Quantifying effects of Earth orbital parameters and greenhouse gases on mid-Holocene climate

Kang, Yibo; Yang, Haijun

doi:https://doi.org/10.5194/cp-19-2013-2023

Articles | Volume 19, issue 10

https://doi.org/10.5194/cp-19-2013-2023

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

https://doi.org/10.5194/cp-19-2013-2023

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

Articles | Volume 19, issue 10

Research article

|

24 Oct 2023

Research article |

| 24 Oct 2023

Quantifying effects of Earth orbital parameters and greenhouse gases on mid-Holocene climate

Yibo Kang and Haijun Yang

Download

Final revised paper (published on 24 Oct 2023)
Preprint (discussion started on 17 Mar 2023)

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2023-380', Anonymous Referee #1, 02 May 2023

Summary

This paper describes the results of mid-Holocene simulations with CESM1 with a focus on separating the impacts of the altered orbit and lower CO2 relative to the pre-industrial. the authors also focus on the change in the AMOC. They find that the orbital forcing at 6kyr BP is much more important than the lower CO2 and that the AMOC strengthens as a result of sea-ice changes.

The text is well-written and clearly laid out. The main issue is in communicating where this paper fits into the established modelling literature and clearly articulating the advance here over other work.

Main comments
Previous studies with multi-model ensembles have already shown that there is no model consensus on the AMOC at the mid-Holocene (Jiang et al., 2023) but also that the changes in AMOC are actually very small between the mid-Holocene and pre-industrial (e.g. figure 9 by Brierley et al 2020). Thus this study’s focus on AMOC could do with additional justification. This could be straightforward if the model used was significantly more advanced than those used before, but that is not the case. For example, Jiang et al. included CESM2 the successor of the model used here. Also, at T31gx3v7 (3 degrees in the ocean?) the present study is pretty coarse resolution. The changes in AMOC reported are also relatively small and might even fall close to the 0.95-1.05 uncertainty bound in Brierley et al (2020)’s figure 9. Given this context, the significance of the present findings need to be clarified.

Technical corrections:

Line 69: Probably worth stating that this is ~3.75 degree resolution?
Line 70: What does “gx3v7 horizontal resolution” mean?
Line 72: Not sure that Smith & Gregory, 2009 should be here?
Lines 74-79: replace “in the X period” with “from the X period”.

FIgure 5: Panel (f) is not described in the caption. Titles above panels d)-f) would improve clarity. Also, timeseries of the AMOC and some quantification of interannual to centennial variability in the model owuld help to assess the significance of the changes shown here.

Figure 6: There are no dashed or dotted lines in the figure? Also I think the vertical scale in panels b and c should be the same.

Line 291: This sounds a bit sceptical, you could say “most proxy data” and cite e.g. Larrosoana et al (2012) or similar.

Line 291: the Jiang et al. study is about China so good to specify that here and please correct the brackets.

Citation: https://doi.org/10.5194/egusphere-2023-380-RC1
- AC1:
  'Reply on RC1', Haijun Yang, 09 May 2023
  Thank you very much for your invaluable comments and suggestions. We will combine the comments of the two reviewers and focus on revising the following parts of the manuscript:
  (1) In Section 1, aiming at the position of this paper in the established modeling literature, we will rewrite the introduction, comprehensively review the past research on coupled model simulation of AMOC in the mid-Holocene, discuss the insufficiency of previous studies, state the necessity of this research, and explain the mechanism of AMOC changes.
  (2) In Section 2, we will provide a detailed description of our experimental setup, including the spin-up phase of the simulation and the resolution of our model. We will also explain the differences between our simulation and those carried out under the PMIP3 protocol, which is commonly used in paleoclimate modeling studies.
  (3) In Section 3, we will increase the content of the time series of the simulated AMOC in order to better understand the changes in AMOC under different forcings. We will provide stronger evidence to support our conclusions by analyzing the time series data in detail.
  (4) Some figures will be re-plotted to improve clarity.
  Main comments
  Previous studies with multi-model ensembles have already shown that there is no model consensus on the AMOC at the mid-Holocene (Jiang et al., 2023) but also that the changes in AMOC are actually very small between the mid-Holocene and pre-industrial (e.g. figure 9 by Brierley et al 2020). Thus, this study’s focus on AMOC could do with additional justification. This could be straightforward if the model used was significantly more advanced than those used before, but that is not the case. For example, Jiang et al. included CESM2 the successor of the model used here. Also, at T31gx3v7 (3 degrees in the ocean?) the present study is pretty coarse resolution. The changes in AMOC reported are also relatively small and might even fall close to the 0.95-1.05 uncertainty bound in Brierley et al (2020)’s figure 9. Given this context, the significance of the present findings needs to be clarified.
  Responses: Thank you very much for this comment. Recent studies on the AMOC during the mid-Holocene indicate that there was no significant change compared to the pre-industrial era. However, the mechanisms underlying this nearly unchanged AMOC remain unclear. Our paper makes two key contributions that distinguish it from other studies. Firstly, we separated the impact of two different forcings on the AMOC under the PMIP4 protocol. Secondly, we emphasize that changes in seasonality are much stronger and more important than changes in annual mean climate. Our study reveals the competitive relationship between the two forcings, which is a significant advancement. It supports the existing conclusion on the mid-Holocene AMOC and sheds light on the underlying mechanisms for the small differences observed during this period.
  The decision to use a low-resolution version of CESM in this paper was based on two main reasons:
  (1) Improved AMOC simulation: The low-resolution CESM demonstrates a more accurate simulation of the AMOC response in real-world conditions compared to the high-resolution version. Li et al. (2021) (AMOC Stability and Diverging Response to Arctic Sea Ice Decline in Two Climate Models. J. Climate, 34, 5443-5460) found that the low-resolution model exhibits better representation of the Atlantic salinity distribution, AMOC mean strength, and AMOC stability, suggesting it could be more realistic for studying the AMOC response to Arctic sea-ice decline. They stated that “A question arises as to which type of AMOC response would occur in nature: a strong AMOC weakening that leads to a new equilibrium state or a modest transient weakening followed by full recovery? Perhaps it is the former type of response since the low-resolution model is more realistic in that it has a better representation of the Atlantic salinity distribution, the AMOC mean strength and AMOC stability.”
  (2) Limited computing resources: Due to the long simulation times required for the experiments (2000 years each, totaling 6000 years), and the constraints of available computing resources, the low-resolution CESM was chosen as it is more resource-efficient while still addressing the research problem effectively.
  
  Technical corrections:
  Line 69: Probably worth stating that this is ~3.75 degree resolution?
  
  Responses: Thank you very much for this suggestion. we will rewrite the sentence. “The atmospheric model has 26 vertical levels and T31 horizontal resolution (3.75° 3 3.75°). ”
  
  Line 70: What does “gx3v7 horizontal resolution” mean?
  
  Responses: The model uses the T31_gx3v7 grid and the marine module POP2 uses the gx3v7 grid, which has 60 vertical levels, and a uniform 3.6°spacing in the zonal direction. In the meridional direction, the grid is nonuniformly spaced. It is 0.6°near the equator, gradually increasing to the maximum 3.4°at 35N/S and then decreasing poleward.
  
  Line 72: Not sure that Smith & Gregory, 2009 should be here?
  
  Responses: Thank you very much for this suggestion. We have removed this and replaced with Smith et al. (2010).
  Smith, R. D., and Coauthors, 2010: The Parallel Ocean Program (POP) reference manual. Tech. Rep. LAUR-10-01853, Los Alamos National Laboratory, 140 pp.
  
  Lines 74-79: replace “in the X period” with “from the X period”.
  
  Responses: Thank you very much for this suggestion. we have replaced all.
  
  Figure 5: Panel (f) is not described in the caption. Titles above panels d)-f) would improve clarity. Also, timeseries of the AMOC and some quantification of interannual to centennial variability in the model would help to assess the significance of the changes shown here.
  
  Responses: Thank you very much for these suggestions. We will correct errors and add titles above the panels. Timeseries of the AMOC in the three simulations will be added to the manuscript.
  Figure R1 Patterns of mean AMOC in (a) Exp MH, (b) Exp MH_ORB, and (c) Exp PI; and (d) the AMOC change in Exp MH, with respect to Exp PI. (e) and (f) show the AMOC changes due to the ORB effect and GHG effect, respectively. The AMOC index is defined as the maximum streamfunction in the range of 0–2000 m of 20°–70°N in the North Atlantic. Units: Sv.
  
  Figure 6: There are no dashed or dotted lines in the figure? Also I think the vertical scale in panels b and c should be the same.
  
  Responses: Thank you very much for these suggestions. We will re-plot Figure 6. Solid, dashed, and dotted for MH, MH_ORB, and PI, respectively, which are nearly overlapped.
  
  Line 291: This sounds a bit sceptical, you could say “most proxy data” and cite e.g. Larrosoana et al (2012) or similar.
  
  Responses: Sorry, we will rewrite this sentence.
  
  Line 291: the Jiang et al. study is about China so good to specify that here and please correct the brackets.
  
  Responses: Sorry, we will rewrite this sentence.
  
  Citation: https://doi.org/10.5194/egusphere-2023-380-AC1
RC2:
'Comment on egusphere-2023-380', Chris Brierley, 05 May 2023

I find it hard to recommend that this article be published in Climate of the Past. This is not because the science, and especially the simulations, are wrong – I see little incorrect with them. But rather because I do not feel it represents a sufficient addition to the discipline to warrant publication. I consider this work to be more like the level of a (good) dissertation instead of being worthy a peer-reviewed article. The main reason for this arises from a lack of engagement with the existing literature and explanation of this study’s contribution to it.
This work documents a series of simulation with the low-resolution version of CESM1. They consist of a preindustrial control run, a mid-Holocene simulation using the PMIP4 protocol and a mid-Holocene simulation using the PMIP3 protocol (although they are not called this). The large scale features of these simulations are presented in a logical fashion, and the orbital and greenhouse gas impacts are estimated. The main findings relate to compensating forced changes in AMOC resulting in little change under PMIP4 protocols. These are associated with changes in buoyancy flux.
Seen in the context of the existing literature, I have strong doubts that these findings are generalisable. Firstly, the AMOC has large internal variability – this even not acknowledged in the piece, let alone investigated. Secondly, it can take a long time for AMOC to equilibrate – my own simulations using this particular model shown millennial response times (Brierley & Fedorov, 2016). Thirdly, if the findings are valid, then one would anticipate a robust AMOC in the PMIP3 ensemble that disappears in PMIP4. But this is not what is seen (Brierley et al, 2020).
I do not doubt that are sufficient data from the simulations to support research of a publishable level. But it will require (a) substantial new analysis, that is (b) focused on a novel research question and is (c) comprehensively placed in the context of the existing literature.

Citation: https://doi.org/10.5194/egusphere-2023-380-RC2
- AC2:
  'Reply on RC2', Haijun Yang, 09 May 2023
  
  Thank you very much for your invaluable comments and suggestions. We will combine the comments of the two reviewers and focus on revising the following parts of the manuscript:
  (1) In Section 1, aiming at the position of this paper in the established modeling literature, we will rewrite the introduction, comprehensively review the past research on coupled model simulation of AMOC in the mid-Holocene, discuss the insufficiency of previous studies, state the necessity of this research, and explain the mechanism of AMOC changes.
  (2) In Section 2, we will provide a detailed description of our experimental setup, including the spin-up phase of the simulation and the resolution of our model. We will also explain the differences between our simulation and those carried out under the PMIP3 protocol, which is commonly used in paleoclimate modeling studies.
  (3) In Section 3, we will increase the content of the time series of the simulated AMOC in order to better understand the changes in AMOC under different forcings. We will provide stronger evidence to support our conclusions by analyzing the time series data in detail.
  (4) Some figures will be re-plotted to improve clarity.
  “Seen in the context of the existing literature, I have strong doubts that these findings are generalisable. Firstly, the AMOC has large internal variability – this even not acknowledged in the piece, let alone investigated.”
  Responses: Thank you very much for your comments. This manuscript delves into the mechanisms responsible for the subtle differences in AMOC between two equilibrium experiments, while quantifying the effects of various forcings. Figure R2 displays a time series of AMOC across the three experiments. To complement our findings, we will include a discussion on the internal variability of AMOC in the manuscript.
  “Secondly, it can take a long time for AMOC to equilibrate – my own simulations using this particular model shown millennial response times (Brierley & Fedorov, 2016).”
  Responses: Thank you very much for your comments. Our three equilibrium experiments are each conducted over a 2000-year period. Figures R2 display the time series of global mean surface temperature and AMOC. The criteria for an equilibrium state are determined by the global mean surface temperature trend (< ±0.05 °C per century) and a stable AMOC (Zhang et al., 2021). It is evident that our simulation has reached an equilibrium state.
  
  Figure R2 Evolutions of (left panels) annual mean globally mean surface temperature (units: K) and (right panels) the AMOC (units: Sv) in three experiments.
  "Thirdly, if the findings are valid, then one would anticipate a robust AMOC in the PMIP3 ensemble that disappears in PMIP4. But this is not what is seen (Brierley et al, 2020)."
  Responses: Thank you for the advices. First, we would like to emphasize that our experiment does not involve a comparison of AMOC during the mid-Holocene under the PMIP3 and PMIP4 protocols. Instead, we separated the impacts of orbital parameters and greenhouse gases within the PMIP4 framework. Our model sets the solar constant at 1360.75 W/m², while the PMIP3 protocol uses a solar constant of 1365 W/m². Additionally, there are minor differences in greenhouse gas concentrations between the two protocols (Kageyama et al., 2017).
  Second, we totally agree with the notion that the AMOC differences between the two periods under the PMIP3 or PMIP4 frameworks are relatively small. Our research aims to further elucidate the mechanisms behind this phenomenon and quantify the influence of each forcing. This work serves as a supplement to previous research and provides additional support for earlier findings.
  
  Citation: https://doi.org/10.5194/egusphere-2023-380-AC2
  - RC3: 'Reply on AC2', Chris Brierley, 11 May 2023
    
    I would like to thank the authors for outlining so clearly how they would improve this. Even with these improvements, I have strong doubts whether the revised manuscript would warrant publication in Climate of the Past. This is because I still find it hard to see what novel insight would be evidenced and conveyed by it.
    If this manuscript can convincingly demonstrate that the changes in orbital forcing experienced during in the mid-Holocene would drive an enhanced AMOC, then that would be worthy of publication as it overturns the prevailing opinion. However, I find it hard to see that being possible from the experiment that has performed. Firstly, this is because previous work has shown AMOC changes to be rather model dependent (e.g. Jiang et al, 2023, shows a decreasing AMOC in EC-Earth, whilst Otto-Bliesner et al, 2020, shows it increasing in CESM2) – but only a single model is deployed here.
    The new figure clearly demonstrates that the runs are sufficiently spun-up, but also show the relatively large amount of internal variability in AMOC within the simulations. Previous authors have already concluded that changes in AMOC seen in midHolocene simulations can arise from internal variability, rather be a forced response (Williams et al, 2020).
    In my mind, internal variability is a simpler and more plausible explanation for the AMOC behaviour seen across the 3 simulations. This would not require invoking a previously undescribed response to orbital forcing compensated by a response to greenhouse gas forcing that operates in the opposite direction than that seen in future projections and assessed by the IPCC. Any revised manuscript would need to comprehensively disprove this simpler explanation, and I do not see how that could be done.
    Under an experimental setup where one forced response is computed as a residual, you would naturally expect to infer a compensation mechanism if internal variability resulted in the sensitivity experiment having a higher AMOC. Using a more sophisticated experimental design with more simulations (such as Lunt et al, 2021) could more robustly deconvolve the various forced responses.
    References:
    Jiang, Z., Brierley, C., Thornalley, D., and Sax, S.: No changes in overall AMOC strength in interglacial PMIP4 time slices, Clim. Past, 19, 107–121, https://doi.org/10.5194/cp-19-107-2023, 2023.
    Lunt, D. J., Chandan, D., Haywood, A. M., Lunt, G. M., Rougier, J. C., Salzmann, U., Schmidt, G. A., and Valdes, P. J.: Multi-variate factorisation of numerical simulations, Geosci. Model Dev., 14, 4307–4317, https://doi.org/10.5194/gmd-14-4307-2021, 2021.
    Otto‐Bliesner, B.L., Brady, E.C., Tomas, R.A., Albani, S., Bartlein, P.J., Mahowald, N.M., Shafer, S.L., Kluzek, E., Lawrence, P.J., Leguy, G. and Rothstein, M., 2020. A comparison of the CMIP6 midHolocene and lig127k simulations in CESM2. Paleoceanography and Paleoclimatology, 35(11), p.e2020PA003957.
    Williams, C. J. R., Guarino, M.-V., Capron, E., Malmierca-Vallet, I., Singarayer, J. S., Sime, L. C., Lunt, D. J., and Valdes, P. J.: CMIP6/PMIP4 simulations of the mid-Holocene and Last Interglacial using HadGEM3: comparison to the pre-industrial era, previous model versions and proxy data, Clim. Past, 16, 1429–1450, https://doi.org/10.5194/cp-16-1429-2020, 2020
    
    Citation: https://doi.org/10.5194/egusphere-2023-380-RC3
    
    AC3: 'Reply on RC3', Haijun Yang, 11 May 2023
    
    Please refer to Pages 8-11 of the attached pdf file for the second round reply to reviewer #2.
    Thank you very much for invaluable comments and suggestions!
    
    Citation: https://doi.org/10.5194/egusphere-2023-380-AC3
    
    RC4: 'Reply on AC3', Chris Brierley, 12 May 2023
    
    If you were to include a new transient simulation from the mid-Holocene to present in your manuscript, then I would harbour no doubts about whether there is sufficient novelty in the manuscript to warrant publication.
    In fact, we had a paper published in GRL that compiles the AMOC trends from a collection of transient simulations yesterday: https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2023GL103078. If I'd have known about your run, we'd have invited you onboard as a co-author.
    Chris
    
    Citation: https://doi.org/10.5194/egusphere-2023-380-RC4
    
    AC4: 'Reply on RC4', Haijun Yang, 13 May 2023
    
    Hi Chris,
    Thank you very much for the comments and suggestions.
    We will include the transient run that was shown to you in the previous reply, and more analysis will be implemented in the revised manuscript. Your new study in GRL is definitely very helpful.
    Best,
    Haijun
    
    Citation: https://doi.org/10.5194/egusphere-2023-380-AC4

Peer review completion

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

ED: Reconsider after major revisions (24 May 2023) by Marisa Montoya

AR by Haijun Yang on behalf of the Authors (07 Jul 2023) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (19 Jul 2023) by Marisa Montoya

RR by Chris Brierley (08 Aug 2023)

Suggestions for revision or reasons for rejection

Overall, I consider this revised version to be much improved. The addition of the new transient simulations definitely increases the novelty and robustness of the work. I would not be disappointed to see the manuscript published as is, but there are few things that I feel could be improved a little if there was an opportunity.

The midHolocene experiment in previous version of PMIP (particularly PMIP3) involved solely altering the orbital configuration. The manuscript currently does not describe that adequately, that all analysis of these simulations has effectively isolated the ORB component. The solar constant in PMIP3 was specified to be same value used for each model’s preindustrial simulation, although that itself was not a single number across all the models. This is highlighted in Otto-Bliesner et al (2017), as it means that the local insolation changes can vary slightly in magnitude between the various PMIP3 experiments.

I still feel that the issue of internal variability could be more explicitly addressed in the manuscript. However, the inclusion of the new simulations make be confident now that internal variability is not the explanation for the changes seen in the MH experiments. This could be achieved by, for example, adding a sentence or two, including error bars on Fig. 7, or stippling on Fig. 6 with a significance test (based on low frequency variability).

• L15: ‘remarkedly’ should be ‘markedly’
• L79: your space is in the wrong place (before T31 rather than 26)
• L105: The sentence starting “Each transient…” would be better placed earlier in the paragraph.
• Around L240, you could mention comparison with the patterns seen in this experiment, and the ‘warming hole’ seen in observations.
• L278. I do not understand what is meant by [no “-“]. I can see no dashes on Fig 10.
• L314. A space is missing before ‘In’
• L353. Add space into “studyfocuses”
• L357. I believe that the Larrosoana references would be better evidencing the statement that N Africa was wetter in the MH period.

Hide

ED: Publish subject to technical corrections (21 Aug 2023) by Marisa Montoya

AR by Haijun Yang on behalf of the Authors (27 Aug 2023) Author's response Manuscript

Short summary

We simulated the climate difference between the mid-Holocene (MH) and the preindustrial (PI) periods and quantified the effects of Earth orbital parameters (ORBs) and greenhouse gases (GHGs) on the climate difference. We think the insignificant difference in the Atlantic meridional overturning circulation between the MH and PI periods has resulted from the competing effects of the ORBs and the GHGs on the climate.