Summary:

This manuscript addresses an important question regarding the propagation of wind-generated near-inertial internal waves. The fate of near-inertial waves' energy is an important topic because the breaking of near-inertial waves serves as a source of energy for vertical mixing in the ocean interior, particularly in the upper ocean. This study investigates the propagation of near-inertial wave energy into the interior and the energy transfer rate between waves and eddies for two winter storm events, based on the data from the mooring system and reanalysis product. Their results highlight the need to account for the interactions between waves and eddies when interpreting the propagating near-inertial energies and their dissipation, leading to turbulent mixing. While the paper is well-motivated and contributes to science, I believe the manuscript requires substantial revision before it can be considered for publication.   

Major comments:

A) The introduction section is well-structured, citing observational studies of wind-generated near-inertial internal waves. However, the novel aspects of the present study compared to these prior studies, as well as the motivation for this research, remain somewhat unclear to readers who are not familiar with these fields. I think it would be better if these points were summarized as the first few sentences of the final paragraph of the Introduction.

B) In Section 4.1, the authors explain that the similar NIKE values between the two events—despite significant differences in the input of wind-generated near-inertial energy—can be understood by taking into account the difference in energy transfer rates between mesoscale eddies and NIW. However, is it reasonable not to consider other energy sinks for the wind-generated NIW's energy inputs? Kunze et al. (1995) assumed that there are three possible energy sinks for wind-generated NIW in the anticyclonic eddy, including i) loss to the mean flow, ii) loss to un-trapped waves that can freely propagate, and iii) instability of shear, leading to turbulence production. Although this manuscript considers only i), is it acceptable to ignore the other two candidates in the interpretation of your observational data? I understand that quantification is difficult—particularly regarding point iii), given the absence of turbulence observations—but the possible effects by ii) and iii) should at least be discussed in the discussion section.

My other comments are mostly on details that could be improved or on which I had minor questions.

Minor comments:

Line 86: Please consider showing the deployment depth of each instrument (ADCPs, SBE 37-SM, and SBE 56) as a figure. It could be helpful for readers to see the vertical resolution of the observed density profile.

Line 93: What value of the damping parameter is applied in the slab model?

Line 99-101: Please clarify how the mixed layer depth (MLD) is defined in the present study.

Line 101-103: To verify the validity of the reanalysis product, please add a depth-time plot of the squared buoyancy frequency N²  and the plots of the mixed layer depth (MLD) of both the observations and the reanalysis product in the supp info. The reproducibility of these variables is important considering that they are used to estimate the near-inertial energy inputs and their modal decomposition.

Line 110: Please check if the citations for this sentence are correct. I guess that the other citations (e.g., Gill (1982)?) would be more appropriate.

Line 127: Please clarify the definition of background horizontal velocity (U, V). Do the authors apply the low-pass filtering spatially and/or temporally to estimate ‘background’ velocity from the daily outputs of the reanalysis product?

Line 137: Is the velocity vector V  the same as the background flow (U, V) used in Eq. (4)? Please clarify it.

Line 138: I think that |V| in this sentence should be V.

Line 140: Do both k_z and m represent the vertical wavenumber? Please use the consistent representation.

Line 142: Please remove ‘horizontal’ in this sentence. Only the vertical group velocity is defined here.

Line 144: I think Equation (9) needs a minus sign. Please check it.

Line 159: It would be helpful to include a comparison between the u_ni derived from the slab model and the bandpass filtered u_ni from the observations, even if it's just in the supplementary information.

Line 177-179: Why does the total NIKE (Fig 5c) appear to be smaller than the decomposed downward energy (Fig 5d) at certain time periods and depths, such as 150 m depth in WS1 and WS2?

Line 182-184: Please clarify how to estimate decay time here. Is the decay time defined as the time it takes for the energy propagating along a ray path determined by the group velocity to decrease to one-tenth of its maximum? It might be helpful to understand if the authors indicate the time and depth of 1) the maximum NIKE and 2) the one-tenth of maximum NIKE in Fig. 5c.

Line 201: Please consider indicating the values of 1.01f₀ and 0.98f₀ as vertical lines in Fig. 6. I would like to check whether the red-shifted near-inertial peak shown in Fig. 6c corresponds to 0.98f₀.

Line 208-210: How is the vertical wavenumber (or vertical wavelength) estimated in the present study? Is the wavenumber estimated as m=d(atan(u_ni/v_ni))/dz, same as in Chen et al. (2023)? Please clarify it.

Table 1: How are the errors (±) in Table 1 estimated?

Line 252-253: Why does the Fig 9a show the estimated P “within” the mixed layer, rather than “below” the mixed layer, where near-inertial internal waves can freely propagate? Does the author assume that the energy transfer between the near-inertial motions and eddies within the mixed layer is dominant for the net energy budget throughout the water column?

Line 256-257: Please consider showing the definition of the OW parameter used in the present study.

Line 264-265: Please also show the time-integrated energy transfer rate ∫∫ρPdzdt during WS1 and WS2 and compare them with the time-integrated energy flux ∫Fdt.

Line 268: I understand that it is not possible to estimate the vertical variation of relative vorticity  from the data obtained by the single mooring site. However, I am concerned about whether there is any reason why  and  can be used in P in Eq. (4) but are not used to estimate relative vorticity?

Line 295: By presenting an equation, please clarify how the authors estimate the modal projection of wind-induced energy flux following Raja et al. (2022).

Anonymous Referee #2, 23 Jul 2026

 Summary

This manuscript contributes to our understanding of atmosphere-ocean energy transfer, a key topic in physical oceanography, by investigating how energy is transferred from the atmosphere into the ocean interior through near-inertial waves and how this process is modulated by mesoscale eddies. Using data observations together with reanalysis products, the authors compare two winter storm events that differ substantially in their wind forcing and NIW characteristics, while exhibiting similar observed near-inertial kinetic energy levels. The study suggests that a combination of eddy-wave interactions and differences in the modal structure of the wind energy input may explain the contrasting behaviour of the two events.

The manuscript is generally well written, flows nicely, with appropriate citations, calculations, and figures to support the reasoning - I appreciate that you are not using a color palette that is difficult to read for colorblind readers. It represents a valid contribution to the field, and the observational dataset is valuable. However, I have a few comments that I believe would help strengthen the interpretation and improve the robustness of some of the conclusions.

I'd like to note that I carefully read the manuscript both in its original form and in the revised version addressing Reviewer 1's suggestions, and in the first reading my main concerns were quite similar to those of Reviewer 1. I therefore really appreciated the revisions that were made, and I took this into account in my part of the review – page numbers are referred to the revised version of 12 July.

Major, addressed

On Major #1, already addressed: From my point of view, Reviewer 1's suggestion is sufficiently addressed in the introduction (though that is ultimately for them to judge), but a brief description of the physical setting of the mooring site would also be helpful. The manuscript focuses on the interaction between NIWs and mesoscale eddies, yet the reader is given little context regarding the regional circulation of the study area. Even a brief discussion or an appropriate reference around lines 69–70 could help provide this context.

On Major #2, already addressed: The first version of the manuscript attributes the observed energy differences primarily to mesoscale eddy-wave energy exchange. While this interpretation is plausible and supported by the diagnosed transfer rate P, other possible sinks and sources of NIW energy (e.g., turbulent dissipation, wave-wave interactions, and energy propagation out of the observed volume) were not considered in the first version, as already pointed out by Reviewer 1. I therefore appreciate the additional discussion of turbulent dissipation and wave radiation introduced in response to Reviewer 1. These additions considerably strengthen the interpretation of the NIW energy budget. Nevertheless, given the observational limitations, I would still encourage the authors to maintain caution when attributing the observed energy differences primarily to eddy-wave interactions. Explicitly acknowledging that multiple mechanisms may contribute to the observed behaviour would both strengthen the interpretation and help motivate future observational studies.

Another way to strengthen the discussion could be the following: as currently written, the manuscript still tends to lead the reader toward an eddy-centered interpretation of the observed differences between the two NIW events. Given the observational limitations, and the fact that other terms of the NIW energy budget cannot be quantified directly, the conclusions may benefit from a more explicit discussion of multiple contributing mechanisms. In particular, the modal projection of wind-generated near-inertial energy appears to provide an additional, and potentially complementary, explanation for the differences observed between the two events. Emphasizing the joint roles of wind forcing and eddy modulation would, in my view, result in a more robust interpretation of the observations.

Minor:

1. 2.1. Data: Adding the mooring line is a great way to help the reader visualize it, but I still suggest a bit more precision in the text — e.g., the depth of the mooring, the positioning of the two ADCPs (I assume one up-looking and one down-looking), and how the raw data from both the ADCPs and the CTDs were processed in the first place.
2. Line 111: In response to Reviewer 1 comments on lines 99-103, you state here that MLD is calculated as the depth where temperature is 0.5°C lower than the SST, and in Supplement material you add more on the reproducibility of the variables. The agreement between reanalysis and Argo-derived MLD also appears satisfactory, nevertheless the manuscript does not clearly state the criterion used to define the MLD. Given the relevance of MLD for the slab-model calculations, the subsequent interpretation of the NIW characteristics, and since temperature-based and density-based criteria may differ substantially, the adopted definition should be explicitly stated and justified.
3. Results, line 172: Maybe "observed" instead of "suggested"?
4. Line 194–198: The use of one-tenth of the maximum NIKE as a criterion for both propagation depth and decay time is still not entirely clear to me. Is this threshold adopted from previous studies, or was it chosen specifically for this analysis? If it is an empirical criterion introduced by the authors, I would suggest briefly stating so and providing a short justification for its use.
5. Figure 6: in the revised version the x-axis label appears to be missing.
6. 4.1. paragraph, general consideration, with the Major #2 discussion in mind: the central conclusion here relies on the diagnosed energy transfer rate P. Therefore, the manuscript would benefit from a clearer description of the processing the variables entering the calculation have been submitted to, even if some details have already been added in 2.3. For example, I wonder whether the filtering procedure applied to the horizontal velocity fields, namely a 5-day temporal low-pass filter, could affect the interpretation of the shorter event (WS2), which is only roughly twice as long as the filter window. A brief discussion of the robustness of the sign and magnitude of P with respect to the chosen filtering and averaging parameters could help clarify this point.
7. In line 275, significant energy exchange is reported to occur primarily within the upper 200 m. However, the vertically integrated P is computed over the upper 300 m, while the profiles shown in Figure 9 extend to 400 m depth. A brief explanation for these choices would help the reader understand the rationale behind the selected integration and display ranges.