the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 03 Feb 2025
The importance of stratocumulus clouds for projected warming patterns and circulation changes
Abstract. Stratocumulus clouds are thought to exert a strong positive radiative feedback on climate change, but recent analyses suggest this feedback is widely under-represented in global climate models. To assess the broader implications of this model error for the modeled climate change responses, we investigate the impact of Pacific stratocumulus cloud feedback on projected warming patterns, equilibrium climate sensitivity and the tropical atmospheric circulation under increased CO2 concentrations. Using the Community Earth System Model with modifications to enhance low cloud cover sensitivity to sea surface temperature (SST) anomalies in Pacific stratocumulus regions, we find increased tropical SST variability and persistence, a higher equilibrium climate sensitivity, an enhanced east–west warming contrast across the tropical Pacific, and a stronger slow-down of the Walker circulation under 4×CO2 conditions. Our findings are supported by inter-model relationships across CMIP6 4×CO2 simulations. These results underscore the importance of accurately representing cloud feedback in climate models to predict future climate change impacts not only globally, but also on a regional scale, such as warming patterns or circulation change.
Competing interests: Paulo Ceppi is a member of the editorial board of ACP.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.- Preprint
(3001 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-221', Anonymous Referee #1, 25 Feb 2025
This study performs cloud feedback experiments using the fully coupled CESM2.1.3 model. By enhancing the sensitivity of low clouds to SST perturbations in the eastern Pacific subsidence regions and comparing them to control simulations, the authors isolate the effects of local low cloud-SST feedbacks in simulations of climate variability and change. The main findings are that the enhanced regional low cloud feedback strength results in: (i) increased SST variability in the eastern tropical and subtropical Pacific, (ii) slightly higher climate sensitivity, and (iii) a weakened east-west tropical Pacific SST gradient and Walker Circulation under 4xCO2.
The study is of high scientific quality, and the paper is generally well written. The most novel and impactful findings relate to tropical SST pattern and atmospheric circulation changes in the future climate, particularly the significant weakening of the Walker Circulation (finding iii). Findings (i) and (ii), while interesting, align with prior studies (e.g., Bellomo et al. 2014, 2015; Brown et al. 2016; Burgman et al. 2017; Loeb et al. 2018; Middlemas et al. 2019; Miyamoto et al. 2023; Myers et al. 2018a,b, 2021; Yang et al. 2023; Zhu et al. 2020). Given this, the authors could strengthen the paper by further contextualizing these results within previous work and clarifying how their experimental setup offers new insights. One distinguishing feature is the separation of fast and slow responses, which the authors might emphasize more. Additionally, a deeper analysis of finding (iii) would be valuable, given the uncertainty surrounding future tropical SST patterns and circulation changes. Implementing these suggestions, along with the specific points below, would improve the paper’s clarity and impact.
Specific Comments:
-
Introduction: The claim that cloud feedback has been studied primarily in terms of its impact on global-mean temperature change (lines 15–17) is somewhat misleading. While global implications have been extensively analyzed, many studies have also investigated regional climate impacts, particularly in the context of internal variability. A broader discussion of previous work in this area would provide better context.
-
Data and Methods: While the low cloud sensitivity to SST anomalies is computed following Ceppi’s approach, additional explanation in the paper would improve clarity. Moreover, since a low cloud cover anomaly proportional to the instantaneous SST anomaly is applied at every radiation time step, this will likely influence sensitivities to other cloud-controlling factors correlated with SST, as seen in Fig. A2a. While these changes appear minor, explicitly noting this effect in the paper would be beneficial.
-
Section 3.2: The paragraph beginning on line 113 discussing cloud feedbacks is somewhat unclear. Explicitly quantifying cloud feedback values in different experiments would allow for more precise comparisons. Additionally, the CRE anomalies with temperature in Fig. A4b are difficult to interpret. Given that CESM2 has a large positive cloud feedback (as quantified by Zelinka et al. 2020 and others), shouldn’t the dCRE/dT slopes be positive overall, not just for the slow responses? Including spatial maps of cloud feedback and low cloud changes would greatly enhance the analysis, making it easier to interpret the influence of enhanced cloud-SST sensitivity on future climate changes.
-
Walker Circulation Slowdown: The amplified Walker Circulation weakening in the 4xCO2 experiments is particularly interesting and warrants further investigation. Why is the change so large? Beyond the enhanced warming in the eastern tropical Pacific, could a reduction in LW radiative cooling at the cloud tops (as stratocumulus clouds decrease) contribute to the decreased SLP in that region? A deeper exploration of this and other possible mechanisms would strengthen the analysis.
-
Section 3.3 - Cloud Feedback Decomposition: The proposed decomposition is an interesting approach, but it lacks a quantification of which cloud-controlling factors drive future cloud changes. Why was this not included? In equation (7), what is the relative contribution of changes in cloud-controlling factors other than SST to the pattern-mediated response? Additionally, equation (4) may need adjustment: since the additional 3% reduction in low cloud cover per unit SST increase modifies dC/dYi for other factors correlated with SST, equation (4) should instead use (dC/dYi)mod. Writing this as dC/dYi + δi, where δi represents the difference between modified and original sensitivities, would ensure accuracy in the derivations.
Specific Questions and Technical Notes:
a) Lines 43-44: Provide a reference or additional justification for the statement regarding CESM sensitivities.
b) Lines 124-126: The discussion on remote "pattern effects" could be clarified with more explicit details.
c) Line 175: Explain explicitly how this quantity is obtained as an output from the model runs.
d) Lines 189-190: Are these ratios to be interpreted as the fraction of the total response driven purely by pattern-mediated changes? Please provide an explanation of the 0.37 and 0.16 values provided.
e) Lines 216-217: The difference in bias between low cloud cover and CRE sensitivities might also have a contribution from low cloud optical depth. A discussion of this possibility would improve interpretation.
f) Line 229: Do the authors mean strong cooling in those outlier models? There could be a sign error here.
Conclusion:
Overall, this paper presents a well-executed set of experiments that offer valuable insights into cloud feedbacks and climate variability. To strengthen the impact of their findings, the authors should emphasize the novel aspects of their approach (e.g., separation of fast and slow responses) and provide additional discussion on key mechanisms, particularly for Walker Circulation weakening. Clarifying the cloud feedback decomposition and further contextualizing findings (i) and (ii) within existing literature will make the study more compelling. With these revisions, the paper would be a strong contribution to the field.
References (not already cited)
Bellomo, K., Clement, A. C., Mauritsen, T., Rädel, G., & Stevens, B. (2015). The influence of cloud feedbacks on equatorial Atlantic variability. Journal of Climate, 28(7), 2725-2744.
Burgman, R. J., Kirtman, B. P., Clement, A. C., & Vazquez, H. (2017). Model evidence for low‐level cloud feedback driving persistent changes in atmospheric circulation and regional hydroclimate. Geophysical Research Letters, 44(1), 428-437.
Loeb, N. G., Thorsen, T. J., Norris, J. R., Wang, H., & Su, W. (2018). Changes in Earth’s energy budget during and after the “pause” in global warming: An observational perspective. Climate, 6(3), 62.
Middlemas, E., Clement, A., & Medeiros, B. (2019). Contributions of atmospheric and oceanic feedbacks to subtropical northeastern sea surface temperature variability. Climate Dynamics, 53(11), 6877-6890.
Miyamoto, A., Nakamura, H., Xie, S. P., Miyasaka, T., & Kosaka, Y. (2023). Radiative impacts of Californian marine low clouds on North Pacific climate in a global climate model. Journal of Climate, 36(24), 8443-8459.
Myers, T. A., Mechoso, C. R., & DeFlorio, M. J. (2018a). Coupling between marine boundary layer clouds and summer-to-summer sea surface temperature variability over the North Atlantic and Pacific. Climate Dynamics, 50, 955-969.
Myers, T. A., Mechoso, C. R., Cesana, G. V., DeFlorio, M. J., & Waliser, D. E. (2018b). Cloud feedback key to marine heatwave off Baja California. Geophysical Research Letters, 45(9), 4345-4352.
Yang, L., Xie, S. P., Shen, S. S., Liu, J. W., & Hwang, Y. T. (2023). Low cloud–SST feedback over the subtropical northeast Pacific and the remote effect on ENSO variability. Journal of Climate, 36(2), 441-452.
Zhu, J., & Poulsen, C. J. (2020). On the increase of climate sensitivity and cloud feedback with warming in the community atmosphere models. Geophysical Research Letters, 47(18), e2020GL089143.
Citation: https://doi.org/10.5194/egusphere-2025-221-RC1 -
-
RC2: 'Comment on egusphere-2025-221', Anonymous Referee #2, 04 Apr 2025
Breul et al. implemented a specialized technique to regionally "fix" the representation of low clouds, enhancing the sensitivity of low cloud cover to SST anomalies in Pacific stratocumulus regions within a CESM model. To my knowledge, a similar approach has been used to investigate the influence of clouds on decadal variability in a slab ocean setting (Bellomo et al., 2014), but not in the context of global warming scenarios. I find this study both relevant and compelling. Combined with additional analysis of CMIP models, it offers timely and insightful findings that improve our understanding of how cloud biases affect climate model projections.Including more information on the background and being specific about the novel findings may help some readers. Some suggestions below:1. Expanded introduction and comparison to previous studies:
The introduction is quite brief. It would benefit from a more in-depth discussion of how clouds influence both SST patterns and climate sensitivity (e.g., Fu and Fedorov 2023; Chalmers et al. 2022). In particular, these earlier studies implement a global cloud-locking technique, which differs significantly from the regional bia correction approach used in the present study. This methodological difference should be explicitly highlighted and discussed in both the Introduction and Discussion sections. This may include additional mechanistic understanding of the potential of observational constraints (since you are "fixing the biases".)2. Emphasis on local vs. global cloud locking and nonlocal SST responses:
Given that a key distinction between this study and existing literature lies in the use of local rather than global cloud locking, the study’s findings on the remote influence of subtropical cloud feedbacks—particularly on the equatorial SST pattern—should be emphasized. This could include additional analysis or discussion explaining the nonlocal effects, potentially drawing on existing literature that addresses atmospheric teleconnections and the propagation of regional perturbations.3. Fast vs. slow response mechanisms:
The distinction between fast and slow responses is intriguing. Do you think similar mechanisms govern the cloud–circulation–SST coupling across both timescales? Alternatively, would you expect additional oceanic processes—such as subsurface adjustment or ocean heat uptake—to become more relevant during the slow response? Clarifying these points could further strengthen the interpretation of the results.Other minor suggestion:I'm a bit concerned about the quantification of the clouds' influence on the fast response with few ensemble members. The regression slopes do not well represent all points in some figures. Consider adding more ensemble members or remove the quantitative statements as in Line 143.Some literature on clouds and inter-annual variability:Rädel, Gaby, et al. "Amplification of El Niño by cloud longwave coupling to atmospheric circulation." Nature Geoscience 9.2 (2016): 106-110.
Middlemas, Eleanor A., et al. "Cloud radiative feedbacks and El Niño–southern oscillation." Journal of Climate 32.15 (2019): 4661-4680.Some literature on clouds' effect under global warming:Fu, Minmin, and Alexey Fedorov. "The role of Bjerknes and shortwave feedbacks in the tropical Pacific SST response to global warming." Geophysical Research Letters 50.19 (2023): e2023GL105061.Chalmers, Jason, et al. "Does disabling cloud radiative feedbacks change spatial patterns of surface greenhouse warming and cooling?." Journal of Climate 35.6 (2022): 1787-1807.Some literature on clouds and remote SST pattern:Bellomo, Katinka, et al. "Simulating the role of subtropical stratocumulus clouds in driving Pacific climate variability." Journal of climate 27.13 (2014): 5119-5131.Kim, Hanjun, et al. "Subtropical clouds key to Southern Ocean teleconnections to the tropical Pacific." Proceedings of the National Academy of Sciences 119.34 (2022): e2200514119.Hsiao, Wei‐Ting, et al. "The role of clouds in shaping tropical Pacific response pattern to extratropical thermal forcing." Geophysical Research Letters 49.11 (2022): e2022GL098023.Citation: https://doi.org/10.5194/egusphere-2025-221-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 126 | 106 | 9 | 241 | 5 | 4 |
- HTML: 126
- PDF: 106
- XML: 9
- Total: 241
- BibTeX: 5
- EndNote: 4
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 112 | 50 |
| Germany | 2 | 16 | 7 |
| United Kingdom | 3 | 13 | 5 |
| Ireland | 4 | 10 | 4 |
| Italy | 5 | 8 | 3 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 112