PARENT SESSION
Contributed Oral Session 27: Marine Populations Under Harvest
Monday, August 8, 1:30 PM - 5:00 PM, Meeting Room 519 A, Level 5, Palais des congrès de Montréal

A Bayesian framework for modelling spatio-temporal oceanic fisheries data.

Edwards, Andrew^*,1, Myers, Ransom¹, ¹ Dalhousie University, Halifax, Nova Scotia, Canada

ABSTRACT- Several patterns emerge from catch-per-unit-effort data across broad spatio-temporal scales. Large long-lived organisms that are caught as bycatch, such as oceanic white-tipped shark, decline to extremely low population levels, while targetted species also suffer population declines, but not so dramatically. Populations of other species may first increase before declining (e.g. Atlantic sailfish), whereas organisms such as stingrays appear to enjoy a population explosion. We construct models to test the hypothesis that such dramatic changes are a direct consequence of reduced predation pressure from the species that declined. We show how to implement data using nonlinear, state-space models in a Bayesian framework. We utilise Markov Chain Monte Carlo (MCMC) methods, so that we can incorporate both process uncertainty and observation error. We use a meta-analytic approach, which allows us to simulataneously include data for multiple species and across multiple scales, and lets us address questions concerning the biodiversity consequences of fishing in the ocean.

Key words: Bayesian, MCMC (Markov Chain Monte Carlo) methods, state-space models, oceanic fish populations