On the ocean's response to enhanced Greenland runoff in model experiments: relevance of mesoscale dynamics and atmospheric coupling

Martin, Torge; Biastoch, Arne

doi:https://doi.org/10.5194/os-19-141-2023

Articles | Volume 19, issue 1

https://doi.org/10.5194/os-19-141-2023

© Author(s) 2023. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/os-19-141-2023

© Author(s) 2023. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 19, issue 1

Research article

| Highlight paper

|

20 Feb 2023

Research article | Highlight paper |

| 20 Feb 2023

On the ocean's response to enhanced Greenland runoff in model experiments: relevance of mesoscale dynamics and atmospheric coupling

Torge Martin and Arne Biastoch

Download

Final revised paper (published on 20 Feb 2023)
Preprint (discussion started on 08 Sep 2022)

Interactive discussion

Status: closed

AC1:
'Amendment to code and data availability', Torge Martin, 09 Sep 2022

Model output from all experiments and the Jupyter notebooks required to reproduce the analysis and figures are available now through GEOMAR at https://hdl.handle.net/20.500.12085/263da22c-247f-4cd1-8080-b221e3f0e2c0 .

Citation: https://doi.org/10.5194/egusphere-2022-869-AC1
- EC1: 'Reply on AC1', Karen J. Heywood, 09 Sep 2022
  
  Thank you for providing this information Torge.
  
  Citation: https://doi.org/10.5194/egusphere-2022-869-EC1
RC1:
'Review on : “On the ocean’s response to enhanced Greenland runoff in model experiments: relevance of mesoscale dynamics and atmospheric coupling”Comment on egusphere-2022-869', Anonymous Referee #1, 31 Oct 2022

Please find comments attached.

Citation: https://doi.org/10.5194/egusphere-2022-869-RC1
- AC2: 'Reply on RC1', Torge Martin, 02 Nov 2022
  
  We thank the reviewer for their overall positive feedback on our manuscript and the detailed comments, which certainly help to improve the paper. The specific comments on forcing, AMOC, internal variability and overall length of the paper are well taken and we certainly will consider all of them for the revised version. In this first reply we just briefly touch upon the main criticism.
  Running the ocean-only experiment with historical (instead of repeated year) forcing but the coupled ones under pre-industrial control conditions was a compromise to have sufficient internal variability in the former and to isolate the impact of fresh water from other global warming signals in the latter.
  We agree that AMOC internal variability exists on multi-centennial time scales, of course. Computing trends for various subsections of the reference runs, we found weak AMOC trends for the two coupled reference runs for periods of 100 to 200 years (coupled: 0.025-0.07 Sv/decade, coupled-nested: -0.05 to +0.03 Sv/dec) but a considerably wider distribution when using periods shorten than 80 years. We thus consider a 100-year long reference period as sufficient to limit the imprint of internal variability. Regarding the AMOC strength at the onset of the freshwater experiment, the coupled-nested run starts close to the long-term mean AMOC strength (+0.5 Sv) whereas the coupled experiment presented here starts in a low phase (-2.2 Sv). However, the coupled run is part of a small ensemble discussed in Martin et al. (2022), who show a robust AMOC weakening over the last 50 years independent of the starting condition.
  The Appendix is intended to address the issue of distinguishing the response to the freshwater and internal variability and we appreciate the additional suggestions to discuss this in a more comprehensive and convincing way in the revised version. Because of the already extensive main part of the paper, we decided to focus this discussion in an appendix but will reconsider its placement (incl. further reviewer comments currently pending) in the revision process.
  
  Citation: https://doi.org/10.5194/egusphere-2022-869-AC2
- AC4: 'detailed response to RC1', Torge Martin, 09 Dec 2022
  
  Please see attached file for our full response to your thoughtful comments.
  
  Citation: https://doi.org/10.5194/egusphere-2022-869-AC4
RC2:
'Comment on egusphere-2022-869', Anonymous Referee #2, 05 Nov 2022

This paper explores how the North Atlantic Ocean responds to enhanced Greenland melt, using a suite of ocean modelling experiments. The authors carefully explore this problem by using twin experiments, with and without Greenland melt, while also examining the role of resolution (through the inclusion of high resolution nests) and forcing (by considering coupled as well as forced ocean model experiments). The authors also run their experiments for a length (100 years) sufficient to allow signals and different behaviors between experiments to develop. Key results include a compensating temperature feedback in the coupled simulations, which also have greater stability. Additionally, mesoscale dynamics, represented in the nests, play a key role, including penetration of freshwater in the sub-tropics.

This is an important topic, and the study nicely examines many aspects. The paper is generally well written and clear, with high quality figures. Thus, it definitely deserves publication, with EGU Sphere being an appropriate journal. That said, there are some ways the manuscript can be improved. There are some minor wording issues (such as un-needed adjectives). The manuscript also feels long, and given that it covers so much space, there are times that it feels like the main big picture goals get lost in the many details. So it might be good to try to tighten up the manuscript and make sure the focus is always on the main ideas and results. There are also a few technical items that could use further discussion.

Salinity Restoring: This is first mentioned at line 102-103 when the authors mention they use a weak restoring. It would be good to explain why this is included. Also, given the authors are looking at salinity signals for Greenland melt, I have concerns about those signals being damped by this term. At the very least this is worth further discussion. Some comparison with other studies that don’t use restoring, or have restoring of different strengths, would be good. Ideally, and even though the experiments with the nests are computationally expensive, it would be good to see what would happen if they were run without restoring, or at least compared to a 10 year integration period with the restoring.

Historical vs Pre-Industrial: This is first discussed for lines 110-112. I know the authors work to justify this choice later in the paper, but I think this choice needs greater justification and discussion.

Averaging Periods: The authors explain why they use different averaging periods, and add Appendix A as a justification. This still feels like a concern in the transient experiments, since a longer averaging period means more Greenland meltwater added to the ocean, and a longer period that potentially means in can propagate farther. I would like to see some comparison with averaging over the same period, to help confirm that the results are not being biased by the variable averaging periods.

Specific points:

L23-25: “leaving the ice sheet at a negative net mass balance” doesn’t read well.

L38: Note sure exactly what “indicate robustly” means

L71: “most critical improvements by the grid refinement” doesn’t read well.

L87: “including entire Greenland” is missing some words/explanation

L88: Might be worthwhile to clear explain what is meant by a strongly eddy ocean

L92: McWilliams

L99: extended

L113: “much more” – much isn’t needed, more pronounced says the same thing

L113: What exactly is “strongly meandering” compared to just meandering?

L117: height

L122-123: “promotes intensified, partly overly pronounced deep convection” doesn’t read well

L131: by “data extending beyond Greenland is not considered though”, do you mean you haven’t included the other non-Greenland glaciers in the dataset? If so, say it directly.

L131: on the annual mean

L133: What does “over 62 and 100” mean?

Figure 3 caption: What is an “Examplary improvements”?

L142: Half of the icebergs melting in fiords – this needs to be referenced.

L144-145: “we find the prescribed freshwater rapidly mixed over the depth of the Greenland shelf by the ocean model also shifting the seasonal peak by a month” – no idea what this statement is trying to say.

L155 “much more pronounced” – more pronounced is good enough – the additional adjective doesn’t really add anything

L169-171: Reference each of the listed process that the authors suggest the overturning is sensitive too.

L175: Would be good to compare the model overturning strengths to observations, such as RAPID and OSNAP. Even if the paper’s focus is understanding responses in different model configurations, it helps to understand the realism of different results/measures.

L187 “much more sensitive” – more sensitive is fine. Also, given the declines in Sv, might it be worth mentioning the percentage changes?

Paragraph ending on L205: How does this propagation compare with other, previous, studies of Greenland melt.

L208: ‘well’ not needed, expressed is fine.

L211: What is meant by the ‘very eastern side’?

L220: coupled experiments – should it be plural? I.e. Is this behavior occurring in all coupled experiments?

L223: “Averaged over the top 200 m representative for the upper ocean” doesn’t read well.

L224: “in some areas on annual mean” doesn’t read well.

L226: “This is except for” doesn’t read well.

L230: Nordic Seas.

L236: What is meant by “the mixed layer rather shoals”?

L238: Local regions of warming – Is this significant? Or just a minor detail?

L243: sites

L243: “For simplification, we only show the spatial means for these areas and averaged over the top 200 m if not mentioned otherwise.” Isn’t clear and doesn’t read well.

L244: Why does this comment about grid cell averaging suddenly appear? Is there any other way to compute averages on model grids where the area/volume spatially varies?

L247: remove “being”

L255: “the large scale we focus on” – be quantitative, which will help this discussion.

L260: “…have approximately the same…”

L264-265: Why is the overflow water warmer if the mixed layers are deeper in the Nordic Seas?

L280: “barely is a density change noticeable” doesn’t read well.

L284: “presents with excessive freshening” isn’t clear.

Figure 10: Why use a line a latitude (60N) instead of the observational OSNAP line – would be useful for readers to look at the model fields where observations exists.

L289: Would a reference to Behrens et al be useful here?

L294: “it does not become clear” doesn’t read well.

Figure 11: Is MLD > 500 m appropriate for the ENA shelf? Additionally, can you try to estimate a formation rate in Sv, to help readers put the numbers in context compared to other studies? Also, how realistic are the areas of deeper MLs in the various simulations?

L307: “enabling properties of the initialization fields still visible” – not sure what the authors are trying to say here.

L312: “show large content of overflow water” doesn’t read well.

L318: “without though the higher resolution is not quite sufficient” – some words or explanation is missing

L330: shown

L339: suffers is probably not the best word here.

L355: Irminger Rings entering the Labrador Sea interior – maybe add some references.

L357: are crucial

L366: dynamically active (highly not needed)

L371: What does “largely improved” really mean?

L375: “well seen” doesn’t strike me as formal scientific wording

L388: In terms of imprints of the different Northwest Corner dynamics on meltwater tracer concentrations, are there any other studies that could be referenced/included in the discussion here?

L418: “the coupled-nested configurations” – plural or singular – I.e. are you meaning the control and melt experiments with this setup?

Figure 14 caption: States magenta lines but I see red and yellow.

L428: “over the entire but mostly eastern SPNA” almost feels contradictory.

L445: Other studies have looked at the role of Ekman transports around the sub-polar gyre. Would be good to reference them.

L450: Is this realistic. Is there a concern of the atmospheric scale being too coarse to look at processes around the narrow boundary currents. This comes up later in the paper, but would be good to mention here. Also, would be good to reference those works that have previously discussed Ekman transport’s role in exchange from the WGC and LC.

L474: “explored in many model studies before” – add some references to those previous studies

L476: “passed decade already” – I think the authors may mean a previous decade?

L484: being

L487: “Potentially in consequence thereof” doesn’t read well

L492: What is meant by “does neither cover”?

L494: the objective of whether…

L501: Don’t like the wording “quite typical” for the salinity bias – maybe explain this in more detail and more clearly.

L503: remove also

L505: “doubting the importance” isn’t a great choice of wording

L509: disagree may be a better word than oppose

L522: I think Schulze-Chretien and Frajka-Williams should also be referred to here.

L527: remove also

L535-540: Some other studies suggest resolutions up to 1/60th degree may now be needed.

L548: remove “a”

L548: Gillard et al, (2022) – Ocean Modelling – looks at the some impacts on this exchange when changing the vertical resolution

L585: Not sure what “presenting with the strongest weakening” is really saying

L636: “present with diverse sensitivity” doesn’t read well.

Citation: https://doi.org/10.5194/egusphere-2022-869-RC2
- AC3: 'Reply on RC2', Torge Martin, 11 Nov 2022
  
  We appreciate the overall positive and constructive comments by the reviewer. We will provide a detailed response together with the revised manuscript and will here just briefly touch upon the main points of the criticism. In our revision we will also try and tighten the manuscript as suggested here and also by Reviewer #1 and clarify the technical issues.
  Salinity restoring is required to stabilize the AMOC in the forced configurations. The atmospheric forcing of the ocean-only model tends to create a fresh bias in the subpolar North Atlantic, which otherwise sends the AMOC on a declining trend. The restoring of sea surface salinity as applied here is standard procedure and we note that we apply a relatively weak correction, which moreover is neither applied under sea ice nor at grid nodes with runoff. This being said, the restoring flux does indeed oppose part of the freshwater perturbation on basin-scale. Integrated over the subpolar North Atlantic between 45˚N and 80˚N, the restoring adds a negative freshwater flux of approx. 0.02 Sv with similar magnitude in both high and low resolution configurations but steeper trend in the latter.This means that about 40% of the freshwater being added along Greenland’s coast is withdrawn over the broad scale of the subpolar seas. Nevertheless, the forced experiments yield a significantly enhanced decline of AMOC strength compared to the coupled ones—despite the restoring. In fact, we may speculate that the restoring actually prevents a runaway positive feedback loop with northward salinity advection declining as the AMOC weakens. We will discuss this aspect in the revised version.
  As in our first response to Reviewer #1, we note that running the ocean-only experiment with historical (instead of repeated year) forcing but the coupled ones under pre-industrial control conditions was a compromise to have sufficient internal variability in the former and to isolate the impact of fresh water from other global warming signals in the latter. This will be further discussed in the detailed response and argued for in the revised version of the paper.
  The effect of the averaging period on the response of the AMOC is, for example, included in Figure 4 and Table 2, where we also show the distribution and mean response for the 20-year period of 43-62 years after onset of the perturbation in the coupled runs. We discuss that this result is more prone to be influenced by multi-decadal variability. We argue against an expansion of the averaging period of the forced experiments to 50 years using Figure A1, which shows a clear trend in AMOC decline prior to year 40 of the simulation. The comment of having more freshwater added and allowing it to spread further by selecting a slightly later and longer averaging period for the coupled experiments is an interesting aspect. This certainly would be an issue in identifying time scales of the responses, which we refrain from doing, and focusing on the large-scale response patterns, it is again the coupled configurations with the later/longer averaging period, which show the weaker responses (despite having more freshwater added in the end). Figure A2 compares the SST response for different averaging periods (e.g. 20 and 50 years), and while there are local differences, the large-scale patterns are robust. Based on such investigations, we concluded that it is rather internal variability than the timing of the averaging period that causes the larger uncertainty and decided for a longer averaging period for the coupled experiments. Nevertheless, we will improve this discussion in the manuscript using the concerns expressed by the reviewer as guidance.
  
  Citation: https://doi.org/10.5194/egusphere-2022-869-AC3
- AC5: 'detailed response to RC2', Torge Martin, 09 Dec 2022
  
  Please see attached file for our full response to your helpful comments.
  
  Citation: https://doi.org/10.5194/egusphere-2022-869-AC5

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

AR by Torge Martin on behalf of the Authors (11 Dec 2022) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (12 Dec 2022) by Karen J. Heywood

RR by Anonymous Referee #2 (04 Jan 2023)

RR by Anonymous Referee #1 (10 Jan 2023)

ED: Publish subject to technical corrections (11 Jan 2023) by Karen J. Heywood

AR by Torge Martin on behalf of the Authors (20 Jan 2023) Author's response Manuscript

Co-editor-in-chief

Increasing Greenland Ice Sheet--melting is anticipated to impact water mass transformation in the subpolar North Atlantic and ultimately the meridional overturning circulation. Greenland meltwater redistribution pathways in and impact on the subpolar North Atlantic and overturning circulation is currently among the most debated topics. The manuscript provides new insights to both physical oceanographic processes and climate modelling. Based on a systematic setup of model configurations the importance of atmospheric feedbacks and mesoscale dynamics for specific regions of the subpolar North Atlantic are emphasized. This reaches beyond aspects of model techniques and also addresses the need for continued and improved observations in critical locations of the subpolar North Atlantic gyre circulation.