Analysis of Otolith Microchemistry Using Bayesian Hierarchical Mixture Models

• Activity Number: 674
• Type: Contributed
• Date/Time: Thursday, August 5, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Bethann Mangel Pflugeisen*+ and Catherine Calder and Elizabeth Marschall
• Companies: The Ohio State University and The Ohio State University and The Ohio State University
• Keywords: Ecology ; Mixture models ; Bayesian Statistics ; Fisheries ; Classificattion

Abstract

The field of fisheries research generally uses classical statistical classification methods to estimate the proportion of fish that return to natal spawning grounds to spawn. With the advent of otolith microchemical analysis, researchers are now able to extract information from fish ear stones (otoliths) about the chemical composition of water in which fish have spent distinct life history stages. Here we present a method of analysis set in the Bayesian framework, which enables explicit incorporation of habitat information into the analysis. The ecological system is seen as arising from a mixture of disparate fish populations and information from the biological relationships inherent in otolith formation is exploited through the hierarchical structure. We present the model and motivation, demonstrate the validity of the model through simulation studies, and conclude with the analysis of a data set from Lake Erie.