Figure - available from: Journal of Geophysical Research: Oceans
This content is subject to copyright. Terms and conditions apply.
Kernel density plot of eddy nonlinearity (r) versus normalized eddy scale Lη/Ld for eddies identified by the automated eddy detection algorithm. The nonlinearity parameter (r) is defined as r=U/c following Chelton et al 2007. Colors blue, red, and gray represent eddies in the 35°, 45°, and 55° latitudinal band, respectively. This regime diagram is adapted from Klocker et al. (2016). This plot combines data corresponding to 4,680 and 3,755 detected eddy structures for (top) NATL60 and (bottom) HYCOM50 respectively.
Source publication
Ocean circulation is dominated by turbulent geostrophic eddy fields with typical scales ranging from 10 to 300 km. At mesoscales (>50 km), the size of eddy structures varies regionally following the Rossby radius of deformation. The variability of the scale of smaller eddies is not well known due to the limitations in existing numerical simulations...
Similar publications
Mars’ transition from an early “warm and wet” to a “cold and dry” environment left fingerprints on the geological record of fluvial activity on Mars. The morphological and mineralogical observations of aqueous activity provided varying constraints on the condition and duration of liquid water on the Martian surface. In this study, we surveyed the m...
During the last 4 decades punctual occurrences of extreme ocean temperatures, known as marine heatwaves (MHWs), have been regularly disrupting the coastal ecosystem of the Peru-Chile eastern boundary upwelling system. In fact, this coastal system and biodiversity hot-spot is regularly impacted by El Niño events, whose variability has been related t...
Gravitational waves (GWs) were observed for the first time in 2015, one century after Einstein predicted their existence. There is now growing interest to extend the detection bandwidth to low frequency. The scientific potential of multi-frequency GW astronomy is enormous as it would enable to obtain a more complete picture of cosmic events and mec...
Citations
... The considered simulation data set relies on a nature run of the NATL60 configuration of the NEMO (Nucleus for European Modeling of the Ocean) model (Madec et al., 2022). This simulation delivers a realistic hindcast simulation of ocean dynamics, including mesoscale-to-submesoscale ocean dynamics (Ajayi et al., 2020), over 1 year from October 2012 to September 2013 for a North Atlantic domain with a 1/60°and hourly resolution. The initial and open boundary conditions rely on GLORYS2v3 ocean reanalysis and the atmospheric forcing is based on DFS5.2 (Dussin et al., 2018). ...
... The initial and open boundary conditions rely on GLORYS2v3 ocean reanalysis and the atmospheric forcing is based on DFS5.2 (Dussin et al., 2018). We refer the reader to (Ajayi et al., 2020) for a more detailed description and analysis of NATL60 simulation. It has been used in numerous studies regarding the reconstruction and observability of sea surface dynamics (see below for the related OSSE data challenge stated in Le Guillou et al. (2020)). ...
Satellite altimetry offers a unique approach for direct sea surface current observation, but it is limited to measuring the surface‐constrained geostrophic component. Ageostrophic dynamics, prevalent at horizontal scales below 100 km and time scales below 10 days, are often underestimated by ocean reanalyzes employing data assimilation schemes. To address this limitation, we introduce a novel deep learning scheme, rooted in a variational data assimilation formulation with trainable observations and a priori terms, that harnesses the synergies between satellite‐derived sea surface observations, namely sea surface height (SSH) and sea surface temperature (SST), to enhance sea surface current reconstruction. Numerical experiments, conducted using realistic simulations, in a case study area of the Gulf Stream, demonstrate the potential of the proposed scheme to capture ageostrophic dynamics at time scales of 2.5–3.0 days and horizontal scales of 0.5°–0.7°. The analysis of diverse observation configurations, encompassing nadir along‐track altimetry, wide‐swath SWOT (Surface Water and Ocean Topography) altimetry, and SST data, highlights the pivotal role of SST features in retrieving a significant portion of the ageostrophic dynamics (approximately 47%). These findings underscore the potential of deep learning and 4DVarNet schemes in improving ocean reanalyzes and enhancing our understanding of ocean dynamics.
... So the same CCN (UNet variant) was retrained on a more realistic training dataset: less random noise than in prelaunch simulations, training on the eNATL60 model with tides (Ajayi et al., 2020;Brodeau et al., 2020) to better simulate the SWOT observations. The re-training was performed with a random wave-modulated white noise at 250-m, with a Hamming filter to 440 mimic the downscaling to 2-km that is performed in the ground segment. ...
The Surface Water and Ocean Topography (SWOT) mission delivers unprecedented swath altimetry products. Despite SWOT’s 2D coverage and precision, its Level-2 products suffer from the same limitations as their counterparts from nadir altimetry missions. Level-2 products are designed in a standalone ground-segment to meet the mission’s primary science objectives. In contrast, some research domains and applications require consistent multi-mission observations such as the Level-3 products provided by the Data Unification and Altimeter Combination System (DUACS) for almost 3 decades, and with 20 different satellites. In this paper, we describe how we extended the Level-3 algorithms to handle SWOT’s unique swath-altimeter data. We also illustrate and discuss the benefits, relevance, and limitations of Level-3 swath-altimeter products for various research domains.
... It is well recognized that mesoscale eddies capture approximately 80% of the total kinetic energy, exerting a significant influence on the ocean circulation, the air-sea interaction, and the marine ecosystem (Dong et al., 2014;Zhang et al., 2014;Zhang & Qiu, 2020). However, the nadir-looking altimeter is limited in its ability to resolve other ocean energetic motions at submesoscale or smaller scales (Amores et al., 2018;Ballarotta et al., 2019;Dufau et al., 2016;Vergara et al., 2019), even though these motions play a crucial role in the cascade and dissipation of ocean energy (Ajayi et al., 2020;Klein et al., 2019;Zhang et al., 2023). To improve the observational capabilities of satellite altimetry, the next generation of wide-swath imaging altimeter has been proposed, including the Surface Water and Ocean Topography (SWOT) satellite launched on 16 December 2022 , as well as the Guanlan mission proposed by China . ...
Extracting balanced geostrophic motions (BM) from sea surface height (SSH) observations obtained by wide‐swath altimetry holds great significance in enhancing our understanding of oceanic dynamic processes at submesoscale wavelength. However, SSH observations derived from wide‐swath altimetry are characterized by high spatial resolution while relatively low temporal resolution, thereby posing challenges to extract the BM from a single SSH snapshot. To address this issue, this paper proposes a deep learning model called the BM‐UBM Network, which takes an instantaneous SSH snapshot as input and outputs the projection corresponding to the BM. Training experiments are conducted both in the Gulf Stream and South China Sea, and three metrics are considered to diagnose model's outputs. The favorable results highlight the potential capability of the BM‐UBM Network to process SSH measurements obtained by wide‐swath altimetry.
... These methods are generally trained on simulated data, and are then evaluated either on simulated data from an excluded test set or on real data such as drifters. System Simulation Experiments (OSSEs) such as [37][38][39] are based on geophysical models and have proven essential in training data-driven methods. An OSSE is a modeling experiment used to evaluate the value of a new observing system when actual observational data are not available. ...
... The OSSE setup comprises a nature run, a data assimilation system, and software tasked with simulating 'observations' derived from the nature run while incorporating realistic observation errors. Numerous studies train models to reconstruct SSH using simulated data from the Gulf Stream [37], such as 4DVarNet (Fablet et al. [8]) and the CNN-based method RESAC (Thiria et al. [40]). Nardelli et al. [7] use simulated data from the Mediterranean Sea [38] to reconstruct SSH and surface currents. ...
Real-time reconstruction of ocean surface currents is a challenge due to the complex, non-linear dynamics of the ocean, the small number of in situ measurements, and the spatio-temporal heterogeneity of satellite altimetry observations. To address this challenge, we introduce HIRES-CURRENTS-Net, an operational real-time convolutional neural network (CNN) model for daily ocean current reconstruction. This study focuses on the Mediterranean Sea, a region where operational models have great difficulty predicting surface currents. Notably, our model showcases higher accuracy compared to commonly used alternative methods. HIRES-CURRENTS-Net integrates high-resolution measurements from the infrared or visible spectrum—high resolution Sea Surface Temperature (SST) or chlorophyll (CHL) images—in addition to the low-resolution Sea Surface Height (SSH) maps derived from satellite altimeters. In the first stage, we apply a transfer learning method which uses a high-resolution numerical model to pre-train our CNN model on simulated SSH and SST data with synthetic clouds. The observation of System Simulation Experiments (OSSEs) offers us a sufficient training dataset with reference surface currents at very high resolution, and a model trained on this data can then be applied to real data. In the second stage, to enhance the real-time operational performance of our model over previous methods, we fine-tune the CNN model on real satellite data using a novel pseudo-labeling strategy. We validate HIRES-CURRENTS-Net on real data from drifters and demonstrate that our data-driven approach proves effective for real-time sea surface current reconstruction with potential operational applications such as ship routing.
... Our primary experiments utilized the eNATL60 configuration of the Nucleus for the European Modelling of the Ocean (NEMO) model (Gurvan et al., 2022), featuring a 1/60°horizontal resolution and 300 vertical levels across the North Atlantic. This highresolution configuration is essential for understanding ocean dynamics, particularly for surface oceanic motions down to 15 km, which aligns with SWOT observations (Ajayi et al. (2020)). We direct readers to this work for a detailed understanding of NATL60's capabilities. ...
Combining remote-sensing data with in-situ observations to achieve a comprehensive 3D reconstruction of the ocean state presents significant challenges for traditional interpolation techniques. To address this, we developed the CLuster Optimal Interpolation Neural Network (CLOINet), which combines the robust mathematical framework of the Optimal Interpolation (OI) scheme with a self-supervised clustering approach. CLOINet efficiently segments remote sensing images into clusters to reveal non-local correlations, thereby enhancing fine-scale oceanic reconstructions. We trained our network using outputs from an Ocean General Circulation Model (OGCM), which also facilitated various testing scenarios. Our Observing System Simulation Experiments aimed to reconstruct deep salinity fields using Sea Surface Temperature (SST) or Sea Surface Height (SSH), alongside sparse in-situ salinity observations. The results showcased a significant reduction in reconstruction error up to 40% and the ability to resolve scales 50% smaller compared to baseline OI techniques. Remarkably, even though CLOINet was trained exclusively on simulated data, it accurately reconstructed an unseen SST field using only glider temperature observations and satellite chlorophyll concentration data. This demonstrates how deep learning networks like CLOINet can potentially lead the integration of modeling and observational efforts in developing an ocean digital twin.
... The horizontal resolution is 1/60°and it has 300 vertical levels. The model configuration was defined to resolve as accurately as possible the surface signature of oceanic motions of scales down to 15 km, in accordance with the expected effective resolution of SWOT observations (Ajayi et al., 2020). This model configuration was implemented and run in two different free-run simulations (no data assimilation): one experiment including explicit tidal forcing and the other experiment without tidal forcing. ...
... Several studies have used the non-extended version of this model, called NATL60 (e.g. Amores et al., 2018;Fresnay et al., 2018;Metref et al., 2019;Ajayi et al., 2020;Metref et al., 2020;Ajayi et al., 2021;Le Guillou et al., 2021). ...
The new Surface Water and Ocean Topography (SWOT) satellite mission aims to provide sea surface height (SSH) measurements in two dimensions along a wide-swath altimeter track with an expected effective resolution down to 15–30 km. In this context our goal is to optimize the design of in situ experiments aimed to reconstruct fine-scale ocean currents (~20 km), such as those that will be conducted to validate the first available tranche of SWOT data. A set of Observing System Simulation Experiments are developed to evaluate different sampling strategies and their impact on the reconstruction of fine-scale sea level and surface ocean velocities. The analysis focuses (i) within a swath of SWOT on the western Mediterranean Sea and (ii) within a SWOT crossover on the subpolar northwest Atlantic. From this evaluation we provide recommendations for the design of in situ experiments that share the same objective. In both regions of study distinct strategies provide reconstructions similar to the ocean truth, especially those consisting of rosette Conductivity Temperature Depth (CTD) casts down to 1000 m and separated by a range of distances between 5 and 15 km. A good compromise considering the advantages of each configuration is the reference design, consisting of CTD casts down to 1000 m and 10 km apart. Faster alternative strategies in the Mediterranean comprise: (i) CTD casts down to 500 m and separated by 10 km and (ii) an underway CTD with a horizontal spacing between profiles of 6 km and a vertical extension of 500 m. In the Atlantic, the geostrophic velocities reconstructed from strategies that only sample the upper 500 m depth have a maximum magnitude ~50% smaller than the ocean truth. A configuration not appropriate for our objective in both regions is the strategy consisting of an underway CTD sampling one profile every 2.5 km and down to 200 m. This suggests that the thermocline and halocline need to be sampled to reconstruct the geostrophic flow at the upper layer. Concerning seasonality, the reference configuration is a design that provides reconstructions similar to the ocean truth in both regions for the period evaluated in summer and also in winter in the Mediterranean.
... Initial conditions were taken from the 1/12 • MERCATOR ocean reanalysis. The output of NATL60 was analyzed in Ajayi et al. (2020). In the present research, we studied the Gulf-stream region (26 • N, 45 • N; 40 • W, 65 • W, Fig. 1) which is the most energetic region of the North Atlantic Ocean. ...
... It relies on an Observing System Simulation Experiment (OSSE) for nadir and wide-swath satellite altimetry data. We exploit one-year NATL60 numerical simulation dataset [3] for a 10 • x10 • area along the Gulf Stream from October 2012 to September 2013 with a daily time resolution and a 1/20 • spatial resolution. Nadir altimetry data involves the space-time sampling of a real 4-altimeter configuration, whereas wide-swath SWOT data rely on SWOT simulator [25]. ...
Due to the irregular space-time sampling of sea surface observations, the reconstruction of sea surface dynamics is a challenging inverse problem. While satellite altimetry provides a direct observation of the sea surface height (SSH), which relates to the divergence-free component of sea surface currents, the associated sampling pattern prevents from retrieving fine-scale sea surface dynamics, typically below a 10-day time scale. By contrast, other satellite sensors provide higher-resolution observations of sea surface tracers such as sea surface temperature (SST). Multimodal inversion schemes then arise as an appealing strategy. Though theoretical evidence supports the existence of an explicit relationship between sea surface temperature and sea surface dynamics under specific dynamical regimes, the generalization to the variety of upper ocean dynamical regimes is complex. Here, we investigate this issue from a physics-informed learning perspective. We introduce a trainable multimodal inversion scheme for the reconstruction of sea surface dynamics from multi-source satellite-derived observations. The proposed 4DVarNet schemes combine a variational formulation involving trainable observation and a priori terms with a trainable gradient-based solver. We report an application to the reconstruction of the divergence-free component of sea surface dynamics from satellite-derived SSH and SST data. An observing system simulation experiment for a Gulf Stream region supports the relevance of our approach compared with state-of-the-art schemes. We report relative improvement greater than 50% compared with the operational altimetry product in terms of root mean square error and resolved space-time scales. We discuss further the application and extension of the proposed approach for the reconstruction and forecasting of geophysical dynamics from irregularly-sampled satellite observations.
... If computational limits do not exist, at which model resolution should one stop? Recently, 0.02 • horizontal resolution models have been shown to substantially improve ocean variability, western boundary currents (Ajayi et al., 2020;Chassignet & Xu, 2021) and the transport of heat (Su et al., 2018) compared to 0.1 • resolution models in the present-day. Since proxy data exhibit large uncertainty and are unlikely to include detailed ocean variability, a higher model resolution should substantially improve large-scale features in order to reduce model-data mismatches (e.g. the East Australia Current that substantially transports more heat southeastward in the eddying compared to the non-eddying simulation; chapter 6). ...
Anthropogenic climate change is one of the greatest challenges for humanity today. How does the Earth system react to the atmospheric greenhouse gas increase that has never happened before with such speed in Earth's history? So far, climate models project an increase of the global average temperature, sea level and the number of `extremely' warm days in the coming decades and centuries. Although it is clear that the Earth is warming, uncertainty remains in the progression of those developments. These climate models are inherently wrong, but are they useful? Climate models give a proper representation of present-day climate, which we know from validations with observations. These models, which contain fundamental physical processes and have been developed over the past decades, cannot be validated in the increased greenhouse-climate of the future. We can however, compare the models with observations of warmer climates in the past, which are similar to the future climate, to get an understanding of these type of `extreme' climates. Although (unfortunately) no direct observations are available for the past hundred millions of years, we do find indirect evidence about the past climate (for example fossils or ice cores). These so-called proxies provide an archive of past climate and can be used to compare with climate model simulations. As a result, the combination of models and proxies of past climate can be used to get a better understanding of how a future climate, which is warmer than today, may look like. A primary part of the Earth's archive to reconstruct past climates is provided by marine sediments, consisting of (fossil remains from) microplankton. The microplankton species in the bottom sediments originated from a location close to the ocean surface before they started sinking to the bottom. Hence, microplankton at the ocean bottom is representative of the ocean surface environment. It is often assumed that these planktonic species sunk vertically downwards. However, the microplankton is transported laterally by ocean currents during its sinking journey. In this thesis, we investigate how sedimentary distributions of microplankton can be explained. We determine how sinking microplankton is advected by ocean currents, which may have great implications for the interpretation of sedimentary microplankton data. For example, subtropical and (sub)polar microplankton species alternate in sediment cores near Antarctica from 34 million years ago until the present-day. If subtropical microplankton species are found near Antarctica in a specific time period, two hypotheses can be tested: (a) Antarctica had a subtropical climate, or (b) Antarctica was not subtropical, but the microplankton were transported laterally by ocean currents and originated from another region with a subtropical climate. We study microplankton particles at the ocean bottom, which got there after a sinking journey, and determine their origin at the ocean surface back again. The ultimate goal is to bridge a gap between the models, which represent the global climate, and the measurements, representing the climate at specific geographic locations. As such, we study past climates back again, to get an idea how we get there in the future.
... Torres et al. (2019) make use of a spatial scale threshold above which BMs dominate and below which ITs dominate. This spectral technique is less effective when the common spatial scale interval extends too much, which is particularly the case in winter due to the emergence of small vortices (<50 km) associated with mixed layer instabilities (Ajayi et al., 2020). Gonzalez-Haro et al. (2019) and Ponte et al. (2017) explore multi-sensor approaches with altimetry and sea surface temperature observations, motivated by the fact that BMs and ITs have distinct signature on both fields. ...
Wide‐swath altimetry, for example, the Surface Water and Ocean Topography mission is expected to provide Sea Surface Height (SSH) measurements resolving scales of a few tens of kilometers. Over a large fraction of the globe, the SSH signal at these scales is essentially a superposition of a component due to balanced motions (BMs) and another component due to internal tides (ITs). Several oceanographic applications require the separation of these components and their mapping on regular grids. For that purpose, the paper introduces an alternating minimization algorithm that iteratively implements two data assimilation techniques, each specific to the mapping of one component: a quasi‐geostrophic model with Back‐and‐Forth Nudging for BMs, and a linear shallow‐water model with 4‐Dimensional Variational assimilation for ITs. The algorithm is tested with Observation System Simulation Experiments where the truth is provided by a primitive‐equation ocean model in an idealized configuration simulating a turbulent jet and mode‐one ITs. The algorithm reconstructs almost 80% of the variance of BMs and ITs, the remaining 20% being mostly due to dynamics that cannot be described by the simple models used. Importantly, in addition to the reconstruction of stationary ITs, the amplitude and phase of nonstationary ITs are reconstructed. Sensitivity experiments show that the quality of reconstruction significantly depends upon the timing of observations. Although idealized, this study represents a step forward towards the disentanglement of BMs and ITs signals from real wide‐swath altimetry data.
