Periodic Reporting for period 1 - EDITO-Model Lab (Underlying models for the European DIgital Twin Ocean - EDITO-Model Lab)
Okres sprawozdawczy: 2023-01-01 do 2023-12-31
EDITO-Model Lab is a Horizon Europe project working to develop the next generation of ocean numerical models to be integrated into the EU public infrastructure of the European Digital Twin of the Ocean, providing access to focus applications and simulations of what-if scenarios for the sustainable management of marine environments. EDITO-Model Lab will demonstrate the feasibility to connect and interoperate, on a common platform, a large variety of ocean and coastal numerical models, allowing for global, regional-to-coastal model configuration and the co-development of new simulations and scenarios for on-demand ocean forecasting and ocean climate prediction.
As a collaborative and user driven initiative, the 8-Work Packages EDITO-Model Lab project will deliver a Virtual Ocean Model Lab, an interactive and co-development environment to operate models, including:
• Core model suite with global high-resolution models and coastal configurations
• Downstream user toolkits
• Developer’s toolkit
The EDITO core model suite is based on modelling and simulation software, artificial intelligence, algorithms, and specialised tools to form a new service capacity for accessing, manipulating, analysing, and understanding marine information, with enhanced ocean forecasting and ocean climate prediction capabilities. Intermediate and downstream users will find digital tools, models, data, and information for:
• FOCUS APPLICATIONS related to the Mission Ocean Lighthouses and Blue Economy sectors, including Marine Protected Areas, ship routing, and pollution
• WHAT-IF SCENARIOS to find solutions to natural and human-induced hazards, such as marine plastic and carbon emissions from aquaculture
• Core model suite with global high-resolution models and coastal configurations
• Downstream user toolkits
• Developer’s toolkit
The EDITO core model suite is based on modelling and simulation software, artificial intelligence, algorithms, and specialised tools to form a new service capacity for accessing, manipulating, analysing, and understanding marine information, with enhanced ocean forecasting and ocean climate prediction capabilities. Intermediate and downstream users will find digital tools, models, data, and information for:
• FOCUS APPLICATIONS related to the Mission Ocean Lighthouses and Blue Economy sectors, including Marine Protected Areas, ship routing, and pollution
• WHAT-IF SCENARIOS to find solutions to natural and human-induced hazards, such as marine plastic and carbon emissions from aquaculture
During Reporting Period 1,
WP1’s coordination was ensured through six project meetings and an internal communication portal. The co-design phase provided important feedback from users, and the components of the EDITO-Model Lab Technical Architecture were validated. NVIDIA withdrew from the consortium and their tasks were redistributed. François Courteille from NVIDIA joined the External Advisory Board. Deliverable D1.1 (Data Management Plan) was submitted.
WP2’s developed AI-based components, particularly deep differentiable emulators (DDEs). Achievements include the emulation of OGCM parameterisation. Neural schemes for simulation have shown favourable outcomes, indicating potential improvements over existing methods. The exploration of DDEs for ocean data assimilation has resulted in advancements in uncertainty quantification and model calibration, showcasing the efficacy of super-resolution techniques and end-to-end neural assimilation.
WP3 developed and prototyped various model improvements. Major advances include the implementation of new time-stepping schemes, testing of mixed precision computing, and improvements in physics parameterisation for models. These model improvements were implemented in ocean configurations for global, regional, and coastal areas across European Seas.
WP4 designed and developed the core model suite with its integration path being refined through co-design. To ensure compatibility between components, the VOML with the EDITO platform were merged. A beta version was released after collaboration between partners, co-design sessions, and continuous integration of feedback.
WP5 progressed in developing the VOML, with co-design workshops also providing user feedback on its architecture. Backend enhancements included integrating Autosubmit 4 and deploying it with Docker. Shared Git environments facilitated code versioning and CI/CD testing exploration. GUIs for model interaction and workflow monitoring were developed, and further user engagement strategies were drawn. Testing and validation procedure were set.
WP6 focused on co-designing Focus Applications, with users’ sessions held for refinement. Computational needs were translated, while model developments were aligned and DDES were integrated. 12 demonstrators for 3 Focus Applications were created. The technical development involved both EDITO-Model Lab and EDITO-Infra teams. FAs 1 and 2 reached TRL6 and is expected to reach TRL7 by 2024. FA 3 is in mature development, aiming for completion in 2024.
WP7 explored What-If scenarios for nature-based solutions and aquaculture. Research into seagrass impact on coastal erosion showcased promising results. Investigations into marine plastic pollution provided insights into plastic distribution in the Baltic and Mediterranean. Aquaculture scenarios demonstrated advancements in modelling organism growth, laying the groundwork for future research. End-users and stakeholders training will facilitate the uptake of project results.
WP8 launched a website, set social media accounts, implemented an editorial plan, and created various communication materials. Eight news articles explaining the project's goals and the contributions to the European Digital Twin of the Ocean were published. Strong presence at key international events further enhanced EDITO’s visibility. In collaboration with EDITO-Infra, all KPIs set for social media and website performance were surpassed.
WP1’s coordination was ensured through six project meetings and an internal communication portal. The co-design phase provided important feedback from users, and the components of the EDITO-Model Lab Technical Architecture were validated. NVIDIA withdrew from the consortium and their tasks were redistributed. François Courteille from NVIDIA joined the External Advisory Board. Deliverable D1.1 (Data Management Plan) was submitted.
WP2’s developed AI-based components, particularly deep differentiable emulators (DDEs). Achievements include the emulation of OGCM parameterisation. Neural schemes for simulation have shown favourable outcomes, indicating potential improvements over existing methods. The exploration of DDEs for ocean data assimilation has resulted in advancements in uncertainty quantification and model calibration, showcasing the efficacy of super-resolution techniques and end-to-end neural assimilation.
WP3 developed and prototyped various model improvements. Major advances include the implementation of new time-stepping schemes, testing of mixed precision computing, and improvements in physics parameterisation for models. These model improvements were implemented in ocean configurations for global, regional, and coastal areas across European Seas.
WP4 designed and developed the core model suite with its integration path being refined through co-design. To ensure compatibility between components, the VOML with the EDITO platform were merged. A beta version was released after collaboration between partners, co-design sessions, and continuous integration of feedback.
WP5 progressed in developing the VOML, with co-design workshops also providing user feedback on its architecture. Backend enhancements included integrating Autosubmit 4 and deploying it with Docker. Shared Git environments facilitated code versioning and CI/CD testing exploration. GUIs for model interaction and workflow monitoring were developed, and further user engagement strategies were drawn. Testing and validation procedure were set.
WP6 focused on co-designing Focus Applications, with users’ sessions held for refinement. Computational needs were translated, while model developments were aligned and DDES were integrated. 12 demonstrators for 3 Focus Applications were created. The technical development involved both EDITO-Model Lab and EDITO-Infra teams. FAs 1 and 2 reached TRL6 and is expected to reach TRL7 by 2024. FA 3 is in mature development, aiming for completion in 2024.
WP7 explored What-If scenarios for nature-based solutions and aquaculture. Research into seagrass impact on coastal erosion showcased promising results. Investigations into marine plastic pollution provided insights into plastic distribution in the Baltic and Mediterranean. Aquaculture scenarios demonstrated advancements in modelling organism growth, laying the groundwork for future research. End-users and stakeholders training will facilitate the uptake of project results.
WP8 launched a website, set social media accounts, implemented an editorial plan, and created various communication materials. Eight news articles explaining the project's goals and the contributions to the European Digital Twin of the Ocean were published. Strong presence at key international events further enhanced EDITO’s visibility. In collaboration with EDITO-Infra, all KPIs set for social media and website performance were surpassed.