Naval Hydrographic and Oceanographic Service

Nomadic ocean forecasting in bandwidth-constrained maritime environments demands innovative solutions to reconcile high-resolution forecasts with sparse onboard observations. Traditional workflows rely on independent data transmission and data assimilation schemes, and struggle to preserve critical features for operational tasks such as underwater acoustics. This study proposes an end-to-end deep learning framework integrating task-aware compression and data assimilation to address these challenges. Using high-resolution NEMO ocean model simulations in a Gulf Stream region, we evaluate two case studies: (1) Convolutional autoencoders (CAEs) for 3D+t sound speed profile (SSP) compression, demonstrating superior performance over principal component analysis (PCA) at compression ratios above 20, with a 31% reduction in reconstruction error; and (2) 4DVarNet-based neural reconstruction of surface duct width (SDW) fields using sparse in situ measurements and sea surface temperature (SST) synergies. The multimodal 4DVarNet-SST achieves superior spatiotemporal resolution (minimum spatial scale: 1.81°, temporal scale: 33.87 days) and a 11% improvement in root-mean-square error (RMSE) over optimal interpolation. By unifying compression, assimilation, and task-aware metrics, this work advances relocatable forecasting for mission-critical applications, emphasizing the role of end-to-end learning in preserving operationally relevant oceanographic features.

Abstract Primitive‐equation models are essential tools for studying ocean dynamics and their ever‐increasing resolution uncovers ever finer scales. At mesoscales and submesoscales, baroclinic instability is one of the main drivers of turbulence, but spurious numerical instabilities can also arise, leading to nonphysical dynamics. This study investigates a spurious instability termed Baroclinic Instability of Computational Kind (BICK), discovered in Arakawa and Moorthi (1988, https://doi.org/10.1175/1520‐0469(1988)045<1688:BIIVDS>2.0.CO;2) and Bell and White (2017, https://doi.org/10.1016/j.ocemod.2017.08.001), through idealized configurations using a vertical (Modified) Lorenz grid. Here, we explore the growth of BICK within quasi‐geostrophic (QG) and hydrostatic primitive‐equation (HPE) frameworks for different setups: the canonical Eady configuration, stratification‐modified Eady configurations, and a surface‐intensified jet configuration. Our results confirm that the emergence of BICK is specific to the vertical staggering of the (Modified) Lorenz grids. Its growth is consistent with linear QG theory, and BICK is confined near the surface and bottom boundaries. In HPE simulations, the nonlinear evolution of BICK generates small‐scale spurious eddies and reduces frontal sharpness. Increasing the number of levels reduces BICK's horizontal scale down to below the model's effective resolution. We illustrate this property using regional HPE simulations with a varying number of levels. BICK is found to significantly affect the vertically under‐resolved simulations by introducing small‐scale noise from both the bottom and surface boundaries. Our recommendation is to keep the ratio between the model horizontal and vertical resolution greater than , where is the Brunt‐Väisälä frequency and the Coriolis parameter, to minimize the impact of BICK on the dynamics.

In order to understand and predict coastal flooding processes in rocky environments, it is necessary to take into account bottom roughness, which plays a key role in wave transformation processes and in general coastal dynamics. The present work aims to implement a parameterisation of roughness-induced dissipation in 3D non-hydrostatic phase-resolved wave models, based on the Symphonie code (MARSALEIX et al., 2019). The modified model is compared with laboratory experiments carried out on a surf zone with a linear slope (DEALBERA et al., 2024). Different irregular waves were tested on various configurations of bottom roughness (represented by block configurations of different sizes and distributions). Wave generation and dissipation induced by breaking were first parametrised with respect to laboratory data on smooth bottoms. The different roughness cases were then studied on the basis of two distinct strategies for parameterising bottom friction, namely the bottom stress approach and the canopy drag approach. The performance of these two approaches is assessed by comparing the model results with measurements of the cross-shore significant wave height profile for different bottom configurations. On the basis of this work, recommendations on the choice of dissipation parameterisations will be made.

International audience

Abstract. An advanced coupling between a three-dimensional ocean circulation model (CROCO) and a spectral wave model (WAVEWATCH-III) is presented to better represent wave-current interactions in coastal areas. In the previous implementation of the coupled interface between these two models, some of the wave-induced terms in the ocean dynamic equations were computed from their monochromatic approximations (e.g., Stokes drift, Bernoulli head, near-bottom wave orbital velocity, wave-to-ocean energy flux). In the present study the exchanges of these fields computed from the spectral wave model are implemented and evaluated. A set of numerical experiments for a coastal configuration of the circulation near the Bay of Somme (France) is designed. The impact of the spectral versus monochromatic computation of wave-induced terms significantly affects the hydrodynamics at coastal scale in the case of storm waves and winds opposed to tidal flows, reducing the wave-induced deceleration of the vertical profile of tidal currents. This new implementation provides current magnitudes closer to measurements than those predicted using their monochromatic formulations, particularly at the free surface. The spectral surface Stokes drift and the near-bottom wave orbital velocity are found to be the most impacting spectral fields, respectively increasing advection towards the free surface and shifting the profile close to the seabed. In the particular case of the Bay of Somme, the approximation of these spectral terms with their monochromatic counterparts ultimately results in an underestimation of ocean surface currents. Our model developments thus provide a better description of the competing effects of tides, winds, and waves on the circulation of coastal seas with implications to the study of air-sea interactions and sediment transport processes.

        Coastal areas host 53% of the world's population increasingly subject to risks related to climate change. In order to understand and prevent these risks, the prediction of coastal hydrodynamic and morphological evolution is essential. Bathymetry is a key geophysical variable in  A key parameter to improve numerical coastal hydrodynamics models is the bathymetry. High resolution satellites now allow to observe coastal areas at a regional to global scale in a most cost-efficient way rather than local traditional echo sounding bathymetry measurements. We present S2Shores (Satellite to Shores), a new state-of-the-art python library developed to estimate wave field characteristics such as wavelength, period, direction, celerity to derive bathymetry from satellites. The core code is optimized to process optical satellite imagery, moreover the library is built in an object-oriented structure allowing efficient and agile manipulation of the modules developed from input and output handling to post-precessing.  S2Shores is optimized, using parallel computing with Dask python library, to compute large spatial scale, or time series evolution bathymetry as well on HPC cluster as on local computer. For example a bathymetry of 4400 km² of ocean, around the French Gironde estuary, can be computed in 90 sec (using 30 cores and 9.5 Go memory usage) with a 500 meters output grid resolution. We present a work in progress new python library for bathymetry estimation that will be open to the public to use and propose collaborative improvements.

The Bay of Biscay (Bob) is a hot spot for the generation of internal tides and nonlinear internal waves (NLIW). However, no studies have focused on internal waves on the continental shelf of the Bob. Here, we present 22 days of collocated temperature, velocity and backscatter profiles within a water depth H of 65 m. The background stratification evolved from two pycnoclines, with the strongest one near the sea bed, to a continuous profile due to wind-driven upwelling.Under the double pycnocline situation, we observed trains of elevation emerging from each internal tidal front with amplitude reaching up to H/4 and propagating at speeds between 0.1 and 0.35 m/s. Sporadically depression waves were measured within the train and can propagate substantially faster (between 0.36 and 0.54 m/s). With the continuous stratification, the trains of NLIWs of elevation and containing opposite polarities were no longer observed.These observations suggest that depression waves can cross the train of elevation waves. Resulting interactions could have significant impacts on sediment dynamics over the shelf. The double pycnocline regime and the impact of the stratification modification due to wind will be investigated numerically in future work.

The versatility, cost-efficiency and easy deployment of seismic sensor nodes facilitate geophysical monitoring in environments that were previously inaccessible for instrumentation, and among them landslides and unstable slopes, most of the time located in remote mountains. Using nodes allows for the setup of dense arrays with sensor inter-trace distances that become compatible with the geometries and dimensions of the geological structures to image. This becomes particularly true for landslides which have complex 3D architectures (hummocky bedrocks, layering, multi-dimensional fractures, diverse geotechnical material, deep and perched aquifers and water circulations) and are shallow processes with respect to the classical investigation depths and sensitivity of most geophysical survey techniques.Here we develop a specific processing workflow to allow the computation of 3D shear-velocity models with Ambient-Noise-based tomography applied to dense arrays of seismic stations. The workflow is applied to a dataset acquired at the Viella shallow landslide (France) developed in altered schists and moraine deposits. We deployed 70 IGU-16HR-3C-5Hz SmartSolo sensors (EOST/PISE service) with inter station distances of 70 m for a period of 25 days.The processing consists in several steps, all of them being tuned to the specific case of shallow depths of investigation. In areas where only few strong (ML>4) earthquakes are triggered, with a low azimuthal distribution, surface-waves velocity fields are complex to estimate with earthquakes. Ambient noise cross-correlation tomography has the advantage of using the ambient noise to model the surface waves velocities by retrieving the interstation Green’s functions. The main hypothesis for retrieving the Green’s functions is a homogeneous noise-source distribution, which is never achieved in a natural environment. Therefore, data filtering and daily stacking are crucial to reduce the effect of non-uniform noise distributions and lead to consistent velocity models. Due to the noisy environment of Viella (torrential flows, farming activity, anthropogenic noise), several procedures were implemented to optimize the processing (reduction of the coherent noises in the processed data, use of a pseudo-topography to estimate as accurately as possible the inter-station distances and travel times). We then computed the dispersion curve diagrams for the surface waves on which we applied a strict selection to only keep the consistent part of the surface waves dispersion curves. The selection parameters were optimized for the Rayleigh and Love waves. Then, we inverted the inter-station travel times to compute group velocities maps at several frequencies. Finally, we proceed to a Markov-Chain-Monte-Carlo inversion of each of the dispersion curves extracted from the group velocity maps. We finally obtained a 3D shear velocity model, which is further combined with geological and borehole information in order to document the 3D structure.The objectives are to present the processing workflow developed specifically for shallow imaging and the retrieval of 3D heterogeneities; effects of the processing parameters will be discussed on the Viella dataset. The approach developed for Viella is generic and has been further applied to other geological processes (permafrost at the Chauvet rock glacier, Marie-sur-Tinée mudslide), and the models will be discussed.

With more than 300 years of existence, Shom is the oldest active hydrographic service in the world. Compiling and deconflicting this much history automatically is a real challenge. This article will present the types of data Shom has to manipulate and the different steps of the workflow that allows Shom to compile over 300 years of bathymetric knowledge. The Téthys project for Shom will be presented in detail. The implementation of this type of process is a scientific, algorithmic, and infrastructure challenge.

This document is MINKE’s Deliverable 9.1: “White paper on advances in Absolute Salinity measurements”. It describes the state-of-the-art in the measurement of Absolute Salinity in seawater.