The Copernicus Marine Service Global Ocean Forecasting Systems: Innovations and Perspectives

Advanced ocean forecasting systems play an essential role in understanding and managing the ever-evolving ocean dynamics, especially in the face of climate and environmental challenges. These systems provide crucial information for maritime safety, marine resource management, and ecosystem protection policies.

This presentation will explore innovations in ocean forecasting, performance improvements, and new perspectives in the simulation and monitoring of oceanographic and biogeochemical processes.

In the framework of Copernicus Marine Service, Mercator Ocean International delivers in real-time weekly analyses and daily 10-day forecasts of the global ocean dynamic, sea ice and biogeochemical components.

The present ocean dynamic and sea ice real-time system is based on a global 1/12° high resolution configuration implemented in NEMO GCM and forced by the 8km/1hour ECMWF IFS system. Oceanic observations are assimilated in the model using a reduced-order Kalman filter method. Along track altimeter Sea Level Anomaly (SLA), satellite sea surface temperature (SST) and sea ice concentration, and in situ temperature and salinity vertical profiles are jointly assimilated to estimate the initial conditions for numerical ocean forecasting.

The present biogeochemical is based on a global 1/4° model implemented in NEMO-PISCES, with an  an offline coupling with the dynamical ocean. This BGC simulation shall benefit from the assimilation of satellite Ocean Colour data (Chlorophyll concentration), and from Machine-Learning-extended-SOCAT-based Carbonates surface data (dissolved inorganic carbon and total alkalinity).

A new 1/36° global configuration (2 to 3 km resolution) has been developed and allows for a better representation of sub-mesoscale processes (1-50 km), enhancing the precision of ocean circulation. It is forced by the 5 tidal components. Thanks to the resolution increase, this model can resolve the Rossby radius in almost all open oceans areas at global scale quite everywhere and to span a large part of the internal wave spectrum. In the framework of the EDITO-Model Lab project, the development of a near real-time demonstrator has been started. The 1/36° global configuration is constrained by a spectral nudging to the CMEMS/MOI global 1/12° real-time system and provides reforecasts over the 2023 year.

Future systems focus on integrating physics-based models with machine learning methods to enhance forecast accuracy and speed. GloNet, a data-driven model developed by Mercator, integrates physics-based principles through neural operators and networks to dynamically capture local-global interactions within a unified, scalable framework, ensuring high small-scale accuracy and efficient dynamics over long forecast intervals. Glonet is trained on Mercator GLORYS12 reanalysis products and evaluated using Mercator’s operational numerical ocean prediction analysis GLO12. Glonet runs daily in a preoperational framework along with other physics-driven numerical models.