Storms Stir Hidden Ocean Eddies Deep Beneath the Baltic Sea

Deep beneath the waves of the Baltic Sea, far below the familiar churn of wind and waves, scientists have discovered a hidden world of swirling, short-lived currents that spring to life every time a storm sweeps overhead. These features, known as submesoscales, span from a few hundred meters to about ten kilometers across and survive for only hours to days. Oceanographers have long studied their counterparts near the sea surface, where they appear as delicate fronts and filaments visible from satellites. But the new research, published in the journal Ocean Science, reveals that these energetic motions are just as abundant in the dark, stratified depths of a semi-enclosed basin, where they are generated not by distant ocean currents but by the interplay of storm-driven winds and the sloping seafloor itself.

The study, led by Evridiki Chrysagi of the University of Hamburg together with colleagues at the Leibniz Institute for Baltic Sea Research Warnemünde and the University of Southampton, used high-resolution numerical simulations of the central Baltic Sea to peer into a layer of the ocean that is notoriously difficult to observe. The team focused on the Eastern Gotland Basin, the largest basin of the Baltic Proper, which reaches depths of roughly 250 meters and is capped by a permanent halocline centered near 70 meters. This sharp salinity gradient acts as a natural lid, isolating the oxygen-poor deep waters from the surface, and it means that processes acting at the basin’s boundaries, rather than at the surface, largely control how the deep water is mixed and ventilated.

What makes the Baltic Sea an ideal natural laboratory for this kind of work is what it lacks. Unlike the open ocean, where powerful, persistent currents such as the Gulf Stream flow over steep continental slopes and generate submesoscale turbulence in their wakes, the Baltic has virtually no tides and no major current systems. Its circulation is instead driven by ephemeral winds that change direction from one week to the next. Previous studies of deep-ocean submesoscales had assumed that strong, steady currents were a prerequisite for their formation. The new simulations demonstrate that this is not the case: even in a basin dominated by transient, wind-driven currents, submesoscale features flourish in the stratified interior below the pycnocline, particularly near the lateral boundaries where the water meets sloping topography.

The dynamical key to these features lies in a quantity called the Rossby number, which compares the local spin of the flow, known as relative vorticity, to the rotation of the Earth. When the Rossby number approaches one, the flow is said to be ageostrophic, meaning it is only weakly constrained by planetary rotation, and this is the hallmark of submesoscale dynamics. The simulations showed that during storms, regions with Rossby numbers of order one were ubiquitous not only in the surface mixed layer but also at depths of 80 and 100 meters, especially close to the basin margins. Crucially, the patterns at depth bore no spatial resemblance to those at the surface, indicating that the deep features are generated by an entirely independent mechanism, most likely through the frictional interaction of along-slope currents with the bottom boundary layer.

That mechanism is elegantly simple. When a current flows along a sloping bottom, the drag it experiences against the slope generates not only vertical shear but also significant horizontal shear, which manifests as vertical vorticity. The sign of this vorticity depends on the direction of the flow relative to the coast: a current flowing with the slope on its left acquires cyclonic, or positively signed, vorticity, while a current flowing the other way acquires anticyclonic, negatively signed vorticity. In the Baltic simulations, the western and eastern margins of the basin typically displayed opposite signs of vorticity, reflecting the direction of the interior flow. As storm-intensified currents swept along the boundaries, these vorticity anomalies grew stronger, and in some cases boundary currents separated from the slope and rolled up into eddies, exporting their anomalous spin into the stratified interior in the form of fronts, filaments, and coherent vortices.

The researchers traced this boundary-to-interior exchange using a passive tracer released into the bottom boundary layer sixteen days before the study period. The tracer, initially confined to the near-bottom waters, was captured by the ageostrophic features and carried away from the boundaries into the basin interior, its dispersion closely mirroring the patterns of sharp horizontal buoyancy gradients. The simulations also revealed anomalies in Ertel potential vorticity, a conserved quantity that combines the flow’s spin with its stratification. Because potential vorticity is nearly conserved in the weakly turbulent interior, these anomalies served as fingerprints of fluid that had detached from the boundary layer, providing direct evidence that waters mixed at the margins are actively exchanged with the interior, a process with major implications for how the Baltic’s oxygen-depleted deep waters are ventilated.

The most striking finding emerged from the analysis of three successive storms that battered the basin in October 2017, with winds exceeding 13 meters per second and peaking near 20 meters per second. Because the storms arrived from different directions, first from the south-southwest, then from the south, and finally from the north, they repeatedly reversed the direction of the along-slope currents. Each reversal flipped the sign of the vorticity generated at the boundaries, so that the submesoscale generation sites and the associated mixing hotspots alternated between the eastern and western sides of the basin. Cross-correlation analysis confirmed the tight coupling: the Rossby number at the eastern boundary correlated strongly with the along-slope current at zero lag, and the wind field led the interior current response by roughly 13 to 16 hours, a timescale comparable to the local inertial period.

Storms did more than relocate the hotspots; they intensified them. The fraction of the boundary region occupied by strong ageostrophic features, defined by Rossby numbers exceeding 0.5, climbed to approximately 27 percent during the final storm episode. Meanwhile, the fraction of the near-bottom zone with negative potential vorticity, which signals conditions favorable to overturning instabilities, reached nearly 30 percent at the eastern boundary during the same event. Among the possible instability types, symmetric instability dominated throughout the study period, satisfying its criteria over roughly a quarter of the eastern boundary during and after the last storm, while centrifugal instability remained rare, likely because the basin lacks the steep bathymetric slopes that favor it. These instabilities drive vigorous turbulent mixing, and the simulated energy dissipation rates along the boundary intensified substantially during storms, coinciding with the low-potential-vorticity regions.

The broader significance of this work extends well beyond the Baltic. Wind-driven current reversals of the kind that shuffle mixing hotspots around the basin are generally absent from the great current systems of the open ocean, where flow directions are steady, but they are expected in any marine or limnic system whose circulation is governed by transient winds. The authors suggest that the same storm-modulated submesoscale dynamics should operate in other semi-enclosed seas such as the Black Sea and the Caspian Sea, and in large lakes including the North American Great Lakes and Lake Geneva, where surface submesoscales have already been observed and gliders have detected strong boundary mixing in the interior. By connecting the frictional generation of vorticity at sloping boundaries with the export of eddies, fronts, and filaments into the stratified interior, the study provides a missing link between observations of enhanced boundary mixing in the Baltic and the recognition that boundary processes control basin-scale deep-water mixing. In doing so, it reveals that intense and variable winds over sloping topography can be a remarkably effective engine for near-boundary turbulence, energy dissipation, and the exchange of waters between the margins and the interior, calling for a re-evaluation of how deep waters are ventilated in wind-driven basins around the world.

Subject of Research: Storm-modulated submesoscale ocean dynamics in the deep layers of the wind-driven, non-tidal Baltic Sea

Article Title: Storm-modulated submesoscale dynamics over sloping topography in a wind-driven, non-tidal basin

Article References: Storm-modulated submesoscale dynamics over sloping topography in a wind-driven, non-tidal basin. (n.d.). https://doi.org/10.5194/os-22-3037-2026

Image Credits: AI Generated

DOI: 10.5194/os-22-3037-2026

Keywords: submesoscale, Baltic Sea, ocean dynamics, bottom boundary layer, storms, vorticity, potential vorticity, symmetric instability, boundary mixing, wind-driven circulation, sloping topography, numerical simulation