Mesoscale eddies, small-scale swirling motions in the ocean, are vital for transporting heat, salt, carbon, and nutrients, influencing air-sea feedbacks and impacting Earth's climate and ecosystems. While studies have shown increased eddy activity in many ocean regions, the Arctic, experiencing rapid warming and sea ice decline, remains poorly understood. Existing climate models struggle to resolve the small-scale Arctic eddies (approximately 10 km or less), hindering projections of future changes. This lack of understanding is particularly concerning given the crucial role eddies play in Arctic halocline properties and marine ecosystems, influencing water mass transport between shelves and the deep basin, and potentially affecting freshwater storage and circulation in the Beaufort Gyre. Existing observational data is limited, leading to conflicting hypotheses about current and future eddy activity changes in the Arctic. This study aims to bridge this knowledge gap by employing a high-resolution model to simulate Arctic eddy activity under a warming climate, providing valuable insights into potential changes and their consequences.

Previous research highlights the importance of mesoscale eddies in global ocean dynamics and their role in various biogeochemical cycles. Satellite altimetry data reveals a global increase in ocean eddy activity in recent decades. High-resolution climate model simulations project continued intensification of surface eddy kinetic energy (EKE) in many regions due to future warming, except for the North Atlantic. However, understanding future changes in Arctic eddy activity remains limited due to challenges in simulating the small-scale eddies found in this region. While observations suggest eddies play a crucial role in Arctic halocline properties and marine ecosystems, they are sparse and lead to conflicting interpretations of past changes. Some studies indicate a strengthening of eddy activity in the western Arctic due to increased freshwater content, while others suggest a relatively small increase in the Beaufort Gyre. The interplay between sea ice decline and eddy activity, with sea ice potentially dampening near-surface eddies, remains a key area of investigation.

This study utilized the Finite-Volume Sea Ice-Ocean Model version 2 (FESOM2) to conduct simulations at two different Arctic resolutions: eddy-present (4.5 km) and eddy-rich (1 km). The eddy-rich simulation, with 1 km resolution in the Arctic and 30 km elsewhere, better resolves mesoscale eddies. The model was forced using atmospheric forcing and river runoff derived from the CMIP6 simulation of the AWI-CM coupled climate model. Two simulations were performed: one eddy-present simulation running from 1958 to 2100 and three time slices from this simulation (2010-2015, 2050-2055, and 2090-2095) used to initialize higher resolution eddy-rich simulations. Eddy kinetic energy (EKE) was calculated using both Reynolds averaging and a coarse-graining method to account for high-frequency wind variability influencing surface currents. Eddy killing, the damping of eddies by wind and sea ice, was assessed using a spatial averaging approach. The energy conversion rate from eddy available potential energy to EKE (Tbc) was calculated using Reynolds averaging. The study also analyzed the energy spectrum to estimate eddy spatial scales and considered the impact of constant ice-ocean drag coefficients and the lack of atmosphere-ocean dynamic coupling on the results.

The study's key findings show a dramatic increase in Arctic Ocean EKE in a four-degree-warmer world by the end of the 21st century. The mean EKE in the upper 200 m tripled, with a significant trend of (1.39 ± 0.07) × 10³ m²/s² per decade in the eddy-rich simulations. This increase is more pronounced in the upper 100 m than in the 100–200 m layer, and is larger in the surface mixed layer than above the halocline base. The normalized surface EKE increased to approximately 15 times the standard deviation of historical surface EKE in both simulations. The regions with high EKE in the current climate experienced the most significant increases, indicating enhanced eddy generation. The increase in EKE is primarily attributed to enhanced baroclinic instability, leading to increased Tbc. Despite sea ice decline, eddy killing was found to strengthen during the ice-covered months (February-May) due to intensified eddy activity, but weakens from August to November because of the near absence of sea ice. The annual average eddy killing increased where EKE became higher, indicating that increased eddy generation overshadows changes in wind speeds as the driving factor behind eddy killing. A regime shift in Arctic Ocean eddies is projected; under the warming scenario, the ratio of surface eddy power to Tbc decreases, suggesting a transition toward a midlatitude-like regime where wind, rather than sea ice, dominates eddy killing. Seasonal sea ice coverage remains a dominant factor influencing surface eddy activity, suppressing surface eddies during ice-covered periods, despite enhanced eddy activity at depth.

The findings demonstrate a substantial and transformative increase in Arctic Ocean eddy activity under a warming climate, surpassing that projected for other eddy-rich regions. The increase in eddy generation due to enhanced baroclinic instability, driven by freshwater accumulation and reduced sea ice, is the primary driver of this surge. The persistence of eddy killing, particularly during ice-covered periods, suggests the continued influence of sea ice friction on surface eddy activity. The projected intensification of eddy transports of heat, carbon, and nutrients, combined with altered air-sea gas exchange, highlights the potential for significant impacts on the Arctic climate and marine ecosystems. This underscores the need for improved representation of Arctic eddy activity in climate models.

This study projects a tripling of upper Arctic Ocean eddy kinetic energy in a four-degree-warmer world, driven by increased baroclinic instability. While eddy killing remains significant, particularly during sea ice-covered months, enhanced eddy generation outweighs its effect. Accurate representation of Arctic eddies in climate models is critical for understanding the impact of this transformative change on the Arctic climate and ecosystem. Future research should focus on reducing model uncertainties by improving the representation of ice-ocean drag coefficients and incorporating atmosphere-ocean dynamic coupling.

The study's simulations used constant ice-ocean drag coefficients and lacked atmosphere-ocean dynamic coupling, potentially underestimating future increases in eddy activity. The relatively short length of the high-resolution simulations may also introduce uncertainties. These limitations should be addressed in future research.