Summary: | Many fisheries in the world are data-moderate, with data types (e.g., total removals, abundance indices, and biological composition data) of varied quality (e.g., limited time series or representative samples) or available data. Integrated stock assessments are useful tools for data-moderate fisheries as they can include all available information, can be updated due to the availability of more information over time, and can directly test the inclusion and exclusion of specific data types. This study uses the simulation testing and systematic data reduction from the US West Coast benchmark assessments to examine the performance of Stock Synthesis with catch and length (SS-CL) compositions only. The simulation testing of various life histories, recruitment variabilities, and data availability scenarios found that the correctly specified SS-CL can estimate unbiased key population quantities such as stock status with as little as 1 year of length data although 5 years or more may be more reliable. The error in key population quantities is decreased with an increase in years and the sample size of length data. The removal of the length compositions from benchmark assessments often caused large model deviations in the outputs compared to the removal of other data sources, indicating the importance of length data in integrated models. Models with catch and length data, excluding abundance indices and age composition, generally provided informative estimates of the stock status relative to the reference model, with most data scenarios falling within the CIs of the reference model. The results of simulation analysis and systematic data reduction indicated that SS-CL is potentially viable for data-moderate assessments in the USA, thus reducing precautionary buffers on catch limits for many stocks previously assessed in a lower tier using catch-only models. SS-CL could also be applied to many stocks around the world, maximizing the use of data available via the well tested, multifeature benefits of SS.
|