|
Activity Number:
|
72
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 12, 2002 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Bayesian Stat. Sciences*
|
| Abstract - #301378 |
|
Title:
|
An Empirical Bayesian Model with Bivariate Normal-log-normal Rregional Prior Distribution for Fish stock-recruitment Analysis
|
|
Author(s):
|
Din Chen*+
|
|
Affiliation(s):
|
International Pacific Halibut Commission
|
|
Address:
|
P. O. Box 95009, Seattle, Washington, 98145, USA
|
|
Keywords:
|
hierarchical Bayesian ; bivariate normal-log-normal prior distribution ; empirical Bayesian ; MCMC ; fish stock-recruitment analysis
|
|
Abstract:
|
A regional meta-model is formulated using a hierarchical Bayesian framework to combine information from multiple populations (fish stocks) of coho salmon (Oncorhynchus kisutch) within two large fisheries management units in southern and northern British Columbia, Canada. The prior distribution is constructed from an analysis of all fish stock-recruitment data, rather than the more usual approach of assuming a prior distribution. The analysis indicated that the prior distribution for the regional parameters of the traditional Ricker stock-recruitment model was bivariate normal-log-normal (NLN), with a high correlation between the two regional parameters. Since this distribution has not been formally discussed, I formulated the density function for the NLN distribution and proved some of its important properties. An empirical Bayesian and MCMC approach were then used to estimate the regional distributions of the Ricker parameters and derived management parameters.
|
