Assessment of Model Uncertainty in the Prediction of Seasonal and Annual Variability in the Black Sea.

A new generation (circulation, biogeochemical and a higher trophic level) model for the Black Sea was developed to provide an integrated analysis of ecosystem attributes that contributes to policy oriented criteria’s of the European Union's ambitious Marine Strategy Framework Directive towards achieving Good Environmental Status by 2020. The biogeochemical model, including the carbonate module, comprises thirty state variables. The higher trophic level model includes thirteen fish species. The model system was used to carry out hindcast analyses for 1980-2013 period and forecast analyses for the 2013-2020 time period assuming different scenarios of anthropogenic pressure acting on the Black Sea. The results of hindcast simulations indicate that the Black Sea model system demonstrates a range of skill, with physical variables having the most skill followed by chemical variables. The HTL model has more skill for small pelagic fish (e.g. Anchovy) than larger pelagic and demersal fish. We show that capturing mesoscale variability is of great importance in such models and directly influences model skill to capture seasonal and regional variability of phytoplankton blooms, specifically the skill to correctly resolve bloom timing and duration.