Warm Greenland during the last interglacial: the role of regional changes in sea ice cover

Merz, Niklaus; Born, Andreas; Raible, Christoph C.; Stocker, Thomas F.

doi:https://doi.org/10.5194/cp-12-2011-2016

Articles | Volume 12, issue 10

https://doi.org/10.5194/cp-12-2011-2016

© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.

https://doi.org/10.5194/cp-12-2011-2016

© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 12, issue 10

Research article

|

28 Oct 2016

Research article |

| 28 Oct 2016

Warm Greenland during the last interglacial: the role of regional changes in sea ice cover

Niklaus Merz, Andreas Born, Christoph C. Raible, and Thomas F. Stocker

Download

Final revised paper (published on 28 Oct 2016)
Preprint (discussion started on 08 Feb 2016)

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC1: 'Review of manuscript by Merz et al. entitled “Warm Greenland during the last interglacial: the role of regional changes in sea ice cover”.', Pepijn Bakker, 24 Feb 2016
- AC1: 'Author comment to Reviewer #1', Niklaus Merz, 17 May 2016
RC2: 'Review of the manuscript by Merz et al.', Anonymous Referee #2, 25 Feb 2016
- AC2: 'Author comment to Reviewer #2', Niklaus Merz, 17 May 2016
RC3: 'Review of manuscript by Merz et al.', Anonymous Referee #3, 14 Mar 2016
- AC3: 'Author comment to Reviewer #3', Niklaus Merz, 17 May 2016

Peer-review completion

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

ED: Reconsider after major revisions (20 May 2016) by Andrea Dutton

AR by Niklaus Merz on behalf of the Authors (01 Jul 2016) Author's response Manuscript

ED: Referee Nomination & Report Request started (12 Jul 2016) by Andrea Dutton

RR by Anonymous Referee #3 (20 Jul 2016)

Suggestions for revision or reasons for rejection

Review of:
Merz et al. "Warm Greenland during the last interglacial: the role of regional changes in sea ice cover"

The authors have taken a lot of care to respond in detail to the reviews and the revised text is in many ways an improvement over the original manuscript. However, the authors have chosen to argue agains most of the major concerns raised by this reviewer. This is of course fine so long as they provide strong arguments against the suggested changes. This is unfortunately not the case for the design of the experiments investigating how the Greenland climate responds to sea ice change in the North Atlantic. My main criticism is therefore the same as in the previous round of review: the sensitivity experiments are highly unrealistic and I am skeptical to how they benefit the scientific community. Also, the fact that the authors use Deser et al. (2010) and the difference between the LGM and PI climates to motivate the design of these experiments suggests that they don't fully understand the criticism.

I don't object to the fact that you see turbulent surface flux anomalies of order 100 Wm-2 when comparing two experiments with different SST/SIC, in fact I would be surprised if you didn't see differences of this magnitude. This is of course true when comparing a future projection of the Arctic surface climate with the PI (Deser et al., 2010) or the LGM and PI climates shown in the rebuttal, as you compare two states where the surface energy balance is totally different and not even described by the same number of terms (sea ice is a good insulator and effectively cuts off the ocean temperature from influencing the surface energy balance). My objection is rather how you arrive at these anomalies and how they influence the circulation. It appears to me that the experiment design has a very limited real world application and I am therefore questioning the robustness of your conclusions.

Fixed SST experiments are great for investigating and isolating the response to different boundary conditions. However, it also comes with a responsibility to design experiments that make sense as the model is more restricted than a slab ocean or a fully coupled GCM. As you know, prescribing the SST field means that the atmospheric circulation is relaxed towards a state that is consistent with the SST boundary condition (a consequence of geostrophic balance, etc.). The problem with your experiments is that you prescribe a climatological collapse/retreat of sea ice (implied to happen over 1000s of years) that is not supported by any of the boundary conditions: not the prescribed SST field, topography, insolation, GHG, etc. This effectively means that you prescribe an infinite heat source (it stays the same no matter what the atmosphere does) in a limited area right next to the region where you investigate the temperature change. Since none of the boundary conditions support the sea ice collapse and the ocean temperature is static, the atmosphere is the only medium that can respond and counteract the heating. This is evident from the ~100 Wm-2 turbulent fluxes where the sea ice has been removed (the ocean is trying to cool by giving up heat to the atmosphere) and that the atmosphere is advecting the warm temperatures away in order to cool the surface and re-form sea ice in this region. Why is this a problem then? It is a problem because something like this would never happen in nature, at least not in the same way as described in the manuscript. In a scenario where the boundary conditions are fixed, the inter-annual variability of sea ice oscillates around a climatological state that is consistent with the mean state of the ocean and atmospheric circulation; this climatological state can only change if there are corresponding changes in the mean state of the ocean and atmospheric circulation. This is ultimately the reason for why you cannot use future projections of the Arctic and/or the LGM to motivate this experimental design as these climate states are globally different due to altered boundary conditions. As you no doubt know, the Arctic sea ice cover is projected to decrease because of elevated concentrations of greenhouse gases and the LGM climate was massively different (as nicely discussed in e.g. Merz et al., 2015) because of changes in: insolation, greenhouse gases, topography, surface albedo, continental outlines, vegetation cover, etc. Even though there are large uncertainties related to both these climate state, it is very unlikely that the climate was/will be the same as in the PI except for in a limited region.

Moreover, I don't agree with the claim that a slab ocean experiment is worse than prescribing the SST/SIC. It is true that the classic implementation of a slab ocean only considers vertical energy fluxes and no horizontal transport. However, what you have done here is arguably much worse as you prescribe an infinite energy source that does not respond at all to changes in the atmospheric circulation. A slab ocean would conserve energy (the surface budget wouldn't be totally outlandish in a limited region) and yield a larger scale response to the implied forcing. This is perhaps not perfect but it is arguably more realistic than what has been done here. It is of course possible that your main conclusions would be the same in a slab ocean experiment but it is also possible that it would be totally different. The only thing we can say with absolute certainty is that a slab ocean experiment would account for interactions and feedback loops that have broken in the current experiment design. It is after all the authors job to prove that the experiments they present are robust and realistic and that we learn something useful from the study. I don't necessarily think that this is the case in the current manuscript.

Also, in future submissions I strongly advice omitting the comparison of the implied 2m temperature over prescribed SSTs. Despite what you wrote in the rebuttal this comparison is not good for several reasons: (i) the strong correlation between the surface and 2m temperature means that you are basically prescribing the phenomena you are investigating (there will be a slightly weaker correlation over land but the SST field will undoubtedly contaminate the results there as well), (ii) you haven't proven that the temperature difference is due to the model physics, it can be due to internal climate variability (see, e.g., Deser et al., 2012, 2012, 2014: DOI 10.1007/s00382-010-0977-x, DOI: 10.1038/NCLIMATE1562, DOI: 10.1175/JCLI-D-13-00451.1), (iii) I argue that any model of the complexity used here will produce similar temperature differences as shown in Fig 2. Though hard to do, it would be better to run two models with less code inherency if you want to prove this point.

Hide

ED: Publish as is (13 Sep 2016) by Andrea Dutton

AR by Niklaus Merz on behalf of the Authors (23 Sep 2016) Manuscript

Short summary

The last (Eemian) interglacial is studied with a global climate model focusing on Greenland and the adjacent high latitudes. A set of model experiments demonstrates the crucial role of changes in sea ice and sea surface temperatures for the magnitude of Eemian atmospheric warming. Greenland temperatures are found highly sensitive to sea ice changes in the Nordic Seas but rather insensitive to changes in the Labrador Sea. This behavior has important implications for Greenland ice core signals.