Covering more than 70% of the Earth’s surface, the oceans, and more particularly the upper oceans (the first few hundred meters below the surface), play a key role in the regulation of the climate, in the processes of climate change and therefore by extension in human societies. There are obvious interests (scientific, economic, etc.) in anticipating oceanic phenomena and monitoring extreme events (plankton blooms, green algae tides, hurricanes, cyclones, etc.).
However, despite the quantitative and qualitative increase in measurements and databases, as well as the ever-increasing development of simulation and observation capabilities, our ability to understand, reconstruct and anticipate the dynamics of the upper layers of the ocean remains limited. The numerical modeling tools currently developed for the prediction of oceanic phenomena do not seem, for the moment, sufficient to face the current societal and environmental challenges.
The OceaniX chair aims to bring new developments in this domain and to move towards a better understanding of the global functioning of the upper ocean.
This general objective is based on bringing together the inherent models of the physical sciences and the database learning approach at the heart of artificial intelligence (AI).
Relying on the leading expertise of Ifremer and the University of Western Brittany in marine sciences and technologies in addition to IMT Atlantique for engineering and data science. OceaniX is a project led by Ronan Fablet, which seeks to explore and develop strategies and systems driven by artificial intelligence. These AI-driven systems, being developed for the next generation of self-adaptive ocean monitoring and surveillance tools and systems, focus on observability and sampling optimality issues for complex dynamics and processes.
Supported by many institutional partners, the OceaniX Chair is, as ISblue, naturally conceived as an interface emphasizing the synergies between university, industry and research. To this end, the Chair integrates two distinct but closely related aspects.
The first one, the scientific aspect, has the ambition to revolutionize the observation systems of ocean simulation. It includes three major components and objectives:
The second component is oriented towards training and allows doctoral students to gain access to a double expertise in both digital and oceanographic fields. In order to achieve this, the OceaniX Chair is putting forward several training opportunities :
The short presentation of R. Fablet is available here : “Physics-Informed AI for Observation-driven Ocean AnalytiX”
You can also find Ronan Fablet’s speech during the One Océan Summit 2022 in Brest :
International collaborators : S. Brunton (Univ. Washington, USA), S. Matwin (Dalhousie Univ., Canada), A. Mahadevan (WHOI, USA), L. Bertino (NERSC, Norvège), F. Doblas-Reyes (BSC, Espagne).
Operating budget : 2M€ over 5 years (including 9 PhD scholarships)