Satellite observations of the last two decades have led to a major breakthrough emphasizing the existence of a strongly energetic mesoscale turbulent eddy field in all the oceans. This ocean mesoscale turbulence is characterized by cyclonic and anticyclonic eddies (with a 100‐ to 300‐km size and depth scales of ∼500–1,000 m) that capture approximatively 80% of the total kinetic energy and is now known to significantly impact the large‐scale ocean circulation, the ocean's carbon storage, the air‐sea interactions, and therefore the Earth climate as a whole. However, ocean mesoscale turbulence revealed by satellite observations has properties that differ from those related to classical geostrophic turbulence theories. In the last decade, a large number of theoretical and numerical studies has pointed to submesoscale surface fronts (1–50 km), not resolved by satellite altimeters, as the key suspect explaining these discrepancies. Submesoscale surface fronts have been shown to impact mesoscale eddies and the large‐scale ocean circulation in counterintuitive ways, leading in particular to up‐gradient fluxes. The ocean engine is now known to involve energetic scale interactions, over a much broader range of scales than expected one decade ago, from 1 to 5,000 km. New space observations with higher spatial resolution are however needed to validate and improve these recent theoretical and numerical results.

Satellite altimeters provide global observations of sea surface height (SSH) and present a unique data set for advancing our theoretical understanding of upper‐ocean dynamics and monitoring its variability. Considering that mesoscale SSH patterns can evolve on timescales comparable to or shorter than satellite return periods, it is challenging to accurately reconstruct the continuous SSH evolution as currently available altimetry observations are still spatially and temporally sparse. Here we explore the possibility of SSH interpolation via Deep Learning by using synthetic observations from an idealized quasigeostrophic model of baroclinic ocean turbulence. We demonstrate that Convolutional Neural Networks with Residual Learning are superior in SSH reconstruction to linear and recently developed dynamical interpolation techniques. Also, the deep neural networks can provide a skillful state estimate of unobserved deep ocean currents at mesoscales. These conspicuous results suggest that SSH patterns of eddies might contain substantial information about the underlying deep ocean currents that are necessary for SSH prediction. Our training data are focused on highly idealized physics and diversification of processes needs to be considered to more accurately represent the real ocean. In addition, methodological improvements such as transfer learning and implementation of dynamically aware loss functions might be necessary to consider before its ultimate use with real satellite observations. Nonetheless, by providing a proof of concept based on synthetic data, our results point to deep learning as a viable alternative to existing interpolation and, more generally, state estimation methods for satellite observations of eddying currents.

The area west of the Kerguelen Islands (20–70°E/45–60°S) is characterized by a weak mesoscale activity except for a standing meander region of the Antarctic Circumpolar Current (ACC) localized between 20 and 40°E. A unique bio-physical dataset at high-resolution collected by a southern elephant seal (Mirounga leonina) reveals a conspicuous increase in foraging activity at the standing meander site up to 5 times larger than during the rest of her three-month trip west of the Kerguelen Islands. Here, we propose a physical explanation for such high biological activity based on the study of small-scale fronts with scales of 5 to 20 km, also called submesoscales. The standing meander is associated with intensified frontal dynamics at submesoscale, not observed in the rest of the region. Results shed new light on the spatial distribution of submesoscale fronts in the under-sampled area west of the Kerguelen plateau and emphasize their importance for upper trophic levels. Despite that most elephant seals target foraging grounds east of the Kerguelen Plateau, our findings suggest that excursions to the west are not accidental, and may be explained by the recurrently elevated physical and biological activity of the site. As such, other standing meanders of the ACC may also act as biological hotspots where trophic interactions are stimulated by submesoscale turbulence.

Jupiter’s atmosphere is one of the most turbulent places in the solar system. Whereas observations of lightning and thunderstorms point to moist convection as a small-scale energy source for Jupiter’s large-scale vortices and zonal jets, this has never been demonstrated due to the coarse resolution of pre-Juno measurements. The Juno spacecraft discovered that Jovian high latitudes host a cluster of large cyclones with diameter of around 5,000 km, each associated with intermediate- (roughly between 500 and 1,600 km) and smaller-scale vortices and filaments of around 100 km. Here, we analyse infrared images from Juno with a high resolution of 10 km. We unveil a dynamical regime associated with a significant energy source of convective origin that peaks at 100 km scales and in which energy gets subsequently transferred upscale to the large circumpolar and polar cyclones. Although this energy route has never been observed on another planet, it is surprisingly consistent with idealized studies of rapidly rotating Rayleigh–Bénard convection, lending theoretical support to our analyses. This energy route is expected to enhance the heat transfer from Jupiter’s hot interior to its troposphere and may also be relevant to the Earth’s atmosphere, helping us better understand the dynamics of our own planet.

Submesoscale ocean processes, characterized by order-1 Rossby and Richardson numbers, are currently thought to be mainly confined to the ocean surface mixed layer, whereas the ocean interior is commonly assumed to be in quasigeostrophic equilibrium. Here, a realistic numerical simulation in the Antarctic Circumpolar Current, with a 1/48° horizontal resolution and tidal forcing, is used to demonstrate that the ocean interior departs from the quasigeostrophic regime down to depths of 900 m, that is, well below the mixed layer. Results highlight that, contrary to the classical paradigm, the ocean interior is strongly ageostrophic, with a pronounced cyclone–anticyclone asymmetry and a dominance of frontogenesis over frontolysis. Numerous vortices and filaments, from the surface down to 900 m, are characterized by large Rossby and low Richardson numbers, strong lateral gradients of buoyancy, and vigorous ageostrophic frontogenesis. These deep submesoscales fronts are only weakly affected by internal gravity waves and drive intense upward vertical heat fluxes, consistent with recent observations in the Antarctic Circumpolar Current and the Gulf Stream. As such, deep submesoscale fronts are an efficient pathway for the transport of heat from the ocean interior to the surface, suggesting the presence of an intensified oceanic restratification at depth.