Simulating long-term dynamics of the coupled North Sea and Baltic Sea ecosystem with ECOSMO II: Model description and validation

• With ECOSMO II a consistent 3d biophysical NPZD model for the coupled North Sea and Baltic Sea system is presented.
• Model results are thoroughly validated against available observational data sets.
• Model performance on seasonal, inter-annual and up to decadal time scales could be demonstrated.
• A consistent multi-decadal simulation for lower trophic level ecosystem dynamics in the coupled ecosystem was analyzed.
• Ecological regime shifts are identified for North Sea and Baltic Sea, but with differences in timing and magnitude.

The North Sea and the Baltic Sea ecosystems differ substantially in both hydrology and biogeochemical processes. Nonetheless, both systems are closely linked to each other and a coupled modeling approach is indispensable when aiming to simulate and understand long-term ecosystem dynamics in both seas. In this study, we present first an updated version of the fully coupled bio-physical model ECOSMO, a 3d hydrodynamic and a N(utrient)P(hytoplankton)Z(ooplankton)D(etritus) model, which is now adopted to the coupled system North Sea–Baltic Sea. To make the model applicable to both ecosystems, processes relevant for the Baltic Sea (e.g. sedimentation, cyanobacteria) were incorporated into the model formulation. Secondly we assess the validity of the model to describe seasonal, inter-annual and decadal variations in both seas. Our analyses show that the model sufficiently represents the spatial and temporal dynamics in both ecosystems but with some uncertainties in the coastal areas of the North Sea, likely related to the missing representation of tidal flats in the model, and in the deep-water nutrient pool of the Baltic Sea. Finally we present results from a 61-year (1948–2008) hindcast of the coupled North Sea and Baltic Sea ecosystem and identify long-term changes in primary and secondary production. The simulated long-term dynamics of primary and secondary production could be corroborated by observations from available literature and shows a general increase in the last three decades of the simulation when compared to the first 30 years. Regime shifts could be identified for both ecosystems, but with differences in both, timing and magnitude of the related change.