Biomathematics seminar
Tuesday 11/26/07
Speaker: Matthew Krachey

Title: Hierarchical Bayesian Tag-Return Models

Abstract:

Fisheries and wildlife managers need reliable estimates of population
abundance and/or mortality rates of exploited stocks. Tag-return
models have long been used as an inexpensive methodology to estimate
 mortality rates. Recent advances have allowed managers to isolate the
main causes of mortality: natural causes and harvest. However,
unbiased tag-return models require the tag-return reporting rate to be
estimated. Recent maximum likelihood tag-return models have had
difficulty estimating this tag-return rate and have likely generated
highly biased variance estimates. Hierarchical Bayesian methods may
allow for models that may make better estimates of tag-return reporting rates
while generating more accurate variance estimates. An overview of
hierarchical Bayes will proceed an investigation of random effects and
changepoint models using Brownie model formulation.