Predicting marine biological hazards to prevent socio-economic impacts

🔥 Bottom O₂ forecasts →

PrimePrevention Modeling Activities

A compact overview of the project’s modeling goal, current state of the art, and ongoing work towards more accurate, higher-resolution forecasts of marine biological hazards along German Baltic coasts.

Goal

We are on a mission to understand and forecast marine biological hazards at the German Baltic coasts for the public good. This is carried out in an cooperative effort, leveraging the best available near-real time observations and operational model products. Specifically we set out to:

  • Translate ocean physics and biogeochemistry modeling expertise into actionable coastal indicators.
  • Predict the occurrence risk of marine biological hazards, like low-oxygen conditions and blooms of potentially toxic algae.
  • Improve forecast accuracy, especially in coastal hotspots, e.g., by fine-scaling spatial models and developing efficient pathways from observations and models to risk prediction.
  • Provide transparent, highly reliable and publically accessible outputs (maps, animations, and documentation).

Explore an early operational-style risk forecast developed within the project based on data and model results provided by the Bundesamt für Seeschifffahrt und Hydrographie BSH and the Swedish Meteorological and Hydrological Institute SMHI and distributed by the Copernicus Marine Environment Monitoring Service CMEMS (CMEMS) here: interactive bottom O₂ forecast map.

The project focuses on the Western Baltic Sea including Kiel Bight, Bay of Lübeck and Bay of Mecklenburg where rising temperatures and biogeochemical changes are suspected to increase the risk of harmful algal blooms and viruses. Processes in these areas are driven by both local processes and export-import dynamics also linked to Major Baltic Inflows. Therefore, our modeling framework comprises the North Sea, the entire Baltic Sea alongside with local high-resolution configurations targeted at resolving small-scale processes in focus areas.

Spatial extent of model configurations used in PrimePrevention

Spatial extent of model configurations used in PrimePrevention. Each configuration is a compromise between area covered and spatial resolution. The BSH, IOW and CAU model feature a horizontal resolution of 1800m, 600m and 100m, respectively.

Existing forecasting efforts already provide key indicators - for example BSH operational products for Oxygen Deficiency or algae blooms and SMHI operational products for cyanobacteria blooms. Within PrimePrevention, forecast models are combined with satellite products and near real-time observations. Such combined products are the foundation for evaluation, visualization, and further improvement.

Operational visualizations are in work—including the linked interactive bottom O₂ forecast map. In PrimePrevention, the goal is to increase forecast accuracy and push towards models with finer spatial resolutions, e.g. in Kiel Bay, Bay of Greifswald or Szczecin Lagoon, in order cover local extremes and impacts. finer coastal resolution in order cover local extremes and impacts.

High-resolution coastal modeling for the Kiel Bight and a workflow toward operational-style outputs (animations and maps).

Movie: bottom velocities (MOMKI / Kiel Bight)

The movie visualizes bottom velocities and highlights that the Kiel Bight is strongly affected by inflowing and outflowing currents, also connected to Major Baltic Inflows (see: Nature Communications Earth & Environment article).

What we have done so far:

  • A one-way nested setup for the Kiel Bight with ~100 m resolution has been implemented (called MOMKI), as well as a model for the whole Western Baltic Sea with ~600m resolution (GETM-WB600m).
  • Atmospheric forcing has been prepared: ERA5 reanalysis (chosen for its up-to-date combination of model and observations).
  • Ocean boundary conditions: observations are insufficient at required coverage, so boundary forcing relies on models. Because coarse models still have known issues representing oxygen-rich North Sea inflows, the approach is to compute an ensemble with different boundary conditions to gain robust model results with low uncertainty.
  • Model pameter sensitivity: numerous parameter combinations were tested to find suitable reference configurations for each model and to design ensembles similar in spirit to modern weather predictions.
  • Additional tracers were implemented in the model to estimate residence times and transport pathways, e.g. of sulphidic waters .
  • Biogeochemical modelling: GETM-WB600m was coupled to the biogeochemical model ERGOM (www.ergom.net) and forced with observed nutrient inputs that allow to study the long-term development of the water quality in the western Baltic Sea and which impact extreme events (such as extraordinary hot or wet summers).

The Next steps are to tie these pieces together: high-resolution process realism in the Kiel Bight, improved boundary-condition handling via ensembles, and user-facing products products that allow stakeholder advice based on newest knowledge, models and data (e.g. by interactive maps + animations). .