The model was fit to the data by minimizing an objective function (O.F.V.) that is the sum of the negative log likelihoods for the CPUE series (), the tagging data () and length compositions in the catches (). We used lognormal error structures for the CPUE time series, a Poisson error structure for the tagging data and a robust normal error structure (Fournier et al. 1990) for the proportions at length in the catch. For each fishery, the log-likelihood for the CPUE component of the model is:
), the tagging data (
) and length compositions in the catches (
). We used lognormal error structures