3D DCSM-FM with water quality: model set up and validation

The 3D DCSM-FM water quality model has been developed over several years within multiple projects addressing water quality and ecological questions, resulting in different model versions based on varying hydrodynamic schematizations. This report presents a comprehensive overview of the most recent version, updated for the 2023 OSPAR applications on river nutrient load scenarios and validated for the years 2015–2017. The model is built on the latest hydrodynamic schematization release and incorporates the most up-to-date representation of water quality processes. The report describes the full model setup, including simulated variables, processes, forcings, and software, and evaluates model performance using maps, time series, and skill metrics, with particular attention to Dutch national monitoring stations (MWTL). Overall, the model performs well in simulating temperature, stratification, and winter nutrient concentrations. However, it underestimates summer phosphate, likely due to excessive phosphorus uptake by phytoplankton, and overestimates nitrate, suggesting overly high nitrogen remineralization rates. Chlorophyll-a is generally well represented in nearshore areas but is overestimated offshore during the growing season. Winter dissolved oxygen is slightly overestimated, while summer minima are captured accurately. The model is considered unsuitable for applications in the Baltic Sea, where it underestimates phosphate and chlorophyll-a and overestimates dissolved oxygen, probably due to underestimated phosphorus concentrations at the Baltic inflow boundaries. The model has been successfully applied to studies on river nutrient impacts on North Sea eutrophication and large-scale low-trophic aquaculture, but it is less suitable for sediment-related management issues or higher trophic level studies, as suspended sediments are not dynamically modeled and zooplankton is not included. For future development, the report recommends expanding validation beyond the Dutch North Sea, for example by aggregating skill metrics across OSPAR regions and exploring improved satellite products for different water types. Additional validation of intermediate variables, such as light extinction, phytoplankton composition, and primary production, is advised to better assess underlying processes. Including zooplankton in future versions is also recommended to improve phytoplankton dynamics and better represent food web interactions. The current model version will serve as the foundation for future schematization releases, with the long-term goal of moving toward a single model version for assessing the ecological effects of future marine uses in the North Sea.