Responses of sub-ice platelet layer thickening rate and frazil-ice concentration to variations in ice-shelf water supercooling in McMurdo Sound, Antarctica

Cheng, Chen; Jenkins, Adrian; Holland, Paul R.; Wang, Zhaomin; Liu, Chengyan; Xia, Ruibin

doi:https://doi.org/10.5194/tc-13-265-2019

Articles | Volume 13, issue 1

https://doi.org/10.5194/tc-13-265-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/tc-13-265-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 13, issue 1

Research article

|

29 Jan 2019

Research article |

| 29 Jan 2019

Responses of sub-ice platelet layer thickening rate and frazil-ice concentration to variations in ice-shelf water supercooling in McMurdo Sound, Antarctica

Chen Cheng, Adrian Jenkins, Paul R. Holland, Zhaomin Wang, Chengyan Liu, and Ruibin Xia

Download

Final revised paper (published on 29 Jan 2019)
Preprint (discussion started on 01 Aug 2018)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Review of tc-2018-135: Response of sub-ice platelet layer thickening rate to variations in Ice Shelf Water supercooling in McMurdo Sound, Antarctica', Anonymous Referee #1, 29 Aug 2018
- AC1: 'Final Response of tc-2018-135', Chen Cheng, 19 Oct 2018
RC2: 'Review of ‘Response of sub-ice platelet layer thickening rate to variations in Ice Shelf Water supercooling in McMurdo Sound, Antarctica’ by Cheng et al, submitted to The Cryosphere 2018', Anonymous Referee #2, 09 Sep 2018
- AC2: 'Final Response of tc-2018-135', Chen Cheng, 19 Oct 2018
SC1: 'Open review by Ken Hughes', Kenneth Hughes, 09 Sep 2018
- AC3: 'Final Response of tc-2018-135', Chen Cheng, 19 Oct 2018

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Chen Cheng on behalf of the Authors (19 Oct 2018) Author's response Manuscript

ED: Referee Nomination & Report Request started (31 Oct 2018) by Christian Haas

RR by Anonymous Referee #1 (13 Nov 2018)

Suggestions for revision or reasons for rejection

Dear Christian,

The authors have revised their manuscript and made significant improvements (including to the quality of the explanations and calculations of the suspension index). The topic of the paper is interesting and worthwhile. However, there remain some substantial issues with the paper, as summarized below. Therefore, I don’t think the manuscript presently merits publication. With further changes, it could potentially be published in some journal in future.

Major issues:

Novelty. The authors now do a better job explaining the novelty of their approach and the contribution of previous studies. However, it remains the case that this paper represents fairly incremental progress on their previous modelling paper. The particular geographical application is interesting but not totally compelling (the other reviewers raise important discrepancies between model and observations). In terms of interpretation, the paper still does not make it clear why the suspension index should be relatively large (greater than one, say). I can only guess that the various velocities that go into u* are very small. I could only see one of these velocities reported in Table 1. It is not clear how widely applicable this situation is. For example, if the ambient currents were (say) 0.1 m/s (10 times larger than the value in Table 1), Z* could be a factor of 10 smaller and the novel process modelled would not be that important. The wider significance of the work is therefore unclear.

Correctness. A major issue relates to the choice of model parameters and the correctness of the numerical solution of the governing equations. In figure R, the authors present the results of calculations with a better-resolved discretization of the crystal size distribution. These differ markedly from the standard discretization presented in the paper, indicating that the standard results are inaccurate. A reader would not be aware of this important issue from the discussion on P6:L9-13. In their response, the authors argue that the poorly resolved discretization is better because it ‘agrees’ with observation. This argument is fallacious. Rather, the evidence suggests that there is some large flaw in another part of the model (most likely some choice of model parameters or, worse, some structural issue). The authors have not addressed the issue regarding the conversion of precipitation rate to ice thickening rate, where there are large uncertainties. While outside the scope of their model, this process greatly influences the comparison of the model with observations. I also think that the neglect of latent heat release during volume doubling violates energy conservation.

Hide

RR by Anonymous Referee #2 (06 Dec 2018)

Suggestions for revision or reasons for rejection

Review of ‘Responses of sub-ice platelet layer thickening rate and frazil ice concentration to variations in Ice Shelf Water supercooling in McMurdo Sound, Antarctica’ following reviewer comments and revision

Summary
The revisions and corrections made to this manuscript have significantly improved both its interest to the community, and clarity for the reader. However, some of my original comments have not been adequately considered, and in my opinion, would need to be properly addressed (both in response to reviewers and by modifying the article text) before the manuscript is ready for publication.

Specific points:
1. All references to ‘temperature’ should be changed to ‘potential temperature’ (assuming this is indeed the quantity used). For example, the text in P2-L15 was modified in response to Reviewer #1’s comments, to support the author’s assertion that temperature can be treated as vertically-uniform within the ISW plume. Unless the qualifier ‘potential’ is used, this is incorrect, and not supported by observations.

2. Distinction between applicability to crystal accretions beneath sea ice and ice shelves remains unclear. In their response to my review, the authors note that “Beneath the platelet ice layers the supercooling is produced by the pressure drop experience by ISW at it emerges from beneath an ice shelf and rise towards the sea surface, while regions of marine ice accretion beneath ice shelves have typically very low basal slopes.” I completely agree with this, and am therefore puzzled as to why the authors seem reluctant to identify this difference to the reader. This is exactly the mechanistic driver that would allow the sub-ice boundary layers to behave differently.
I am not suggesting that the work (tested for platelet layers beneath sea ice, and potentially applied to marine ice layers beneath ice shelves) is invalid. However, I do think that a statement clarifying the differences between the regimes is necessary for the reader (and may protect the authors of the present study in the event that their model is naively applied to a regime for which it has not been validated). Specifically, the potential sources of divergence are:
a. The difference in basal slope is likely to result in the ISW plume existing much closer to the in-situ freezing temperature than is observed in McMurdo Sound, with the result that in-situ supercooling is likely to be much smaller (and potentially similar to the resolution of present-day instruments. i.e. unobservable except by implication) beneath ice shelves;
b. The difference in basal slope also has implications for generating buoyancy-induced momentum, which is the implicit source of the background current in Hughes et al. (2014), and therefore would need to be excluded (or vastly reduced) for application to an ice shelf cavity.
c. The authors allude to another difference with their statement “The one indirect effect on the ISW plume might be an increased drag coefficient beneath the platelet layer, where more rapid freezing of the deposited crystals may create more irregularity in the form of the ice-ocean interface.” This is true, and will affect the sedimentation process.
d. In addition to the above, the length of time over which crystal accretion may occur is vastly different (i.e. about 1-3 years in McMurdo Sound vs tens-hundreds of years beneath ice shelves). Combined with the likely difference in degree of supercooling, this could potentially lead to very different internal structures of the crystal layers (e.g. marine ice layers more likely to collapse under accreted buoyancy), and hence present different physical boundaries to ocean flow, and an entirely different source of effective hydrodynamic drag (as suggested by Robinson et al., 2017, which the authors cite).
e. Finally, the observed supercooled plume in McMurdo Sound, having only recently experienced the step-change in pressure, is still adjusting to the change through active ice formation (onto both suspended frazil and to accreted platelet ice), and will therefore come to a point of equilibrium at some point beyond where the observations to date have been made. This represents a significantly different thermodynamic regime to the general situation of an ISW plume beneath an ice shelf.

This specifically relates to P11L8-10 of the revised manuscript, since in the general sub-ice shelf regime, the supercooled layer will almost certainly not approach the thickness observed in McMurdo Sound. This may have implications for the regime for which ‘the efficiency of converting ISW supercooling into frazil concentration … is determined by the suspension index’, since this is true only when ‘the thickness of a supercooled layer of ISW is large enough’ (P11L8-10) (i.e. greater than 65 m for the McMurdo Sound parameters).

I suggest that the addition of a well-crafted paragraph of text outlining the potential differences between the regimes would be sufficient to both demonstrate that the authors understand the implications of these differences (and I am convinced they do), and highlight to the reader where caution (and/or improved understanding) is required in applying this model to the sub-ice shelf regime.

3. Justification for values of parameters used, and acknowledgement of available observations. In their response, the authors point to the lack of ‘observational guidance’ as justification for the extensive tuning of specific parameters. I agree, there are very little data available to constrain the models. However, the values they have chosen do find support in the literature, and it would strengthen the paper to acknowledge these. In particular (P7L9-10):
a. ISW outflow properties: the chosen values for temperature and salinity coincide with those reported by Hughes et al., 2014;
b. Platelet layer basal drag coefficient: the value chosen fits appropriately within the range identified by Robinson et al., 2017;
c. Frazil ice crystal size distribution: unknown, but presumably these are chosen from somewhere – perhaps previous modelling studies? Similarly for the Shields criterion?
d. Ambient current speed: The chosen value is consistent with the lowest speeds reported by Robinson et al., 2014 (although lower than their reported residual flow).
e. In addition, the observations in both Hughes et al. (2014) and Robinson et al. (2014) papers show the homogeneous ISW layer (observed in the centre of the modelled plume flow) as being 150 - 200 m thick, and the supercooled portion extending to 60/70 m. I would have thought these would be useful reference points for this manuscript.

4. My concern about the apparent resolution in figures 5 & 6 still stands: the separation of the contour lines, especially around the core of the plume, implies greater resolution than the model contains. A potential solution may be to plot only every second contour line.

Hide

RR by Anonymous Referee #3 (15 Dec 2018)

Suggestions for revision or reasons for rejection

The paper is logically and clearly written. It describes the application of a two-dimensional ISW plume model to a set of observations of an ISW under sea ice. The results are an improvement over the model of Hughes et al (2014), particularly in terms of their two-dimensional nature that illustrates the Coriolis effect along the centre line of the plume. In addition the authors show (as naively might be expected) that the vertical distribution of supercooling and frazil concentration determine the growth of the sub-ice platelet layer.

Comment 1: Some details of the process of formation of sea ice need to be outlined and are ignored in the paper. There are two consequences

(i) The term “platelet layer” is confusing. The term “sub-ice platelet layer” was coined by Gow et al (1998) and later authors for the high porosity layer that principally forms by accumulation of ice from the water column. I think the authors use “platelet layer” to mean this friable layer beneath the more consolidated “incorporated platelet ice”. Incorporated platelet ice can be identified easily in an ice core by its crystal structure, but its other physical properties (salinity, porosity, permeability) are very similar to the usual columnar sea ice. It is formed by the freezing of the sub-ice platelet layer due to heat loss to the atmosphere. The term “platelet layer” could be understood to mean the sum of the incorporated platelet ice and the sub-ice platelet layer. In my view “sub-ice platelet layer” or “SIPL” (as in the original version) should be used to avoid confusion.
Gow, A. J., S. F. Ackley, J. W. Govoni, and W. F. Weeks (1998), Physical and structural properties of land-fast sea ice in McMurdo Sound, Antarctica, in Antarctic Sea Ice: Physical Processes, Interactions and Variability, Antarct. Res. Ser., vol. 74, edited by M. O. Jeffries, pp. 355–374, AGU, Washington, D. C., doi:10.1029/AR074p0355. 

(ii) The data used from HU14 is the “sub-ice platelet layer” thickness in November. However the top portion of this layer will have become incorporated into the consolidated sea ice above. Thus the value used will be an underestimate of the true value had the freezing due to heat loss to the atmosphere not taken place. Better agreement might be obtained betweem model and observation if the thickness of the incorporated platelet ice was included. There are some values of the incorporated platelet ice thickness in HU14 (Fig 8b) but it would be better derived from Figs 2 & 3 and the Supplementary material of Langhorne et al (2015).

Comment 2: The authors have conducted a significant sensitivity study. I would like to know what can be learned from this sensitivity study about the physics controlling the process of deposition of ice crystals; and how could this inform the design of observational campaigns. For example Fig 6 shows much steeper gradients in sub-ice platelet layer thickness than demonstrated in the observations. Why? Could there be post-depositional processes taking place that are not accounted for in the model; or a mismatch in the deifnition of between modeled and observed "platelet layer" (see next comment) or undersampling of the observations.

The authors are rather glib in simply saying that small scale features were not resolved by relatively coarse scale spatial distribution of the sea ice thickness measurements (incidently these were not ice core samples). Experimental evidence of steep gradients in the sub-ice platelet layer thickness have been published (Hunkeler et al, 2015) but these were associated with a sea ice breakout and an abrupt change in sea ice plus snow thickness. A significant change in sea ice thickness was not observed in the McMurdo Sound observations.

Hunkeler, P. A., M. Hoppmann, S. Hendricks, T. Kalscheuer, R. Gerdes (2016), A glimpse beneath Antarctic sea ice: Platelet layer volume from multifrequency electro- magnetic induction sounding, Geophys. Res. Lett., 43, 222–231, doi:10.1002/2015GL065074.  

Further if the position of the 3 km wide plume outflow (an unknown in the model) was moved east or west by a small amount then there would be a big discrepancy between model and experiment.

Comment 3: Fig 1 is incorrect. A pre-2011 ice shelf front has been used and this means that the purple box has been placed too far to the north. The samples stations (red dots) are in the wrong locations.

Technical Corrections
p. 2, line 15-24: I was surprised that the instabilities derived by Jordan et al (2014) were not discussed in relation to the vertical distribution of frazil. Are they not relevant?
p. 3, Line 2-3: “The main objective is to quantify for the first time the response of the platelet layer thickening rate to variations in ISW supercooling.” I don’t think this can be claimed as a first as HU14 considered supecooling and sub-ice platelet layer thickness, as did Buffo et al (2018). Please delete “for the first time”
p. 8, line 13-18: The authors are rather glib in simply saying that small scale features were not resolved by relatively coarse scale spatial distribution of the sea ice thickness measurements (BTW these were not ice core samples). Experimental evidence of steep gradients in SIPL thickness have been published (Hunkeler et al, 2015) but these were associated with a breakout and an abrupt change in sea ice plus snow thickness. A significant change in sea ice thickness was not observed in the McMurdo Sound observations.
Further if the position of the 3 km wide plume outflow (an unknown in the model) was moved east or west by a small amount then there would be a big discrepancy between model and experiment.
p. 8, line 26-30: “we know of no studies to date of the response of marine ice (or platelet layer) thickening rate beneath ice shelves (or sea ice) to variations in supercooling,” Again I would argue that HU14 and Buffo et al (2018) considered this relationship. On the other hand I think the authors could expand on the relationship and provide insights and explanations.
p. 10, line 21-26: This is exactly the sort of explanation that I find a useful outcome of the detailed modeling.

Hide

ED: Publish subject to minor revisions (review by editor) (27 Dec 2018) by Christian Haas

AR by Chen Cheng on behalf of the Authors (17 Jan 2019) Author's response Manuscript

ED: Publish as is (20 Jan 2019) by Christian Haas

Short summary

The sub-ice platelet layer (SIPL) under fast ice is most prevalent in McMurdo Sound, Antarctica. Using a modified plume model, we investigated the responses of SIPL thickening rate and frazil concentration to variations in ice shelf water supercooling in McMurdo Sound. It would be key to parameterizing the relevant process in more complex three-dimensional, primitive equation ocean models, which relies on the knowledge of the suspended frazil size spectrum within the ice–ocean boundary layer.