Benchmark workshop on selected haddock and saithe stocks (WKBGAD)

WKBGAD 2024 met to revise the assessments for Northeast Arctic (NEA) saithe (pok.27.1-2), North Sea saithe (pok.27.3a46), and Rockall haddock (had.27.6b). In addition, the meeting was to consider how ICES could provide advice on catches of saithe west of Ireland (ICES areas 7 through 10). For all three stocks with existing category one assessments the data, life history parameters, and internal state space assessment model (SAM) settings were evaluated. The final models produced assessments with low to moderate retrospectives and patterns in the residuals, and all three were therefore evaluated as providing a valid basis for ICES advice.

The approved NEA saithe model was based on refining the existing SAM assessment model. Key improvements included moving to survey-based stock weights and revising the maturation function. For the reference points, the decision was made to use an MSYB_trigger (set to 226 430 tonnes) rather than the previous B_pa. The stock has been fished mostly at or below F_MSY for several decades, and previous rapid changes in stock size would mean that breakpoints in the HCR set at B_pa would not have time to react to prevent the stock from falling below B_lim. F_MSY was estimated to be 0.315. This stock uses a harvest control rule (HCR), based on the ICES advice rule (AR), but with the inclusion of banking and borrowing and a stability constraint on the catches. Analysis concluded that the existing HCR (with control parameters at or below F_target = 0.32 and MSYB_trigger at or above B_pa = 131 492 tonnes) is precautionary and can we recommend that the existing HCR (but using the new MSYB_trigger = 226 430 tonnes) remains the basis of ICES advice until the Managing Party were to decide to request a revision.

The North Sea saithe was also based on an existing SAM model. Key discussions here revolved around the length of the times series to use for recruitment and different life history parameters (including Lorenzen mortality and a time-varying maturation function), given the competing needs of long time series and known changes in North Sea productivity. The assessment now uses survey-based stock weights, and the models behind the survey and commercial CPUE indices were also revised. Reference points were set, using a B_pa at 1.4 times (B_lim = 130 090 tonnes, B_pa = 180 770 tonnes), B_lim in the absence of information to set an MSYB_trigger, and F_MSY of 0.316. This stock has been declining, and in the absence of good incoming recruitment could be expected to be heading towards B_lim in the medium term.

For Rockall haddock, a new SAM model was constructed. In contrast to the NEA saithe stocks, Lorenzen natural mortality was used. The haddock model (and fishery) starts at a younger age than the two saithe stocks, and therefore it was felt that the extra realism of higher mortality on the younger ages was required. The model includes using externally estimated CVs to set weights for individual years of data within the model-based survey (1999–present) to account for variable quality of the data over time. Density-dependent catchability was included, to account for differences in several high recruitment peaks. The key uncertainty around this stock is that the highest observed recruitment is only now entering the fishery. The benchmark therefore recommends that model performance be monitored closely and adjusted if necessary as the dynamics of this year class become better understood. In particular, the method for estimating mean weight in the short-term forecast should be chosen at the assessment working group meeting. Reference points were estimated as B_lim = 8542 tonnes, MSYB_trigger = 12 877 tonnes, F_MSY = 0.28.

The benchmark felt the process of dealing with the catches west of Ireland was not ideal, and this should have gone through the ICES Stock Identification Methods Work Group (SIMWG) before ever coming to benchmark. In the absence of a judgement from the SIMWG, the benchmark found it very difficult to decide what to do with the catches in areas 7–10, which are not currently assigned to any existing ICES stock. Had there been an ICES existing stock then there would not have been sufficient evidence to change the existing structure. There is weak, though somewhat inconclusive, evidence that the population west of Ireland is derived from the North Sea stock, but it is unclear if the west of Ireland is a sink population or if it contributes to the spawning of the overall stock. In the absence of any current stock structure and advice, the benchmark evaluated two options. One using a category 5 approach (where there is evidence to support advice of status quo catch) and one using the approach of including the catches into the larger North Sea area (in which case ICES would lack evidence on how to split advice between areas). Both of these options were considered viable options for ICES management advice, with weak evidence to decide between them. We therefore recommend that SIMWG (who have more expertise in stock structure than the benchmark) should be making this decision. Until such a judgement is made, the ICES advice should be based on a management unit using category 5 advice with no precautionary buffer as described in this report. The benchmark would stress that there is no judgment made that 7–10 is biologically distinct stock, this would be purely management advice. It would make sense for this stock to be included in WGNSSK given the evidence that the stocks do have links.