Stochastic virtual population analysis

Abstract

The Metropolis-Hastings algorithm was applied to a standard Virtual Population Analysis using a simulated data set containing catch data from 10 age groups over a 10-year period. Posterior distributions of fishing mortalities and population sizes are studied under different prior assumptions. The results demonstrate the usefulness of obtaining distributions of plausible parameter values instead of point estimates. Computational aspects of the algorithm are also discussed, as well as the estimation of the natural mortality.