Connrmation, Revision, and Sensitivity Analysis of the 1994 Scientiic Committee Assessment of the Bering-chukchi-beaufort Seas Stock of Bowhead Whales

Results from the approximate reweighting method used by the Scientiic Committee for its 1994 assessment of the Bering-Chukchi-Beaufort Seas stock of bowhead whales are connrmed using the full Bayesian synthesis approach. Sensitivity trials are examined to investigate several areas of interest identiied during this assessment. The results show that the full analysis gives results which are very close to those obtained by the Scientiic Committee in 1994, and which are not very sensitive to changing the distributions of model parameters in ways which are reasonable for bowheads. The Scientiic Committee obtained an estimated 5 th percentile for replacement yield of 104 whales; four independent runs of the full analysis give values of 104, 108, 104, and 115. The results show no reason to question the conclusions of the 1994 Scientiic Committee assessment. In its 1994 bowhead assessment, the Scientiic Committee agreed to use a stock abundance estimate based on data from 1988 because the estimate from the more recent 1993 data was preliminary at that time. Analyses of these recent data have now been revised to produce updated population abundance and trend estimates. We present a revised assessment, like that completed by the SC in 1994, but using these revised abundance and trend estimates. The updated results show little change from the 1994 assessment. Borel's paradox has been identiied as a concern in Bayesian synthesis analyses. Simple sensitivity analysis methods, like those in this paper, can be used to examine the degree to which the problem may occur in an application. We show how analysis of sensitivity to alternative priors can be reinterpreted as investigations of sensitivity to alternative reparameterisations and to Borel's paradox.