Abstract
The American lobster, Homarus americanus, is the most important fishery in the north‐east United States. Over the last 20 years, the Collie‐Sissenwine (catch‐survey) model has been used in estimating fishing mortality in the fishery. This is then compared with a biological reference point derived from an egg‐per‐recruitment model for determining the status of the stock. However, the complexity of the fishery and population biology for American lobster, data limitations, and uncertainty about underlying parameters call for the development of stock assessment models with more biological detail. Complex and simple modelling approaches are complementary and both are likely to remain useful for such an important and complex fishery. The objective of this study is to develop a size‐structured stock assessment model for the American lobster. The proposed stock assessment model includes a set of size‐structured, seasonal, and sex‐specific population dynamics models and a statistical approach incorporating data of different sources. Using a fishery simulation based on the information on the lobster fishery in the Gulf of Maine, we evaluated the performance of the proposed model in the presence of biased errors in various model variables. This study suggests that the proposed stock assessment model performs well in retrieving the population dynamics of the simulated lobster and is rather robust to the biased errors in growth, fishing selectivity, and landings. Future studies need to focus on the evaluation of the model performance under different scenarios with both biased and random errors.