Incorporating an Otolith-Derived Environmental Index into Growth for Stock Assessment Model

Growth is one of the key demographic processes for population dynamics. There is a growing body of literature showing that growth in fish is plastic over time, and affected by climate variability. However, time-varying indices of growth variation are not often incorporated into stock assessment models, on which fishery management advice is based, and growth is often estimated with time-invariant parameters or incorporating annual variation without an explicit growth model. This is due to a lack of a general framework for deciding when and whether to incorporate indices of time-varying individual growth into an assessment model, and of a general understanding of its mechanistic drivers. In this thesis, I developed a framework to evaluate the suitability of incorporating a time-varying growth index into a stock assessment model, and use risk analysis to evaluate its management-related advantages and shortcomings. I then developed a nonlinear mixed-effects model that estimates a time-varying growth index from otolith increment data, while also allowing for random individual effects. Including an accurate time-varying growth index generally is expected to improve estimates of spawning stock biomass and recruitment in a stock assessment, while a relatively small number of otoliths is required to attain such a level of accuracy when estimating the random year effects. The simulation framework described in this thesis can be applied across multiple stocks, to further our understanding of the effects of modelling time-varying growth on population dynamics.