Dear Ann-Sofie and co-authors,

Thank you for submitting your replies to the reviewer reports. Reviewer 1 (R1) has raised a number of important concerns, and has not recommended your work for publication in its current form. Reviewer 2 (R2) has requested minor revisions, although some of their comments echo the concerns from R1. Based on my own reading and your replies, I would like to provide you with the opportunity to submit a revised manuscript. To help you target your efforts, I suggest focussing on the following aspects:

1. Timeseries

You aim to explore the causal relationship between channel emergence and various external factors, such as ocean forcing. However, without a fully diagnostic model that resolves the relevant processes at the appropriate spatial and temporal scales (as noted by R1), it will not be possible to definitively demonstrate causality. To avoid unrealistic expectations for the reader, I suggest to remove or add nuance to claims in this regard. That said, I think you can provide a more robust analysis of possible correlations. I feel that the concerns raised by both reviewers likely stem from a lack of robust analysis of these correlations. This can be addressed by making more effective use of the time series data you already have. In particular, to assess the relationship between channel emergence and ENSO variability, consider plotting the ENSO index (as you allude to in your response to R2) against MITgcm melt rates averaged over a region encompassing the channel. If ENSO—and its associated impacts on ocean circulation and heat/salt transport—plays a role, this should be evident in the melt rate variations, which are more directly relevant to ice-ocean interactions than heat/salt or flow fields. It is then important to demonstrate confidence in the MITgcm-derived melt rates and their variability, which you can do by comparing them with the independent BURGEE (and potentially, LADDIE) melt estimates. Do the latter datasets exhibit similar variability linked to ENSO? While you correctly note the difficulty in directly comparing 2D melt rate patterns due to topographic advection, you can still examine whether broader-scale averages show consistent changes tied to large-scale climate variability. The BURGEE and LADDIE time series will be relatively short, but the ENSO index and MITgcm data span a much longer period (e.g., from 1979 onward). This extended record allows you to evaluate whether the 2015 ENSO event stands out against the background of long-term variability. Doing so may strengthen your argument for why the channel emerged in 2015 rather than, for example, during earlier ENSO events. This would strengthen the message of the paper.

2. Spatial maps

Some aspects of the story are best illustrated using spatial maps. However, in line with R1's comments, I am not convinced that maps of MITgcm velocities at fixed depths for two different years provide meaningful insights. Simulated ocean currents at multi-kilometer scale exhibit substantial variability across a wide range of timescales, and comparing two yearly-averaged snapshots does not offer information about the significance or persistence of the simulated differences between both. What is more relevant for ice–ocean interactions are the melt rates. This is why I recommend focusing on time series of MITgcm-derived melt rates rather than on heat, salinity, or current speed. Additionally, it is important to remember that the observed channel is not stationary; it advects downstream over time and may encounter varying oceanic conditions along its path. These changing conditions could influence plume–ice interactions and potentially lead to re-routing of the channel, without the need to invoke changes in large-scale ocean conditions linked to climate variability.

The other key aspect of the study (fracture or not?) is the correlation between channel formation and changes in ice surface velocity. Again following R1, I am not convinced that your current method of calculating velocity divergence across the channel captures the full complexity of the evolving flow field. To better analyse this, I suggest presenting 2D maps of effective strain rates—or principal strain rates—derived from the surface velocity fields at different times before and after the channel’s formation. In fact, the BURGEE method likely performs a similar analysis internally via the divergence term ∇⋅u in the conservation equations, as R2 noted. These strain rate maps will allow you to separately analyse across-channel (tensile) and along-channel (shear) deformation patterns. To reduce noise and obtain more reliable derivatives, it is common to apply a local quadratic (or cubic) regression to the velocity field over a fixed-size spatial window (see, for example, Fig. 7b in De Rydt et al. (2018) and many other references in the literature).

3. Disparity of scales

One possible way to address R1’s concerns about scales is to use MITgcm output in a aggregated, large-scale manner—specifically, to assess the influence of climate variables on melt rates using time series, as suggested above, rather than presenting spatially detailed maps. In this context, LADDIE serves as a complementary tool to MITgcm, as it enables investigation of smaller-scale processes. You make a convincing case that melt rates could be elevated within the channel. However, it is important to acknowledge that ocean dynamics and melt physics in narrow channels with steeply sloping sidewalls (exhibiting ageostrophic flow—shown by e.g. Millgate et al.; https://doi.org/10.1002/2013JC009402) are fundamentally different from those in broad channels with gentle slopes (geostropich flow—where LADDIE has previously been shown to perform well). Apart from the limited resolution, LADDIE may not capture key processes in the steeper, more complex geometries. It would strengthen the manuscript if you could clarify these limitations of more directly. Similarly, it would be helpful to reiterate the limitations of the BURGEE approach for calculating basal melt rates at spatial scales smaller than a few ice thicknesses, where the hydrostatic assumption may not be valid.

Taking into account the reviewer comments and my additional notes above, I invite you to submit a revised version of your manuscript. If you need more time than the standard revision period then please don’t hesitate to contact the editorial office to request an extension.

All the best,
Jan De Rydt – topical editor

Dear Anne-Sofie and co-authors,

Following your replies to the earlier reviewer reports and the submission of a revised version of the manuscript, we have received a further report from reviewer 2. The reviewer agrees with your revisions and supports the publication of your work, subject to a few minor corrections. To proceed with the publication of your manuscript, please upload a point-by-point reply to the reviewer's comments, together with a new version of the manuscript with tracked changes.

Best regards,
Jan De Rydt - topical editor

Dear Anne-Sophie and co-authors,

Thank you for providing a point-by-point response to the reviewer report and for implementing the corresponding changes in the manuscript. I believe all outstanding concerns have now been addressed, and I am pleased to accept your work for publication in The Cryosphere. I would also like to take this opportunity to thank both reviewers for their thorough and constructive feedback. Before you submit the final version for copyediting, please address the following two comments from the editorial office:

1. Your reference list includes works “in review”. Such works can be cited upon submission if being available to the reviewers. They should not be cited in the final, accepted manuscript, unless published, accepted for publication, or available as preprint with a DOI.
2. Some sections and figures are misplaced. Please pay attention to the manuscript composition: https://www.the-cryosphere.net/submission.html#manuscriptcomposition

Congratulations, and best regards,
Jan De Rydt - topical editor