A new Bayesian catch curve stock assessment model for the analysis of length data from multi-gear fisheries

Abstract A new Bayesian length-based catch curve stock assessment model is described that estimates mortality-at-length using length composition data grouped into gear categories, which can be used to calculate spawning potential ratio and yield per recruit indicators. The method will be particularly useful for data-limited, multigear fisheries where management advice is sought not only on stock status, but also on technical measures that affect selectivity. The model assumes that the population is in a steady state, length-at-age has a mean defined by the von Bertalanffy growth model with growth variation described by gamma distributed asymptotic lengths (${{L}_\infty }$), and that mortality is constant within each length interval. Any selectivity model could be used to describe the length specific fishing mortality, but it is suggested to use the logistic, normal, and double-sided normal selectivity functions or minimal mixtures of these, which should cover even quite complex situations. The model has been implemented in an R Package (fishblicc), which was examined using simulations and length-frequency data from the Indian Ocean yellowfin tuna multi-gear fishery.