Report of the Workshop on integrated multispecies advice for Baltic fisheries (WKMULTBAL)

The results of four models for cod, sprat and herring were examined (Kiel Model, BALMAR, stock production models and Stochastic Multispecies Model (SMS)). SMS results were scrutinised in more detail as they were considered for use in management advice. The distribution of the three species in the Baltic has changed in recent years and the estimation of the trade-offs between species at different exploitation rates should take the overlap between the species into account. The present distribu-tion pattern implies that increasing F on cod will not necessarily result in increased Baltic wide clupeid stock sizes, and conversely a decrease in F on cod will not neces-sarily result in a decrease of the Baltic clupeid stock sizes. A reduction of clupeid F in Subdivision 25 will likely improve growth of cod and reduce cod cannibalism. An increase in clupeid F in northern areas (SDs 27–32) will likely not have a negative effect on cod. A higher F on clupeids in northern areas might reduce density-dependence and improve growth of the clupeid stocks. The ICES/HELCOM Working Group on Integrated Assessments of the Baltic Sea (WGIAB) evaluated various single species and models using an ensemble approach. This showed the importance of evaluating (1) alternative species interactions set-ups, (2) multiple stock–recruitment relationships, and (3) multiple possible future climate realizations, when evaluating alternative management strategies and deriving reference points.
There are indications that sprat biomass above 1.5 million tonnes could have negative effects on cod and herring. The possibility of establishing an upper limit on sprat biomass was discussed, but there were concerns with the approach if assessment uncertainties are large.
To identify potential reference points and candidates for target Fs and trigger points, SMS forecasts scenarios were made using an array of target Fs, for the three species, cod, herring and sprat.  FMSY estimates for cod is 0.60, for herring 0.24 and for sprat 0.42. These values were obtained assuming that cod predation affects the whole stocks of clupeid (i.e. does not take into account the limited spatial overlap between cod and clupeids, and that cod growth or recruitment is not affected by the amount of clupeid as food. They should therefore be considered as preliminary. As indicated above, they have not been derived testing for alternative species interactions and multiple recruitment relationships, and they do not take into account important eco-logical processes such as density-dependent growth and fecundity and changed spatial overlap between stocks.
Multispecies simulations to determine reference points accounting for intra- and interspecific interactions depend, as in single species simulations, on the assumptions about future recruitment. The relationships applied for cod in the SMS simulations implicitly assume that the eastern spawning areas in the Gdańsk Deep and the Got-land Basin will not contribute to reproductive success. This will restrict the values of SSB and yield in the model outputs to values far below those observed in the 1970s and early 1980s. However, the reference points for the management of the Eastern Baltic cod stock and its prey species, derived by simulations with this type of stock–recruitment relationship, will have to be revised in a situation of a series of major inflows within a few years, which would change the carrying capacity of the system for cod, and potentially also for sprat and herring.