Evaluation of the skill of length-based indicators to identify stock status and trends

Abstract

Abstract In data-poor situations, length-based indicators (LBIs) and reference points based on life history parameters have been proposed to classify stocks according to conservation status and yield optimization. Given the variety of potential LBIs, life history traits, and fisheries, it is necessary to evaluate the robustness of length-based advice to ensure that despite uncertainty that management objectives will still be met. Therefore, a simulation procedure was employed where an Operating Model conditioned on life history parameters was used to generate pseudo data. Receiver operator characteristics and the true skill score were then used to screen LBIs based on their ability to identify overfishing and recovery. It was found that LBIs performed better for long-lived species with low individual growth rates, those aimed at ensuring the conservation of mature fish performed better than those aimed at the conservation of immature fish, are better at indicating trends than at quantifying exploitation level, and in general were robust to uncertainty about dynamic processes.