Review of “Sea ice transport and replenishment across and within the Canadian Arctic Archipelago: 2016-2022” by Howell et al.

Summary:

This study focuses on quantifying sea ice transport and replenishment across and within the Canadian Arctic Archipelago (CAA), and particularly along the critical segment of the Northwest Passage shipping route, spanning from the Queen Elizabeth Islands to the Parry Channel. Results indicate that the CAA functions as both a source and sink for sea ice, exporting significant amounts to the Arctic Ocean and Baffin Bay. The study underlines the resilience of multi-year ice (MYI) replenishment within the CAA, with ongoing import from the Arctic Ocean and retention of first-year ice (FYI). The authors emphasize the persistent risk that sea ice poses to key shipping routes in the CAA, including the Northwest Passage, due to substantial ice flux and sustained MYI replenishment.

This study makes use of high-resolution drift data from SAR as well as CryoSat-2 altimetry and can serve as a baseline study of sea ice fluxes in the CAA. While the method itself (sea ice flux estimation) is not new, the study region, especially inside the CAA is rather understudied compared to Fram Strait for example.

General Comments:

To my knowledge there is no comparable study for ice fluxes within the CAA currently available. Therefore, I think it potentially deserves publication. I also had no problems to follow the text in general. The applied methods seem generally solid, but there are some details and decisions that I find disputable and that need clarification and more in-depth descriptions and discussion. Here and there, the paper lacks explanations. I also find the figures could partly contain more information. Major comments are:

1. More information on the input data is needed. The authors use two sea ice thickness data sets to compute volume fluxes. The all-year CryoSat-2 summer sea ice thickness retrieval from Landy et al. (for the outer gates) and the proxy-record from Glissenaar et al. (for the inner gates of the CAA). I think it should be discussed how consistent these data sets are. How do they compare at intersections? Because any inconsistency might introduce biases here. I also suggest introducing acronyms or at least make it clearer when and where which data set is used (See also in the specific comments). It is also not entirely clear over which period data have been used. Under “2) Data”, it is stated that volume fluxes have been calculated until 2021, but later in Figure 4, it is only until 2020, while the caption mentions until 2022. This is confusing. It should be possible to provide volume fluxes until 2022, I assume.
2. I find the method description of how area and volume fluxes are computed lacks information. How are sea ice motion, concentration, and thickness co-registered along the gate? Are you using a nearest-neighbor scheme? How are data gaps handled? Are you filling those by interpolation? I also recommend adding a figure showing the drift, concentration, and thickness along a few gates, to get an impression of spatial resolution of input data and how thickness is distributed along the gates.
3. I find the approach of using a linear trend to bridge the summer gap for the thickness at the inner gates is daring. Are there in situ measurements like from buoys or other observations that support this approach? At least uncertainties should be significantly higher.
4. The authors calculate uncertainties for both area and volume fluxes, and provide estimates in a table, but it is quite difficult to relate them to the flux estimates in the figure. I strongly recommend adding error bars in the figures where you provide flux estimates.  
5. There is the study of Agnew et al. (2008) that provides area fluxes across the CAA. Can you compare area fluxes with those of Agnew et al.?

More specific comments:

L96: I thought the Landy et al. data set already combines summer and winter sea ice thickness from CryoSat-2: “A year-round satellite sea-ice thickness record from CryoSat-2”? Did you use these data? In the text it sounds like you did the combination of summer and winter data yourself.

L118: The information on the apertures can be provided either in a table or better in Fig. 1, avoiding listing it in the text.

L154: “was determined from the product of the monthly ice area flux and the monthly average CryoSat-2 sea ice thickness” – In the method section you write that you use different thickness data sets (The Landy record and the proxy record) for the inner and outer gates. Please clarify.

L155: This is because you use the “proxy record” here, right? Perhaps mention that. However, I wonder how robust it is to just interpolate between April and October. How do you estimate the volume flux uncertainties in the CAA in summer? Are there any in-situ data to compare with?

Figure 2: I suggest adding error bars with the calculated uncertainties. Moreover, in the caption and/or the figure itself, it should be written more clearly which flux and gates/regions are considered here.

Figure 3: I suggest to also add error bars here. Moreover, in the caption, I assume it should be “NET export/import”?

Figure 4: Same comment as for Fig. 3. Moreover, the caption says “for 2017 to 2022” – but only 2017-2020 is shown?

The study explores ice area and volume fluxes into, around, and out of the Canadian Arctic Archipelago (CAA) using a combination of remote sensing datasets. Ice motion data are derived from SAR imagery for deriving area fluxes and thickness estimates from CryoSat-2 around the primary gates and an ML random forest model for the inner gates. The paper follows on from a study by the same author in 2022 deriving the new ice area flux estimates from Sentinel-1 and RADARSAT RCM (Howell et al., 2022), a paper in 2022  looking at MYI conditions in the CAA (Howell et al., 2023a), then a paper in 2023 looking at similar ice area/volume fluxes of the CAA and Nares Strait by integrating the new CryoSat-2 thickness data with the previous area flux estimates (Howell et al., 2023b).

Overall, the analysis in the paper was well presented and relatively easy to follow and the science aligns well the scope of The Cryosphere. My comments are provided below, I hope these make sense and help improve the manuscript. 

Alek

General points:

1. The first thing I think worth flagging is how closely aligned this paper is with the 2023 JGR study (Howell et al., 2023) that derived similar estimates of ice area and volume fluxes using the same method and over the same time-period but also comparing to Nares Strait. The methods sections of both papers are virtually the same, which I think is fine, but I was surprised there wasn’t a clearer note to this effect in this paper. The addition in this study seems to be the CAA inner gate area/volume flux estimates that were not included in the 2023 study and the generally increased focus on the CAA results. Still, it took me a bit of a while to realize that and I think this paper should tie together much more with that study and make clearer what the relative goals of each study are and what exactly is new in this study. Similarly, the MYI replenishment section didn’t refer to the author’s recent paper on CAA MYI replenishment which seemed a little odd too. Put another way, the author and author team have been looking closely at area/volume/MYI fluxes around the CAA for a while now, so what was the gap that this study needed to respond to and how did it build on the preceeding efforts?
2. I’m not a fan of the uncertainty bounds approach, especially as they aren’t really used (hard to plot an uncertainty bounds in a time-series plot!). I really think you should just pick your best guess uncertainty estimate and justify it as best you can – I don’t particularly believe the bounds truly represent realistic bounds anyway. It would be good to then use those values on the time-series plots you show to get a sense of how important the uncertainties are for assessing seasonal/annual variability. However, I do have some additional concerns about the uncertainty quantification:
   1. What about the errors in ice concentration (L158)? I would guess they are not negligible considering how small some of the channels are, but maybe this was addressed in one of the recent papers.
   2. Seems quite odd to assume the ice motion errors are uncorrelated if derived based on the same image pair? But then you later add the errors to generate the monthly uncertainty estimates so you assume they are correlated? Are these assumptions justified better in previous papers?
   3. The use of the linear trend to fill in the summer months for the inner gates seems very crude, especially as this seemed to be one of the big differences with the 2023 JGR study. At the very least I think the paper would benefit from showing what the raw and interpolated thickness values actually look like at each gate and how justifiable they are. Any way you could tie it together with the all-season CS-2 data outside the CAA?
   4. It would be good to get a better sense of how important the results are to the thickness estimates, I’m guessing there is some skill in the seasonal thickness cycle in the input datasets but not much de-seasonal skill beyond that considering the errors.
3. L266-266 on QEI area vs volume import/export I view as the most interesting idea from the paper but think it should be explored in much more detail to help justify this paper. How thick does the ice need to be north of QEI for the assumptions of net volume sink to be true? Do we think we’re approaching an inflection point of this not being true anymore? I think it would be easy and quite illuminating to run a little sensitive test here changing the ice thickness values north of QEI and re-running the analysis. I don’t know the cited Melling 2022 study that well but the claim that the thickness isn’t changing north of QEI is a little surprising to me, but eventually I think we can agree it’s quite likely to change after a big MYI flushing event. In general, I think the paper would be much improved if you could test some hypotheses out in this framework rather than just showing raw data and discussing ideas.
4. Finally, on a similar theme, it’s quite hard as a reader to intake all the different flux estimates and get a sense of what it all means. Most of the figures don’t really much of a compelling story or scientific result. The discussion and conclusion sections do help but they are not very visual. A map schematic showing the area and volume flux estimates for your study period I think could help a lot?

Specific comments:

L33 – are goods actually transported through the NWP?!

L60 – but then at L66 you say people have done this also using AMSR-E. So, what resolution do we truly need for this kind of analysis?

L64 – what exactly do you mean by images not being consistent in space and time?

L68. – The relative benefits/merits of S-1/RCM vs R1/R2 for doing this area flux analysis is still a little unclear to me, was this explained better in the earlier papers?

Figure 1 – I think it would be good to highlight Baffin Bay/Beaufort Sea as they are mentioned in the text too, appreciate the figure may need to be zoomed out a little more.  

Generally confused by the discussion of Parry Channel but Peary Channel being indicated on the map?!

L96 - this discussion of the all-season CS-2 data is a little odd, you are just using the data from the 2022 paper right, so shouldn’t that be cited first? The rest is background to that study.

L106 – what are the open seas of the CAA?

The CAA thicknesses from the Isolde 2023 paper basically show a seasonal cycle of ~60 cm to 160 cm, errors in the proxy data of 30 cm?

L28 - Not quite sure what you mean by buffer region, this seems maybe a bit colloquial. Can you be more explicit?

L143 – what are the units of that? Km/day??

L160 – this is confusing as you don’t reference the proxy data which I believe is what you are using for the inner gates here. I think in general it would help to show what these data look like for a given gate as a case study – show the area flux, the thickness then the fluxes for a given season with those applied uncertainties. Would help us visualize the variability in the source terms and how it relates to the variability in the flux terms.

Figure 2 - How well correlated are the area and volume fluxes? Could also show the thickness/are variability too. As stated in the general points, also unsure why the uncertainties are not shown. I think just take your best guess uncertainty and include that in the figures, would help to visualize how they compare to the variability of the signal.

L260 – Shackleton would be turning in his grave over that comment! But seriously, I would guess there is a good chance there could be thicker ice in the Weddell Sea..?

L318 onwards - I feel like the MYI replenishment ideas could benefit from knowing how much MYI is lost too? Struggling a bit too put the replenishment numbers in context.

359 – again I think here is where you could benefit from a better understanding of how sensitive this result is to the underlying thicknesses.

Figure 3 and 4 – it would be quite easy and I think much better to read if you combined these, put the volume flux alongside the area flux bars with a twinned y-axis on the other side. Ideally it would be good to see the area and thickness numbers too.

In several figures you should add the exponent multiple to the label and make the figures consistent in this regard.

References:

Howell, S. E. L., Brady, M., and Komarov, A. S.: Generating large-scale sea ice motion from Sentinel-1 and the RADARSAT constellation mission using the environment and climate change Canada automated sea ice tracking system. The Cryosphere, 16(3), 1125–1139. https://doi.org/10.5194/tc-16-1125-2022, 2022.

Howell, S. E. L., Babb, D. G., Landy, J. C., and Brady, M.: Multi-year sea ice conditions in the Northwest Passage: 1968-2020. Atmosphere-Ocean, 61:4, 202-216, DOI:10.1080/07055900.2022.2136061, 2023a.

Howell, S. E. L., Babb, D. G., Landy, J. C., Moore, G. W. K., Montpetit, B., and Brady, M.: A comparison of Arctic Ocean sea ice export between Nares Strait and the Canadian Arctic Archipelago. Journal of Geophysical Research: Oceans, 128, e2023JC019687. https://doi.org/10.1029/2023JC019687, 2023b.