amstrup et al. 1986). Other populations also were suffering from overharvest at that time (Prestrud and Stirling 1994). Managers need reliable estimates of population size and trend to prevent overharvests from recurring. Despite over 30 years of polar bear capture, however, estimates of population size and trend have been elusive. Failures of past capture­recapture efforts to provide reliable estimates apparently resulted from biases caused by heterogeneity in capture or survival probabilities (Seber 1982; Hwang and Chao 1995). Amstrup et al. (2001) and McDonald and Amstrup (2001) improved estimates of numbers of polar bears by modeling heterogeneity in capture probability with covariates. Their analyses also pointed out many shortcomings of past efforts as well as weaknesses in available data. Estimates of population size based on mark and recapture data are dependent on the number of animals marked and the number of occasions during which marking occurs. Larger numbers of marked animals increase probabilities of capture and reduce the variance of the estimated population size. Increased numbers of occasions allow selection of covariates that help to compensate for heterogeneity. The lessons learned from past efforts must be applied to future studies if more accurate and precise estimates are to be obtained.