EGU25-8976, updated on 14 Mar 2025
https://doi.org/10.5194/egusphere-egu25-8976
EGU General Assembly 2025
© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.
Poster | Thursday, 01 May, 10:45–12:30 (CEST), Display time Thursday, 01 May, 08:30–12:30
 
Hall X5, X5.219
Enhancing wave and ocean forecasts with Artificial Intelligence in the North-East Atlantic and Shelf Region – The Copernicus Marine Service Evolution CERAINE project
Manuel Garcia-Leon1, Marcos Portabella2, Jose María Garcia-Valdecasas1, Evgeniia Makarova2, Breogán Gómez1, Lotfi Aouf3, Stefania Ciliberti1, Alice Dalphinet3, Víctor Aquino1, Axel Alonso1, Carlos Fernández4, Roland Aznar1, and Marcos Sotillo1
Manuel Garcia-Leon et al.
  • 1NOLOGIN OCEANIC WEATHER SYSTEMS, S.L.U., Santiago de Compostela (A Coruña), Spain (manuel.garcia@nowsystems.eu)
  • 2ICM - CSIC, Barcelona (Barcelona), Spain
  • 3Meteo-France, Departement Marine et Oceanographie, Toulouse, France
  • 4CESGA, Santiago de Compostela (A Coruña), Spain

Copernicus Marine Service Monitoring and Forecasting Centres (MFCs) are improving their models to resolve finer-scale oceanographic features, driven by a growing need for high-resolution, short-term ocean forecasts. A key limitation to forecast accuracy, however, stems from errors in the model forcings. These errors can be mitigated with Artificial Neural Networks (ANNs) that are trained with the increasing volume of remote sensing observations. ANNs allow to extract spatio-temporal patterns from these measurements, enabling the generation of enhanced forcings by integrating these correction patterns with existing operational forcings.

The Copernicus Marine Service Evolution CERAINE project (2024 – 2026) aims to improve short-term ocean and wave model forecasts within the European North-East Atlantic (NEA) region by enhancing the accuracy of their model forcings using ANNs. Two distinct ANN methodologies will be implemented. The first one will focus on correcting wind forcings, using Synthetic Aperture Radar (SAR) data for coastal zones and scatterometer data for offshore areas. These improved wind fields will subsequently be used as forcing inputs for both ocean physics and wave models. A second type of ANNs will be developed to correct surface ocean currents, which are important inputs for spectral wave models, using data acquired from High Frequency Radar deployed at coastal sites.

The NEA region, which encompasses the Copernicus Marine Service IBI (Iberian-Biscay-Ireland) and NWS (North-West-Shelf) products, has been chosen as the project's pilot area due to two key reasons: (i) the expected significant impact of the proposed forcing corrections on both coastal and offshore waters, and (ii) the availability of a comprehensive observational network in this region. The project will assess the impact of these ANN-derived forcings on the IBI-MFC NEA ocean and wave models through a series of sensitivity tests. CERAINE holds the potential for direct integration of its results into the IBI-MFC operational service and the subsequent extension of this approach to other Copernicus Marine MFC target regions.

This contribution will show the on-going development of the wind and surface currents ANNs, and their updated validation under a set of recent events.

How to cite: Garcia-Leon, M., Portabella, M., Garcia-Valdecasas, J. M., Makarova, E., Gómez, B., Aouf, L., Ciliberti, S., Dalphinet, A., Aquino, V., Alonso, A., Fernández, C., Aznar, R., and Sotillo, M.: Enhancing wave and ocean forecasts with Artificial Intelligence in the North-East Atlantic and Shelf Region – The Copernicus Marine Service Evolution CERAINE project, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8976, https://doi.org/10.5194/egusphere-egu25-8976, 2025.