Improving age-based fish stock assessments, with an application to American Plaice on the Grand Bank of Newfoundland

Abstract

Age-based fish stock assessment models use available data to provide fisheries managers with estimates of population processes and to determine sustainable harvest rates. However, the data to inform these models is often complex and ignoring or oversimplifying these complexities can result in unsustainable or sub-optimal harvesting advice. This thesis improves age-based stock assessment models by accounting for important variability in the data. I first simulation tested nine methods that aim to account for the commonly used length-stratified age sampling design when estimating growth parameters. My results showed that commonly used methods had poor accuracy and the empirical proportion approach was optimal. Secondly, I developed a state-space stock assessment model for American plaice that allowed for errors in the underlying population processes and provided improvements in the retrospective plots. This research improved our understanding of American plaice population dynamics and growth and is a step forward in fitting more realistic integrated age-based stock assessment models.