Persistent climate model biases in the Atlantic Ocean's freshwater transport

van Westen, René M.; Dijkstra, Henk A.

doi:https://doi.org/10.5194/os-20-549-2024

Articles | Volume 20, issue 2

https://doi.org/10.5194/os-20-549-2024

© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/os-20-549-2024

© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 20, issue 2

Research article

|

12 Apr 2024

Research article |

| 12 Apr 2024

Persistent climate model biases in the Atlantic Ocean's freshwater transport

René M. van Westen and Henk A. Dijkstra

Download

Final revised paper (published on 12 Apr 2024)
Preprint (discussion started on 17 Jul 2023)

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2023-1502', Anonymous Referee #1, 23 Aug 2023

Review of: “Persistent Climate Model Biases in the Atlantic Ocean’s Freshwater Transport”, by Van Westen and Dijkstra.

In this paper the authors study biases in the AMOC stability metric Fov, by analyzing two CESM simulations with different resolutions, as well as a large number of CMIP6 simulations. The authors conclude that the biases that existed in CMIP3 and CMIP5 persist in CMIP6. Furthermore, they point to several biases in the freshwater budget as likely culprits for these biases in Fov.
This is a very thorough analysis, and I commend the authors for the work they have done. That said, the depth of the analysis has gone at the expense of the readability of the manuscript; I have to admit --with some embarrassment-- that I have not been able to get past the first pages of the Results section, despite several attempts. In my mind, the information density is far too high to make this a comfortable read. To illustrate this point, page 4 alone refers to Fig. 1 (4 panels plus 7 insets), Fig. 2 (8 panels, each with two insets), and 5 figures in Supplemental. The total number of panels + insets covered on page 4 is 70. That is a lot of information to get one’s head around in the space of 30 lines.
I hope that the authors will reconsider simplifying the paper and improve its readability. The paper can be slowed down significantly, simply by taking more time to develop the material. Not by adding more information, but by more carefully walking the reader through the argumentation following the key results. I understand that it is a challenging task, but the authors should make an effort to boil down the figures to those that are most critical to the storyline. Relegating more figures to Supplemental would be an option, but it only works if they are indeed treated as being of secondary importance, with limited referencing in the main text to avoid distracting from the main storyline. Although the insets might be useful in some cases (after careful study, Fig. 2 started to make sense), in others they are definitely a distraction (Figs. 1, 3). The insets that are critical to the narrative deserve their own figure and should be described and referenced in the proper order.
It is possible that there is simply too much ground to cover for one paper, in which case the authors might consider splitting it up in two companion papers.
I apologize that I don’t have anything more substantial to offer at this point.

Citation: https://doi.org/10.5194/egusphere-2023-1502-RC1
- AC1: 'Reply on RC1', René van Westen, 25 Oct 2023
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1502/egusphere-2023-1502-AC1-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-1502-AC1
- RC2:
  'Reply on RC1', Anonymous Referee #2, 31 Oct 2023
  Review of ‘Persistent Climate Model Biases in the Atlantic Ocean’s Freshwater Transport’
  
  In this manuscript van Westen and Dijkstra take a detailed look into what is causing the model biases in Fov in the Atlantic at 34S. They investigate the bias by looking at the different water masses at 34S and how they change immediately after the model spins up. These responses are compared in both a high (0.1 degree) and low (1 degree) CESM model and later compared to CMIP6 models and changes in future projections. In the CESM models the surface fresh bias can be related to the impact the Indian Ocean has on the Atlantic Surface waters, while slightly deeper the North Atlantic Deep Water biases are related to issues with surface fluxes in the North Atlantic Subpolar gyre. The manuscript furthermore investigates Fov in CMIP6 models and how it changes in future climate projections.
  
  Having begun the review of this manuscript after reviewer 1 posted their response I agree with them on the manuscript. The authors have approached the Fov bias issue from a from the perspective of different water masses, which is a very nice and informative way of investigating the model bias. Therefore, I believe this work is of interest to the community. They have also completed and presented a large amount of analysis. However, there is a large amount of information packed densely into one manuscript and it would benefit from streamlining it and/or splitting the manuscript. Similarly with the figures, some panels could be combined instead of having separate panels for the two models allowing a few of the insets to become their own panels, as opposed to small postage stamps.
  
  A few smaller points:
  
  The introduction seems short and could benefit from being expanded, discussion of the usefulness of Fov as an indicator would nice. See Yin and Stouffer 2007 and Mecking et al. 2016 for a discussion on using the divergence around the subtropical gyre. Also, the role of bias correction using flux adjustment (i.e. Liu et al. 2014, Liu et al. 2017, Jackson 2013). The paper Mignac et al. 2019 also worth mentioning.
  
  line 80 – see Menary et al. 2020 figure S1 for a comparison between computing own AMOC and AMOC provided by CMIP6 models.
  
  What are the initial S&T conditions used in this study?
  
  How are the freshwater transports computed in Fig.2 for the different water masses?
  
  There isn’t very much mention about Faz in the manuscript despite being defined. Interestingly, looking at the inset in Figure 1b it is clear that Faz also makes a quick adjustment. One thing that is very noticeable in Figure 5 d,e and f is that there is an azonal structure in ASW.
  
  In figure 5 and A6 it would be nice to see the plots as biases as opposed to absolute values.
  
  There is no mention in the abstract about the future projection results.
  
  I believe there are several nice results in this manuscript, and I would be happy to provide a more detailed review of this after the above mentioned comments have been considered.
  
  Refs:
  
  Yin, J. and Stouffer, R.J., 2007. Comparison of the stability of the Atlantic thermohaline circulation in two coupled atmosphere–ocean general circulation models. Journal of Climate, 20(17), pp.4293-4315.
  
  Mecking, J.V., Drijfhout, S.S., Jackson, L.C. and Graham, T., 2016. Stable AMOC off state in an eddy-permitting coupled climate model. Climate Dynamics, 47, pp.2455-2470.
  
  Liu, W., Liu, Z. and Brady, E.C., 2014. Why is the AMOC monostable in coupled general circulation models?. Journal of Climate, 27(6), pp.2427-2443.
  
  Liu, W., Xie, S.P., Liu, Z. and Zhu, J., 2017. Overlooked possibility of a collapsed Atlantic Meridional Overturning Circulation in warming climate. Science Advances, 3(1), p.e1601666.
  
  Jackson, L.C., 2013. Shutdown and recovery of the AMOC in a coupled global climate model: the role of the advective feedback. Geophysical Research Letters, 40(6), pp.1182-1188.
  
  Mignac, D., Ferreira, D. and Haines, K., 2019. Decoupled freshwater transport and meridional overturning in the South Atlantic. Geophysical Research Letters, 46(4), pp.2178-2186.
  
  Menary, M.B., Robson, J., Allan, R.P., Booth, B.B., Cassou, C., Gastineau, G., Gregory, J., Hodson, D., Jones, C., Mignot, J. and Ringer, M., 2020. Aerosol‐forced AMOC changes in CMIP6 historical simulations. Geophysical Research Letters, 47(14), p.e2020GL088166.
  
  Citation: https://doi.org/10.5194/egusphere-2023-1502-RC2
  - AC2: 'Reply on RC2', René van Westen, 06 Nov 2023
    
    The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1502/egusphere-2023-1502-AC2-supplement.pdf
    
    Citation: https://doi.org/10.5194/egusphere-2023-1502-AC2
EC1: 'Comment on egusphere-2023-1502', Bernadette Sloyan, 19 Oct 2023

As editor I suggest the authors post a reply detailing how they plan to address the issued raised by the reviewer 1.

Citation: https://doi.org/10.5194/egusphere-2023-1502-EC1
EC2: 'Comment on egusphere-2023-1502', Bernadette Sloyan, 01 Nov 2023

Both reviewers suggest that the paper be streamlined to remove non-critical information or split into two manuscript. The authors reply to reviewers 1 indicate that they will remove information/subplots rather than split the study into two manuscripts. I encourage the authors to undertake the suggested revisions to the manuscript.

Citation: https://doi.org/10.5194/egusphere-2023-1502-EC2

Peer review completion

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

AR by René van Westen on behalf of the Authors (11 Nov 2023) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (14 Nov 2023) by Bernadette Sloyan

RR by Anonymous Referee #1 (05 Jan 2024)

Suggestions for revision or reasons for rejection

Review of: “Persistent Climate Model Biases in the Atlantic Ocean’s Freshwater Transport”, first revision, by Van Westen and Dijkstra.

First of all, apologies for having taken more than the allotted time for this review.

In this paper the authors study biases in the AMOC stability metric Fov, by analyzing two CESM simulations with different resolutions, as well as a large number of CMIP6 simulations. The authors conclude that the biases that existed in CMIP3 and CMIP5 persist in CMIP6. Furthermore, they point to several biases in the freshwater budget as likely culprits for these biases in Fov.

I recognize that the authors have significantly improved the readability of the manuscript by removing many figures, primarily from supplemental information. I commend them for making this effort. I did not find any fault with the analysis, so in principle the paper is publishable, after considering a few minor comments detailed below.

That said, despite the reduction in the number of figures, I still found the paper a tedious read, and I’m concerned that this might be detrimental to the impact of the paper. I think that the main culprit is the fact that this paper drowns in details, making it hard to discern any overarching purpose of the paper. In fact, the introduction lacks an explicit goal altogether, missing an opportunity to provide direction for the subsequent analysis. What is the research question that this paper wants to answer? And what analysis will enable us to answer that question? I’m not recommending a major revision of this paper, but it would help the reader if the authors could include some summary statements here and there, as well as introductory statements that explain why a certain analysis is being done, and how it helps to address the main thrust of the paper.

Minor comments:
l. 26: I think collapses have been found in models with every level of complexity except for eddy-resolving models, but including the eddy-permitting studies of Mecking et al.
l. 33: salinity -> saline
l. 63: climatological -> climatologies
l. 78: I find it hard to reconcile the use of the term ‘baroclinic’ with the traditional use, even when simply meant to indicate a velocity profile with the barotropic component removed. Removing a section-averaged velocity surely retains a significant barotropic component of the resulting flow field.
ll. 106-107, “In this section we focus…”: It’s probably best to start the section with this sentence, and it wouldn’t hurt to be more explicit about the goal of the section.
l. 107: simulation -> simulations
l. 112: This applies to LR only.
l. 113: fairy -> fairly
l. 114, caption of Fig. 1: Aghulas -> Agulhas
l. 149: “…both models, the AMOC…”: insert ’while’.
l. 159, 311: The statement about the horizontal mixing is not obvious to me. On what evidence (or literature) is this based? How does this relate to the use of GM in LR? And is there no role for vertical (or isopycnal) mixing?
l. 182: “…and is related to…”: please rephrase.
l. 195: I don’t quite understand this: Fig. 5 g, h, and i show that the NADW contribution to FovS is consistently small in both the reanalysis and the models.
l. 230: more -> further
l. 274: freshwater -> fresh
l. 280: Is it relevant that MCM-UA-1-0 seems to have the lowest resolution of all models?
l. 324: needs -> need
l. 335: Do you mean Weijer et al. (2020) here?
Section 3.4: Does this analysis shed light on why the current generation of climate models cannot have a correct AMOC and FovS at the same time? This would be incredibly useful to know.

Hide

RR by Anonymous Referee #2 (07 Feb 2024)

Suggestions for revision or reasons for rejection

I commend the authors on greatly improving the manuscript and enhancing the readability. I think this is a very interesting manuscript and deserves publication. However, I still have a few concerns/suggestions:

1) Abstract - The manuscript mainly focuses on the results from LR-CESM and HR-CESM and how they relate to the salinity bias at 34S and only the last section of the results talks about CMIP6. The second sentence does not seem correct, the tipping point behaviour is likely caused by freshwater capping and/or warming of high latitudes as opposed to salt-advection feedback, while the idealised studies have shown that positive salt advection feedback is related to whether the AMOC collapse is stable or the AMOC will strengthen again quickly. Also, the abstract makes it seem like the manuscript is exclusively a CMIP6 study. Finally, since the bias at 34S is the main topic of the manuscript it would be useful to be more specific about the biases, i.e. surface too fresh and too salty at depth.
2) Line 18 – recent study on future projections and ocean heat transport: Mecking and Drijfhout 2023
3) Line 24 – The study Lobelle et al. 2020 has estimated that 29-67 years are required to properly detect a weakening of the AMOC.
4) Line 28 – Again here I would say that the transitions are not caused by salt advection feedback but rather the stability of them are (or a transition back to an AMOC-on state)
5) Line 33 – Note that the sign of Fov is not dependent on the reference salinity, the reference salinity just acts as a scaling, since the section averaged velocity is subtracted at 34S before Fov is computed.
6) Line 39-45 – Since the salinity bias is talked about in detail it would be worth introducing the bias in more detail since this is described in several papers,
7) Line 40-42 – The salinity bias is often corrected using flux adjustment as can clearly be seen in Yin and Stouffer 2007, Jackson 2013, and Liu et al., 2017 while in Mecking et al. 2016 the salinity bias wasn’t corrected directly but the bias was lower in this model (i.e. Figure 8c in Mecking et al 2017).
8) Line 45-46 – This sentence is very strongly worded and it has not been shown that the positive Fov is the reason for the models not having a tipping AMOC and I see it more as an indicator for the stability of an AMOC off-state as opposed to the model’s ability to tip.
9) Line 39-45 – A new multi-model study show, Jackson et al. 2013, shows that there is no clear link between whether the AMOC recovers and Fov, however, in Figure 9 it hints that the models with negative Fov at 34S might be more likely to recover. I think this paper deserves a mention.
10) Line 58 – Mention that it’s the ocean component that has 60 vertical levels
11) Figure 1c,d – In the reanalysis data for Faz there is a very big spike around 2000 or just before. Given how large this is it should be mentioned. Is there a known explanation for it? If it is caused by salinity can it be seen in other data sets? Also, Figure 5d there is a lot less zonal gradient in salinity in the ASW than in e and f, but this is likely related the freshwater from the Indian Ocean.
12) Figure 7e,f – the inset is very narrow, that it took me a while to notice them, I would suggest making them wider
13) Line 267 – this is similar to Mecking et al. 2017
14) Figure 8 – The figure label says the CMIP6 dots are black but they are actually grey and would be helpful to add the two different symbols for CMIP6 to the legend
15) Line 292-294 – Maybe I missed something, but how was it shown that the trends are salinity versus velocity driven?
16) Figure 9- Panels b,c,d would be easier to see differences if they were plotted as anomalies w.r.t. the Reanalysis
17) Figure A1 – The insets are very narrow and difficult to see. This figure would also benefit from plotting as anomalies w.r.t. reanalysis. As stated in the response to the previous reviews, I disagree that much information will be lost in the regridding process since this is a 2D field and not a computation of transport, where more care must be given to conserve the transport in the regridding process. Also, the different water masses can be seen with the dashed lines
18) Response to Reviewers - Reviewer 2 response 2 – This change has not been implemented into the manuscript

Mecking, J.V. and Drijfhout, S.S., 2023. The decrease in ocean heat transport in response to global warming. Nature Climate Change, 13(11), pp.1229-1236.

Lobelle, D., Beaulieu, C., Livina, V., Sevellec, F. and Frajka‐Williams, E., 2020. Detectability of an AMOC decline in current and projected climate changes. Geophysical Research Letters, 47(20), p.e2020GL089974.

Jackson, L.C., 2013. Shutdown and recovery of the AMOC in a coupled global climate model: the role of the advective feedback. Geophysical Research Letters, 40(6), pp.1182-1188.

Jackson, L.C., Alastrué de Asenjo, E., Bellomo, K., Danabasoglu, G., Haak, H., Hu, A., Jungclaus, J., Lee, W., Meccia, V.L., Saenko, O. and Shao, A., 2022. Understanding AMOC stability: the North Atlantic hosing model intercomparison project. Geoscientific Model Development Discussions, 2022, pp.1-32.

Hide

ED: Publish subject to minor revisions (review by editor) (14 Feb 2024) by Bernadette Sloyan

AR by René van Westen on behalf of the Authors (20 Feb 2024) Author's response Author's tracked changes Manuscript

ED: Publish as is (05 Mar 2024) by Bernadette Sloyan

AR by René van Westen on behalf of the Authors (05 Mar 2024)

Short summary

The Atlantic Meridional Overturning Circulation (AMOC) is an important component in the global climate system. Observations of the present-day AMOC indicate that it may weaken or collapse under global warming, with profound disruptive effects on future climate. However, AMOC weakening is not correctly represented because an important feedback is underestimated due to biases in the Atlantic's freshwater budget. Here we address these biases in several state-of-the-art climate model simulations.