Workpackage 4 - Numerical Modelling

Research Objective: Essential transformation processes in shallow coastal waters are not resolved and transports into the Baltic Proper are not sufficiently reproduced in current state-of-the-art Baltic Sea models. This includes processes as waves, currents, sediment and their interaction with the biogeochemistry. The aim of WP4 is to develop or improve parameterizations relevant for the dynamics in the coastal zone. This includes

• Consistent coupling of waves, currents, turbulence and sediment dynamics in shallow water to reproduce shallow water effects of surface waves
• Establishment of a coupled pelagic-benthic biogeochemical model for shallow coastal waters to quantitatively reproduce the coastal filter
• Process-resolving scenario simulations
• Development of a near-real-time hindcast system with resolved coastal scales

Example Activities

ROBOELF: We are developing roboelf, a GPU-accelerated marine biogeochemistry modelling framework particularly well-suited for shallow-water research. By using a hybrid Lagrangian-Eulerian advection scheme, the model avoids the very short time steps typically required at high spatial resolution, making high-resolution simulations of coastal and estuarine environments far more efficient. Because roboelf runs on a single consumer-grade GPU rather than a supercomputer, it is accessible to a much wider research community. We hope this will open the door to genuine interdisciplinary collaboration, enabling marine scientists, ecologists, and environmental managers to work together interdisciplinarily on pressing questions in shallow-water ecosystem research.

Hypoxic Upwelling: In September 2025 an upwelling event of hypoxic water at parts of the german Baltic Sea coast caused a fish kill event in front of the doorsteps of Warnemünde. For the first time such an event has been sampled in high quality with the newly developed S2B mooring. To analyse the drivers and prerequisites that lead to this event we use a combination of this unprecedented dataset together with a numerical model.

Q10 (Temperature dependency): Using the coupled physical–biogeochemical model MOM-ERGOM, we investigate how temperature-driven sediment and water-column processes influence nutrient cycling across nine Baltic basins. To evaluate the impact of temperature-dependent processes, we performed sensitivity experiments by increasing q10 values across the entire Baltic Sea and separately within coastal zones. Results show that even modest increases in detritus recycling drive disproportionate, nonlinear shifts in nitrogen and phosphorus cycling in deep basins. Responses vary strongly among basins, with anoxic regions showing pronounced ammonium accumulation and nitrate depletion, indicating enhanced denitrification relative to nitrification under warmer conditions.

Near-realtime modelling: We operate a coastal ocean modelling system for the southwestern Baltic Sea based on the hydrodynamic model GETM, designed to reconstruct the most recent state of the system on a daily basis. The workflow runs automatically and provides a high-resolution deterministic hindcast (200 m) together with an ensemble of lower-resolution simulations (600 m) to quantify uncertainty. Model outputs include water level, temperature, salinity, and biogeochemical variables such as oxygen concentration and saturation. At present, the system serves as a proof of concept and a foundation for further applications, including ecosystem analysis, early event detection, and the potential planning of research cruises.