Order-Restricted Bayesian Estimation of the Age Composition of a Population of Atlantic Cod

• DOI: 10.2307/2291447
• Publication Date: 2006-05-10T13:18:46Z
• Authors: Balgobin Nandram; J. Sedransk; Stephen J. Smith
• Keywords: 0504 sociology; 05 social sciences; 14. Life underwater; 0101 mathematics; 01 natural sciences

Abstract The management of a commercial fishery often requires estimates the age composition. These are typically based on and length data obtained from sampling landings catches research vessel survey. We use annual cruises Georges Bank conducted by Canadian Department Fisheries Oceans to show how inferences for Atlantic cod (Gadus morhua) can be improved. Traditionally, two-phase stratified is used with fish (or weight) as stratification variable. Letting Pi . denote proportion belonging stratum i, π ij class j in we Bayesian methods estimate P.j = i πij , that j. Specifications smoothness, expressed unimodal order relations among (within between strata), incorporated into prior distributions. Uncertainty about both locations modes unimodality itself included part probabilistic specification. With computations facilitated using Gibbs sampler, smoothness conditions provide very large gains precision. For analyzed this article, one obtain similar precision (a) conventional analysis or (b) an restrictions sample half size (a). also better composition improved quantities fisheries managers forecast cohorts fish. Key Words: managementGibbs samplerMultinomialPrior specificationStructural zerosUnimodal