Inter-benchmark Workshop on West of Scotland Cod (6.a) (IBPCod6.a)

The Inter-benchmark Process for West of Scotland Cod in Division 6.a (IBPCod.6.a 2019), chaired by Poul Degnbol (Denmark) took place by correspondence during seven WebEx meetings spread over several weeks (18 February–29 March 2019). Participants included scientists from the UK and Ireland, and external reviewers from Denmark.

The main focus of the IBP was to remedy the poor diagnostics in the TSA assessment and to identify the reasons for differences between the 2018 WG TSA assessment and an alternative peer-reviewed assessment which shows markedly different trends to the 2018 WG assessment, modifying the TSA configuration as necessary.

This IBP investigated the possible reasons for this discrepancy including exploring the sensitivity of the TSA assessment to selectivity parameter assumptions, weighting of different data sources and inclusion of another surveys in the assessment. The IBP has agreed on that configuration of the TSA assessment which in the opinion of the group, within the options in the TSA assessment, best reflects the developments which may have taken place in the stock and the fisheries.

Reference points have been updated on the basis of this final agreed assessment. The revised TSA shows a significant improvement in the model diagnostics: the tendency of the model to over-estimate terminal F (as apparent in the 2018 WG assessment retrospective) has been resolved and the heterogeneity observed in model residuals is now also much less of an issue.

However, during the work it became clear that there were some issues which could either not be addressed or not be resolved in the current process. The main issue which could not be addressed was the possibility for other assessment models or the use of multiple models as basis for the assessment. This limitation was a given in the ToRs and would also not have been practically possible within a brief IBP by correspondence as the present.

It appears that this stock assessment may be particularly sensitive to model bias or model error as different models produce quite different perspectives, an issue which does not occur when the same models are applied to similar stocks in the region or neighbouring regions. As a result there was an extensive discussion in the IBP regarding the need for independent evidence to support various interpretations which may emerge from different models.

The question of process uncertainty (along with other unresolved issues) is discussed in the report in the context of future benchmark planning.