Mülmenstädt et al. study how changes in cloud-condensation nuclei (CCN) aerosols lead to adjustments in the liquid water path (LWP) of warm liquid clouds. The two prevailing hypotheses are that more CCN lead to stronger entrainment drying and precipitation suppression, which act to decrease and increase the LWP, respectively. The authors perform global-climate model (GCM) and single-column model experiments and analyze boundary layer heat and moisture budgets to identify causal relationships related to the proposed mechanisms. The results indicate that both proposed mechanisms are active in GCMs, but the enhanced-entrainment mechanism has a negligible effect on global-mean LWP. The authors conclude by interpreting these results and posing guiding questions for future research.

            I believe that this topic is highly relevant to the aerosol-cloud-climate community, the analysis is well done, and the paper is clearly written. I have very little to say regarding criticisms of the current manuscript, but I offer a few ideas for additional analysis and discussion that I think could improve the paper. If the authors deem that these additional pieces would be beyond the scope of the study, then I believe the paper would also be publishable without them. I therefore recommend minor revision.

General Comments

• The authors conclude with a substantial discussion section in which they pose six guiding questions for future research. This is helpful for the community to think about next steps. However, it would be even more helpful if the authors could explicitly connect the questions to concrete examples from their analysis. For example, one guiding question is “What complexity is required (to simulate the global LWP adjustment)?” The authors suggest looking for the minimal set of parameterizations that capture relevant process understanding. Can the authors perform single-column experiments with a range of complexity to give some guidance about what this minimal set of parameterizations might look like in practice? The authors also pose the question “how representative are susceptibilities in small ensembles of individual cases?” Can the authors identify any cases with the single-column model in which the LWP adjustment differs substantially from the canonical LES cases that are widely studied? Where do we need to look to find this differing behavior? If the authors can provide specific, concrete examples from their analysis to motivate the six questions in the discussion, then I think the discussion would be more useful to the community.
• If the enhanced-entrainment mechanism is in fact negligible for the global-mean LWP adjustment, as suggested by the results in the study, then what are the implications for the historical effective radiative forcing from aerosol-cloud interactions (ERFaci)? The enhanced-entrainment mechanism was used to justify a positive radiative adjustment from LWP changes in the Bellouin et al. (2020). If this positive radiative LWP adjustment is in fact negligible, then doesn’t that imply an even stronger negative ERFaci? There is already a tension between ERFaci estimates derived from process understanding and aerosol-cloud relationships (“bottom-up estimates”) and ERFaci estimates from global energy-budget constraints (“top-down estimates”), with the former predicting a stronger negative ERFaci. Do the current findings exacerbate this tension? How do we interpret this, and how do we more forward? Given that the paper concludes with a substantial forward-looking discussion, I was hoping that the authors would have discussed this topic.

Specific Comments

• Line 44: I suggest changing “cloud the…” to “complicate the…” or “contradict the…” to avoid confusion because the noun form of “cloud” is used often in the preceding text.
• Line 122: “The idealizations active in the baseline experiment are:” A colon should not be used here because colons should follow complete sentences, not sentence fragments (apologies for my obsession with grammar)
• Line 134: “top-of-atmosphere flux” -> “top-of-atmosphere radiative flux”
• Line 368: consider changing “is small enough to require far longer model runs” to “is small enough to require far longer model runs to detect”

Please find my review attached below.

Please find my review attached.

Summary

This manuscript is a timely investigation of the ability of GCMs and their underlying single-column models to qualitatively capture adjustments of liquid water path in stratocumulus clouds (Sc) to an increase in aerosol, relative to the behavior of liquid water path in response to aerosol changes known from large eddy simulations (LES). The question is addressed whether GCMs are incorrect due to parametric or base-state representation errors,  or whether relatively few canonical cases simulated with LES are sufficiently representative to serve as a benchmark for the behavior of Sc liquid water path in response to aerosol changes as simulated by GCMs globally.

The first interesting insight provided by this work is that SCMs are qualitatively able to represent, via the mechanisms known from LES (precipitation suppression and entrainment drying), the liquid water path adjustments.

The second interesting insight provided is that GCMs can produce different liquid water path adjustments relative to canonical LES cases, possibly for the right reasons such as environmental variability, and possibly due to limitations such as limited model resolution, uncertainty in process representations, and uncertainty in the representation of the atmospheric base state.

A set of approaches to improve the ability of GCMs to represent the liquid water path adjustment to aerosol changes is presented and discussed, a valuable contribution to advance the science.

The section explaining the entrainment diagnostics is very nicely written. Other parts of the manuscript would benefit greatly from a similarly linear and didactic writing style, and certainly from much shorter sentences.

This reviewer recommends this manuscript for publication after minor revisions.

Specific Points

Title: "Can GCMs represent cloud adjustments to aerosol–cloud interactions?"

The title is a bit misleading in that it overstates the scope of this work. After all, only the adjustment of liquid water path to aerosol-cloud interactions is investigated, and for a very limited subset of stratocumulus clouds.

Please change the title to something more representative, e.g.,

"Can GCMs represent the liquid water path adjustment to aerosol–cloud interactions?"

Section 2.1: Please specify that E3SM and ModelE3 are both used in SCM mode, but only E3SM is used in GCM mode.

Line 260: "We may be mitigating the E3SM artifact by averaging over two full deepening cycles, effectively averaging over the dependence of E on position in the deepening cycle. We attempt to mitigate the ModelE3 artifact by only averaging E until the first PBL deepening occurs."

Please justify in the text that this subsetting is not cherry-picking the results that produce a particular outcome.

Lines 277 to 281: "Whether the effect of varying the aerosol concentration on L is expected depends on our Bayesian prior. If our expectation for the L response is based on the RA L results of Mülmenstädt et al. (2024), we would predict the causal effect of increased Nd to be an increase in L."

The results of Mülmenstädt et al. (2024) represent the behavior of a cloud population simulated in that work, whereas the clouds simulated with SCMs in this work represent one particular set of conditions. Why is it valid to formulate an expectation based on the former and expect it to hold for the latter?

"If our expectation is based on the LES-based process understanding, then we would predict the causal effect of Nd to be a decrease in L." ... "The surprising result is that the SCM sides with the LES, not the response of the 3D GCM with which the SCM shares its model physics."

Again, why is this surprising, given that the 3D GCM simulates a population of clouds whereas the SCM and LES runs simulate only one particular case? It is perfectly possible that the population of clouds behaves differently from the particular SCM case.

Line 281: Please add a reference to Appendix A.

Line 332: "The L values at which entrainment turns on are in reasonable agreement with recent LES (Hoffmann et al., 2020) and observational (Zhang et al., 2022) results."

Too unspecific. Please detail how the agreement is reasonable with Hoffmann et al. (2020) and Zhang et al. (2022).

Line 336: "The main conclusion from these plots is that the model produces greater entrainment in response to higher Nd in Sc clouds with strong entrainment."

The response of entrainment to Nd is not shown as a function of entrainment. Please explain how one can draw the given conclusion.

Lines 348-356: This paragraph gives the impression of a stream of consciousness reflecting the authors' familiarity with their work. It is presented without references to figures or other supporting material, and invoking not-shown results, all of which makes it very hard to untangle. It is not obvious that it is needed, but if it is, it requires rewriting.

Line 392: The evidence for this causal relationship comes from SCM studies, where, like Guo et al. (2011), we find that increased aerosol, while holding all other boundary conditions fixed, leads to liquid-water loss.

This is not true for all Nd values, is it? Please qualify this finding accordingly.