Using production models with new and developing fisheries: A case study using the sea cucumber Parastichopus californicus in British Columbia, Canada

Data from four experimental sea cucumber (Parastichopus californicus) fishing areas in British Columbia are used to estimate productivity that has occurred during the experiments. The estimates are made by incorporating the data into a Bayesean model describing both productivity and survey-error. The Bayesean nature of the analyses enables the estimated productivity to be expressed probabilistically.The estimated productivities that have occurred in the data can serve as a guide for managing commercial sea cucumber fisheries for stocks with similar characteristics to those in the experiment. Since productivities are presented probabilistically, they can, under certain conditions, be used to generate a level of confidence that a particular harvest-management strategy is sustainable.Thus far, the applicability of the analyses is limited by the range of data that has occurred during the experiment. All the stocks began the experiment in a near-virgin state. The lowest level of abundance that has occurred so far is approximately 28% of the virgin-value. To estimate productivity for lower levels of abundance would rely on mathematical models to extrapolate the current results. As the experiments continue, it is expected that data for stocks that are more depleted will be collected and that the revised analyses will be applicable over a wider range of conditions.

► Measures latent productivity that has occurred for sea cucumbers (Parastichopus californicus) at four experimental fishing areas in British Columbia, Canada.
► Demonstrates a framework for implementing a productivity model that is flexible in the choice of model and in the timing of data.
► Presents a novel production model that is similar to, but more flexible than the commonly used Schaefer model.
► Develops a precautionary strategy for applying a productivity model where there is no data available that represents severely depleted stocks.
► Demonstrates a limited ability to make a probabilistic prediction of the response of a commercial stock to a harvest strategy.