Skip to main content
SLU:s publikationsdatabas (SLUpub)

Sammanfattning

The benchmark workshop on sea bass (Dicentrarchus labrax) (WKBSEABASS) was conducted over an extended period, spanning from 2023 to 2025, to agree on the stock identities, stock structure, and assessment methodologies in future assessments of sea bass in the North Sea, Northeast Atlantic, and the Bay of Biscay. The benchmark had a strong focus on the investigation of spatial boundaries of the population within the area and the migratory behaviour and mixing between units.

As part of this benchmark process, the evidence on stock identity for sea bass was reviewed in a dedicated stock ID workshop, and plausible stock structure scenarios for the ICES sea bass benchmark were proposed. This generated several potential stock identities that included: a single stock covering the whole area; the current Northern (ICES divisions 4.b–c, 7.a, and 7.d–h) and Biscay (ICES divisions 8.a–b) stocks with mixing; and three equally plausible scenarios with three subpopulations and mixing.

At the data workshop, it was identified that there was insufficient data to model the Irish Sea (7.a) separately, and catch was low in this area, so it was merged with the Celtic Sea for modelling. In addition, there were insufficient resources to test all scenarios, so two were chosen: the current Northern and Biscay stocks with exchange of catches; and a single meta-population model containing North Sea (4.b–c, 7.d), Celtic and Irish Seas (7.a, 7.e–h), and Biscay (8.a–b) subpopulations with mixing. At the modelling workshop, it was decided to start with the simpler of the two scenarios (i.e. Northern and Biscay stocks with exchange of catches due to the complexity of the full mixing model.

As a result, only scenario A, Northern and Biscay stocks with exchange of catches, was taken forward to the benchmark.

The data workshop was conducted from 29 January to 2 February 2024 in Copenhagen, followed by intersessional work in groups to address and produce the missing data identified at the workshop. Catch numbers at length, natural mortality-at-age, recreational catches and survey indices were recalculated based on the latest insights and tools provided and additional data becoming available. Key to this process was also the advanced understanding and estimation processes of the recreational fishery data, while tagging data provided a quantity for the mixing of the stocks.

For both regions, new SS3 models were created, with a proportion of the catches, according to data from tagging studies, being reallocated from the Northern to the Southern stock. Baseline models for both areas were created, which had to pass a set of diagnostics, including diagnostics on catch and survey residuals, parameter uncertainty estimates, trends in spawning-stock biomass (SSB), Fishing mortality (F), Recruitment (R) and retrospective analyses, set out at the beginning of the benchmark workshop. Setting out a set of diagnostics before the benchmark will considerably improve the transparency and comparability of such a process and improve the flow of the workshop. Discussion focused on some poor fits to specific fleets, which is outlined under the individual headings of the stocks. Once the base models had been finalized, each of them was assigned several sensitivity tests and alternative scenarios to find the best model to describe the state of the stock:

Northern stock

  • Different time-series of recreational fisheries catches.
  • Turn off the extra standard deviation for juvenile surveys.
  • Discard as retention function for OTB and SEINE as catch fleet.
  • Mirror OTB discard fleet to OTB landings fleet and remove discard length information.
  • Recreational fleet as landed and discarded fraction.
  • Use of Lorenzen M.
  • M profiling.
  • Steepness profiling.
  • Associate ALK to a different fleet than OTB.
  • Estimate all growth parameters.
  • Assume no movement of fish from the Southern to the Northern stock.
  • Leave one out (surveys).


Southern stock

  • Use of Lorenzen M.
  • M profiling.
  • Steepness profiling.
  • Estimate all growth parameters setting A for L2 at 16 years old.
  • Assume no movement of fish from the Southern to the Northern stock.
  • Leave one out (surveys).
  • Random walk for GTR and LL instead than blocks.
  • Switch off extra standard deviation for LPUE.
  • Switch off the random walk for GNS and substitute with blocks.
  • Decrease CV of the last 2 years of the OTB discards.
  • Associate ALK to a different fleet than OTB.


All scenarios were then tested with a script comparing the previously created diagnostics to find the models with the best fit. Some of the scenarios were purely done for the sensitivity of the model (such as the scenario not to transfer the catches from the northern model to the southern model to account for mixing). In contrast, others focused on the selectivities of the various time-series, the setting of natural mortality, etc. It was agreed to treat some of the dataseries the same way in both models, such as the re-calculation of the recreational data back in time; both models were successfully completed, and the reference points were estimated.

Publicerad i

ICES scientific reports
2025, nummer: 7:3
Utgivare: International Council for the Exploration of the Sea

SLU författare

UKÄ forskningsämne

Vilt- och fiskeförvaltning

Publikationens identifierare

  • DOI: https://doi.org/10.17895/ices.pub.25002488

Permanent länk till denna sida (URI)

https://res.slu.se/id/publ/143398