A Bayesian analysis of NSW eastern king prawn stocks (Melicertus plebejus) using multiple model structures

The eastern king prawn (Melicertus plebejus) is a valuable target species for commercial fisheries operating on the Australian east coast. The ayesian analysis presented here aims to determine the current state and productivity of the NSW component of the eastern king prawn stock nd analyse the possible consequences of altering commercial catches in the future. The Bayesian approach is well suited to both these aims, articularly given the significant uncertainty about the true population dynamics of the stock, and the multiple sources of information available. he sampling/importance resampling method was applied as it is numerically robust and straightforward to implement. Various types of uncertainty ere incorporated into this analysis including: process and observation error, uncertainty in model structure, and uncertainty associated with the arameter values (captured with prior probability distribution functions). A delay–difference model was used with four different representations f recruitment. Each of the four models examined provided differing results for stock depletion since 1984/1985. Despite this uncertainty, none of he models suggested that the stock has been heavily depleted since 1984/1985. The analysis also identifies 2003/2004 as a particularly poor year or production (as was 1984/1985) but that such events lie within the limits of historically observed variability. Projections of the modelled stock ynamics into future years indicate that the stock does not appear to be at high-risk in the near future. Finally, the results of the decision analysis uggest that significant changes in the future catch are not expected to have a large impact on catch rates or the stock depletion ratio. These results, owever, are dependent upon the assumption of continued and robust recruitment from Queensland. 2006 Elsevier B.V. All rights reserved.