Working Group on Multispecies Assessment Methods (WGSAM)

The Working Group on Multispecies Assessment Methods (WGSAM) works on the development of multispecies and ecosystem models that have potential or actual application for supporting advice on management of fish stocks and fisheries and assessing the implications of multispecies interactions. This report describes key-runs (standardized model runs updated with recent data) of multispecies and ecosystem models for the Baltic and Irish Sea. Model key-runs are used in ICES advice processes, and WGSAM provides critical expert review of these key-runs to recom-mend appropriate use of results.

WGSAM first formalized a consistent set of review criteria to conduct key-run reviews (https://ices-eg.github.io/wg_WGSAM/ReviewCriteria.html), as these are increasingly requested. WGSAM then applied these review criteria to three potential key-runs: two for the Baltic Sea ecosystem and one for the Irish Sea ecosystem. Any difficulties with the review process were noted to further refine the review criteria and to make future key-run reviews more efficient and effective.

WGSAM notes that streamlining processes for updating data would facilitate future key-runs. From 2020, WGSAM will require draft key-run results and documentation 2–4 weeks before the meeting to allow more thorough review prior to the meeting. This will increase available meeting time for model comparisons, ensemble modelling, and coming to agreement on recommenda-tions.

For the Baltic Sea, multispecies model key-runs estimate predation mortality to provide time-series of natural mortality (M) for use in single species stock assessments for herring and sprat. The review of key-runs from an SMS model (used in the previous 2012 key-run) and a newly developed Gadget model demonstrated that both models provided consistent time-series of M for herring and sprat when using the same assumptions regarding residual natural mortality, despite different representations of cod population dynamics.

WGSAM recommends the use of natural mortality estimates from the Baltic SMS key-run for use in single species stock assessment models of Baltic herring and sprat. Due to issues of stability in the historical period the Gadget model was not selected as a key-run. WGSAM notes that the results of the SMS key-run depend on the outcome of the ICES Eastern Baltic cod assessment. Any bias in this assessment directly influences the predation mortality estimates. Uncertainty estimates for M from SMS should not be used as these assume that the cod population abundance is known without error.

For the Irish Sea, an an Ecopath with Ecosim (EwE) model was reviewedto suggest methods by which some of the outputs could be used to provide an indicator to help inform the choice of Ftarget within the pre-defined Fmsy ranges. This allows for the incorporation of ecosystem information when setting Ftarget for individual stocks, while remaining within the existing pre-cautionary fisheries management framework and reference point ranges used by ICES. An EwE model key-run was accepted as a basis for generating the indicator(s) but directly use of mod-elled Fmsy values in other models or for management was not recommended.