Coastal ecosystems show variations on different time scales, but particularly pronounced variations are found on multi decadal time scales. Climate variations on a range of time scales mediate the hydrodynamic and biogeochemical processes responsible for nutrient supply to the euphotic zone and impact marine primary production, biogeochemical cycling and foodweb structure. Here we investigate these variations in the land-ocean transition using a coupled physical-biogeochemical model. We will present new, consistent and improved reconstructions of hydro- and ecosystem dynamics for the regional seas North Sea, Baltic Sea and Barents Sea by coupled physical-biological models and investigate long term variations in all three systems. The model reconstructions will be compared to a wide range of available observations and validated in great detail. Based on the model hindcasts and data we aim at resolving multi-decadal variations in hydrodynamic, sea ice and biogeochemical characteristics on the Barents Sea during the past 60 years. We would like to quantify the multi-decadal changes and answer the following questions: (i) Have the dynamic regimes in all three seas undergone changes in the past 60 years? (ii) Can we attribute the identified changes in primary production to changes in hydrodynamic conditions? (iii) Is there predictability inherent in the systems?
Using multivariate statistical analysis we have identified basic modes of variability in primary production in all three seas, which correlate with dynamic characteristics such as atmospheric wind changes and corresponding changes in ocean circulation. Moreover statistical relations to thermodynamic characteristics are identified for the different system. To identify, whether correlation also implies causality, we complement our statistical analysis with model sensitivity experiments and thereby identified the dominant impact of the wind field for anomalies in primary production over thermal effects and variations of nutrient supply from terrestrial sources. Finally we explore the predictability in the systems and found significant potential to forecast primary production anomalies based on previous year hydrodynamic conditions.
