Benchmark workshop on sandeel (ammodytes spp.) (Outputs from 2022 and 2023 meetings) (WKSANDEEL)

The Benchmark Workshop on Sandeel (Ammodytes spp.) (WKSandeel) was conducted over an extended period, spanning 2021-2023, to agree the assessment methodology to be used in future update assessments of sandeel in areas 1–4 (Figure 1). The benchmark had a strong focus on investigating spatial stock boundaries within the North Sea, bringing catch and survey data compilation and standardization up to standard with the latest insights in ICES and updating the assessments accordingly.

The data compilation workshop was preceded by several informal sessions to present the latest findings from an extensive study on stock structure identification, including tagging, otolith micro-chemistry, drift modelling and genetics research. The workshop proposed two alternative stock structure delineation scenarios as presented in Figure 2.

The different research projects provided contrasting views on appropriate stock structure delineation, with evidence for residence within the delineated sandeel areas but also migration across the a and b borders as denoted in Figure 2. The group reached consensus that, as a pragmatic way forward, only scenario WKSAND2022a would be put forward for evaluation in the assessments, together with the existing WKSAND2016 delineation as a baseline.

Catch numbers at age, fishing effort, maturity at age, weight at age, natural mortality at age and survey indices were recalculated based on the latest insights and tools provided through the ICES RDBES and statistical standardization. Key to this process was also the re-definition of sandeel banks, which are used directly to derive standardized survey indices for the different assessments (Figure 3).

For all four sandeel areas, updated assessments were prepared for the WKSAND2022a and WKSAND2016 (baseline) area delineation scenarios. There is no change in area delineation for area 4 under both scenarios.

Diagnostics on catch and survey residuals, parameter uncertainty estimates, trends in Spawning Stock Biomass (SSB), Fishing mortality (F), Recruitment (R) and retrospective analyses were prepared and evaluated by the participants. The discussions focused on introducing or removing changes in selectivity to account for changes in the fishing fleet operation including effort creep, as well as on retrospective performance and uncertainty of the estimated SSB, F and R outputs.

Assessments for areas 1, 2 and 4 were associated with appropriate diagnostics under both area delineation scenarios while for area 3 the WKSAND2022a scenario showed a large increase in uncertainty, not previously demonstrated for the area 3 stock under scenario WKSAND2016 (baseline). The change in area delineation, resulting in the inclusion of catch and survey stations in the Skagerrak/Kattegat area, likely contributed to a mismatch in cohort signal between catch and survey data giving rise to an increase in uncertainty in the assessment outputs. The WKSAND2022a scenario was therefore considered to be inappropriate for advice purposes and led to the conclusion that the WKSAND2016 baseline scenario would need to be applied for area 3. Owing to this decision, the other sandeel area delineations had to align leading to no update in sandeel delineation from the 2016 benchmark meeting. The group reflected on this decision indicating that the 2016 area delineation was based on management considerations rather than best available biological information on stock structure but did not see room to reconcile biological understanding with fit for purpose stock assessment and management advice. Although the area delineation remains the same, improvements in catch data preparation, survey data standardization, biological parameter improvements and statistical modelling of these data in the assessment all improved compared to the previously agreed assessment methodology, and hereby resulting in lower retrospective patterns in especially recruitment which were previously found to be too large. These reductions in retrospective patterns should ensure more robust management.

Biomass reference points for all sandeel areas were defined based on spasmodic (Type 1) stocks with occasional large year classes. A Management Strategy Evaluation (MSE) framework was then used to search for the maximum value of F that increases baseline risk by no more than 5% when used in an escapement strategy (F_cap). Given spiky and, for some sandeel areas, declining risk trajectories over time, the reference period for calculations was defined based on generation length. The MSE for sandeel area 4 produced inconsistent results, suggesting a high value of F_cap despite a relatively high baseline risk. Given a need for precaution, the F_cap for sandeel area 4 was set equal to that for sandeel area 1, the lowest of the other three sandeel areas. Sensitivity analyses showed small increases to risk when including banking and borrowing and a monitoring TAC of 5000 tonnes.