the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 27 Jan 2023
Impact of the Nares Strait sea ice arches on the long-term stability of the Petermann Glacier Ice Shelf
Qin Zhou
Tore Hattermann
Nina Kirchner
Abstract. One of the last remaining Greenland Ice Sheet (GrIS) glaciers featuring a floating tongue – the Petermann Glacier Ice Shelf (PGIS) is seasonally shielded by the formation of sea ice arches in the Nares Strait. However, continued decline of the Arctic sea ice extent and thickness suggest that arch formation is likely to become anomalous, necessitating an investigation into the response of PGIS to a year round mobile and thin sea ice cover. We use a high-resolution 3-D ocean-sea ice-ice shelf setup featuring an improved sub-ice shelf bathymetry and a realistic PGIS geometry, to investigate in unprecedented detail, the implications of transitions in the Nares Strait sea ice regime; from Thick Landfast to Thick Mobile and Thin Mobile, on the PGIS basal melt. Across all three regimes, basal melting increases occur during summer, and under the deeper (> 250 m) regions of the ice shelf. Diagnosing this variability via melt rate drivers suggest that a higher thermal driving under the deeper regions causes higher melt rates, which, as a secondary effect, increases the friction velocity slightly downstream. The increased meltwater production and a stronger melt overturning in the PGIS cavity deliver more meltwater from depth to the shallower regions which lowers the thermal driving and basal melt in these regions; with the winter season showing a converse pattern. Modulations in surface forcing under a mobile and thin sea ice cover act to enhance the heat transport in the cavity, enhancing the thermal driving and friction velocity at the ice shelf base, and thereby, the basal melt. Thermodynamically, under mobile sea ice, wind upwelled Atlantic Water (AW) from the Nares Strait enter the cavity. Additionally, when sea ice thins, convective overturning drives further upwelling of AW in winter. Mechanically, wind driven inflow intensifies, and is most pronounced under a (negligibly thin) mobile summer sea ice cover; and where it acts in concert with the stronger melt overturning to enhance the friction velocity which predominantly drives the basal melt under the deeper regions. These results suggest that the projected continuation of the warming of the Arctic Ocean until the end of the 21st century and the decline in Arctic sea ice extent and thickness will amplify the basal melt, impacting the long term stability of the Petermann Glacier and its contribution to the future GrIS mass loss and sea level rise.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(39051 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(39051 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Abhay Prakash et al.
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-73', Anonymous Referee #1, 23 Feb 2023
Review, EGUsphere-2023-73
Impact of the Nares Strait sea ice arches on the long-term stability
of the Petermann Glacier Ice ShelfPrakash et. al. 2023
Overview
This paper presents results from a set of experiments examining the impact of changing sea ice cover in Nares Strait on the transport of Atlantic Water towards the Petermann glacier ice shelf. The results show that changing from a landfast sea ice coer to a thin and mobile one in Nares Strait causes upwelling of Atlantic Water at the mouth of Petermann Fjord, thus increasing heat transport towards the glacier. The increased heat transport into the fjord in turn increases the melt of the ice shelf and increases circulation in the fjord. This study demonstrates how changes in the water column along the coast of Greenland propagate into the fjord and directly impact glacier melt.
This study makes a useful contribution to the literature on circulation in the glacial fjords of Greenland, particularly since the model setup enables to study the impact of relatively far-field changes to the fjord. However, I think there are four main issues that need to be addressed to in order to clearly communicate and confidently interpret the results. In addition, I have quite a few minor comments and suggestions on how to improve the text and the figures.
Major comments
Major comment 1: Description of the model:
The model setup is not accurately described in the Methods-section. The reader should get sufficient information of the included model physics from this paper, not having to refer to the author’s previous publication describing the development of the setup. Currently, the Methods-section does not describe the included model physics, particularly which processes regarding sea ice and the ice shelf are included or excluded. Is the melt parameterized with the conventional three-equation formulation by Jenkins? Is surface runoff excluded? The fact that subglacial discharge is missing is only coming across in the Discussion. That is a key piece of information that should be explained and justified in the Methods, and thoroughly discussed in the Discussion, see the next major comment.
The flow of the experiments in this study should be (and partly is but not consistently), landfast to mobile to thin mobile, which represents the order of expected future changes in the sea ice cover (see a minor comment on the naming of the experiments). However, the model is initialized from the thin mobile experiment, apparently since this experiment was available from a previous study. This raises the question if the model has stabilized after changing the sea ice conditions, or if the thin mobile -regime still has an imprint on the hydrography. Model stability should be demonstrated, and this could be achieved, for example, by showing domain average temperature, flow speed and specific density through the entire run time of the experiments. This stabilization could be in the supplement and only briefly referred to in the main text, however, given how the experiments are initialized, stability should be demonstrated.
The paper is also lacking a comprehensive table of used model parameters, particularly the used turbulent transfer parameters and diffusivity and viscosity values. It is fine to have such a table in the supplement, but the information should nevertheless be available. In general, I suggest that the Methods should flow in the order of: Description of model, included physics (and justify excluded physics), domain and grid. Then the initial state and boundary conditions for the experiments with justification. Then present the differences between experiments and how long the experiments are run and the result metrics.
Major comment 2: Lack of subglacial discharge
Since subglacial discharge is known to be a critical driver of the fjord circulation during summer (see reference list, particularly Carroll et.al., 2017), its exclusion should be clearly noted and justified in describing the model in the Methods, and implications thoroughly discussed in the Discussion. At the very least, the typical subglacial discharge volume flux should be compared to meltwater fluxes from the ice shelf (see for example Ehrenfeucht et.al., 2022 for the drainage of the Peterman subglacial system). I do see that since the ice shelf of Petermann is so extensive, subglacial discharge is likely to have a smaller role in the melt, but some attempts should be made to communicate the significance. A particularly interesting point to discuss is the implication of subglacial discharge to the modelled seasonal cycle; since discharge is only present in significant degree during summer, what does omitting it mean for the modelled range of the seasonality?
The experiments use an enhanced turbulent transfer coefficient to improve the model result. This process is not described in the paper, but only referred to the author’s previous publication. However, since it seems like the turbulent transfer coefficient is used to some extent replace the missing subglacial discharge, this process and its meaning to the results should be examined in the Discussion. Overall, due to the uncertainties arising from the missing subglacial discharge, I suggest taking the focus of the results away from the numerical values of the melt rate, and focus more on the changes in the heat transport towards the glacier, and the processes driving those changes. These are in any case more directly related to the change in the sea ice cover, and relevant regardless of the questions relating to subglacial discharge.
Major comment 3: Missing analysis of density and presentation of the results
Since previous work on the glacial fjords of Greenland has so clearly demonstrated the role of buoyancy in forcing circulation (see ref. list), it is surprising that buoyancy forcing has not been considered at all in the analysis of the model results. The lack of isopycnals in the plots and the lack of fjord stratification as a background makes it difficult to understand the processes driving the changes within the fjord, to an extent that the drivers seem to be somewhat uncertain to the authors as well (line 373). I suspect buoyancy is a key forcing in the fjord, and I think careful analysis of the changes in density both between seasons and between the experiments should be made. See previous work on circulation in high-silled glacial fjords, such as Carroll et.al. 2017, Hager et.al., 2022, Kajanto et.al., 2023.
I suggest that the authors review the presentation of the model results, since there are a lot of interesting results that are not shown in the figures; seasonal temperature and salinity in absolute terms, both in Nares Strait and the fjord, isopycnals and flow rates over the sills were not presented. I suggest including all of these, and in the minor comments below, I have included detailed suggestions on how to improve the text and particularly the figures.
Major comment 4: Discussion in a broader context
Currently, the Discussion considers a rather narrow range of references, and the results from this study are not put to context with the extensive literature on the circulation and heat transport in the glacial fjords of Greenland. Furthermore, model uncertainties are not clearly communicated and discussed, and without putting the uncertainties to context, the reader is left to wonder at the validity of the presented results. I suggest restructuring the Discussion thematically, clearly communicating the uncertainties, and making attempts at quantifying these uncertainties. It will add to the value of the results. I have made detailed suggestions in the minor comments below
Minor comments
General comments regarding writing style:
- Please review the use of tilde in the text. Tilde should in general be avoided and be replaced with text whenever possible (with words such as about or approximately). In addition, tilde is used in the manuscript in front of numerical values with several significant digits. It is unclear what is the meaning of the tilde in front of a number that is not rounded up. Also, you should not use tilde in phrases such as “up to 100 m”. Review the use of tilde throughout the text and do not use it unless necessary. I have indicated some of the instances in the line-by-line comments below, but not all.
- The authors tend to describe the results from summer and winter within the same sentence by placing values and descriptive words in parenthesis. This is impossible for the reader to follow, please replace these formulations with proper sentences, such as “Conversely, during winter the heat transport...” or “, while in winter the increase in heat transport...”
- Use active language to make clear what are model results from your study (“Our model results show...” etc.). When discussing results from previous studies, always indicate if the results are from model studies or observations.
- Please review throughout the manuscript that you use a systematic notation of terms such as “high-resolution 3-D” etc., so that the hyphens and capital letters are always spelled similarly.
Line by line comments:
Line 1: Title “Abstract” should have its own line.
Lines 1–2: The glacier is called Petermann Glacier, not PGIS. Also, it is “seasonally shielded” from what?
Line 6: Do not use the manes of the experiments in the abstract, rather describe the changes “from a thick and landfast sea ice regime to a mobile and further thin and mobile sea ice regime” etc.
Line 7: “the implications ... on the PGIS basal melt”. Consider if you should rephrase this to heat transport to the PGIS, see major comment.
Lines 7–8: “Across all three regimes ...” What do you want to say with this sentence? Are you describing seasonality? That basal melting increases during the summer compared to winter? But then what do you mean with the under deeper regions part when does melt rate increase there and relative to what?
Lines 8–9: “Diagnosing this variability ...” The purpose of this sentence is unclear. I suppose you want to say you disentangled the thermal driver from the velocity driver in your study? In this case, say something like: “Our diagnosis of the drivers of the change in melt rate ... suggests that the increase is primarily a result of a higher thermal driving, while the increase in the friction velocity are a secondary effect” etc. But it is still unclear to me which change you are talking about.
Lines 9–12: I assume you are still describing seasonal changes in the model as opposed to differences between the experiments, but this is unclear so please add the word summer somewhere. You should highlight more that you are first describing seasonality in lines 7–12, in contrast to the changes caused by the changing sea ice cover, as you will do later. You could start the description of seasonality on line 7 with something like: “In all three sea ice regimes, the basal melt in the deep regions of the ice shelf presents a seasonal increase during summer. This seasonal increase in melt rate is primarily driven by ...”
Lines 12–15: “Modulations in surface forcing under a mobile and thin sea ice cover act to enhance the heat transport in the cavity”, this is very difficult to read, I suggest rather saying “A thin and mobile sea ice cover enhances heat transport into the ice shelf cavity, by increasing wind-driven upwelling of Atlantic Water from the Nares Strait into the cavity”
Line 15: Skip the “Mechanically”, it is confusing.
Lines 17–20: I suggest adding “the accompanying decline in the Arctic sea ice extent...” and “the basal melt of PGIS”
Lines 26–28: This sentence is really packed, consider simplifying or breaking it into two. Also, the dynamic thinning and acceleration is a consequence of the loss of buttressing forces, so it should come after.
Line 31: Is “sectorial” the North or Northeastern sector?
Lines 36–43: Review the use of tilde, and replace with words when possible. You could refer to your map figure. Use en-dash when presenting a numeric range.
Line 39: Unclear what is meant by “it has been studied through various lenses”.
Lines 43–44: This sentence is overly complicated, you could just say “Heat and freshwater exchange between the ... through the Nares Strait, and the exchange is influenced by ...”
Lines 45–49: Indicate the approximate locations of the sea ice arches to Fig. 1. The arches are even in your paper title, and it is hard to grasp where they are located.
Line 50: What do you mean by “short”?
Lines 50–52: Sounds like the study by Shroyer et al., was very similar to this one. Consider somehow hinting here what is the added value of your study. At least, highlight the differences in the discussion, is it only the bathymetry? Or are there differences in model forcings and physics?
Line 53: “outflow” is not well defined here, say rather something like “Sea ice area and the flow of ice volume through the Robeson Channel...”
Line 55: Omit “flux gate”
Line 56: “ice volume flux was smaller”
Line 61: “formation duration” this is unclear
Line 70–72: See earlier comment, are there any other improvements compared to earlier work that you might want to highlight? Improved model physics?
Line 80–82: It is clear that the author wants to advertise the new package, but it is important that the study of this paper is comprehensible and thus the description of the model in this paper should be independent. Summarize here what are the included model physics (see Major Comment 3 on the suggested flow of the Methods-section), and refer to the previous paper for details (not for basics or results). Do not use the term “standard run” when referring to the previous paper, it is very confusing, particularly given that what is called “standard” is the Thin Mobile experiment, while what could be called the standard run in this paper is the Thick Landfast experiment.
Lines 88–89: Justify (just shortly here), why you use the pre-2010 draft of the ice shelf, and consider discussing the impact of this choice in the Discussion, or the impact of a change in the ice shelf geometry in more general terms.
Line 91: So, is one of the key novelties in this study that you have a second deep sill in the domain? If yes, then this should come across in the results more clearly, and you should discuss the impact of the sill to the circulation and heat transport in the cavity.
Line 91: Here at the latest you should explain how you include the ice shelf in the model. How do you parameterize the ice-ocean interaction? Explain and justify that you don't have subglacial discharge.
Line 92: You should move the description of initialization and boundary conditions here, before you describe the different experiments
Line 94: Remove the subclause starting with which, you have already introduced the module.
Line 95: For general readability, I suggest calling the experiments simply “Landfast”, “Mobile” and “Thin Mobile”, and to use the full names throughout the paper and also in the figures. The abbreviations are too similar and impossible to follow.
Lines 95–107: Introduce the experiments as a clear progression towards expected future conditions: Landfast represents the 90’s conditions, Mobile when the sea ice arches fail to form and Thin Mobile the expected future, etc. Keep the same order you have in Table 1. throughout the paper.
Lines 97–99: This belongs to the paragraph where you present the initialization of the model, and again, do not use the name “Standard run”.
Lines 102–103: This sentence is unclear. Use active language. It is not possible to follow the subclause “, which...”, I suggest writing the point in a separate sentence.
Line 105: “sea ice velocities are retained from the Thin Mobile run”. This is very confusing, and the reader does not know the forcing or initialization procedure at this point. You should include a section describing the initialization and boundary conditions, before describing the experiments, as already commented earlier.
Line 108–111: As I said before, you should carefully present and justify your initialization procedure, and demonstrate that the runs have stabilized after these perturbations. However, it should be made clear that the progression of the modelling procedure is separate from the evolution the experiments aim to present. Currently, these are muddled together. Since these progressions are opposite, you should make very clear that it is ok to initialize the runs in this way.
Lines 115–134: Now you finally describe the initialization and boundary conditions, this is critical information to know before describing the experiments, move up and make its own subsection.
Line 115: Start by saying that you run the model from 2014 to 2017, and only then say when the model results are presented.
Line 117: I do not understand the subclause starting with “for which ...”
Lines 119–122: I’m not sure I get this right, so you get atmospheric temperature from RACMO and maybe winds? And which sea ice conditions from the A4 ROMS? State these clearly! And then you use the IceNudge module to transfer these to boundary conditions for the ocean? Be clearer, what variables do you get from which model? What are the time steps? Do you have some downscaling procedure?
Lines 124–125: I don’t follow, se is the model set up to create the arches? Or you have chosen a period when this happens in the model?
Lines 125–126: I’m even more confused, so a thin and mobile sea ice cover agrees well with what and when?
Lines 126–127: Remove the reference to the standard run, I think it is at the root of the confusion.
Lines 129–131: Which fields? Temperature, salinity, velocity?
Lines 130–132: I’m confused, are these the hourly conditions you just mentioned? Why this period for the average?
Lines 115–134: Ignoring the previous paper, is the general picture that the model was calibrated to reproduce a thin and mobile year since there was quality forcing data and observations for these types of conditions? And then you create the experiments with more sea ice by changing sea ice parameters indicated in Table 1? If yes, please make sure that this idea is clearly communicated. I still don’t know at which point of the Landfast and Thick Mobile experiments the sea ice conditions were changed, if the model stabilized in reference to this change and if any other parameters or forcings were changed. Be sure to communicate these clearly. And show in the supplement that the model stabilizes.
Line 137: How is the melt rate computed?
Lines 138–142: I don’t see how this information is necessary
Lines 143–182: This is a very detailed description of how to process model output, which is unnecessary information for most of the readers interested in the results. I suggest moving this diagnostics part into a supplementary. You don’t really need to have the subsection of “Model output and diagnostics”.
Lines 184–186: In reference to Major Comment 2: I suggest reformulating this to more generally examining the circulation and heat transport from Nares Strait to the fjord, and further into the PGIS cavity. And the yes, you will show the melt rate as well, but put more focus on the processes in between.
Line 190: This is for the whole domain, right? Fig. 3 is also partly of Nares Strait. I recommend showing the seasonal temperature and salinity vs depth of the inflow to the fjord.
Line 191: Remind the reader here how you define the seasons. Have you defined a “wider PF area” somewhere?
Line 195: “lateral inflow and outflow” of which directions? Write it in the sentence, don’t use the “, respectively”.
Lines 196–198: I find the references here confusing, just refer to the discussion.
Line 200: You should also show the flow along the fjord.
Lines 205–206: Write separate proper sentences to each of the seasons, this formulation is very difficult to read.
Lines 206–208: When?
Lines 208–212: It’s nice to see the across-fjord flow, but since we have not seen how temperature and salinity change between the seasons in Nares Strait, it is difficult to put the change here in the fjord to context. Also, along-fjord isopycnals are essential to see what role density-driven inflow over the sills play.
Lines 216–218: Where do I see this? Again, along-fjord profile with isopycnals is missing.
Lines 219–22: This is very dense and difficult to follow. Why do we need to know the annual mean? I think summer and winter inflow and outflow values would be enough. Percentages of increase relative to annual mean have little value, summer increase relative to winter would be more informative. Write proper sentences, and do not use parentheses.
Line 225–228: Rephrase. The modelled annual mean of what? Moreover, why do we need to look at the annual mean, I don’t see its value.
Line 225: See suggestions to Figure 5
Lines 230–235: You should start by describing the stratification within the fjord. You have a strong pycnocline separating the Polar Water layer from the Atlantic Water. Due to the difference in density, the seasonal response of these layers is different. Throughout the results you should describe the changes relative to the layers, not the depth. This way you will avoid clumsy formulations like “beneath the shallower ice base”. Along-fjord study of the density is key to understand what drives the circulation. Seasonal along-fjord flow patterns would also be of key importance, right now they are not shown. See suggestions to Figure 5.
Line 238: First go through the circulation in the fjord properly, and describe the melt rate in a new paragraph.
Lines 239–244: Write separate sentences for winter and summer, and give the values for mean melt rates for winter and summer, not the anomalies.
Lines 245–249: What is the difference of the “regions of strong melt aligned with the PGIS flow direction” and the “substantial lateral variability”. Aren’t they the same thing? What is meant by “and by up to 50 m/yr”?
Lines 251–253: Why the “presumably” and “likely”? Shouldn’t you know from the model if it due to AW inflow? Note that you have not defined friction velocity.
Lines 254–256: Rephrase this in terms of buoyancy and the pycnocline.
Line 257: If you would have an along-fjord plot of flow streamlines, then you would know and would not need to guess.
Lines 257–259: This sentence belongs to Section 3.2
Line 260: Rename this section, and do not include the name of the package. “Response to decreasing sea ice” etc.
Line 261: Don’t go to the basal melt first. Keep the structure as in 3.1, and start from far field and work your way in towards the melt rate. This way it is possible to track how the impact of changing sea ice conditions propagate into the fjord.
Lines 275–277: Rewrite without the parentheses, remove “calculations show”. Instead of “deeper regions”, say “below the pycnocline” or “within the Atlantic Water”.
Line 281: “primarily driven by the strengthening of the fjord circulation”. This is very interesting, and yet we know nothing of the fjord circulation and along-fjord transport
Lines 283–285: This sentence is difficult to follow, rewrite without the parentheses.
Lines 291–293: “In winter, the relative increase ...” This sentence is hard to follow.
Lines 293–295: Again, this sentence is difficult to follow. State first the general trend before giving numbers: “As the sea ice becomes mobile, the deep water masses seaward of the sill get warmer and saltier both in winter and summer.” etc.
Lines 295–298: “As the sea ice thins in addition to becoming mobile, ...”
Lines 298–299: “The colder water masses are advected into the PGIS cavity”. This is too vague. At this point, you have not discussed the transport processes into the fjord, and it really should be done. It is not clear what is the process that delivers the change into the cavity.
Lines 305–310: Finally, we get to this, but this needs a better figure to clarify the process, see comments to Fig. 10.
Line 316: Have you defined the upper ocean layer?
Lines 322–325: This is probably the most important result, and yet you say “it is likely”. Shouldn’t you see what happens in the model? You will once you study the isopycnals relative to sill depth.
Lines 325–327: These are important results, write more explicitly. What is “this mechanism”. In the following sentence you switch to discussing cold water. Which cold water and how does that follow from the previous sentence?
Lines 329–332: Are you sure about this? The cavity is well sheltered from the impact of winds, and I presume what you see in the cavity is density-driven inflow.
Lines 334–335: Where do you discuss the heat budget?
Line 336: “As the sea ice becomes mobile, the inflow during winter along the western fjord sector strengthens and intensifies with depth, being up to ...”
Line 338: What is an “increment in the inflow”?
Lines 345–347: This looks like a numerical error, are you sure that this is real?
Lines 349–350: Do not use the exact same titles as you had in results. I recommend rethinking the structure of the Discussion, perhaps organize thematically or in the order of your most significant findings.
Lines 351–353: modelled mean flow where? Is consistent in terms of what?
Line 353: What do you mean by “was seen”?
Line 354–356: I cannot follow this sentence, please rephrase.
Lines 353–359: I don’t get it, why the “likely”? “it is thus speculated”, who speculates? Why do you need to speculate if you have the results from all experiments, as you say in the next sentence?
Line 361: I don’t think it is due to air-sea momentum transfer, the density gradient should be driving this, but since you results lack the analysis of density, we cannot see it from the figures. See Carroll et.al., 2017, Kajanto et.al., 2023 and Hager et.al., 2022 for similar work in other glacial fjords.
Line 362: Who suggests? Are you saying that you find that there is water mass exchange and renewal in the fjord throughout the year? This a very interesting result, but you should state what processes are driving the water mass renewal, and discuss the role of missing subglacial discharge.
Line 364: What do you mean by “can be”?
Line 368: Why the “may be”?
Lines 369–372: Rephrase, and no need to repeat all the result values. I think you are trying to say that the inflowing water over the sill is warmer and thus causes more melt, which in turn creates a stronger buoyancy forcing that drives a stronger circulation in the cavity.
Line 373: “Irrespective of the causes”. How can you say this? Understanding the causes is the key!
Line 375: Quantify the cold bias. Is the bias systematic around the year or does it have seasonality? Is the bias connected to the missing subglacial discharge?
Line 377: Remember to explain this modification of turbulent transfer already in Methods. Remember to include a table of all model parameters into the supplement. Write here the values and compare them to the values suggested in Jackson et.al. 2020. Discuss the implications of having an increased melt over the entire ice shelf as opposed to having the subglacial discharge plume.
379–380: Who assumes? You can’t say this without the comparisons and discussion of the previous comments.
Lines 381–382: And subglacial discharge? Reference?
Lines 385–386: I don’t follow. Is this your model result you are describing?
Lines 386–388:
Are you describing your model results? Or observations? Why the “likely”? Then you apparently switch to discuss observational analysis of melt channels mid-sentence? Subglacial discharge plumes often erode channels under ice shelves, but how does that connect to your model since you don’t have plumes? These are interesting points but you should elaborate and carefully consider what your model can say. The connection with the ice shelf shape is interesting and merits more thought.
Line 392: Were these both observational studies?
Line 395–397: Which one are you talking about now?
Line 397–398: This is a very harsh sentence after the comparison you just did.
Line 401: It is unclear what the reference to Holland et.al. Is for here.
Lines 401–404: This is the general shape you get with the Jenkins’ 99 three-equation parameterization that I supposed this model also uses (which is information that should appear in Methods).
Lines 405–410: This paragraph is just listing a number of effects that could have an impact, some less interesting (like the well-known pressure-dependency of the melting point), and some very interesting, like potentially plume-eroded channels in the shelf. I recommend reviewing a bit more of the relevant literature on subglacial discharge plumes (some suggestions in the reference list below) and revisiting your melt rate parameterization, and carefully considering what are the clear factors impacting the distribution of melt in this model, and what are factors that are excluded in the model, but could potentially still impact the melt rate.
Lines 413–415: How does this relate to depth/thermocline?
Lines 416–417: Elaborate on what is the significance of no subglacial discharge. What sort of error does that cause in the results and why are the melt rates given in this study relevant?
Lines 413–414: what inconsistencies, and how is the Table relevant? Elaborate on what is the meaning and impact of these inconsistencies to this study.
Line 420: What novel mechanisms do you mean?
Lines 420–421: I think you want to say that even though you are not calculating the melt rate correctly, the relative changes caused by a change in sea ice are robust. You could also point out that your estimates of the change in the Atlantic Water properties at the mouth of the fjord are only little impacted by the lack of subglacial discharge. It would be really good if you could quantify how much you're tweaking the turbulent transfer and neglecting subglacial discharge impacts the circulation regime. I suspect you have runs, at least in the Thin Mobile configuration with different turbulent transfer coefficient? How much does the change in the coefficient mean in terms of meltwater flux from the glacier? Or in terms of inflow over the outer sill? It is hard to compare to the buoyancy flux from a plume, but some attempts at quantifying this uncertainty should be made. You could just attempt to make an estimate of the relative increase in the plume melt rate based on the increase in Atlantic Water temperature you project, see Ezhova et.al., 2018.
This is exactly why I suggest on taking the focus more towards the change in the inflowing Atlantic Water properties and heat transport into the fjord, since you are on a more solid ground in terms of reliability of the results.
Line 423: What kind of estimates are these, model?
Lines 440–444: This is too long, break into shorter sentences. You're talking about basal temperatures and calving front temperatures within the same sentence. Be clear in what you mean. Give references to undercutting.
Line 446: Are you talking about future projections? Are these references predicting a future warming?
Line 448: “the upwelling mechanisms” this is too vague. I know by know what you refer to with this, but be more specific. Upwelling where?
Line 448: Speculate also what would the impact from increased freshwater flux from Greenland be
Lines 452–460: If you analyze the results in density space as I have suggested, you can add here that your model is able to show the density-driven inflow over the outer sill from Nares Strait into the fjord that fills the deep basin and reaches the grounding line.
Line 452 Wind-driven upwelling, here and elsewhere
Line 464: Be very careful, I don’t think you have shown that wind drives the circulation in the PGIS cavity. Wind drives the upwelling of AW in Nares Strait and I presume that density drives the circulation in the fjord.
Lines 469–470: Did you not prescribe this in the model? Then it’s hardly surprising that there is agreement.
Lines 470–475: You have set up the model to represent the conditions so this is not really discussion. You should rather move this to where you describe the runs and say that the experiments represent approximately the conditions of this and that period.
Lines 475–477: I like the sentence, but again, it is not clear that wind drives the inflow into the cavity, see Carroll et.al., 2017.
Line 478: Remove the “interpreted”, and reformulate to state that this is your model result.
Lines 481–482: Reformulate, “not all of the heat is taken up by ice melt”, etc.
Lines 486–488: Again, I don’t think you have properly analyzed the drivers to say this.
Line 493: The summary is very long and slightly repetitive, you should compress, and highlight the results that are new from this study.
Lines 500–534: you don’t have a bullet point explaining your findings on the impact of a change in sea ice to Nares Strait, and you don’t have a bullet point dedicated to circulation in the fjord. Several bullet points describe similar points regarding melt rate, and are thus quite repetitive.
Lines 535–539: References?
Lines 535–548: This is borderline Discussion, and too long, write a short and clean paragraph on how your results describe the potential future, uncertainties are for the Discussion. Skip the a and b.
Figures and Tables
Figure 1:
I suggest putting quite a bit more effort to this figure. In 1a, I suggest cropping much closer to the domain, now much of the figure are is out of the domain, and the area of interest is too small. You can add a small inset figure indicating the location on a larger scale map. The blue and green colors of the background map are too dark, the red and yellow are too hard to read. The caption says that the sea ice arch locations are marked in the figure, but I cannot see them. In 1b, use different, brighter colors in the lines, they cannot be distinguished from this figure. I suggest including an along-fjord profile and indicating the sills, the location of the cross-sections and all of the other features you refer to later in the paper (for example Fig. 3a). This is important since bathymetry and draft are the key novelties in this study, and the inner sill is a new(?) feature. In all figures be more consistent with fonts and font sizes, the figure looks busy with so many different fonts and sizes.
Figure 2:
This figure is difficult to follow, but that can be fixed easily. I suggest the following changes: First column should be the Landfast experiment, since that represents earlier, “historical or pre-industrial”, conditions in this paper. Do not use the abbreviations, use the full names of the experiments. Consider adding the words summer, winter and year-round as column headers.
Table 1:
Once you have revised the text describing the initialization of the experiments, remember to revise the caption accordingly. It is now impossible to grasp which run starts from when and where. Remove the abbreviations, and update the experiment names. Give the duration of the summer and winter seasons also in days or months so that the reader does not need to calculate, and would be better to give the time in calendar days. Remove the “Standard run”.
Figure 3:
Review the use of space in this figure. White space and labels take more space than the actual results you attempt to show. Seasonal mean speed of what? You should say that these are ocean currents. Consider adding a small inset in the empty space of 3b) that would show a line plot of the mean depth along the fjord vs the Y-coordinate. Check that all of the features indicated in red in 3a) will be also indicated in the along-fjord profile of Fig. 1c). Indicate the gyres discussed in the text with arrows, and refer to the figure when describing the gyres.
Figure 4:
Why are the inner profiles (4a,b) above the outer ones (4a,d)? I think it would be more intuitive the other way around. I do not know how these cross sections relate to the fjord bathymetry (See comments to Fig. 1). I the black color on top of the water column the ice shelf? If yes, mark it with a different color than the terrain, and use the same color for the ice shelf throughout the figures. What is the gap in the ice shelf in 4a,b? Increase the font size of the temperature values of the isotherms. Blue and red are not good font colors on a blue-red background. Why the cropped width? There is some white space in between. Include in the caption which way is positive (towards the glacier?)
Figure 5:
This is the first along-fjord figure of the paper, much overdue, and overall, quite unhelpful. Differentiate the ice shelf from the terrain by using the same color as in Fig. 4, now it is difficult to see the GL. I don’t think the annual means are valuable at all, I would just remove them. Similarly, showing anomaly relative to annual mean is not meaningful. Also, the salinity anomaly is also rather uninformative, density should be salinity-dominated here anyway. I recommend showing isopycnals for both seasons, as well as temperature for winter and summer. Plot the isopycnals at even intervals, in 5a) it is hard to see the pycnocline since the contours jump from 27.1 to 25.9. You can explain the range of annual variability in the text. Consider also including the along-fjord heat transport overlaid with streamlines. Otherwise, we are not seeing along-fjord flow at all. This would mean four panels, which would make it possible to increase the panel size a bit.
Figure 6:
There is a lot to unpack in this figure, and again, I do not see the point of showing anomalies relative to annual mean. If you want to plot anomalies, why not plot the summer increase relative to winter? However, I would prefer to just show melt rate in winter and melt rate in summer with the same color range, on a much bigger figure. I find panels d—i deeply unhelpful, and I do not see the point. If you are trying to show the relative contributions of temperature increase and the increase in friction velocity to the melt rate, it is not working. Why are these anomalies? Could you just show the water temperature at the base of the shelf in summer and winter, and then the friction velocity. So as in e,f,h and i but in absolute terms? I can’t imagine if it will be better, but something should be done. Also, note that you have not defined anywhere the friction velocity. Turn panels d and g so that ice draft is along the y-axis and increases downwards, it’s more intuitive. Also, I recommend plotting temperature and friction velocity in absolute terms for both winter and summer. There is a significant amount of white space in the figure so should be possible to reorganize the figure so that the shelves appear bigger, right now they are too small.
Figure 7:
I recommend to plot in panels b and c the absolute annual mean melt rate in the Thick mobile and thin mobile experiments. This way your first row is comparison of annual mean melt rates. If you have included summer and winter melt rates (not anomalies) in Fig. 6, you can show panels e,f,h,i as they are. However, write on the panel if they are winter or summer, and use the full names of the runs instead of abbreviations (landfast to mobile etc.). Flip panels d and g to have the draft as y-axis increasing downwards. Reduce white space and make the shelves appear bigger. Reorganize to have seasons in columns, as you have in other figures. Use different colors to plot the substractions in panels d and g to avoid confusion, and keep the colors also in other figures.
Figure 8:
Flip axes and change color as in the previous figures. Here I think the difference in annual mean makes sense.
Figure 9:
A lot of the same comments as earlier. Change the color of the ice shelf, keep the seasons in columns to be systematic in your figures and write summer and winter on top of the columns, do not use abbreviations of the experiment names but spell out the names. Again, I think in order to understand to flow under the ice shelf, we need to see isopycnals for both seasons and both, and since intensification of the circulation was a key result, you should plot the flow somehow, for example along-fjord velocity in blue-red. You could save space by skipping the difference in salinity if you show isopycnals, and only show the seasonal difference in temperature.
Figure 10:
In panels a—d) is there a black (sea ice) layer on top, or is it the darkest color of the color scale? If it is ice, plot it with a different color, if it is the darkest shade of the color scale, plot the terrain in brown (also in other figures). Consider writing Nares Strait on the figure, so that it is clear without reading the caption that this is not the fjord. Write winter and summer on top of the columns. Do not use the abbreviations of the experiments, and do plot isopycnals on this figure. 10 e and f are really informative and important, it is important that this figure comes earlier.
Figure 11:
Keep seasons in columns, write winter and summer on the figure, do not use abbreviations, use a different color for ice to separate it from terrain. State in the caption with words where this cross section is from. Add the colored circle indicating the section, as you had in the previous fjord section figure. Check from the model what happens in the narrow return-flow column in 11d, and state that in the caption, it draws a lot of attention. It seems to be related to the gap in the ice shelf.
Table 2:
Write the experiment names, no abbreviations. It is not clear where this is calculated from, is it the fjord mouth corss-section?
References:
Carroll et.al., 2015, “Modeling Turbulent Subglacial Meltwater Plumes: Implications for Fjord-Scale Buoyancy-Driven Circulation”, J. Phys. Oceanography
Carroll et.al., 2017, “Subglacial discharge-driven renewal of tidewater glacier fjords”, J. Geophys. Res.-Oceans
Ehrenfeucht et.al., 2022, “Seasonal Acceleration of Petermann Glacier, Greenland, From Changes in Subglacial Hydrology”, Geoph. Res. Lett.
Ezhova et.al., 2018, “Dynamics of three-dimensional turbulent wall plumes and implications for estimates of submarine glacier melting”, J. Phys. Oceanography
Hager et.al., 2022, “Subglacial Discharge Reflux and Buoyancy Forcing Drive Seasonality in a Silled Glacial Fjord”, J. Geophys. Res.-Oceans
Jackson et.al., 2020, “Meltwater Intrusions Reveal Mechanisms for Rapid Submarine Melt at a Tidewater Glacier”, Geophys.Res.Lett.
Jenkins, A. 1999, “The impact of melting ice on ocean waters”, J. Phys. Oceanography
Kajanto et.al., 2023, “Impact of icebergs on the seasonal submarine melt of Sermeq Kujalleq”, The Cryosphere
Straneo F. And Cenedese C., 2015, “The Dynamics of Greenland’s Glacial Fjords and Their Role in Climate”, Ann. Rev. Mar.Sc.
Citation: https://doi.org/10.5194/egusphere-2023-73-RC1 - AC1: 'Reply on RC1', Abhay Prakash, 27 May 2023
-
RC2: 'Comment on egusphere-2023-73', Anonymous Referee #2, 10 Mar 2023
The manuscript describes the output from a regional, three-dimensional numerical model of the ocean. It includes two lateral boundaries in the north and south where water properties (temperature and salinity) and velocity vectors are specified and complex coastlines. The model geometry includes a floating ice shelf (Petermann Gl.) as well as realistic bottom topography both under the floating glacier and everywhere else within the modeling domain. The presence of ice (sea ice and glacier ice) is prescribed (not modeled) as a surface boundary condition with a surface stress (vertical momentum flux) condition that depends on the prescribed (not modeled) ice concentration and ice velocity. The ice is allowed to melt because of conductive and turbulent heat flux from the ocean to the ice. Much of this model set-up and selective comparison to limited data is published as Prakash et al. (2022). Reading this publication, I consider essential to critically evaluate the present manuscript. I recommend against publication for the following three major at least 10 minor reasons:
A_ Prior ice-ocean models of the region exist (Shroyer et al., 2015; Shroyer et al. 2017) and substantial overlap exists with few new or original results. Unlike the present model, the prior models include a prognostic (as opposed to a prescribed) sea ice model. Basal melt rates are predicted under the floating portion of the glacier by both present and prior models. Both present and prior models evaluate the differences between sea ice in the adjacent ocean and fjord that is mobile vs sea ice that is not mobile. I cannot discern fundamentally new results or insight in the present model application that basically repeats prior modeling with a different model, that, I feel, has more weaknesses than strength over prior models.
B_ The present model uses temperature and salinity at the boundaries that are both poor and unrealistic. This “cold bias” is prominently discussed in Prakash et al. (2022), but it is introduced in the present manuscript almost as an after-thought in passing on Line-375. Moored and synoptic observations indicate salinity and temperature of about 34.7 psu +0.2 C for the Atlantic-influenced waters both in Petermann Fjord and below the ice shelf (Muenchow et al., 2016; Washam et al, 2018). The model provides “… annual mean temperature and salinity of the water masses that overflow the inner sill…” that are 34 psu and -0.7 C (Line-230). So, the modeled ocean is almost 1.0 C too cold and 0.7 psu too fresh at a depth where seasonal and interannual variations are less than 0.1 C and 0.01 psu! So, the model includes huge biases that I find unacceptable. Note that Prakash et al. (2022, page-27, Line-6/7) “ … suggest applying a depth-dependent bias correction to the upstream A4 T-S fields such as in Shroyer et al. (2015).” Please follow your own suggestion.
C_ The present model attempts to compensate for the “cold and fresh bias” by artificially increasing the vertical turbulent exchange coefficient to produce melt rates that are somewhat agreeable to observations of melt rate. So, the model includes a wrong (northern) boundary condition (from the nested A4 model) and a wrong “turbulence model” to make things right. From my observational perspective two wrongs rarely make a right and I thus loose trust in this model as, it appears, the model can “nudge” or “fix” anything as there is always a parameter or dial that one can change to obtain any desired result. I do not endorse this practice, especially since the authors know the proper way to remove these biases and perhaps use realistic turbulent exchange coefficients.
In addition to the above major weaknesses, there are a number of more minor concerns
1_ The manuscript is too long and unnecessarily complex. I could not discern much substantial difference between the “Thick-Mobile” and the Thin-Mobile” case. What differences there are, I feel, may fall into the domain of model uncertainty, noise, and poorly constrained observational and/or model parameters.
2_ The manuscript is too long and contains trivial co-ordinate transformations. Just state that co-ordinates are rotated into along- and across-channel components rather than spelling out what I perceive as trivial algebra (Lines-143 through Line-182). I understand that the present model is finite element that that this may not be as trivial computationally as it is in finite-difference models or observations.
3_ I am confused and disturbed by streamlines that start at boundaries and end in the interior. Figure-3 offends my sense of mass conversation and lateral boundary conditions. Would you not expect a zero velocity along the coast?
4_ Figure-4 has large spatial oscillations that result from a poorly used graphics package. Please learn how to draw contours properly. Furthermore, I prefer distances in km as opposed to Longitude. The strong bottom-intensified slope current under the western ice shelf during the summer (Fig.-4b) caught my attention. Strangely, no such current exists in the fjord where bottom slopes may be similar, well, I cannot tell, because Longitude rather km is used for distance.
5_ On Line-310 it is stated that “… upwelled AW [Atlantic Water] from the adjacent Nares Strait enter the fjord…” How does this work? Would not winds from north to south in Nares Strait that may cause the AW to upwell along Greenland in the east also lower sea level along Greenland relative to Canada? Would then Petermann Fjord not respond with a large outflow the way an estuary would with lower sea level at is mouth?
6_ Please provide uncertainty estimates to your estimates of heat fluxes such as summarized in Table-2 as it is done with observational estimates of these same fluxes.
7_ In my view the authors mistakenly equate heat flux into the fjord and/or glacier cavity with basal melting (Line-368, Line-373). First, most of the heat flux into the fjord leaves it. Only a small fraction of the heat is actually used to melt the ice which in the present model is largely caused by the artificially increased vertical turbulent exchange coefficient. Second, did the authors actually check, if mass is conserved? Does the volume flux add to the amount of melting?
8_ The comparison of model predicted basal melt rates with observations appears to me less systematic than it could be. The authors appear to pick whatever value from whatever paper for whatever season that fits their purpose. Perhaps this part of the discussion can be strengthened by more clearly delineate seasons, space, and vastly different observational techniques (snapshots vs. moored observations or remotely satellite vs fixed radar stations).
9_ Line-438/439/440: The Rueckamp et al. (2019) reference clearly states that the “still attached” new ice island has already “dynamically detached,” that is, from a practical or physical perspective, this segment of the ice shelf is already gone.
10_ Line-511: The “Nares Strait sea ice arches” (mobile sea ice) add 2 m/y basal melt to the 24 m/y (Line-243), so the entire paper is about a 10% effect. The supply of heat to the fjord or glacier cavity matters little, but both the vertical turbulent mixing and the grounding line discharge of freshwater (not included in this model) may dominate over this 10% effect.
In summary, the manuscript should not be published without properly fixing both water mass bias and the vertical exchange. Even with these fixes, substantial overlap with prior modeling work may limit new and original insights that warrant publication.
Muenchow, A., L. Padman, P. Washam, and K. Nicholls: The ice shelf of Petermann Gletscher, North Greenland and its connection to the Arctic and Atlantic oceans, Oceanography, 29 (4), 84-95, 2016.
Prakash, A., Q. Zhou, T. Hattermann, W. Bao, R. Graversen, and N. Kirchner: A nested high-resolution unstructured grid 3-D ocean-sea ice-ice shelf setup for numerical investigations of the Petermann ice shelf and fjord, MethodsX, Vol 9, https://doi.org/10.1016/j.mex.2022.101668, 2022.
Rückamp, M., Neckel, N., Berger, S., Humbert, A., and Helm, V.: Calving induced speedup of Petermann glacier, Journal of Geophysical Research: Earth Surface, 124, 216–228, 2019.
Shroyer, E.L., R.M. Samelson, L. Padman, and A. Muenchow: Modeled ocean circulation in Nares Strait and its dependence on land-fast ice cover, J. Geophys. Res., doi:10.1002/2015JC011091, 2015.
Shroyer, E.L., L. Padman, R.M. Samelson, A. Muenchow, and L.A. Stearns: Seasonal control of Petermann Gletscher ice-shelf melt by the ocean's response to sea-ice cover in Nares Strait, J. Glac., pp. 1-7, doi:10.1017/jog.2016.140, 2017.
Washam, P., A. Muenchow, K. Nicholls: A decade of ocean change impacting the ice shelf of Petermann Gletscher, Greenland. J. Phys. Oceanogr., 48, 2477-2493, 2018.
-
CC1: 'Reply on RC2', Tore Hattermann, 24 Mar 2023
Since there is a function for open discussion, we might as well use it. Usually, an author's reply starts by thanking the reviewer for their constructive criticism. Although Anonymous Referee #2 points out a few relevant aspects, those appear being put forward as an exaggerated criticism (of mainly technical aspects) and lack of goodwill to see an intellectual contribution that this work may have for the understanding of the response of the Greenland Ice Sheet to future climate change.
The results of this study are based on a hypothesis on how long-term changes of sea ice properties might affect the future stability of PGIS. The premises for this deduction are largely based on previous modelling work that discovered the impact of seasonal varying sea ice cover. This is clearly marked and acknowledged in the manuscript. However, the intellectual transfer of how this mechanism may play out to a concretely observed interannual change of large-scale sea ice properties in the region is a novel contribution. To my knowledge, there exists no peer-reviewd literature that has explicitly addresses this aspect.
Then, the study uses a moderately sophisticated numerical tool to lend support for this hypothesis and further detail its manifestation and impacts, and with arguable advantages and disadvantages compared to previous modeling works. There are certainly shortcomings in the model that is being used, and all models are wrong, but in this case, the primary use is (i) prescribing the sea ice to specific conditions (which is not possible with a fully coupled sea ice module, as might have been misunderstood by the reviewer) and (ii) illustrating and characterize the qualitative and quantitative impacts and characteristics on the ice shelf. In particular, the latter is undertaken with unprecedented spatial detail in PGIS fjord, also investigating aspects that have previously never been addressed, such as providing high resolution 2-D maps of the basal melt pattern, investigating the response of thermal driving vs. friction velocity, comparing (not equating) those to melt rate changes to heat flux anomalies. Admittedly, there is a bias and likely a high uncertainty on the absolute value of the simulated basal melt rates, which should be addressed in a revised version of the manuscript. However, the primary findings of the qualitative changes stand independent of this and have not yet been explored in this form.
So, while I partially agree with the Anonymous Referee #2 editorial comments and on assessing technical shortcomings of the model (which, however, are already openly communicated in the manuscript and altogether only have minor impacts on the main conclusions of this study), I am surprised over its harsh judgement that this work should not be published due to its lack of relevance and/or novelty. Even if possibly not providing a ground-breaking advance, this study reflects the solid result of a hard-working ECR's professional development that should not be defeated like this. Much less solid of impactful work has been published and celebrated, and in an educational system that is based on peer-review publications as a standard, there should be room for incremental advance like this in the literature.
Personally, I think that publicly posting such exaggerated criticism (for whatever reason) protected by anonymity is simply not fair. Hoping that the editor will see though the mist, I will encourage my student to reasonably address the points raised in a revised version of the manuscript and live up to high ethical standards in in his future scientific career himself.
Sincerely,
Tore Hattermann
Citation: https://doi.org/10.5194/egusphere-2023-73-CC1 - AC2: 'Reply on RC2', Abhay Prakash, 27 May 2023
-
CC1: 'Reply on RC2', Tore Hattermann, 24 Mar 2023
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-73', Anonymous Referee #1, 23 Feb 2023
Review, EGUsphere-2023-73
Impact of the Nares Strait sea ice arches on the long-term stability
of the Petermann Glacier Ice ShelfPrakash et. al. 2023
Overview
This paper presents results from a set of experiments examining the impact of changing sea ice cover in Nares Strait on the transport of Atlantic Water towards the Petermann glacier ice shelf. The results show that changing from a landfast sea ice coer to a thin and mobile one in Nares Strait causes upwelling of Atlantic Water at the mouth of Petermann Fjord, thus increasing heat transport towards the glacier. The increased heat transport into the fjord in turn increases the melt of the ice shelf and increases circulation in the fjord. This study demonstrates how changes in the water column along the coast of Greenland propagate into the fjord and directly impact glacier melt.
This study makes a useful contribution to the literature on circulation in the glacial fjords of Greenland, particularly since the model setup enables to study the impact of relatively far-field changes to the fjord. However, I think there are four main issues that need to be addressed to in order to clearly communicate and confidently interpret the results. In addition, I have quite a few minor comments and suggestions on how to improve the text and the figures.
Major comments
Major comment 1: Description of the model:
The model setup is not accurately described in the Methods-section. The reader should get sufficient information of the included model physics from this paper, not having to refer to the author’s previous publication describing the development of the setup. Currently, the Methods-section does not describe the included model physics, particularly which processes regarding sea ice and the ice shelf are included or excluded. Is the melt parameterized with the conventional three-equation formulation by Jenkins? Is surface runoff excluded? The fact that subglacial discharge is missing is only coming across in the Discussion. That is a key piece of information that should be explained and justified in the Methods, and thoroughly discussed in the Discussion, see the next major comment.
The flow of the experiments in this study should be (and partly is but not consistently), landfast to mobile to thin mobile, which represents the order of expected future changes in the sea ice cover (see a minor comment on the naming of the experiments). However, the model is initialized from the thin mobile experiment, apparently since this experiment was available from a previous study. This raises the question if the model has stabilized after changing the sea ice conditions, or if the thin mobile -regime still has an imprint on the hydrography. Model stability should be demonstrated, and this could be achieved, for example, by showing domain average temperature, flow speed and specific density through the entire run time of the experiments. This stabilization could be in the supplement and only briefly referred to in the main text, however, given how the experiments are initialized, stability should be demonstrated.
The paper is also lacking a comprehensive table of used model parameters, particularly the used turbulent transfer parameters and diffusivity and viscosity values. It is fine to have such a table in the supplement, but the information should nevertheless be available. In general, I suggest that the Methods should flow in the order of: Description of model, included physics (and justify excluded physics), domain and grid. Then the initial state and boundary conditions for the experiments with justification. Then present the differences between experiments and how long the experiments are run and the result metrics.
Major comment 2: Lack of subglacial discharge
Since subglacial discharge is known to be a critical driver of the fjord circulation during summer (see reference list, particularly Carroll et.al., 2017), its exclusion should be clearly noted and justified in describing the model in the Methods, and implications thoroughly discussed in the Discussion. At the very least, the typical subglacial discharge volume flux should be compared to meltwater fluxes from the ice shelf (see for example Ehrenfeucht et.al., 2022 for the drainage of the Peterman subglacial system). I do see that since the ice shelf of Petermann is so extensive, subglacial discharge is likely to have a smaller role in the melt, but some attempts should be made to communicate the significance. A particularly interesting point to discuss is the implication of subglacial discharge to the modelled seasonal cycle; since discharge is only present in significant degree during summer, what does omitting it mean for the modelled range of the seasonality?
The experiments use an enhanced turbulent transfer coefficient to improve the model result. This process is not described in the paper, but only referred to the author’s previous publication. However, since it seems like the turbulent transfer coefficient is used to some extent replace the missing subglacial discharge, this process and its meaning to the results should be examined in the Discussion. Overall, due to the uncertainties arising from the missing subglacial discharge, I suggest taking the focus of the results away from the numerical values of the melt rate, and focus more on the changes in the heat transport towards the glacier, and the processes driving those changes. These are in any case more directly related to the change in the sea ice cover, and relevant regardless of the questions relating to subglacial discharge.
Major comment 3: Missing analysis of density and presentation of the results
Since previous work on the glacial fjords of Greenland has so clearly demonstrated the role of buoyancy in forcing circulation (see ref. list), it is surprising that buoyancy forcing has not been considered at all in the analysis of the model results. The lack of isopycnals in the plots and the lack of fjord stratification as a background makes it difficult to understand the processes driving the changes within the fjord, to an extent that the drivers seem to be somewhat uncertain to the authors as well (line 373). I suspect buoyancy is a key forcing in the fjord, and I think careful analysis of the changes in density both between seasons and between the experiments should be made. See previous work on circulation in high-silled glacial fjords, such as Carroll et.al. 2017, Hager et.al., 2022, Kajanto et.al., 2023.
I suggest that the authors review the presentation of the model results, since there are a lot of interesting results that are not shown in the figures; seasonal temperature and salinity in absolute terms, both in Nares Strait and the fjord, isopycnals and flow rates over the sills were not presented. I suggest including all of these, and in the minor comments below, I have included detailed suggestions on how to improve the text and particularly the figures.
Major comment 4: Discussion in a broader context
Currently, the Discussion considers a rather narrow range of references, and the results from this study are not put to context with the extensive literature on the circulation and heat transport in the glacial fjords of Greenland. Furthermore, model uncertainties are not clearly communicated and discussed, and without putting the uncertainties to context, the reader is left to wonder at the validity of the presented results. I suggest restructuring the Discussion thematically, clearly communicating the uncertainties, and making attempts at quantifying these uncertainties. It will add to the value of the results. I have made detailed suggestions in the minor comments below
Minor comments
General comments regarding writing style:
- Please review the use of tilde in the text. Tilde should in general be avoided and be replaced with text whenever possible (with words such as about or approximately). In addition, tilde is used in the manuscript in front of numerical values with several significant digits. It is unclear what is the meaning of the tilde in front of a number that is not rounded up. Also, you should not use tilde in phrases such as “up to 100 m”. Review the use of tilde throughout the text and do not use it unless necessary. I have indicated some of the instances in the line-by-line comments below, but not all.
- The authors tend to describe the results from summer and winter within the same sentence by placing values and descriptive words in parenthesis. This is impossible for the reader to follow, please replace these formulations with proper sentences, such as “Conversely, during winter the heat transport...” or “, while in winter the increase in heat transport...”
- Use active language to make clear what are model results from your study (“Our model results show...” etc.). When discussing results from previous studies, always indicate if the results are from model studies or observations.
- Please review throughout the manuscript that you use a systematic notation of terms such as “high-resolution 3-D” etc., so that the hyphens and capital letters are always spelled similarly.
Line by line comments:
Line 1: Title “Abstract” should have its own line.
Lines 1–2: The glacier is called Petermann Glacier, not PGIS. Also, it is “seasonally shielded” from what?
Line 6: Do not use the manes of the experiments in the abstract, rather describe the changes “from a thick and landfast sea ice regime to a mobile and further thin and mobile sea ice regime” etc.
Line 7: “the implications ... on the PGIS basal melt”. Consider if you should rephrase this to heat transport to the PGIS, see major comment.
Lines 7–8: “Across all three regimes ...” What do you want to say with this sentence? Are you describing seasonality? That basal melting increases during the summer compared to winter? But then what do you mean with the under deeper regions part when does melt rate increase there and relative to what?
Lines 8–9: “Diagnosing this variability ...” The purpose of this sentence is unclear. I suppose you want to say you disentangled the thermal driver from the velocity driver in your study? In this case, say something like: “Our diagnosis of the drivers of the change in melt rate ... suggests that the increase is primarily a result of a higher thermal driving, while the increase in the friction velocity are a secondary effect” etc. But it is still unclear to me which change you are talking about.
Lines 9–12: I assume you are still describing seasonal changes in the model as opposed to differences between the experiments, but this is unclear so please add the word summer somewhere. You should highlight more that you are first describing seasonality in lines 7–12, in contrast to the changes caused by the changing sea ice cover, as you will do later. You could start the description of seasonality on line 7 with something like: “In all three sea ice regimes, the basal melt in the deep regions of the ice shelf presents a seasonal increase during summer. This seasonal increase in melt rate is primarily driven by ...”
Lines 12–15: “Modulations in surface forcing under a mobile and thin sea ice cover act to enhance the heat transport in the cavity”, this is very difficult to read, I suggest rather saying “A thin and mobile sea ice cover enhances heat transport into the ice shelf cavity, by increasing wind-driven upwelling of Atlantic Water from the Nares Strait into the cavity”
Line 15: Skip the “Mechanically”, it is confusing.
Lines 17–20: I suggest adding “the accompanying decline in the Arctic sea ice extent...” and “the basal melt of PGIS”
Lines 26–28: This sentence is really packed, consider simplifying or breaking it into two. Also, the dynamic thinning and acceleration is a consequence of the loss of buttressing forces, so it should come after.
Line 31: Is “sectorial” the North or Northeastern sector?
Lines 36–43: Review the use of tilde, and replace with words when possible. You could refer to your map figure. Use en-dash when presenting a numeric range.
Line 39: Unclear what is meant by “it has been studied through various lenses”.
Lines 43–44: This sentence is overly complicated, you could just say “Heat and freshwater exchange between the ... through the Nares Strait, and the exchange is influenced by ...”
Lines 45–49: Indicate the approximate locations of the sea ice arches to Fig. 1. The arches are even in your paper title, and it is hard to grasp where they are located.
Line 50: What do you mean by “short”?
Lines 50–52: Sounds like the study by Shroyer et al., was very similar to this one. Consider somehow hinting here what is the added value of your study. At least, highlight the differences in the discussion, is it only the bathymetry? Or are there differences in model forcings and physics?
Line 53: “outflow” is not well defined here, say rather something like “Sea ice area and the flow of ice volume through the Robeson Channel...”
Line 55: Omit “flux gate”
Line 56: “ice volume flux was smaller”
Line 61: “formation duration” this is unclear
Line 70–72: See earlier comment, are there any other improvements compared to earlier work that you might want to highlight? Improved model physics?
Line 80–82: It is clear that the author wants to advertise the new package, but it is important that the study of this paper is comprehensible and thus the description of the model in this paper should be independent. Summarize here what are the included model physics (see Major Comment 3 on the suggested flow of the Methods-section), and refer to the previous paper for details (not for basics or results). Do not use the term “standard run” when referring to the previous paper, it is very confusing, particularly given that what is called “standard” is the Thin Mobile experiment, while what could be called the standard run in this paper is the Thick Landfast experiment.
Lines 88–89: Justify (just shortly here), why you use the pre-2010 draft of the ice shelf, and consider discussing the impact of this choice in the Discussion, or the impact of a change in the ice shelf geometry in more general terms.
Line 91: So, is one of the key novelties in this study that you have a second deep sill in the domain? If yes, then this should come across in the results more clearly, and you should discuss the impact of the sill to the circulation and heat transport in the cavity.
Line 91: Here at the latest you should explain how you include the ice shelf in the model. How do you parameterize the ice-ocean interaction? Explain and justify that you don't have subglacial discharge.
Line 92: You should move the description of initialization and boundary conditions here, before you describe the different experiments
Line 94: Remove the subclause starting with which, you have already introduced the module.
Line 95: For general readability, I suggest calling the experiments simply “Landfast”, “Mobile” and “Thin Mobile”, and to use the full names throughout the paper and also in the figures. The abbreviations are too similar and impossible to follow.
Lines 95–107: Introduce the experiments as a clear progression towards expected future conditions: Landfast represents the 90’s conditions, Mobile when the sea ice arches fail to form and Thin Mobile the expected future, etc. Keep the same order you have in Table 1. throughout the paper.
Lines 97–99: This belongs to the paragraph where you present the initialization of the model, and again, do not use the name “Standard run”.
Lines 102–103: This sentence is unclear. Use active language. It is not possible to follow the subclause “, which...”, I suggest writing the point in a separate sentence.
Line 105: “sea ice velocities are retained from the Thin Mobile run”. This is very confusing, and the reader does not know the forcing or initialization procedure at this point. You should include a section describing the initialization and boundary conditions, before describing the experiments, as already commented earlier.
Line 108–111: As I said before, you should carefully present and justify your initialization procedure, and demonstrate that the runs have stabilized after these perturbations. However, it should be made clear that the progression of the modelling procedure is separate from the evolution the experiments aim to present. Currently, these are muddled together. Since these progressions are opposite, you should make very clear that it is ok to initialize the runs in this way.
Lines 115–134: Now you finally describe the initialization and boundary conditions, this is critical information to know before describing the experiments, move up and make its own subsection.
Line 115: Start by saying that you run the model from 2014 to 2017, and only then say when the model results are presented.
Line 117: I do not understand the subclause starting with “for which ...”
Lines 119–122: I’m not sure I get this right, so you get atmospheric temperature from RACMO and maybe winds? And which sea ice conditions from the A4 ROMS? State these clearly! And then you use the IceNudge module to transfer these to boundary conditions for the ocean? Be clearer, what variables do you get from which model? What are the time steps? Do you have some downscaling procedure?
Lines 124–125: I don’t follow, se is the model set up to create the arches? Or you have chosen a period when this happens in the model?
Lines 125–126: I’m even more confused, so a thin and mobile sea ice cover agrees well with what and when?
Lines 126–127: Remove the reference to the standard run, I think it is at the root of the confusion.
Lines 129–131: Which fields? Temperature, salinity, velocity?
Lines 130–132: I’m confused, are these the hourly conditions you just mentioned? Why this period for the average?
Lines 115–134: Ignoring the previous paper, is the general picture that the model was calibrated to reproduce a thin and mobile year since there was quality forcing data and observations for these types of conditions? And then you create the experiments with more sea ice by changing sea ice parameters indicated in Table 1? If yes, please make sure that this idea is clearly communicated. I still don’t know at which point of the Landfast and Thick Mobile experiments the sea ice conditions were changed, if the model stabilized in reference to this change and if any other parameters or forcings were changed. Be sure to communicate these clearly. And show in the supplement that the model stabilizes.
Line 137: How is the melt rate computed?
Lines 138–142: I don’t see how this information is necessary
Lines 143–182: This is a very detailed description of how to process model output, which is unnecessary information for most of the readers interested in the results. I suggest moving this diagnostics part into a supplementary. You don’t really need to have the subsection of “Model output and diagnostics”.
Lines 184–186: In reference to Major Comment 2: I suggest reformulating this to more generally examining the circulation and heat transport from Nares Strait to the fjord, and further into the PGIS cavity. And the yes, you will show the melt rate as well, but put more focus on the processes in between.
Line 190: This is for the whole domain, right? Fig. 3 is also partly of Nares Strait. I recommend showing the seasonal temperature and salinity vs depth of the inflow to the fjord.
Line 191: Remind the reader here how you define the seasons. Have you defined a “wider PF area” somewhere?
Line 195: “lateral inflow and outflow” of which directions? Write it in the sentence, don’t use the “, respectively”.
Lines 196–198: I find the references here confusing, just refer to the discussion.
Line 200: You should also show the flow along the fjord.
Lines 205–206: Write separate proper sentences to each of the seasons, this formulation is very difficult to read.
Lines 206–208: When?
Lines 208–212: It’s nice to see the across-fjord flow, but since we have not seen how temperature and salinity change between the seasons in Nares Strait, it is difficult to put the change here in the fjord to context. Also, along-fjord isopycnals are essential to see what role density-driven inflow over the sills play.
Lines 216–218: Where do I see this? Again, along-fjord profile with isopycnals is missing.
Lines 219–22: This is very dense and difficult to follow. Why do we need to know the annual mean? I think summer and winter inflow and outflow values would be enough. Percentages of increase relative to annual mean have little value, summer increase relative to winter would be more informative. Write proper sentences, and do not use parentheses.
Line 225–228: Rephrase. The modelled annual mean of what? Moreover, why do we need to look at the annual mean, I don’t see its value.
Line 225: See suggestions to Figure 5
Lines 230–235: You should start by describing the stratification within the fjord. You have a strong pycnocline separating the Polar Water layer from the Atlantic Water. Due to the difference in density, the seasonal response of these layers is different. Throughout the results you should describe the changes relative to the layers, not the depth. This way you will avoid clumsy formulations like “beneath the shallower ice base”. Along-fjord study of the density is key to understand what drives the circulation. Seasonal along-fjord flow patterns would also be of key importance, right now they are not shown. See suggestions to Figure 5.
Line 238: First go through the circulation in the fjord properly, and describe the melt rate in a new paragraph.
Lines 239–244: Write separate sentences for winter and summer, and give the values for mean melt rates for winter and summer, not the anomalies.
Lines 245–249: What is the difference of the “regions of strong melt aligned with the PGIS flow direction” and the “substantial lateral variability”. Aren’t they the same thing? What is meant by “and by up to 50 m/yr”?
Lines 251–253: Why the “presumably” and “likely”? Shouldn’t you know from the model if it due to AW inflow? Note that you have not defined friction velocity.
Lines 254–256: Rephrase this in terms of buoyancy and the pycnocline.
Line 257: If you would have an along-fjord plot of flow streamlines, then you would know and would not need to guess.
Lines 257–259: This sentence belongs to Section 3.2
Line 260: Rename this section, and do not include the name of the package. “Response to decreasing sea ice” etc.
Line 261: Don’t go to the basal melt first. Keep the structure as in 3.1, and start from far field and work your way in towards the melt rate. This way it is possible to track how the impact of changing sea ice conditions propagate into the fjord.
Lines 275–277: Rewrite without the parentheses, remove “calculations show”. Instead of “deeper regions”, say “below the pycnocline” or “within the Atlantic Water”.
Line 281: “primarily driven by the strengthening of the fjord circulation”. This is very interesting, and yet we know nothing of the fjord circulation and along-fjord transport
Lines 283–285: This sentence is difficult to follow, rewrite without the parentheses.
Lines 291–293: “In winter, the relative increase ...” This sentence is hard to follow.
Lines 293–295: Again, this sentence is difficult to follow. State first the general trend before giving numbers: “As the sea ice becomes mobile, the deep water masses seaward of the sill get warmer and saltier both in winter and summer.” etc.
Lines 295–298: “As the sea ice thins in addition to becoming mobile, ...”
Lines 298–299: “The colder water masses are advected into the PGIS cavity”. This is too vague. At this point, you have not discussed the transport processes into the fjord, and it really should be done. It is not clear what is the process that delivers the change into the cavity.
Lines 305–310: Finally, we get to this, but this needs a better figure to clarify the process, see comments to Fig. 10.
Line 316: Have you defined the upper ocean layer?
Lines 322–325: This is probably the most important result, and yet you say “it is likely”. Shouldn’t you see what happens in the model? You will once you study the isopycnals relative to sill depth.
Lines 325–327: These are important results, write more explicitly. What is “this mechanism”. In the following sentence you switch to discussing cold water. Which cold water and how does that follow from the previous sentence?
Lines 329–332: Are you sure about this? The cavity is well sheltered from the impact of winds, and I presume what you see in the cavity is density-driven inflow.
Lines 334–335: Where do you discuss the heat budget?
Line 336: “As the sea ice becomes mobile, the inflow during winter along the western fjord sector strengthens and intensifies with depth, being up to ...”
Line 338: What is an “increment in the inflow”?
Lines 345–347: This looks like a numerical error, are you sure that this is real?
Lines 349–350: Do not use the exact same titles as you had in results. I recommend rethinking the structure of the Discussion, perhaps organize thematically or in the order of your most significant findings.
Lines 351–353: modelled mean flow where? Is consistent in terms of what?
Line 353: What do you mean by “was seen”?
Line 354–356: I cannot follow this sentence, please rephrase.
Lines 353–359: I don’t get it, why the “likely”? “it is thus speculated”, who speculates? Why do you need to speculate if you have the results from all experiments, as you say in the next sentence?
Line 361: I don’t think it is due to air-sea momentum transfer, the density gradient should be driving this, but since you results lack the analysis of density, we cannot see it from the figures. See Carroll et.al., 2017, Kajanto et.al., 2023 and Hager et.al., 2022 for similar work in other glacial fjords.
Line 362: Who suggests? Are you saying that you find that there is water mass exchange and renewal in the fjord throughout the year? This a very interesting result, but you should state what processes are driving the water mass renewal, and discuss the role of missing subglacial discharge.
Line 364: What do you mean by “can be”?
Line 368: Why the “may be”?
Lines 369–372: Rephrase, and no need to repeat all the result values. I think you are trying to say that the inflowing water over the sill is warmer and thus causes more melt, which in turn creates a stronger buoyancy forcing that drives a stronger circulation in the cavity.
Line 373: “Irrespective of the causes”. How can you say this? Understanding the causes is the key!
Line 375: Quantify the cold bias. Is the bias systematic around the year or does it have seasonality? Is the bias connected to the missing subglacial discharge?
Line 377: Remember to explain this modification of turbulent transfer already in Methods. Remember to include a table of all model parameters into the supplement. Write here the values and compare them to the values suggested in Jackson et.al. 2020. Discuss the implications of having an increased melt over the entire ice shelf as opposed to having the subglacial discharge plume.
379–380: Who assumes? You can’t say this without the comparisons and discussion of the previous comments.
Lines 381–382: And subglacial discharge? Reference?
Lines 385–386: I don’t follow. Is this your model result you are describing?
Lines 386–388:
Are you describing your model results? Or observations? Why the “likely”? Then you apparently switch to discuss observational analysis of melt channels mid-sentence? Subglacial discharge plumes often erode channels under ice shelves, but how does that connect to your model since you don’t have plumes? These are interesting points but you should elaborate and carefully consider what your model can say. The connection with the ice shelf shape is interesting and merits more thought.
Line 392: Were these both observational studies?
Line 395–397: Which one are you talking about now?
Line 397–398: This is a very harsh sentence after the comparison you just did.
Line 401: It is unclear what the reference to Holland et.al. Is for here.
Lines 401–404: This is the general shape you get with the Jenkins’ 99 three-equation parameterization that I supposed this model also uses (which is information that should appear in Methods).
Lines 405–410: This paragraph is just listing a number of effects that could have an impact, some less interesting (like the well-known pressure-dependency of the melting point), and some very interesting, like potentially plume-eroded channels in the shelf. I recommend reviewing a bit more of the relevant literature on subglacial discharge plumes (some suggestions in the reference list below) and revisiting your melt rate parameterization, and carefully considering what are the clear factors impacting the distribution of melt in this model, and what are factors that are excluded in the model, but could potentially still impact the melt rate.
Lines 413–415: How does this relate to depth/thermocline?
Lines 416–417: Elaborate on what is the significance of no subglacial discharge. What sort of error does that cause in the results and why are the melt rates given in this study relevant?
Lines 413–414: what inconsistencies, and how is the Table relevant? Elaborate on what is the meaning and impact of these inconsistencies to this study.
Line 420: What novel mechanisms do you mean?
Lines 420–421: I think you want to say that even though you are not calculating the melt rate correctly, the relative changes caused by a change in sea ice are robust. You could also point out that your estimates of the change in the Atlantic Water properties at the mouth of the fjord are only little impacted by the lack of subglacial discharge. It would be really good if you could quantify how much you're tweaking the turbulent transfer and neglecting subglacial discharge impacts the circulation regime. I suspect you have runs, at least in the Thin Mobile configuration with different turbulent transfer coefficient? How much does the change in the coefficient mean in terms of meltwater flux from the glacier? Or in terms of inflow over the outer sill? It is hard to compare to the buoyancy flux from a plume, but some attempts at quantifying this uncertainty should be made. You could just attempt to make an estimate of the relative increase in the plume melt rate based on the increase in Atlantic Water temperature you project, see Ezhova et.al., 2018.
This is exactly why I suggest on taking the focus more towards the change in the inflowing Atlantic Water properties and heat transport into the fjord, since you are on a more solid ground in terms of reliability of the results.
Line 423: What kind of estimates are these, model?
Lines 440–444: This is too long, break into shorter sentences. You're talking about basal temperatures and calving front temperatures within the same sentence. Be clear in what you mean. Give references to undercutting.
Line 446: Are you talking about future projections? Are these references predicting a future warming?
Line 448: “the upwelling mechanisms” this is too vague. I know by know what you refer to with this, but be more specific. Upwelling where?
Line 448: Speculate also what would the impact from increased freshwater flux from Greenland be
Lines 452–460: If you analyze the results in density space as I have suggested, you can add here that your model is able to show the density-driven inflow over the outer sill from Nares Strait into the fjord that fills the deep basin and reaches the grounding line.
Line 452 Wind-driven upwelling, here and elsewhere
Line 464: Be very careful, I don’t think you have shown that wind drives the circulation in the PGIS cavity. Wind drives the upwelling of AW in Nares Strait and I presume that density drives the circulation in the fjord.
Lines 469–470: Did you not prescribe this in the model? Then it’s hardly surprising that there is agreement.
Lines 470–475: You have set up the model to represent the conditions so this is not really discussion. You should rather move this to where you describe the runs and say that the experiments represent approximately the conditions of this and that period.
Lines 475–477: I like the sentence, but again, it is not clear that wind drives the inflow into the cavity, see Carroll et.al., 2017.
Line 478: Remove the “interpreted”, and reformulate to state that this is your model result.
Lines 481–482: Reformulate, “not all of the heat is taken up by ice melt”, etc.
Lines 486–488: Again, I don’t think you have properly analyzed the drivers to say this.
Line 493: The summary is very long and slightly repetitive, you should compress, and highlight the results that are new from this study.
Lines 500–534: you don’t have a bullet point explaining your findings on the impact of a change in sea ice to Nares Strait, and you don’t have a bullet point dedicated to circulation in the fjord. Several bullet points describe similar points regarding melt rate, and are thus quite repetitive.
Lines 535–539: References?
Lines 535–548: This is borderline Discussion, and too long, write a short and clean paragraph on how your results describe the potential future, uncertainties are for the Discussion. Skip the a and b.
Figures and Tables
Figure 1:
I suggest putting quite a bit more effort to this figure. In 1a, I suggest cropping much closer to the domain, now much of the figure are is out of the domain, and the area of interest is too small. You can add a small inset figure indicating the location on a larger scale map. The blue and green colors of the background map are too dark, the red and yellow are too hard to read. The caption says that the sea ice arch locations are marked in the figure, but I cannot see them. In 1b, use different, brighter colors in the lines, they cannot be distinguished from this figure. I suggest including an along-fjord profile and indicating the sills, the location of the cross-sections and all of the other features you refer to later in the paper (for example Fig. 3a). This is important since bathymetry and draft are the key novelties in this study, and the inner sill is a new(?) feature. In all figures be more consistent with fonts and font sizes, the figure looks busy with so many different fonts and sizes.
Figure 2:
This figure is difficult to follow, but that can be fixed easily. I suggest the following changes: First column should be the Landfast experiment, since that represents earlier, “historical or pre-industrial”, conditions in this paper. Do not use the abbreviations, use the full names of the experiments. Consider adding the words summer, winter and year-round as column headers.
Table 1:
Once you have revised the text describing the initialization of the experiments, remember to revise the caption accordingly. It is now impossible to grasp which run starts from when and where. Remove the abbreviations, and update the experiment names. Give the duration of the summer and winter seasons also in days or months so that the reader does not need to calculate, and would be better to give the time in calendar days. Remove the “Standard run”.
Figure 3:
Review the use of space in this figure. White space and labels take more space than the actual results you attempt to show. Seasonal mean speed of what? You should say that these are ocean currents. Consider adding a small inset in the empty space of 3b) that would show a line plot of the mean depth along the fjord vs the Y-coordinate. Check that all of the features indicated in red in 3a) will be also indicated in the along-fjord profile of Fig. 1c). Indicate the gyres discussed in the text with arrows, and refer to the figure when describing the gyres.
Figure 4:
Why are the inner profiles (4a,b) above the outer ones (4a,d)? I think it would be more intuitive the other way around. I do not know how these cross sections relate to the fjord bathymetry (See comments to Fig. 1). I the black color on top of the water column the ice shelf? If yes, mark it with a different color than the terrain, and use the same color for the ice shelf throughout the figures. What is the gap in the ice shelf in 4a,b? Increase the font size of the temperature values of the isotherms. Blue and red are not good font colors on a blue-red background. Why the cropped width? There is some white space in between. Include in the caption which way is positive (towards the glacier?)
Figure 5:
This is the first along-fjord figure of the paper, much overdue, and overall, quite unhelpful. Differentiate the ice shelf from the terrain by using the same color as in Fig. 4, now it is difficult to see the GL. I don’t think the annual means are valuable at all, I would just remove them. Similarly, showing anomaly relative to annual mean is not meaningful. Also, the salinity anomaly is also rather uninformative, density should be salinity-dominated here anyway. I recommend showing isopycnals for both seasons, as well as temperature for winter and summer. Plot the isopycnals at even intervals, in 5a) it is hard to see the pycnocline since the contours jump from 27.1 to 25.9. You can explain the range of annual variability in the text. Consider also including the along-fjord heat transport overlaid with streamlines. Otherwise, we are not seeing along-fjord flow at all. This would mean four panels, which would make it possible to increase the panel size a bit.
Figure 6:
There is a lot to unpack in this figure, and again, I do not see the point of showing anomalies relative to annual mean. If you want to plot anomalies, why not plot the summer increase relative to winter? However, I would prefer to just show melt rate in winter and melt rate in summer with the same color range, on a much bigger figure. I find panels d—i deeply unhelpful, and I do not see the point. If you are trying to show the relative contributions of temperature increase and the increase in friction velocity to the melt rate, it is not working. Why are these anomalies? Could you just show the water temperature at the base of the shelf in summer and winter, and then the friction velocity. So as in e,f,h and i but in absolute terms? I can’t imagine if it will be better, but something should be done. Also, note that you have not defined anywhere the friction velocity. Turn panels d and g so that ice draft is along the y-axis and increases downwards, it’s more intuitive. Also, I recommend plotting temperature and friction velocity in absolute terms for both winter and summer. There is a significant amount of white space in the figure so should be possible to reorganize the figure so that the shelves appear bigger, right now they are too small.
Figure 7:
I recommend to plot in panels b and c the absolute annual mean melt rate in the Thick mobile and thin mobile experiments. This way your first row is comparison of annual mean melt rates. If you have included summer and winter melt rates (not anomalies) in Fig. 6, you can show panels e,f,h,i as they are. However, write on the panel if they are winter or summer, and use the full names of the runs instead of abbreviations (landfast to mobile etc.). Flip panels d and g to have the draft as y-axis increasing downwards. Reduce white space and make the shelves appear bigger. Reorganize to have seasons in columns, as you have in other figures. Use different colors to plot the substractions in panels d and g to avoid confusion, and keep the colors also in other figures.
Figure 8:
Flip axes and change color as in the previous figures. Here I think the difference in annual mean makes sense.
Figure 9:
A lot of the same comments as earlier. Change the color of the ice shelf, keep the seasons in columns to be systematic in your figures and write summer and winter on top of the columns, do not use abbreviations of the experiment names but spell out the names. Again, I think in order to understand to flow under the ice shelf, we need to see isopycnals for both seasons and both, and since intensification of the circulation was a key result, you should plot the flow somehow, for example along-fjord velocity in blue-red. You could save space by skipping the difference in salinity if you show isopycnals, and only show the seasonal difference in temperature.
Figure 10:
In panels a—d) is there a black (sea ice) layer on top, or is it the darkest color of the color scale? If it is ice, plot it with a different color, if it is the darkest shade of the color scale, plot the terrain in brown (also in other figures). Consider writing Nares Strait on the figure, so that it is clear without reading the caption that this is not the fjord. Write winter and summer on top of the columns. Do not use the abbreviations of the experiments, and do plot isopycnals on this figure. 10 e and f are really informative and important, it is important that this figure comes earlier.
Figure 11:
Keep seasons in columns, write winter and summer on the figure, do not use abbreviations, use a different color for ice to separate it from terrain. State in the caption with words where this cross section is from. Add the colored circle indicating the section, as you had in the previous fjord section figure. Check from the model what happens in the narrow return-flow column in 11d, and state that in the caption, it draws a lot of attention. It seems to be related to the gap in the ice shelf.
Table 2:
Write the experiment names, no abbreviations. It is not clear where this is calculated from, is it the fjord mouth corss-section?
References:
Carroll et.al., 2015, “Modeling Turbulent Subglacial Meltwater Plumes: Implications for Fjord-Scale Buoyancy-Driven Circulation”, J. Phys. Oceanography
Carroll et.al., 2017, “Subglacial discharge-driven renewal of tidewater glacier fjords”, J. Geophys. Res.-Oceans
Ehrenfeucht et.al., 2022, “Seasonal Acceleration of Petermann Glacier, Greenland, From Changes in Subglacial Hydrology”, Geoph. Res. Lett.
Ezhova et.al., 2018, “Dynamics of three-dimensional turbulent wall plumes and implications for estimates of submarine glacier melting”, J. Phys. Oceanography
Hager et.al., 2022, “Subglacial Discharge Reflux and Buoyancy Forcing Drive Seasonality in a Silled Glacial Fjord”, J. Geophys. Res.-Oceans
Jackson et.al., 2020, “Meltwater Intrusions Reveal Mechanisms for Rapid Submarine Melt at a Tidewater Glacier”, Geophys.Res.Lett.
Jenkins, A. 1999, “The impact of melting ice on ocean waters”, J. Phys. Oceanography
Kajanto et.al., 2023, “Impact of icebergs on the seasonal submarine melt of Sermeq Kujalleq”, The Cryosphere
Straneo F. And Cenedese C., 2015, “The Dynamics of Greenland’s Glacial Fjords and Their Role in Climate”, Ann. Rev. Mar.Sc.
Citation: https://doi.org/10.5194/egusphere-2023-73-RC1 - AC1: 'Reply on RC1', Abhay Prakash, 27 May 2023
-
RC2: 'Comment on egusphere-2023-73', Anonymous Referee #2, 10 Mar 2023
The manuscript describes the output from a regional, three-dimensional numerical model of the ocean. It includes two lateral boundaries in the north and south where water properties (temperature and salinity) and velocity vectors are specified and complex coastlines. The model geometry includes a floating ice shelf (Petermann Gl.) as well as realistic bottom topography both under the floating glacier and everywhere else within the modeling domain. The presence of ice (sea ice and glacier ice) is prescribed (not modeled) as a surface boundary condition with a surface stress (vertical momentum flux) condition that depends on the prescribed (not modeled) ice concentration and ice velocity. The ice is allowed to melt because of conductive and turbulent heat flux from the ocean to the ice. Much of this model set-up and selective comparison to limited data is published as Prakash et al. (2022). Reading this publication, I consider essential to critically evaluate the present manuscript. I recommend against publication for the following three major at least 10 minor reasons:
A_ Prior ice-ocean models of the region exist (Shroyer et al., 2015; Shroyer et al. 2017) and substantial overlap exists with few new or original results. Unlike the present model, the prior models include a prognostic (as opposed to a prescribed) sea ice model. Basal melt rates are predicted under the floating portion of the glacier by both present and prior models. Both present and prior models evaluate the differences between sea ice in the adjacent ocean and fjord that is mobile vs sea ice that is not mobile. I cannot discern fundamentally new results or insight in the present model application that basically repeats prior modeling with a different model, that, I feel, has more weaknesses than strength over prior models.
B_ The present model uses temperature and salinity at the boundaries that are both poor and unrealistic. This “cold bias” is prominently discussed in Prakash et al. (2022), but it is introduced in the present manuscript almost as an after-thought in passing on Line-375. Moored and synoptic observations indicate salinity and temperature of about 34.7 psu +0.2 C for the Atlantic-influenced waters both in Petermann Fjord and below the ice shelf (Muenchow et al., 2016; Washam et al, 2018). The model provides “… annual mean temperature and salinity of the water masses that overflow the inner sill…” that are 34 psu and -0.7 C (Line-230). So, the modeled ocean is almost 1.0 C too cold and 0.7 psu too fresh at a depth where seasonal and interannual variations are less than 0.1 C and 0.01 psu! So, the model includes huge biases that I find unacceptable. Note that Prakash et al. (2022, page-27, Line-6/7) “ … suggest applying a depth-dependent bias correction to the upstream A4 T-S fields such as in Shroyer et al. (2015).” Please follow your own suggestion.
C_ The present model attempts to compensate for the “cold and fresh bias” by artificially increasing the vertical turbulent exchange coefficient to produce melt rates that are somewhat agreeable to observations of melt rate. So, the model includes a wrong (northern) boundary condition (from the nested A4 model) and a wrong “turbulence model” to make things right. From my observational perspective two wrongs rarely make a right and I thus loose trust in this model as, it appears, the model can “nudge” or “fix” anything as there is always a parameter or dial that one can change to obtain any desired result. I do not endorse this practice, especially since the authors know the proper way to remove these biases and perhaps use realistic turbulent exchange coefficients.
In addition to the above major weaknesses, there are a number of more minor concerns
1_ The manuscript is too long and unnecessarily complex. I could not discern much substantial difference between the “Thick-Mobile” and the Thin-Mobile” case. What differences there are, I feel, may fall into the domain of model uncertainty, noise, and poorly constrained observational and/or model parameters.
2_ The manuscript is too long and contains trivial co-ordinate transformations. Just state that co-ordinates are rotated into along- and across-channel components rather than spelling out what I perceive as trivial algebra (Lines-143 through Line-182). I understand that the present model is finite element that that this may not be as trivial computationally as it is in finite-difference models or observations.
3_ I am confused and disturbed by streamlines that start at boundaries and end in the interior. Figure-3 offends my sense of mass conversation and lateral boundary conditions. Would you not expect a zero velocity along the coast?
4_ Figure-4 has large spatial oscillations that result from a poorly used graphics package. Please learn how to draw contours properly. Furthermore, I prefer distances in km as opposed to Longitude. The strong bottom-intensified slope current under the western ice shelf during the summer (Fig.-4b) caught my attention. Strangely, no such current exists in the fjord where bottom slopes may be similar, well, I cannot tell, because Longitude rather km is used for distance.
5_ On Line-310 it is stated that “… upwelled AW [Atlantic Water] from the adjacent Nares Strait enter the fjord…” How does this work? Would not winds from north to south in Nares Strait that may cause the AW to upwell along Greenland in the east also lower sea level along Greenland relative to Canada? Would then Petermann Fjord not respond with a large outflow the way an estuary would with lower sea level at is mouth?
6_ Please provide uncertainty estimates to your estimates of heat fluxes such as summarized in Table-2 as it is done with observational estimates of these same fluxes.
7_ In my view the authors mistakenly equate heat flux into the fjord and/or glacier cavity with basal melting (Line-368, Line-373). First, most of the heat flux into the fjord leaves it. Only a small fraction of the heat is actually used to melt the ice which in the present model is largely caused by the artificially increased vertical turbulent exchange coefficient. Second, did the authors actually check, if mass is conserved? Does the volume flux add to the amount of melting?
8_ The comparison of model predicted basal melt rates with observations appears to me less systematic than it could be. The authors appear to pick whatever value from whatever paper for whatever season that fits their purpose. Perhaps this part of the discussion can be strengthened by more clearly delineate seasons, space, and vastly different observational techniques (snapshots vs. moored observations or remotely satellite vs fixed radar stations).
9_ Line-438/439/440: The Rueckamp et al. (2019) reference clearly states that the “still attached” new ice island has already “dynamically detached,” that is, from a practical or physical perspective, this segment of the ice shelf is already gone.
10_ Line-511: The “Nares Strait sea ice arches” (mobile sea ice) add 2 m/y basal melt to the 24 m/y (Line-243), so the entire paper is about a 10% effect. The supply of heat to the fjord or glacier cavity matters little, but both the vertical turbulent mixing and the grounding line discharge of freshwater (not included in this model) may dominate over this 10% effect.
In summary, the manuscript should not be published without properly fixing both water mass bias and the vertical exchange. Even with these fixes, substantial overlap with prior modeling work may limit new and original insights that warrant publication.
Muenchow, A., L. Padman, P. Washam, and K. Nicholls: The ice shelf of Petermann Gletscher, North Greenland and its connection to the Arctic and Atlantic oceans, Oceanography, 29 (4), 84-95, 2016.
Prakash, A., Q. Zhou, T. Hattermann, W. Bao, R. Graversen, and N. Kirchner: A nested high-resolution unstructured grid 3-D ocean-sea ice-ice shelf setup for numerical investigations of the Petermann ice shelf and fjord, MethodsX, Vol 9, https://doi.org/10.1016/j.mex.2022.101668, 2022.
Rückamp, M., Neckel, N., Berger, S., Humbert, A., and Helm, V.: Calving induced speedup of Petermann glacier, Journal of Geophysical Research: Earth Surface, 124, 216–228, 2019.
Shroyer, E.L., R.M. Samelson, L. Padman, and A. Muenchow: Modeled ocean circulation in Nares Strait and its dependence on land-fast ice cover, J. Geophys. Res., doi:10.1002/2015JC011091, 2015.
Shroyer, E.L., L. Padman, R.M. Samelson, A. Muenchow, and L.A. Stearns: Seasonal control of Petermann Gletscher ice-shelf melt by the ocean's response to sea-ice cover in Nares Strait, J. Glac., pp. 1-7, doi:10.1017/jog.2016.140, 2017.
Washam, P., A. Muenchow, K. Nicholls: A decade of ocean change impacting the ice shelf of Petermann Gletscher, Greenland. J. Phys. Oceanogr., 48, 2477-2493, 2018.
-
CC1: 'Reply on RC2', Tore Hattermann, 24 Mar 2023
Since there is a function for open discussion, we might as well use it. Usually, an author's reply starts by thanking the reviewer for their constructive criticism. Although Anonymous Referee #2 points out a few relevant aspects, those appear being put forward as an exaggerated criticism (of mainly technical aspects) and lack of goodwill to see an intellectual contribution that this work may have for the understanding of the response of the Greenland Ice Sheet to future climate change.
The results of this study are based on a hypothesis on how long-term changes of sea ice properties might affect the future stability of PGIS. The premises for this deduction are largely based on previous modelling work that discovered the impact of seasonal varying sea ice cover. This is clearly marked and acknowledged in the manuscript. However, the intellectual transfer of how this mechanism may play out to a concretely observed interannual change of large-scale sea ice properties in the region is a novel contribution. To my knowledge, there exists no peer-reviewd literature that has explicitly addresses this aspect.
Then, the study uses a moderately sophisticated numerical tool to lend support for this hypothesis and further detail its manifestation and impacts, and with arguable advantages and disadvantages compared to previous modeling works. There are certainly shortcomings in the model that is being used, and all models are wrong, but in this case, the primary use is (i) prescribing the sea ice to specific conditions (which is not possible with a fully coupled sea ice module, as might have been misunderstood by the reviewer) and (ii) illustrating and characterize the qualitative and quantitative impacts and characteristics on the ice shelf. In particular, the latter is undertaken with unprecedented spatial detail in PGIS fjord, also investigating aspects that have previously never been addressed, such as providing high resolution 2-D maps of the basal melt pattern, investigating the response of thermal driving vs. friction velocity, comparing (not equating) those to melt rate changes to heat flux anomalies. Admittedly, there is a bias and likely a high uncertainty on the absolute value of the simulated basal melt rates, which should be addressed in a revised version of the manuscript. However, the primary findings of the qualitative changes stand independent of this and have not yet been explored in this form.
So, while I partially agree with the Anonymous Referee #2 editorial comments and on assessing technical shortcomings of the model (which, however, are already openly communicated in the manuscript and altogether only have minor impacts on the main conclusions of this study), I am surprised over its harsh judgement that this work should not be published due to its lack of relevance and/or novelty. Even if possibly not providing a ground-breaking advance, this study reflects the solid result of a hard-working ECR's professional development that should not be defeated like this. Much less solid of impactful work has been published and celebrated, and in an educational system that is based on peer-review publications as a standard, there should be room for incremental advance like this in the literature.
Personally, I think that publicly posting such exaggerated criticism (for whatever reason) protected by anonymity is simply not fair. Hoping that the editor will see though the mist, I will encourage my student to reasonably address the points raised in a revised version of the manuscript and live up to high ethical standards in in his future scientific career himself.
Sincerely,
Tore Hattermann
Citation: https://doi.org/10.5194/egusphere-2023-73-CC1 - AC2: 'Reply on RC2', Abhay Prakash, 27 May 2023
-
CC1: 'Reply on RC2', Tore Hattermann, 24 Mar 2023
Peer review completion
Journal article(s) based on this preprint
Abhay Prakash et al.
Abhay Prakash et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 452 | 225 | 28 | 705 | 15 | 10 |
- HTML: 452
- PDF: 225
- XML: 28
- Total: 705
- BibTeX: 15
- EndNote: 10
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(39051 KB) - Metadata XML