A Bayesian Age-structured State-space Model for the Torres Strait Rock Lobster ( Panulirus Ornatus ) Fishery, Australia

摘要

The ornate rock lobster ( Panulirus ornatus ) in Torres Strait supports locally important artisanal and commercial fisheries in both Australia and Papua New Guinea. In this paper we developed a new age-structured population model for the lobster fishery, cast the model into the framework of dynamic state-space modelling of modern time series methodology and used a Bayesian approach to produce posterior distributions of the model parameters. A stock recruitment relationship was incorporated into the age-structured model and its parameters were fitted together with other model parameters. The mean estimates of natural mortality rate (M) range from 0.47 to 0.77 year-1. Using the mode values of natural mortality rate and the two parameters defining the stock recruitment relationship, the maximum sustainable yield was estimated to be around 265 t (tail weight) with a corresponding fishing mortality rate (F) 0.4 year-1. The estimated fishing mortality rate exceeded this level in 1994, 1996, 1997 and 1999. To consider the natural variation in these parameters and the uncertainty involved in their estimates, Monte Carlo simulation was used to quantify the uncertainty in the relationship between sustainable yield and fishing mortality. The results show that the 50% credible interval of the maximum sustainable yield ranges from 110 to 390 t and its corresponding F ranges from 0.2 to 0.5 year-1. State-space models can take into account both measurement and process errors and the Markov chain Monte Carlo techniques produce posterior probability density functions for all the parameters of interest. All these should provide more information about the uncertainty involved in stock assessment and improve risk management in fisheries.