Using Simulation Results And Bayesian Belief Networks To Assist Fisheries Planning: A Case Study Of The Patagonian Toothfish Fishery At Macquarie Island

No abstracts are to be cited without prior reference to the author.

In this paper we consider two key areas of fisheries management. Firstly, we demonstrate how knowledge related to biological hypotheses and observations can be brought into fisheries management modelling in a relatively simple way. Secondly, we show how value-of-information and value-of-control can be used when planning fisheries management systems. We apply Bayesian belief network methodology to the fishery for Patagonian toothfish at Macquarie Island, Australia. In this fishery, fishers’ observations, as well as the scientific data, suggest that large-sized transient fish, increasing the total catches and CPUE values in a remarkable but also unpredictable way, might occasionally occupy the area. We use a model to simulate the daily fishing decisions of the fishers. We show that the reaction of the fishers to the variability in daily income includes a strong self-regulatory mechanism, which serves to protect the stock. The probabilistic output results of this system are further analysed in a belief network model, including biological observations from genetics, tagging experiments and fatty acid compositions. We demonstrate that in addition to catch limits as management tools, economic tools focusing on the behaviour of the fishermen can be effective in fisheries planning, especially in increasing the robustness of management systems. Moreover, changes within the utility function can reveal potential conflicts between various interest groups.