I recommend this paper be accepted, subject to addressing the following concerns:

Yes, the paper shows that the emulator can produce output for some variable that are linearly related to forcing.  But I think the entire project needs to be framed more clearly.  Some of these points are addressed in the discussion, but they need to be added to the abstract.  What can the emulator not do?  What is its purpose?  Because it does not include a seasonal cycle in the resulting maps of temperature and precipitation, it cannot be used for many potential impacts of climate change that matter, such as food production and water resources.  It cannot simulate monsoon impacts.  It cannot simulate the diurnal cycle, which is important for many impacts, including agriculture and tropospheric ozone.  It does not include downward diffuse vs. direct radiation, or UV.  And because the two GCMs used as examples differ quite a bit in some aspects, any actual use of the emulator would require multiple climate models so as to add a probability distribution to the resulting emulated climate. 

Will it be possible to also emulate impacts of SRM in addition to the standard climate variables of temperature, precipitation, and evaporation?  What would it take?  And can the authors add warnings about the usage of emulators prominently in their abstract and introduction?

The way the emulator works is to include CO₂-equivalent and SO₂ time series of forcing, but no mention is made of tropospheric aerosols, land use, and other forcings included in SSPs.  How are they handled?

Please also address the 40 comments in the attached annotated manuscript.

Technical issues:

Cider is not brewed.  It is fermented.  So the title to section 2 needs to be changed.  See the comments in the attached annotated manuscript.

Your model nomenclature is confusing.  You have CESM2(WACCM), CESM2-WACCM6, and CESM2.  Are these all the same?  Then use the same notation.  The same for UKESM1 and UKESM1.0.

There are several acronyms that are not defined.

You use “validation” multiple times, when you mean “evaluation.”  “Validation” means that you are proving that the model works, that it is valid.

Multiple references in the text do not use parentheses around the years.

Farley et al. present in their manuscript CIDER, which is a climate emulator designed to emulate regional and global responses to a SAI deployment. The emulator is trained by using existing simulations from two Earth System models and is then evaluated with a novel climate model simulation scenario of a multi-actor uncoordinated SA deployment. Their results show that CIDER can be used to estimate multiple climate variables of interest and thus can be a valuable tool for exploring the climate implications of various deployment scenarios.

General comments:

The study by Farley is quite interesting and the presented emulator is a quite impressive tool. This study certainly deserves to be published, but the manuscript currently lacks clarity and needs major revisions. First of all, it did not become clear to me what the main purpose of this study is. Is it the presentation of the climate emulator itself or the validation of the climate emulator. The intention or purpose of this study should be clearly communicated in the manuscript.

The  current version of the manuscripts lacks important information on the model simulations. Farley et al. refer to other publications, however, they cannot expect that every referee or reader reads three additional papers to understand what actually has been done in this study. Therefore, some repetition of information is unavoidable. For example, the simulation runs used for training and validation should be shortly described, e.g. the information should be added for which time period  the simulations have been performed and what set-up has been used.

Specific comments:

P3, L71: How many years have been used or are needed for training?

P3, L74: Which one or two climate models? Are you referring here to specific climate models or do you mean one or to models in general?

P3, L77ff: Which algorithm has been used for training? How does the input data for CIDER look like (e.g. dimension?)

P3, L80: “computes global mean climate variables”? How have these been computed? Which physics are behind these computations? Has a model been used?

P4, L90-92: I cannot follow you what actually has been done here.

P4, L94: What is the impulse response?

Section 2.1: It would be quite helpful if Section 2.1 could be split into two or three subsections to be more concise.

P6, L135: How many years have been used for training?

P6, L136: What exactly do you mean with simulation types? Simply to different simulations (set-ups)?

P6, L157-P7, L168: Move this paragraph higher up. This actually should be the first subsection in Section 2 since the description of the used climate models is essential for understanding what you have done.

P7, L168: The abbreviation PI has not been introduced.

P7, L172: For which years  has the simulation been performed? Provide this information as well as the most important settings used.

P7, !79: Are these the training scenarios or has all this done in one scenario? I thinks it needs to be more clearly be pointed out what has been done.

P8, Table 1: Are the latitudes given here the latitudes where SO2 has been injected? More details should be added in the table so that the reader can understand what has been done without reading three other papers. Last row: Add also here more information than just referring to Sect. 2.3.

P9, L205: Do you mean AOD, temperature and precipitation are shown in Sect. 3, but evaporation only in Appendix A1? Could this be more clearly stated?

P9, L206: In Fig. 2 time series are shown. This should be mentioned in the text and also what time period is shown should be mentioned.

P9, L207: validation of what?

P9, L207: Which SAI scenarios have been considered should be explicitly be mentioned.

P10, Figure 10 caption: How can this be an average over 2020-2039 when a time series is shown?

P11, Figure 3: What is meant with SE normalized?

P11, Figure 3 caption: Which simulation period has been considered?

P12, Figure 4 caption: It should be clearly mentioned in the caption that time series are shown.

P13, Figure 5 caption: Mention that a time series is shown.

P14, Figure 6 and Figure 6 caption: Description in the caption and lines shown in the legend do not agree.

P14, L262: This is rather a discussion than a conclusion section.

P15, L271-278: This is a very important point and should be much earlier in the manuscript be mentioned. I would suggest to move this (or additionally mention this) In the introduction.

P17, L344-346: This does not belong to the conclusion section and should rather be provided in the code and data availability section.

P17, L347: Since this section is a discussion a short conclusion section summarizing the major results should be added.

P19, L352 and L353: doubling of “assisted in”.

Technical corrections:

P2, L66-69: Section should be abbreviated as Sect.

P5, units: Check the Copernicus style for units, e.g. mm/day is written as mm day -1 and no dots are used between units.

P7, L183: Temperature -> temperature

P12, L246: Section should be abbreviated as Sec. (unless it appears at the begin of the sentence). The same holds for Figure which should be abbreviated as “Fig.”.

P13, L248: Add “Fig.” before %b, 5c, 5d and 5f.

P15, L291: can’t -> cannot

References: Check the reference style for Copernicus publication, to my knowledge journal names should be abbreviated.