The firn meltwater Retention Model Intercomparison Project (RetMIP): evaluation of nine firn models at four weather station sites on the Greenland ice sheet

Vandecrux, Baptiste; Mottram, Ruth; Langen, Peter L.; Fausto, Robert S.; Olesen, Martin; Stevens, C. Max; Verjans, Vincent; Leeson, Amber; Ligtenberg, Stefan; Kuipers Munneke, Peter; Marchenko, Sergey; van Pelt, Ward; Meyer, Colin R.; Simonsen, Sebastian B.; Heilig, Achim; Samimi, Samira; Marshall, Shawn; Machguth, Horst; MacFerrin, Michael; Niwano, Masashi; Miller, Olivia; Voss, Clifford I.; Box, Jason E.

doi:https://doi.org/10.5194/tc-14-3785-2020

Articles | Volume 14, issue 11

https://doi.org/10.5194/tc-14-3785-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/tc-14-3785-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 14, issue 11

Research article

|

06 Nov 2020

Research article |

| 06 Nov 2020

The firn meltwater Retention Model Intercomparison Project (RetMIP): evaluation of nine firn models at four weather station sites on the Greenland ice sheet

Baptiste Vandecrux, Ruth Mottram, Peter L. Langen, Robert S. Fausto, Martin Olesen, C. Max Stevens, Vincent Verjans, Amber Leeson, Stefan Ligtenberg, Peter Kuipers Munneke, Sergey Marchenko, Ward van Pelt, Colin R. Meyer, Sebastian B. Simonsen, Achim Heilig, Samira Samimi, Shawn Marshall, Horst Machguth, Michael MacFerrin, Masashi Niwano, Olivia Miller, Clifford I. Voss, and Jason E. Box

Download

Final revised paper (published on 06 Nov 2020)
Supplement to the final revised paper
Preprint (discussion started on 02 Mar 2020)
Supplement to the preprint

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC1: 'Vandecrux review', Anonymous Referee #1, 14 Apr 2020
- AC1: 'Response to reviewer 1', Baptiste Vandecrux, 10 Jul 2020
SC1: 'Vandecrux comment', Kendall FitzGerald, 28 Apr 2020
- AC2: 'Response to Dr. FitzGerald', Baptiste Vandecrux, 10 Jul 2020
RC2: 'Review of Vandecrux et al.', Samuel Morin, 07 May 2020
- AC3: 'Response to Dr. Morin', Baptiste Vandecrux, 10 Jul 2020

Peer-review completion

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

ED: Publish subject to revisions (further review by editor and referees) (10 Jul 2020) by Xavier Fettweis

AR by Baptiste Vandecrux on behalf of the Authors (16 Jul 2020) Author's response Manuscript

ED: Referee Nomination & Report Request started (24 Jul 2020) by Xavier Fettweis

RR by Anonymous Referee #1 (12 Aug 2020)

Suggestions for revision or reasons for rejection

The revised version of ‘The firn meltwater Retention Model Intercomparison Project (RetMIP): Evaluation of nine firn models at four weather station sites on the Greenland ice sheet’ by Vandecrux et al., is an improvement from the original submission. The manuscript will make a meaningful contribution to the field. I do, however, provide here some comments which I believe will improve the manuscript prior to publication. Edits to a number of the figures would improve their clarity and communication of the results. I believe the authors need to clarify to the reader if (and if so, how) preservation of the initial conditions in short model runs may skew the results. I also believe the important manuscript points regarding model runoff/retention agreement could be better communicated by presentation of expanded output in a specific results section. I have detailed these and other comments below.

Line 64: consider replacing ‘quality’ with ‘fidelity’

Line 75: should be ‘nine firn models’.

Line 182: I don’t believe ‘runoff’ can be used as verb. Break in to two words.

Line 205-206: Why is ‘more recent’ synonymous with higher quality? Maybe more appropriate to state that you use this forcing dataset because it is co-located with the other observations you have for evaluation metrics?

Table 3: The forcing datasets all span very different periods of time, which has a direct bearing on the results. For instance, the authors compare the firn density at depth ranges up to 20m. It will take many years to decades for the initial conditions to be buried beyond these depths, but none of the forcing datasets are longer than ~17 years, and at the firn aquifer site (where the climate forcing only spans 8 months), the model results may be largely dictated by the initial conditions. Consequently, it’s unclear how much of the presented results reflect performance of the models, and how much is simply due to the continued presence of initial conditions in the modeled temperature and density. As an aside, it remains unclear what density profiles the authors use as the initial condition; both Herron and Langway and measured profiles are presented in the SI.

The authors should include some treatment of this issue in the manuscript. For the sites with longer forcing datasets (Summit and Dye-2), is it safe to say that the initial conditions have been pushed through the model domain? If so this should be stated (based on Figure 3 it appears this may not be the case?). At the other sites, KAN_U and FA, some statement regarding the influence of relict initial conditions needs to be made.

Line 236: This ‘deep firn temperature’ is presumably the firn temperature presented in Table 3, but at some sites it is in fact not deep at all (5 m at KAN_U). Yet, in the case of KAN_U, Figure 6 shows that temperature measurements extend to 10m. So why this discrepancy? How much of the deviation between measured and modeled temperatures simply results from the prescription of a boundary condition that is far from reality?

Figure 2: I recognize the challenge in presenting these spatial/temporal model results, but the embedded statistics in panel (C) make it quite difficult to visually assess the results. This is similarly true for the figures at the other sites. Consider moving these statistics in to a separate table.

Figure 3: In panel (A), consider changing each subpanel so that y-axes span the same density range but are focused on different intervals. As presented, the panels do not communicate much. In panel (D), the results appear to mostly display the model initial conditions and not a comparison to observation since the measured densities are limited to such shallow depth. Consider only plotting from 0-7 or 8 m.

Line 318: I don’t believe a comparison ‘responds’, as is stated. Consider alternate wording: ‘tracks closely’?

Line 345: Consider changing ‘allows to assess..’ to ‘allows assessment of’.

Line 346: See my earlier comment re: line 205-206. Why do these data present ‘the best conditions’? Consider different wording.

Lines 392-394: Does the increased deviation with time have something to do with the fact that the signal from the initial conditions is slowly being evacuated from the domain?

Figure 9: In panel A, Why not extend the model results through the observed density marker in 2017? There are clearly model results during this time period, as shown in panel H.

Figure 10: This appears to be a repeat of KAN_U results, and not the Firn Aquifer.

Section 6.2: This section actually honors most closely the title of the manuscript -- it presents an intercomparison of meltwater retention model results. I would strongly consider presenting this output in the results, including a panel for the single model year of firn aquifer conditions, and include the DTU model results. This would give the reader valuable insight in to the model agreement with respect to meltwater retention over a range of climate conditions. After all, perhaps the central motivation for developing these models (as described in the Introduction) is to assess retention and runoff. A discussion section focused on these results would still be quite useful for identifying why some models (e.g. DTU) are outliers, why the models are collectively challenged under certain climate conditions (e.g. those resulting aquifer generation), and for discussing the increasing model spread under increasing melt conditions. The latter point is perhaps the paper’s most important, as it illustrates the lack of certainty in any estimation of runoff/retention for the purposes of ice sheet mass balance.

Line 676-677: This final sentence I think is a misguided direction. Are you setting the stage for another intercomparison project? Rather than tips for future model intercomparisons, summarize the inter-model agreement at high melt sites --> it appears to be poor.

Hide

ED: Publish subject to minor revisions (review by editor) (13 Aug 2020) by Xavier Fettweis

AR by Baptiste Vandecrux on behalf of the Authors (20 Aug 2020) Author's response Manuscript

ED: Publish subject to minor revisions (review by editor) (08 Sep 2020) by Xavier Fettweis

AR by Baptiste Vandecrux on behalf of the Authors (11 Sep 2020) Author's response Manuscript

ED: Publish as is (22 Sep 2020) by Xavier Fettweis

AR by Baptiste Vandecrux on behalf of the Authors (22 Sep 2020) Author's response Manuscript

Short summary

In the vast interior of the Greenland ice sheet, snow accumulates into a thick and porous layer called firn. Each summer, the firn retains part of the meltwater generated at the surface and buffers sea-level rise. In this study, we compare nine firn models traditionally used to quantify this retention at four sites and evaluate their performance against a set of in situ observations. We highlight limitations of certain model designs and give perspectives for future model development.