Compétences pour réussir :

Bac + 5 en informatique, si possible appliquée aux géosciences, mathématiques, mathématiques appliquées.
Connaissance de langages de programmation (python, Fortran…) et de l’environnement linux.
Connaissances en océanographie.
Connaissance du traitement du signal et analyse numérique.
Bonne maîtrise de l'anglais.

Encadrants :

Giovanni RUGGIERO, Océanographe spécialisé assimilation de données au sein du département R&D, Mercator Océan.

Pauline RIBOT, Maîtresse de conférences à l'Université Paul Sabatier - Toulouse III, LAAS-CNRS.

Elodie CHANTHERY, Maîtresse de conférences, HDR à l’INSA de Toulouse, LAAS-CNRS.

Quentin GAUDEL, Architecte de services cloud au sein du département Digital Ocean, Mercator Océan.

Angélique MELET, Océanographe spécialisée étude du climat au sein du département R&D, Mercator Océan.

[English version]

Description

Mercator Ocean International has been developing operational oceanography activities for nearly 30 years, as part of its public interest mission to preserve the ocean.

Many scientific and societal challenges must be met to ensure a sustainable ocean, whether they concern the environment, biodiversity, climate change, the blue economy or education. To meet these challenges, Mercator Ocean designs, develops, operates and maintains state-of-the-art digital systems capable of describing, analysing and forecasting the state of the ocean in 3D, continuously and in real time. The scientific information is then translated to be accessible to all, whether they are public or commercial services, political decision makers, industrialists, associations, NGOs, teachers or citizens. Mercator Ocean International thus combines scientific excellence and social commitment on a daily basis.

As a non-profit company under multinational governance (ES, FR, GB, IT, NO), we work in a climate of trust with our ten shareholder partners, all key players in the development of European oceanography.

Responsibilities

About the project:

Mercator Ocean implements the European marine environment observation and forecasting service Copernicus Marine Service, while also coordinating the development of the EDITO plateform, the European digital twin of the ocean. In this context, Mercator Ocean develops and operates several global ocean analysis and forecasting systems (high-resolution deterministic, low-resolution ensemble, machine learning). These systems use complex models whose trajectory is adjusted to take into account various observations from satellites or on-site observations (surface temperature, sea level anomalies, salinity, etc.). Each system has its own purpose, and together they complement each other to meet the needs of users who need to know the current environmental conditions (e.g., temperature, salinity, currents, etc.) on a day-to-day basis, at every point in the ocean from the surface to the bottom, how they have changed over the past decades, and how they will change in the future.

The detection of feared events (environmental and socio-economic), caused by the occurrence of extreme physical or biological events, is a key issue and a critical feature of the digital twin of the ocean, one of the benefits of which is to enable users to better measure the impacts of ocean change.

As part of a collaboration between Mercator Ocean and LAAS-CNRS, a research laboratory specializing in systems analysis and architecture, the DISCO (Diagnostics, Supervision, and Control) team contributes its expertise in the detection and identification (diagnostics) and prediction (prognosis) of feared events. Through an application on a system representing simplified ocean dynamics, the aim of the internship is to integrate and test a tool for detecting feared events based on:

Ensemble modeling, capable of quantifying the uncertainties of results through the manipulation of state distributions.
Hybrid modeling, i.e., containing both discrete and continuous information.
Artificial Intelligence (machine learning), enabling the acceleration of ocean simulation.

The intern will be required to use digital tools developed in Python, perform calculations on a supercomputer, and create visualizations on the EDITO cloud platform. EDITO is the European public platform contributing to the digital twin of the ocean, which offers, among other things, basic tools (applications, models, notebooks) for analyzing data directly on a computing platform as close as possible to the data (data lake). The intern will develop a workflow integrating existing Python tools. They will contribute to defining the service offering of the EDITO digital twin of the ocean on the subject of detecting extreme events in the ocean.

Goal:

Conduct a bibliographic study on the subject of extreme oceanic events and their modelling.
Develop a demonstrator based on ensemble and hybrid techniques (discrete and continuous).
Integrate the demonstrator in the EDITO plateform.
Acquire proficiency in a Python ‘search’ tool and gain experience in transferring it to operational production on EDITO.

Profile

Skills to succeed:

Master's degree in computer science, preferably applied to geosciences, mathematics, applied mathematics.
Knowledge of programming languages (Python, Fortran, etc.) and the Linux environment.
Knowledge of oceanography.
Knowledge of signal processing and numerical analysis.
Proficiency in English.

Supervisors:

Giovanni Ruggiero, Océanographe spécialisé assimilation de données, département R&D

Pauline Ribot, Maîtresse de conférences à l'Université Paul Sabatier - Toulouse III, LAAS-CNRS

Elodie Chanthery, Maîtresse de conférences, HDR à l’INSA de Toulouse, LAAS-CNRS

Quentin Gaudel, Architecte de services cloud, département Digital Ocean

Angélique Melet, Océanographe spécialisée étude du climat, département Ocean Climate

CANDIDATURES SPONTANEES - UNSOLLICITED APPLICATIONS

Détection et prévision d’évènements océaniques redoutés sous incertitude / Detection and forecast of feared ocean events under uncertainty