Wave-ice interaction strengthens eddy activity in Fram Strait

Liu, Yi; Yang, Yunhan; Dai, Haijin; Zhao, Jun; Li, Qiang

doi:10.5194/egusphere-2026-153

Preprints

https://doi.org/10.5194/egusphere-2026-153

Preprints

09 Apr 2026

| 09 Apr 2026

Status: this preprint is open for discussion and under review for The Cryosphere (TC).

Wave-ice interaction strengthens eddy activity in Fram Strait

Yi Liu, Yunhan Yang, Haijin Dai, Jun Zhao, and Qiang Li

Abstract. A mesoscale eddy was detected by satellite in the MIZ of Fram Strait and verified by reference to the barotropic instability of the East Greenland Current (EGC). According to the reanalysis data, the eddy originated from a mother eddy that grew and diminished in a branch of EGC during the summer. After September 12, this branch of the EGC strengthened and became equally strong as the main body of the EGC by the end of September. As a result, the eddy grew into a strong mesoscale eddy, which was captured on October 4 by satellite. The strengthening of the branch may be attributed to the influence of wave-ice interactions. In September, sea ice expanded toward the open ocean as a result of the seasonal cycle and covered the branch of the EGC. Wave-ice interactions and eddy genesis were revealed by numerical simulations. When waves propagated into the ice zone, they dissipated quickly at the ice edge and produce an ice-edge jet, thus strengthening the background flow. The resulting enhanced barotropic instability helped small turbulence grow into large eddies. During the same period, an ocean front grew due to ice formation and dense water sinking, thus indicating that baroclinic instability may not play an important role in eddy genesis.

Received: 12 Jan 2026 – Discussion started: 09 Apr 2026

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Yi Liu, Yunhan Yang, Haijin Dai, Jun Zhao, and Qiang Li

Status: open (until 28 Jun 2026)

Post a comment Subscribe to comment alert

RC1:
'Comment on egusphere-2026-153', Anonymous Referee #1, 22 May 2026 reply
General comments:
This manuscript addresses a potentially important mechanism for eddy generation at the edge of the ice zone (namely, wave-sea ice interactions), and the combination of satellite observations, reanalysis products, and idealized modeling would have the potential to make for a valuable contribution.
However, in its current form the paper is difficult to evaluate scientifically because the presentation lacks sufficient clarity, structure, and methodological precision. A central issue is organization: the manuscript needs much clearer signposting at the outset so that the reader can distinguish between what is inferred from observations and reanalysis, what is being tested in the idealized model, and how these different lines of evidence support the main argument. Related to this, the descriptions of the datasets, model framework, assumptions, parameter choices, and excluded mechanisms are often too incomplete to assess reproducibility or interpretation with confidence. Several scientific statements also need to be made more precise, including terminology around mesoscale versus submesoscale dynamics, process attribution, and the distinction between suggestive consistency and direct mechanistic proof. In addition, the figures and captions should be revised to make the evolution of the eddy and the logic of the experiments easier to follow.
But perhaps more fundamentally, the current evidence appears insufficient to establish the proposed wave–ice interaction pathway as the definitive mechanism for the observed eddy. The observations, reanalysis, and idealized experiments are certainly suggestive and point in an interesting direction, but at present they demonstrate consistency with the mechanism more than they provide direct process attribution or mechanistic proof.
To establish a stronger mechanistic case, I think the manuscript would need evidence that more directly resolves the full causal chain from wave forcing to ice-edge jet formation and then to eddy growth via barotropic instability. In practice, this would likely require some combination of independent current observations at the ice edge, quantitative instability diagnostics, explicit tests against plausible alternative mechanisms, and ideally an event-based modeling framework that can reproduce the observed timing, location, and evolution of the eddy under realistically constrained conditions. In particular, what is currently missing is a demonstration not just that the observations and simulations are consistent with the proposed mechanism, but that competing explanations can be ruled out and that the inferred jet and its instability are directly supported by data rather than primarily by reanalysis interpretation.
I thought the paper very interesting, but there are too many things which bring down the quality and obfuscate the science. I recommend rejection but would encourage the authors to resubmit with revisions.

Specific comments:
Text from the manuscript will be shown like this.
Reviewer comments will be shown below like this.

When waves propagated into the ice zone, they dissipated quickly at the ice edge and produce an ice-edge jet, thus strengthening the background flow
Language: tense mismatch between “dissipated”, “produce” and “strengthening”.

The Fram Strait [64°N-82°N, 30°W-20°E], which represents the only channel for deep water exchange between the Arctic and other regions, is located between Greenland and the Svalbard Archipelago. On the western side, the East Greenland Current (EGC) originates from Fram Strait and moves equatorward until it reaches Cape Farewell at a speed of 0.3-0.5 cm/s. On the eastern side, the Norwegian Atlantic Slope Current (NwASC) and Norwegian Atlantic Front Current (NwAFC) move poleward anticyclonically and cyclonically, respectively. The NwASC and the NwAFC merge into the West Spitsbergen Current before passing through Fram Strait and the inflow from the Atlantic Ocean. EGC moves equatorward in the whole layer,
A diagram would be helpful for setting up the different currents in the reader’s mind

Ocean eddies are also called mesoscale eddies (Zhang et al 2016) or submesoscale eddies.
Consider rephrasing to something like “Ocean eddies, depending on their size, are also called mesoscale eddies (Zhang et al 2016) or submesoscale eddies.”

Ocean eddies are typically larger at mid–low latitudes than at high latitudes due to the reduced Rossby radius (Manley and Hunkins 1985), which leads to submesoscale eddies appearing more frequently at high latitudes (Iakovlev 2018)
This second clause is inaccurate.
The chapter you quoted is actually much more nuanced than the manuscript summary. In particular, the chapter is not saying simply:
“high latitudes have more submesoscale eddies because the Rossby radius is smaller.”
What it is saying:
the mesoscale shrinks dramatically in the Arctic because (L_R) is small;

observed submesoscale eddies are only slightly smaller than mesoscale eddies there;

therefore the mesoscale/submesoscale separation becomes blurry (“gray zone”).

The key sentence:
“the spatial scales of the observed submesoscale eddies at high latitudes are close to the mesoscale eddies (2 and 5 km, respectively). Hence, these two processes cannot be separated…”
That is very different from saying submesoscale eddies “appear more frequently.”
“submesoscale” is a relative/dynamical category tied to the local deformation radius.
The chapter defines mesoscale through the Rossby deformation radius:
R_d = NH / f
and then defines submesoscale dynamically through order-one Rossby number:
Ro = U / fL ~ O(1)
not through a universal horizontal length scale.
So the logic is:
Arctic mesoscale becomes very small because (R_d) is small.

Submesoscale phenomena therefore overlap strongly with mesoscale scales.

This creates a modeling “gray zone.”

That is a scale-collapse argument, not necessarily an occurrence-frequency argument.
Something like the following would be more accurate
“At high latitudes, the reduced Rossby deformation radius shifts mesoscale variability toward smaller scales, such that motions classified as submesoscale constitute a larger fraction of the energetic spectrum.”
most Arctic eddies occur in the Chukchi–Beaufort Sea, Fram Strait and Barents-Kara Sea.
Probably better phrasing would be “Chukchi–Beaufort Sea, Fram Strait and Barents-Kara Sea have the highest occurrence of Arctic eddies”

Among them, Dai et al. (2019) proposed a mechanism of barotropic instability whereby wave-ice interactions generate along-ice-edge jets and mesoscale eddies.
Are you saying that this is an additional one proposed which was not considered in the original classification of Johannnesen et al. (1987)? Or is at a variation on the third mechanism “(3) Horizontal velocity shear, which occurs between the jet and surrounding flow, produces eddies via the barotropic instability mechanism (Liu et al 1993, Dai et al 2019)”
More generally to this paragraph, it feels a bit anachronistic how the referencing is done here. If you are reproducing the classification of Johannnesen et al. (1987), I would say either do no additional references therein, or only use the references that the original paper provides. Then, if you would like to introduce additional references (such as Dai et al 2019), do it after the initial classification.

Unfortunately, observational evidence to support this mechanism remains scarce. This raises two critical questions: (2) Can the entire process be observed in the real world? (3) What are the limiting factors for the occurrence of this mechanism? To solve the three issues above,
Is number one “observational evidence to support this mechanism remains scarce”. If so, annotate accordingly

Section 3 describes the eddy genesis process
Is this theory/results? specify

Usually, in situ photography and/or satellite detection can be used to identify surface eddies since sea ice moves following (anti)cyclonic flows. However, eddies covered by frozen ice or occurring in the subsurface (Hunkins 1974, Dai et al. 2025) can be detected only by mooring. In this study, we investigate a surface eddy detected in the Fram Strait, where the topography is relatively flat and the water depth is greater than 250 m. The topographic controlled mechanism and nonlinear advection mechanism are excluded in this study.
Feels strange to have this after your sign posting. I suggest either do the sign posting after this, or move this to the methodology part
Additionally, to the point of structural clarity: later on I got lost between the various methods. This needs to be signed posted more clearly at the start. For example, first we will look at ocean and wave reanalysis data to get an idea of the governing conditions of the eddy formation. Then we will test in an idealized setup.

Usually, in situ photography and/or satellite detection can be used to identify surface eddies since sea ice moves following (anti)cyclonic flows. However, eddies covered by frozen ice or occurring in the subsurface (Hunkins 1974, Dai et al. 2025) can be detected only by mooring.
It is unclear the role you are saying sea ice should play here. Is it the following? when there is partial covering of the ocean by sea ice, then the eddy can be well detected by motion of the sea ice, but when the ocean surface is fully covered by sea ice then you cannot?
The first part also does not really feel accurate that in situ photography is used to identify surface eddies. Firstly, it’s the time dependent motion tracking of the sea ice (which acts as a passive tracer) across hours-days; Secondly, in situ photography is rarely used for this (at least to my knowledge), and I believe remote sensing would be much more common.
Lastly, whilst it is true that subsurface eddies are hard to observe, and you generally need in situ or indirect 3D information it is then a very strong statement to say that they can only be detected by mooring (this is just one observational technique in as of many, others that could possibly measure subsurface eddies include autonomous platforms [e.g. ARGO, gliders] or hydrographic sections)

In this study, we investigate a surface eddy detected in the Fram Strait, where the topography is relatively flat and the water depth is greater than 250 m. The topographic controlled mechanism and nonlinear advection mechanism are excluded in this study.
Ok, so because of flat topography you are eliminating the topographic controlled mechanism, but what is your reason for not considering the nonlinear advection mechanism?

To investigate the mechanism underlying the genesis of the submesoscale eddy detected by SAR imagery and characterize its evolving environmental context, oceanic reanalysis data are employed to demonstrate how the eddy and the corresponding environment evolved. Daily oceanic reanalysis data obtained from the Mercator Center include temperature, current, salinity, and sea ice data with a horizontal resolution of 1/12˚ and are used to demonstrate eddy evolution.
I wouldn’t expect an ocean reanalysis to capture any specific eddy? They get general eddy statistics correct, but not individual eddies. Furthermore, I wouldn’t expect them to get the eddie correct in such complex conditions (apart from not having the physics within there, there is simply not enough data to constrain them).

This dataset is generated by a global ocean eddy-resolving model (Copernicus Marine Environment Monitoring Service, CMEMS), which is driven by atmospheric forcing (ERA-Interim and ERA5) and assimilates observations via a reduced-order Kalman filter.
Insufficient information given here. CMEMS is the access portal, but we do not know the product. One can infer that it is most likely GLORYS12, but this should be stated, and the reference should be given (and also the version). It is unlikely that it is driven by both ERA-Interim and ERA5 (ERA5 is the successor of ERA-Interim), and if so, please specify why or how.

Wave reanalysis data are derived from a wave spectrum model (Météo-France wave model, MFWAM), with a horizontal resolution of 1/5°.
The use of the word “derived” here suggests an active process on the side of the authors in getting the data from the model. I suggest replacing it with “comes” from. Secondly, this is not enough information about the data. For example, hosting platform (Copernicus? CMEMS? Météo-France archive?), dataset name/version (e.g., global reanalysis vs forecast archive). Does it have wave-sea ice interactions in it? If so, which?
Hourly atmospheric reanalysis data obtained from the ECWMF (Hersbach et al. 2023) include surface wind information.
For this, you provide the following reference:

Hersbach, H., Bell, B., Berrisford, P., Biavati, G., Horányi, A., Muñoz Sabater, J., Nicolas, J., Peubey, C., Radu, R., Rozum, I., Schepers, D., Simmons, A., Soci, C., Dee, D. and Thépaut, J-N: ERA5 hourly data on single levels from 1940 to present. Copernicus Climate Change Service (C3S) Climate Data Store (CDS) [Dataset]. (Accessed on DD-MMM-YYYY)

This is ERA5 access via the CDS
ERA5 should be referenced in the following way:
Primary reanalysis system paper

Hersbach et al. (2020, revised later versions exist):
The ERA5 global reanalysis

published in Quarterly Journal of the Royal Meteorological Society

This is the main scientific reference for ERA5.
Dataset citation (CDS)

Then separately:
Hersbach et al. (CDS dataset citation, like 2023 version you showed)

used for reproducibility and data access

The in-text reference should be:
Hourly atmospheric reanalysis data from ERA5 (ECMWF; Hersbach et al., 2020; Copernicus Climate Change Service) provide surface wind information.
and then in references:
Hersbach, H. et al. (2020). The ERA5 global reanalysis. QJRMS.

Hersbach, H. et al. (CDS, 2023). ERA5 hourly data on single levels…

This should be done the first time you mentioned ERA5 in the paper (which is not here, but a few sentences earlier)

“Accessed on DD-MMM-YYYY”
This should not be there

In this study, we employed a coupled modeling framework to investigate eddy genesis. This framework integrates a sea ice model (Simple Sea Ice Model, SSIM; Dai et al. 2019), a wave model (Uchiyama et al. 2010), and an ocean model (Regional Ocean Modeling System, ROMS; Shchepetkin and McWilliams 2005).
Is this a framework that you integrated newly for this study, or is it a preexisting framework? If preexisting, please clarify before you go into the details of the individual components.

a wave model (Uchiyama et al. 2010)
Not enough information here, there are many different types of wave models.
From Uchiyama et al. (2010):
“In some realistic cases, this model is applied with k = kp the spectrum-peak wavenumber and [A=f(Hs)] the equivalent wave amplitude in terms of the wave energy E, the so-called significant wave height Hsig, or a wave height H commonly used in breaking parameterizations. The simplest extension from a monochromatic/spectrum-peak model to a multi-component model is based on superposition of components with spectrum G. For more general wave dynamics including nonlinear spectrum evolution and wind generation, a wave simulation model is used (e.g., SWAN; Booij et al., 1999) to provide G and w.”
In this manuscript, you end up taking the monochromatic assumption, and taking the relevant integrated parameter information k=kp from the MFWAM reanalysis rather than using the whole spectrum from MFWAM. This would be better signposted here.

Where k and k are the amplitude and the vector of the wave number, respectively; cg and cg are the amplitude and the vector
Something wrong with your notation here, k and k, cg and cg.

Figure 1.
Looks quite impressive that the ocean reanalysis actually captures this..! Does panel b match exactly the inner domain of panel a? Would be good to make this clearer and to have the inner domain of panel a much larger, and for the two to be of comparable width x height ratios so one can more easily compare by eye.
Same comment regarding each of the panels in figure 2.

The anticyclonic eddy is located on the northwestern side of the EGC (Fig. 1b, vector),
Unclear. Which vector?
Figures would be much clearer if you wrote the date above like “September 12, 2020” instead of 0912

At this point in time, it was not yet an eddy, since its vorticity (𝜁𝑓, Fig. 2a, shading) was only -0.15 and no more than 0.1 smaller than the surrounding vorticity.
Is this following some definition? If so, please provide the reference

After splitting, the eddies continued to grow from September 24 (Fig. 2d, shading) to October 4, 2020, with the second eddy strengthening (September 24 to October 4) and outgrowing the first (September 12 to 18).
I know you have the October 4 plot in figure 1, but for consistency and ease of following it by eye, I would recommend to include as another panel in figure 2

The consistency between the jet strength and the eddy strength suggested the contribution of barotropic instability during eddy evolution.
Needs more elaboration. This is seemingly a very important part of the paper

First, Mercator Center provide the wave information, although they seem to neglect the wave in the ice-covered region (Fig. 3, white color) instead of describing how the wave propagates in the ice zone in the model. Second, different resolutions between the ice model (1/8°) and wave model (2°) entail that the ice zone exhibits a smaller area in the wave model (Fig. 3, white color) than in the ice model (Fig. 3, pink line).
I am still not sure what wave product you are using from Mercator, but I suspect they might do something like “killing” all wave energy at some particular ice threshold (possibly 30%), and then masking all output. Before this threshold it looks like the waves are unaware of the ice (i.e. not attenuated). This was also what ECMWF was doing until the launch of their recent cycle (CY50R1) on May 12, 2026 (for reference, see https://www.ecmwf.int/en/newsletter/185/news/introduction-waves-sea-ice ). This would be helpful information to include for understanding of the reader.
An equivalent hindcast is available from ECMWF which does include waves for all conditions of sea ice for the dates you are interested in (available for public access here: https://apps.ecmwf.int/ifs-experiments/rd/ijxo/ ). Using instead this hindcast would provide a clearer picture of what is going on here.

We think the ice-edge current acceleration is captured by the reanalysis data because of the using of data assimilation, since wave‒ice interactions were not described in the model.
A description of how this is possible would be necessary.
Assuming that the assimilation of sea ice concentration is likely doing most of the work here. This means that the sea ice boundary geometry is constrained (i.e. ice edge position is corrected toward observations ). This changes: surface stress partitioning (ice–ocean vs wind–ocean) , momentum transfer in marginal ice zones and effective friction felt by currents. So even without wave physics, the dynamic boundary condition (ice edge) is much more realistic than a free-running model. That alone can strongly improve coastal jets , ice-edge currents & baroclinic instability locations.

Although the data analysis seem to support the third hypothesis, there are still three primary issues that the reanalysis data fail to reproduce:
Please restate what the third hypothesis is

wave period is 2.2-11.62 s-1, thus indicating that the wave number and wave length are 0.001-0.021 m-1 and 300-6000 m, respectively.
Something very wrong here. Using basic open ocean relations these wave periods would convert to 8m and 200m, not 300m and 6000m. Also wrong units, should be seconds not seconds^-1 for period.

Figure 5
It is hard to follow what is happening in your idealized simulations. So your initial jet is well within the open ocean. And by day 30 in experiment one I think I am to see that it has stayed in the same location but destabilized and decomposed more or less into eddies. Same for experiment to, but it has migrated all the way over to the sea ice edge. I would like to see a few more steps along the way to really understand this. Additionally, I would like to also see the velocity fields shown. I think it would also be helpful to show the direction of the incident waves.
What wasn’t clear until I looked at figure 6 is that the sea ice field is initialized much further to the right of the domain, and due to the wave-coupled interactions, the sea ice is pushed to the left.
I was confused as to why you would initiate your jet so far to the right of the sea ice.
Once studying figure 6 long enough I saw that the black line did indeed scoop down on the right hand side of the domain. This took my eye quite a while to see but I think it would be much clearer if it is shown in figure 5.

The red line indicates the energy transferred from the eddy available potential energy (EAPE, black line) to the EKE, which is the result of time integration of the buoyancy flux work (BSW) on the eddy.
This doesn’t exist in the plot and is not mentioned in the text.

This energy is transferred into the sea ice via radiation stress and is alternatively transferred into the ocean via ocean-ice interfacial stress
Worth noting that it is highly unlikely that all energy is transferred into the sea ice via radiation stress, but this is a simplifying assumption. In reality energy could be lost to a whole myriad of processes, such as under-ice turbulence, ice floe collisions, overwash, etc. There is a rich body of literature on this. As of yet, none of them have come out as “the dominant process”, but rather they are all believed to play a role. A good starting point would be the Royal Society special issue on MIZ dynamics (available at https://royalsocietypublishing.org/rsta/article/380/2235/20210265/112287/Theory-modelling-and-observations-of-marginal-ice ). I believe it is worth noting this to show the complexity of this region.

Thompson et al. (2021) observed that wave‒ice interactions led to jets in the Chukchi Sea.
Typo, Thomson. This error appears multiple times

In this study, we discussed the eddy genesis process in further depth and in more ways than previous studies have achieved. First, we solved three issues outlined in the introduction. (1) Wave‒ice interactions constitute a significant contributor to the EGC. (2) Wave-ice interactions produce eddies via barotropic instability, which was first proposed on the basis of numerical simulations in previous studies; however, this point has hitherto remained entirely theoretical (Dai et al. 2019). Subsequently, Thompson et al. (2021) observed that wave‒ice interactions led to jets in the Chukchi Sea. Employing satellite observations and reanalysis data, this study successfully documented the complete process within the Fram Strait. (3) Wave energy, the extent of the marginal ice zone, and water depth are identified as critical constraints that govern the occurrence of this mechanism. In Chukchi Sea, the water depth is small, which may largely decrease the jet strength via bottom drag and hardly produce visible eddies via barotropic instability (Zhang et al. 2020). At higher latitudes, wave energy is typically insufficient to trigger this mechanism, rendering it unobservable. With these three constraints, the wave-ice interaction induce eddy activity may also be found in Barents-Kara Sea in cold season, when there is strong wave energy, marginal ice zone and large water depth.
This is somewhat hard to follow.
My interpretation is as follows:
In basic words, the theory (Dai et al. 2019): wave–ice forcing → jet formation → barotropic instability → eddies
We will only see this full chain if wave energy is strong enough, Marginal ice zone (MIZ) exists and is extensive and water depth is sufficiently large.
Thomson et al. (2021) had observed the first part of this in the Chukchi Sea, namely: waves → jets. But due to bottom friction the jets were not strong enough and could not lead to eddies.
In this study these conditions have been satisfied in Fram Strait and now we see full chain of the Dai et al. (2019) theory.
An “aside” of your theory: we should also expect to see the same in winter conditions for Barents–Kara Sea
If this interpretation is correct, I think it is very interesting, and evidence in the good direction, but not necessarily definitive that waves play the defining role. You have shown
There is a strong eddy present in both satellite observation and high resolution ocean reanalysis

Wave and ice conditions seem right such that waves could play a key role here.

Testing in an idealized configuration show that waves can provide the mechanism for such eddy generation

That is co-occurrence, not necessarily mechanistic proof. Ocean reanalysis is not equal to independent evidence of wave-driven jets, mainly because data assimilation ensures dynamical consistency, not process attribution as outlined above.
The study of Thomson et al. (2021) uses a variety of buoy- and ship-based measurements to measure currents at the ice edge, ice drift and under ice turbulence, and their evolution as one goes into the ice. They then show directly how their model can simulate these observations.
You have showed one satellite image. I don’t necessarily rule out that such things could be established using remote sensing, but I think there would be a lot more work needed in this direction before you could say you have rigorous evidence for this mechanism here.
Further, in Thomson et al. (2021), I see the following:
“In the present study of a marginal ice zone, the total water depth is large (∼3,000 m) so τb represents ice-ocean stresses, instead of an actual bottom stress.”
And
“Observations were collected on September 18, 2018 as part of the Stratified Ocean Dynamics in the Arctic (SODA) campaign in the southern portion of the Beaufort Sea, approximately 300 km north of Alaska (USA).”
This contradicts with your statements about it being shallow and Chukchi (or am I looking in the wrong place?)

we employed a shorter wavelength
As far as I can see, never specified which wavelength was actually used in your model.

The ocean front subsequently begins to appear, which contributes little to eddy generation in comparison with the barotropic instability mechanism.
This is a statement that would require quantification. I haven’t seen quantification to this point (although I think you probably could do it in your idealized model). With softer phrasing though it could pass e.g. “which likely contributes little to eddy generation due to the low gradients in comparison with the barotropic instability mechanism.”

Reply
Citation: https://doi.org/10.5194/egusphere-2026-153-RC1
RC2:
'Comment on egusphere-2026-153', Anonymous Referee #2, 06 Jun 2026 reply
This manuscript reveals that wave-ice interaction may strengthen eddy activity in Fram Strait. The authors use SAR imagery, ocean reanalysis data, and numerical experiments to argue that wave-ice interaction is the main mechanism responsible for enhanced eddy activity during September and October, when sea ice is expanding. The use of multiple datasets and process-oriented numerical experiments is valuable and helps support the proposed mechanism. However, the observational evidence is based on a single event/year, and the exclusion of other potential eddy-strengthening mechanisms is largely qualitative. In particular, the manuscript does not provide sufficient quantitative evidence to rule out alternative mechanisms such as topographic effects, nonlinear advection, or baroclinic instability. Therefore, I recommend major revision.
Major comments
Line 158–165: The discussion of the second hypothesis is not fully convincing. The authors argue that ice acceleration should induce stronger ocean current acceleration where sea-ice concentration is larger. However, the relationship is not this simple. Momentum transfer depends on wind-ice drag, ice-ocean drag, the relative velocity between sea ice and ocean, sea-ice internal stress, drag coefficients, wind direction, mixed-layer depth, stratification, and background current shear. In addition, ice-ice interaction can constrain sea-ice motion and modify the effective stress transferred to the ocean. Therefore, ocean acceleration beneath sea ice does not necessarily scale monotonically with sea ice concentration. The authors should provide more quantitative estimates to support the argument before dismissing this mechanism.

Line 177–188: The authors argue that baroclinic instability is unlikely to be the primary mechanism because the salinity front strengthens while the eddy also strengthens. However, this argument needs more concrete validation. A strengthening front can indicate the accumulation of available potential energy, and baroclinic instability may develop with a time lag rather than requiring an immediate opposite phase relationship between front strength and eddy strength. Quantitative diagnostics, such as EAPE, buoyancy-flux conversion, mixed-layer stratification, etc., are needed to support the conclusion that baroclinic instability plays only a minor role.

Section 3.2: he numerical experiments are useful as process-oriented simulations showing that wave-ice interaction can strengthen an ice edge jet and enhance barotropic instability. However, in EXP2, the simulated jet reaches approximately 1.1 m/s, which is much stronger than the branch current of approximately 0.24 m/s inferred from the reanalysis. The authors should clarify that the simulations demonstrate the dynamical plausibility of the proposed mechanism rather than quantitatively reproducing the observed event.

Section 3.2: The authors use a relatively short wavelength and a large incident angle, both of which favor strong wave dissipation and along-ice-edge momentum transfer. Because these choices strongly influence the simulated jet strength and eddy growth, the authors should provide sensitivity tests showing whether the proposed mechanism remains effective under longer wavelengths, smaller incident angles, and weaker wave forcing conditions that may be more representative of the observed event.

Section 3.2: The authors may consider additional numerical experiments to better isolate the proposed mechanism. For example, a no-initial-jet case would be helpful. In addition, an experiment that can assess the competition between baroclinic instability and wave-enhanced barotropic instability can be useful.

General comments
Line 30-33: The manuscript states that ocean eddies have typical horizontal scales of 5-100 km, while mesoscale and submesoscale motions are later defined as 50-300 km and 1-50 km, respectively. This may confuse readers about the scale classification of the observed eddy.

Line 58-60: The manuscript states that the topographic-control mechanism and nonlinear advection mechanism are excluded, but the justification is very brief. A relatively flat bathymetry and water depth greater than 250 m may reduce the importance of topographic effects, but this does not automatically rule them out. The authors should provide more evidence or analysis to justify the exclusion of these mechanisms.

Line 143-144: The manuscript states that the second eddy strengthens, but this is not clearly demonstrated in Figure 2. Also, SAR images need to be provided in Figure 2.

Line 144: The sentence “Jet strengthening is a significant feature” appears abruptly and is not well connected to the preceding discussion of eddy evolution.

Formatting, language
Line 52-53: The manuscript refers to “the three issues above,” but only (2) and (3) are stated in this paragraph. (1) appears earlier in the Introduction and is not clearly connected to this sentence(maybe?). The authors should revise the wording to make the numbering of the research questions consistent.

There are several typos that the author needs to be careful of
Title: activity -> “a” has a different font

Line 28: EGC? -> EGC.

Line 86: writing as -> written as

Line 123: Need space before “Since”

Line 122: Subscript – u1, u2, u3

Line 123: Grammatical error: Single sentence with “Since”

Line 130: Double check the definition of R_d

Line 177: using -> usage

Line 195: vmax -> subscript max.

Line 217: ky and kx -> subscript.

Reply
Citation: https://doi.org/10.5194/egusphere-2026-153-RC2

Yi Liu, Yunhan Yang, Haijin Dai, Jun Zhao, and Qiang Li

Viewed

Total article views: 356 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
254	81	21	356	18	18

HTML: 254
PDF: 81
XML: 21
Total: 356
BibTeX: 18
EndNote: 18

Views and downloads (calculated since 09 Apr 2026)

Month	HTML	PDF	XML	Total
Apr 2026	129	42	14	185
May 2026	107	35	5	147
Jun 2026	18	4	2	24

Cumulative views and downloads (calculated since 09 Apr 2026)

Month	HTML	PDF	XML	Total
Apr 2026	129	42	14	185
May 2026	107	35	5	147
Jun 2026	18	4	2	24

Viewed (geographical distribution)

Total article views: 356 (including HTML, PDF, and XML) Thereof 356 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 14 Jun 2026

Short summary

The East Greenland Current (EGC) typically flows equatorward at high speeds. In autumn, a branch of EGC strengthened, reaching an intensity comparable to that of the main current. Sea ice expands as temperatures drop, while ocean waves continue to propagate toward the ice zone and dissipate at the ice edge via wave–ice interactions, transferring energy to the sea ice and ocean. This mechanism accounts for strengthening of the EGC branch and enhances eddy activity through barotropic instability.


Total:	0
HTML:	0
PDF:	0
XML:	0