Cloud properties and their projected changes in CMIP models with low to high climate sensitivity

Bock, Lisa; Lauer, Axel

doi:https://doi.org/10.5194/acp-24-1587-2024

Articles | Volume 24, issue 3

https://doi.org/10.5194/acp-24-1587-2024

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

https://doi.org/10.5194/acp-24-1587-2024

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

Articles | Volume 24, issue 3

Research article

|

05 Feb 2024

Research article |

| 05 Feb 2024

Cloud properties and their projected changes in CMIP models with low to high climate sensitivity

Lisa Bock and Axel Lauer

Download

Final revised paper (published on 05 Feb 2024)
Preprint (discussion started on 12 Jun 2023)

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2023-1086', Anonymous Referee #1, 29 Jun 2023

Review of “Cloud properties and their projected changes in CMIP models with low/medium/high climate sensitivity”

by Lisa Bock and Axel Lauer
This paper presents an intercomparison of the simulation of clouds in the CMIP5 and CMIP6 multi-model ensembles. The models are grouped in three different categories: low, medium and high Effective Climate Sensitivity. In general, high-sensitivity models tend to perform better in the metrics analysed in this study. The paper is well written and the content is adequate for publication in ACP. It provides a valuable intercomparison, making it a useful addition to the scientific literature, and I recommend publication subject to minor revision. Please see my specific commments below.
GENERAL COMMENTS

I believe the results need to be put in the context of other intercomparisons that use different types of metrics. Studies like Brunner at al. (2020) reach very different conclusions by using a metrics that incorporate information about trends. I think this different approach needs to be critically discussed.
A more detailed description of the caveats in the comparisons of the IWP is needed. The model variable used (clivi) includes precipitating frozen hydrometeors only if the precipitating hydrometeor is seen by the model's radiation code. This is model-dependent and can introduce significant biases in the comparisons. Also, I wonder if the observational datasets chosen are representative of the diversity in observational estimates. Both ESACCI and MODIS are based on passive retrievals, and therefore will share similar caveats and biases (Waliser et al., 2009). I'd suggest using an alternative reference datased based on a different remote sensing technology like CloudSat.

SPECIFIC COMMENTS

-L41-42: there are other studies that looked into the reasons for the increased in sensitivity in specific models, like Gettelman et al. (2019) and Bodas-Salcedo et al. (2019). It's worth noting that coupled feedbacks (e.g. sea-ice albedo) can play a significant role in some models (Andrews et al., 2019).
- Table 2. Please specify which CERES-EBAF version you've used. Also, the reference for ERA5 is missing.

REFERENCES

Andrews, T., Andrews, M. B., Bodas-Salcedo, A., Jones, G. S., Kuhlbrodt, T., Manners, J., et al. (2019). Forcings, feedbacks, and climate sensitivity in HadGEM3-GC3.1 and UKESM1. Journal of Advances in Modeling Earth Systems, 11, 4377– 4394. https://doi.org/10.1029/2019MS001866.
Bodas-Salcedo, A., Mulcahy, J. P., Andrews, T., Williams, K. D., Ringer, M. A., Field, P. R., & Elsaesser, G. S. (2019). Strong dependence of atmospheric feedbacks on mixed-phase microphysics and aerosol-cloud interactions in HadGEM3. Journal of Advances in Modeling Earth Systems, 11, 1735– 1758. https://doi.org/10.1029/2019MS001688.
Brunner, L., Pendergrass, A. G., Lehner, F., Merrifield, A. L., Lorenz, R., and Knutti, R.: Reduced global warming from CMIP6 projections when weighting models by performance and independence, Earth Syst. Dynam., 11, 995–1012, https://doi.org/10.5194/esd-11-995-2020, 2020.
Gettelman, A., Hannay, C., Bacmeister, J. T., Neale, R. B., Pendergrass, A. G., Danabasoglu, G., et al. (2019). High climate sensitivity in the Community Earth System Model Version 2 (CESM2). Geophysical Research Letters, 46, 8329– 8337. https://doi.org/10.1029/2019GL083978.
Waliser, D., et al. (2009), Cloud ice: A climate model challenge with signs and expectations of progress, J. Geophys. Res., 114, D00A21, doi:10.1029/2008JD010015.

Citation: https://doi.org/10.5194/egusphere-2023-1086-RC1
- AC1: 'Reply on RC1', Lisa Bock, 29 Sep 2023
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1086/egusphere-2023-1086-AC1-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-1086-AC1
RC2:
'Comment on egusphere-2023-1086', Anonymous Referee #2, 18 Aug 2023

The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1086/egusphere-2023-1086-RC2-supplement.pdf

Citation: https://doi.org/10.5194/egusphere-2023-1086-RC2
- AC2: 'Reply on RC2', Lisa Bock, 29 Sep 2023
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1086/egusphere-2023-1086-AC2-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-1086-AC2

Peer review completion

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

AR by Lisa Bock on behalf of the Authors (29 Sep 2023) Author's tracked changes Manuscript

EF by Polina Shvedko (04 Oct 2023) Author's response

ED: Referee Nomination & Report Request started (06 Oct 2023) by Paulo Ceppi

RR by Anonymous Referee #3 (26 Oct 2023)

RR by Anonymous Referee #2 (30 Oct 2023)

Suggestions for revision or reasons for rejection

In response to my concerns, the authors have added caveats to the paper that the model evaluation of cloud properties is qualitative rather than quantitative. However, they largely leave Section 3.2 (Evaluation of Cloud Properties) unchanged. In this section, numerical values of regional averages are compared between models and observations, global mean values are compared between models and observations, root mean square differences between models and observations are computed, and spatial pattern correlations between models and observations are computed, in all cases for the three groups of models. Hence Section 3.2 remains almost entirely a quantitative evaluation of modeled cloud properties against observations.

The authors acknowledge “This study is therefore not possible with the available CMIP model output generated by satellite simulators.” I agree that rigorously evaluating models against observations is not possible across the whole suite of CMIP5 and CMIP6 models as attempted here because most models lack COSP output that makes for a meaningful and reliable comparison. This is why I originally raised this concern, and I remain confused as to why the authors continued down this path. Moreover, I disagree that one can “qualitatively” evaluate models against observations while relying on comparing fundamentally different quantities. This could be misleading at best and simply wrong at worst. Therefore I don’t believe the authors can confidently say make statements like “we found that models with a high climate sensitivity typically have a better representation of observed cloud properties than models with a low or medium ECS.” I recommend removing Section 3.2 entirely, or focusing only on fields for which model-observations differences in the definition of the field can be minimized.

My concern about comparing models and observations from different time periods remains, but is secondary to the more fundamental concern above. Arguably the least ambiguous way of comparing models and observations would be to use COSP output from atmosphere-only AMIP simulations in which the observed radiative forcings, SSTs, and sea ice concentrations are prescribed, and choosing the period of perfect temporal overlap between models and observations.

I do not have a problem with Sections 3.1 or 3.3, though I am still not sure we learn much from these sections that are not already well established. If Section 3.2 were removed or trimmed to focus only on fields that can rigorously be compared between models and observations, I could support publication of this paper.

Hide

ED: Reconsider after major revisions (07 Nov 2023) by Paulo Ceppi

AR by Lisa Bock on behalf of the Authors (23 Nov 2023) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (23 Nov 2023) by Paulo Ceppi

RR by Anonymous Referee #3 (06 Dec 2023)

Suggestions for revision or reasons for rejection

Review of “Cloud properties and their projected changes in CMIP models with
low/medium/high climate sensitivity” by L. Bock and A. Lauer.

Round 2.
The authors have taken care of most of my concerns in this new version. Some minor comments remain, but should be fairly easy to address. I’m listing them below

A follow up comment to one of the authors’ response:
“The increase in ECS between CMIP generations opened up a highly discussed topic about possible reasons and how realistic this is. This motivated us to look in more detail into the models. But not all CMIP6 models have higher ECS values than the CMIP5 models, there are also CMIP5 models in the high ECS group. As we were interested in whether there are systematic differences in projected cloud properties among the models contributing to differences in ECS, we sorted the CMIP models in three ECS groups. Most of the present-day evaluation with satellite data has been removed (see detailed comments above). The focus is now on projected changes in cloud properties in the models and the differences in present-day fields among the model groups that are a starting point for the following discussion of the projections.”
I understand your point of view now that you have elaborated. However, you didn’t change the conclusion opening and the idea that you convey in the above statement doesn’t really appear in this opening.
The direct link between change in cloud properties with respect to climate change may have contributed to ECS is not done in the conclusion. I would recommend implementing some form of this sentence in it “As we were interested in whether there are systematic differences in projected cloud properties among the models contributing to differences in ECS, we sorted the CMIP models in three ECS groups.”

Other minor comments (line numbers of the highlighted version):
L55-56: Not sure this sentence still makes sense now that the comparison with observations has been removed.

L60: I would remove the comma after both.

Cloud feedback computations: The method used to compute cloud feedbacks works well for the SW component between 60˚S/N, but it is not accurate in the polar regions and everywhere for the LW component (Shell et al., 2008). Therefore, it should be clearly mentioned that this cloud feedbacks are not adjusted for the effect of non-cloud feedbacks.

L264-269: I don’t quite follow the logic. The high ECS models produce larger IWP values over the SO and more clouds, therefore it proves that more realistic supercooled clouds lead to less negative cloud phase feedback? If anything, I would think the opposite. Higher amount of ice in the high-ECS model would suggest a potential for more negative cloud phase feedback, since the ice reservoir to be turn into liquid clouds in response to warming is larger than in the other model groups.

L275-280: Interesting to see that there is no increase in LWP over the Sc decks although there is an increase in their cloud fraction as shown by the authors, which seems to suggest that the clouds of the group are probably less bright (also consistent with Cesana et al., 2023).

L308-312: Could you please quantify and share the results in the supplementary?

L450-452: “in the high ECS group”, not the low, unless I don’t understand Fig. 3.

References
Cesana, G. v., Ackerman, A. S., Črnivec, N., Pincus, R., & Chepfer, H. (2023). An observation-based method to assess tropical stratocumulus and shallow cumulus clouds and feedbacks in CMIP6 and CMIP5 models. Environmental Research Communications, 5(4). https://doi.org/10.1088/2515-7620/acc78a
Shell, K. M., Kiehl, J. T., & Shields, C. A. (2008). Using the radiative kernel technique to calculate climate feedbacks in NCAR’s Community Atmospheric Model. Journal of Climate, 21(10), 2269–2282. https://doi.org/10.1175/2007JCLI2044.1

Hide

RR by Anonymous Referee #2 (16 Dec 2023)

Suggestions for revision or reasons for rejection

The authors have substantially revised the paper and I am largely satisfied with the changes. I have a few remaining comments:

L114-116: This does not make sense. It is exactly because coupled models simulate their own distinct transient evolution that a given time period will be characterized by a climate state different from each other and from observations. To do the most apples-to-apples comparison of cloud properties between model groupings and between models and observations, one should use AMIP simulations from the overlapping period. That way, differences are unambiguously attributable to model physics, not the state of the atmosphere (notably, SST pattern differences and aerosol loading differences). I am not asking you to do this, but I think this statement should be rephrased for accuracy.

L 167: Consider citing Zelinka et al (2022) here, as they recently investigated this connection.
https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2021JD035198

L169-171: Here it is shown that high ECS models tend to have higher IWP in the midlatitudes. This is stated to be consistent with the hypothesis that improved representation of supercooled liquid leads to higher ECS. This statement is not right on two levels. Historically GCMs have had too much ice and not enough liquid, so higher IWP would typically not be considered as an improvement. Secondly, models with larger mean-state ice tend to have a larger negative phase feedback as that ice transitions to liquid with warming. In isolation, this would mean high IWP models would tend to have lower ECS.

L214-215: Is it really necessary to reproduce this statement via direct quote?

L343-345: What does this statement have to do with the results shown here? There is no evaluation of whether the models over- or under-estimate cloud feedback in stratocumulus regions. Moreover, inter-model differences in cloud feedback strength is governed by inter-model differences in representation of cloud physics, not in pattern effects (especially in response to a large warming signal like analyzed here). Suggest deleting this statement.

Hide

ED: Publish subject to minor revisions (review by editor) (18 Dec 2023) by Paulo Ceppi

AR by Lisa Bock on behalf of the Authors (20 Dec 2023) Author's response Author's tracked changes Manuscript

ED: Publish as is (22 Dec 2023) by Paulo Ceppi

AR by Lisa Bock on behalf of the Authors (22 Dec 2023) Manuscript

Short summary

Climate model simulations still show a large range of effective climate sensitivity (ECS) with high uncertainties. An important contribution to ECS is cloud climate feedback. We investigate the representation of cloud physical and radiative properties from Coupled Model Intercomparison Project models grouped by ECS. We compare the simulated cloud properties of today’s climate from three ECS groups and quantify how the projected changes in cloud properties and cloud radiative effects differ.