SCIENTIFIC CONTEXT AND OBJECTIVES

This PhD aims to analyze the dynamics of internal tides (ITs), focusing on high modes and solitons (IT HM&S) and their interactions with mesoscale circulation (MS). These interactions modulate the structure, propagation, and intensity of ITs, making them highly variable in space and time (Goret et al., 2025; Kouogang et al., 2025b). They are also associated with enhanced turbulent dissipation (Kouogang et al., 2025a) and the redistribution of hydrological properties, nutrients, and chlorophyll (M’Mamdi et al., 2025; de Macedo et al., 2025), thus playing a major role in ocean variability, productivity, and climate. A key scientific challenge is to separate the signatures of ITs and associated waves from those of MS, which often overlap in altimetric and satellite observations. Improving their detection and understanding is a major priority for the oceanographic community, especially in the context of SWOT and the preparation of S3NG. A first test was done in Goret et al. (2025), and the main objective of this PhD is to extend this work toward automatic detection in SWOT and ocean color data to improve chlorophyll mapping.

The PhD will focus on three contrasting IT regions: the Amazon slope and the Indonesian seas, where strong ITs interact with energetic MS over complex topography, and the Pernambuco Plateau, characterized by weak IT activity, serving as a benchmark for validating detection methods and assessing algorithm robustness.

These regions benefit from strong collaborations and extensive in situ and modeling resources within the CNES MIAMAZ-ETI and SWOT-SWATI projects. The AMAZOMIX mooring (2.5 years) offers a unique observatory at the intersection of strong IT flux and MS (Koch-Larrouy et al., in prep.; Artana et al., in prep.). The SWOT-SWATI cruise (2026) on the Pernambuco Plateau and BRIN SWOT under-track campaigns (2026–2027) in Indonesia will provide precious below SWOT track. All in situ data will be archived in the CNES ODATIS repository for long-term access. Each region also benefits from dedicated high-resolution model configurations (AMAZON36, TAPIOCA36, INDO36; PI: Koch-Larrouy) including tides (Assene et al., 2023; Kouogang et al., 2025; Nugroho et al., 2025). A 500 m non-hydrostatic simulation, available in Indonesia (Bardot et al.) and under development for the Amazon region (PI: Koch-Larrouy), will reproduce IT high modes and nonlinear ISWs.

Within these CNES projects and regional tools, the PhD will develop AI-based methods for detecting IT HM&S and perform multi-platform analyses to investigate their dynamics and signatures in multi-sensor data (CHL, SST) across regions.

MAIN OBJECTIVES and METHODOLOGY

The main goal is to design and test a deep-learning algorithm for automatic IT HM&S detection in SWOT data, building on initial detections in eddy-contrasted cases (Goret et al., 2025). Both supervised and unsupervised approaches will be explored across contrasting IT–eddy regimes. The method will combine high-frequency filtering and along-track analyses, and will be compared with the VARDYN algorithm from which we might combine our techniques (Le Guillou et al., 2025; Ubelmann et al., 2018) and other international initiatives (Magalhaes et al. 2025). Non-hydrostatic simulations (~500 m) will support the training and validation of the AI model.

Beyond detection, a key objective will be to identify the nature of the detected waves—whether they correspond to IT HM&S or instabilities attached to IT crests. Determining the underlying processes will provide estimates of their wave length, velocity and width, helping to predict their evolution and energy pathways. This step is essential to distinguish them from mesoscale variability. To achieve this, the PhD will exploit extensive in-situ datasets and regional model outputs to analyze their dynamics and vertical structure.

Since these waves may imprint strong signals on chlorophyll and SST (M’Hamdi et al. 2025, de Macedo et al. 2025), the next step will extend AI detection to IR SST, MODIS, and Sentinel-2 data. Detected signatures will be cross-correlated with environmental descriptors to evaluate their impact on the water column structure and primary production.

SCIENTIFIC ENVIRONMENT and CO-FUNDING

The PhD will be hosted at CERFACS, which offers strong expertise in artificial intelligence and high-performance computing applied to ocean and climate sciences. CERFACS will co-fund the PhD, covering half of the fellowship, while the present CNES request concerns the remaining half. The work will involve CLS, LEGOS, LOCEAN, and partners in Indonesia, Brazil, the USA, and Portugal.

=================

For more Information about the topics and the co-financial partner (found by the lab!); contact Directeur de thèse - ariane.koch-larrouy@ird.fr

Then, prepare a resume, a recent transcript and a reference letter from your M2 supervisor/ engineering school director and you will be ready to apply online  before March 13th, 2026 Midnight Paris time!