Extracting abundance indices from longline surveys: method to account for hook competition and unbaited hooks

To estimate fish population trends and abundance, fisheries scientists commonly apply dynamic population models fitted to relative abundance indices. Populations are often monitored using longline fishing gear and the most commonly used relative abundance index in this case is the catch per unit effort (CPUE), defined as the number of fish of the targeted species caught per hook and minute of soak time. Longline CPUE can be affected by interspecific competition and the retrieval of unbaited or empty hooks, and this can lead to biases in the apparent abundance trends. Interspecific competition has been previously studied but the return of empty hooks is ignored in all current treatments of longline CPUE. This work proposes and compares different stochastic models to define indices to address both issues simultaneously. Maximum likelihood estimators and their asymptotic covariance matrices are obtained. Simulating different joint scenarios for interspecific competition and the empty hooks, we show that CPUE behaves badly in every scenario. Information about the source of the empty hooks is required to select the appropriate identifiability constraint and therefore derive the appropriate abundance index.
The above methods are applied to build relative indices from 2003 to 2009 for quillback rockfish (Sebastes maliger) in British Columbia from longline survey data. Due to variation in the incidence of non- target species, the index trend obtained is moderately sensitive to the choice of the estimator. The proposed methodology permits the building of reliable abundance indices for all populations monitored using longline fishing gear.