[our replies to reviewer comments are included in bold below]

This paper examines 4 LEs for many metrics regarding the forced response. It is well written, the science is sound and I recommend for publication after minor revisions.

Thanks for the time spent reviewing our work and for your constructive comments.

Minor comments:

Abstract - FR is confusing in the abstract - I recommend spelling it out in this section.

Agreed – it is now spelled out.

45/46 - needs citations. Here are possibilities: https://esd.copernicus.org/articles/12/401/2021/ OR https://www.nature.com/articles/s41558-020-0731-2.epdf?author_access_token=OnhiIlhZkm3p58t3OHjV3tRgN0jAjWel9jnR3ZoTv0NnBpoO0B2TlXC-0PoMuTZLSw6_YFL5sMa_yz8wJSmNsSAiVurA3amG421TAO0symp2l6aSmTSddEE13z7RMtzL7vdbK7d6lErsCSwQhQgY9Q%3D%3D

Agreed. Thanks for the suggestions. These are important citations to include and we have added a few others.

Introduction - this is very short. It could include another paragraph on studies looking at FR in LEs - there are quite a few of these that could easily be discussed

Agreed. We have now expanded the introduction to include mention of various references above to illustrate the diverse contexts in which the FR has been examined using LEs.

92 CESM2 is not yet introduced

Good catch. We’ve moved this statement to the CESM2 section.

Section 2.5.1 and 2.5.2 what is the observational uncertainty? This should be stated and/or included in the figures

By including multiple temperature datasets in Fig. 2, which are characteristic generally of inter-dataset disagreements, we have made an effort to address observational uncertainties. Given that the CERES values are already a challenge to see in Fig 3, due to their brevity, and depend on context (absolute values vs drift), we’d prefer to mention these uncertainties in the data section rather than add additional time series to Fig. 2.

Table 1 - in the title can you explain what is in the table please? It is unclear what each metric means

Good point, as these metrics are likely to be unfamiliar to most readers. Text has been added to provide a general overview of the metrics and their significance.

192 - how long a record of CERES would be needed to resolve this

It is not quite clear what needs to be “resolved” here but perhaps the reviewer is suggesting that this statement should be expanded to account for observational uncertainty.  We have expanded it to address this.  

314 though should be although

Thanks, changed.

328 is 1 a maximum score? There is not enough detail for the reader to interpret this without reading the cited manuscript

Agreed that more detail is needed. Context has been added to this section to make it clearer to readers that are unfamiliar with the package.

Conclusions - lastly is there twice

Good catch. Revised.

The conclusions should be explaned to talk about future ideas as per the EGUSphere guidelines. Some of this could be as simple as expanding the analysis to all avaiable LEs

 Text has been added to provide this perspective on future efforts.

[Our responses indicated in bold below.]

The authors described the E3SM2 LE and compared it with the other 3 LEs. This paper is well written and would be useful for users of the output data of these LEs. Although the excessive aerosol cooling in the 20th century is unfortunate, I hope that the authors would fix this problem in a future work. I recommend publication after a minor revision.

Thank you for your effort reviewing our work and for your constructive comments. Indeed, a successive version of E3SM that reduces biases in simulated aerosol-cloud interactions is planned, along with a large ensemble based on it. We now mention this in the conclusions.

Minor comments.

*Fig 3c: Why do the top-of-model net radiation flux in the Southern Hemisphere has negative trends in the 21st century for E3SM1 and E3SM2? 

Good question. We now highlight these unexpected trends and include an expanded discussion of the potential drivers. We find that both E3SM1 and 2 have such flat trends in the 21st Century in R_T for different reasons. In E3SM1 the R_T trends are flat because OLR and temperature trends in the SH are stronger than in the other ensembles (and thus offset albedo reductions). In v2 the flat trend is because shortwave (SW_TOA) trends are weaker than in the other ensembles and are thus fully offset by OLR trends, which are in line with the other ensembles. This in turn contributes to a slower rate of warming in the SH.  

*It is better for readers to note whether positive values of MHTatm and MHTocn are northward or southward.

Good point. Now defined.