Providing scientific advice to fisheries managers is a risky activity. It is not uncommon that a model which was performing well suddenly fails to properly fit an additional year of data, or projections made in the past did not materialize when more recent information became available. The inclusion of previously unmodelled processes, changes in processes that are assumed constant, or conflicting information and data revisions, have the insidious tendency to ruin what had been a perfectly acceptable assessment fit, invalidating one's advice and weakening confidence in future advice efforts.
To deal with the potential lack of robustness in fisheries advice, WKENSEMBLE will explore the potential to expand the current assessment modelling basis to integrate multiple sources of uncertainty using ensemble models.
The workshop will focus on practicalities of providing advice based on model ensembles. To do this, the group will:
- expand on previous work about how to implement ensembles of stock assessment models
- consider how model ensembles could affect reference points
- investigate how scientific advice using ensemble models would look like.
This work builds on lessons learned from existing examples of ensemble based modelling in ICES, NOAA and RFMOs, and the conclusions and recommendations from the JRC Exploratory Workshop on Model Ensembles.
The workshop takes place on 11–15 May at ICES Headquarters in Copenhagen, Denmark. For registration, contact Lise Cronne at ICES Secretariat.