SEAwise Report on improved predictive models of recruitment under different environmental scenarios

This report investigates how key environmental variables influence the recruitment process of target fish stocks. Understanding how the environment affects recruitment may allow more accurate predictions of fish stock dynamics under scenarios of environmental change and in particular their response to global warming, supporting the development and implementation of robust management policies. Case studies from the four Seawise case study regions have been analysed, and the main results obtained so far are summarized below. In the Baltic Sea, the Gulf of Riga spring spawning herring showed effects of spawning stock biomass on individual weight of age-1 fish, with prey abundance in May and previous year feeding period temperature also playing significant roles. The explanatory power of the final model was moderate. Higher weight of herring is achieved at higher prey densities, lower SSB levels and lower temperatures during the main feeding season of age-0 fish. Recruitment of Western Baltic cod and herring showed decreasing reproductive potential at increasing temperature. In the North Sea, the effects of temperature, salinity, currents, chlorophyll and zooplankton on recruitment of cod, haddock, saithe, whiting, plaice, sole, sprat and herring were investigated using a semi-automated, machine learning framework. The incorporation of environmental signals in recruitment predictions showed improved predictions over a stock recruitment model without environmental effects in six out of the eight stocks. For small pelagic foirage fish, four stock-recruitment models were fitted for three sandeel stocks and the North Sea sprat stock Linear regressions revealed various relationships between recruitment and environmental variables. Short-term time scales based on monthly averages produced a noisier and less consistent pattern for most stock. In the Western Waters, Bayesian online change point detection models were applied to thr