On the use of tagging data in statistical multispecies multi-area models of marine populations

Abstract

The use of multispecies models of marine stocks along with recognition of the importance of temporal differences in spatial overlap has resulted in migration rates playing an increasingly important role in models of fish stocks. Moreover, traditional estimates of growth based on samples from fishing gear are confounded with the selection pattern, which is exacerbated when multispecies issues are considered. For these and other reasons, there is a need to include explicit tagging data as a component of fisheries models. A statistical multispecies multi-area framework has been extended to predict tag returns and subsequently to incorporate tagging data in likelihood components to be used when estimating migration rates. The information content of such data is not clear a priori, but simulations indicate the point estimates to be quite reasonable. A bootstrap method is proposed, based on bootstrapping entire tagging experiments (rather than individual fish). The resulting bootstrapped uncertainty estimates are generally applicable and are found to be close to the true values in the simulated examples. © 2008 International Council for the Exploration of the Sea. Published by Oxford Journals. All rights reserved.