Visioni et al. describe the plans for the contribution of a geoengineering MIP to CMIP7. This is a very laudable community effort, and the modelling community evidently organised the process of defining the simulations and simulation protocols in a very open and transparent way. The paper states that much of the definitions is based on the outcomes of a workshop to which the community was invited.

I have a few somewhat overarching suggestions for improvements, and then a number of smaller suggestions.

– Some of the propositions/requests in my view are too vague to be a meaningful MIP contribution. Of course each modelling group can (and should) do their own simulations, and smaller groups can always come up with own ideas. The very meaning of a MIP is that it is a strict protocol and that a meaningfully large number of groups submit simulations that exactly follow this protocol. So my suggestion is to omit everything that is only vaguely defined. There could be for example an appendix to the paper in which such suggestions – potentially these are minority opinions from the workshop that didn’t make it to the Tier-1 MIP proposal? – can be listed as examples.

– A bit related comment: The real MIP proposal is the Tier-1 simulations. However, they are impossible to do without some of the Preparatory experiments. These Preparatory experiments, in turn, are not understandable really before one understood the Tier-1 experiments. To the reader of the paper, a much more logical structure would be to describe the Tier-1 experiments, and include in these the Preparatory experiments (except for the third block, Inj-Sulf-15N etc., which seem to be of rather little use except for the zoo of Tier-2 experiments). Of course I acknowledge it might be difficult to change such things after the workshop, though. In the Tier-2 experiments, to me only the experiments listed in Table 3 are well-enough defined to be part of CMIP7. The other experiments are very vague and would better be done and consolidated within the community. They could be described here in an Appendix making clear such ideas were discussed but then not included as being well-defined enough for a large MIP effort as part of CMIP7. Again I acknowledge that the community may have discussed such arguments in detail and the harm done in listing such ideas here is probably minor since groups will carefully select what they contribute anyway.

– The integration of GeoMIP into CMIP7 should be discussed. In particular, what are the synergies with AerChemMIP and DAMIP?

– There are very important aspects of the Earth system models that may have a large influence on the outcomes (e.g. the cloud microphysical schemes for MCB and CCT). The protocol should define a way to document these particularly relevant ESM aspects in an easily accessible and transparent manner.

– Several figures show results but these rather distract from the MIP description. Fig. 3, 5, 7, 8 in my view contribute little to the paper but are rather confusing. They could go to an appendix or to a separate paper.

L12 – “that all participating models are encouraged to run”: the corollary is that models are discouraged to run tier-II simulations

L15 – “propose their own experiments” links to my main comment above: this is not a contribution to a MIP but the bread-and-butter of every modelling team. It should be omitted from this paper or at best be moved to an appendix

L26 – “increase of Earth’s albedo through _increased_ cloud brightness”

L26 – what could be “other cloud changes”? cloud coverage?

L47 – the “potential asymmetries” should be explained

L111 – “this could include metrics”: I would expect from a MIP definition a definition of specific metrics. Maybe the text could point here to Section 5 (Data request) and take up this point in that section 5.

L125 – “Models can also perform…” this seems to be a redundant statement, of course they could. If there is a reason to state this here, this reason should be made transparent (e.g. “The GeoMIP community has also discussed simulations at other latitudes… but decided not to include those in the Tier-1 request.” or so).

L141 – Only later the emission height and size distribution are defined, it would be useful to point to this here.

l147 – The community uses the term “aerosol-cloud interactions” since AR5, “aerosol indirect effects” is outdated.

l155 – It should be clearly defined how this careful documentation should look like.

L160 – The Table said it is run for 20 years (not 40 years)

L202 – Only one ensemble member run to 2150, was stated in the table.

L225 – “as little as possible” is a difficult statement, maybe simply “little”?

L232 – I don’t understand the “Legendre polynomials” explanations. A clear definition of who T1 and T2 are computed should be provided.

L248 – For the MCB experiment, a definition for “maintain the PI+1.5°C temperature target” is provided, in line 332, at ±0.2°C. This seems meaningful.

L254 – “it is possible” rather than saying what is possible the protocol should describe how it should be done exactly.

L258 – “can be computed” similar point, it should be said what should be computed and how.

L267 – “can request a dataset” – how exactly, to whom to address?

L327 – “lowest model layer” seems to be too vague. Models have vastly different vertical resolutions. The request should clarify a height range instead.

L329 – I would find a map useful.

L334 – Only one ensemble member to 2150, right?

L363 – Also the number is evenly distributed since the size distribution is given.

L411/Figure 6 – The “intervention magnitude” curve for G7-0.5K-SAI seems to be wrong to me. The magnitude should be much larger very early after 2035.

L421 – If “ramped up linearly”, the temperature evolution would look very different from the one depicted in Fig. 6.

L424 – But here no overshoot is considered?

L435 – Same question, there is no overshoot in this scenario.

L517/Section 4.3 – This entire section deviates a lot from the other Tier-2 descriptions. As stated above, in my view best would be to move these very vaguely defined simulations out of the actual MIP description (e.g. to an appendix). If the authors choose they want to include these, the description of the mixed-phase cloud thinning experiments should be harmonised with the description of the other Tier-2 simulations.

L567 – specifically this section should be incorporated into Section 5.

L660 – “technique is” or “techniques are”

This manuscript presents a well-written and comprehensive overview of the next generation of GeoMIP experiments, and the authors have (mostly) done a good job with producing what will likely be an important reference documents for years to come. However, there are a few major and several minor points that should be addressed prior to publication, as follows:

Major points:

1. There is very little attention to model evaluation in the manuscript, and assessment of whether models are fit for purpose is largely ignored. For the SRM methods considered (SAI, MCB, MCT) I would as a minimum expect a couple of sentences for each about what variables or benchmarking cases models should simulated well for us to have confidence in them (as now briefly mentioned for MCT, where cloud phase is stated to be a key variable)
2. The choice of the ML scenario as the only baseline scenario in the Tier 1 experiments needs better justification.  Arguably, with its very steep drop in emissions starting almost immediately, this scenario appears on the optimistic side given that global CO2 emissions are still going UP, not down. As such, it could have been valuable to place some more emphasis on the M scenario. I was very surprised that when a second baseline scenario was proposed in the Tier 2 experiments, it was the LN scenario (which I don’t think anyone really thinks is going to materialize), and not M, which from my reading of Visioni et al. (2025) was the alternative scenario being considered by the community. 
3. The paper discusses preliminary model results, which is very nice, but many of the related figure panels are never referred to. An example is Figure 3, which is never referred to in its entirety. There are references to Figs. 3b, d and f but there are no figures 3b, d and f (Figure 3 lacks sub-panel labeling). Likewise, Figure 5 is not referred to in its entirety, only sub-panel 5a (which does at least exist). This is sloppy - either figures should be referred to or they should not be in the paper.
4. Cirrus Cloud Thinning is completely ignored in the manuscript, after being mentioned alongside MCT in the introduction. I am not necessarily calling for a CCT experiment (although there appears to be some renewed interest in this method), but a few sentences about why there are no experiments exploring CCT further this time around would be good.

Minor points:

1. Line 58-59: climate modelers are not distinct from climate researchers, but rather a subset of climate researchers
2. Line 376: This is an interesting point regarding the non-linearity of MCB. This sentence refers to how efficiency may go down as sea salt emissions go up, but in the baseline scenario anthropogenic aerosol emissions also go down quite rapidly, could this matter?
3. Line 534: These INP perturbations would be expected to over-seed Arctic mixed-phase clouds. Why such high concentrations?
4. Table 6: It would be very relevant to add information about whether models include secondary ice production, if possible
5. Line 605: ODS abbreviation is never introduced
6. Line 660: This should be «timing», not «timeliness», I think? Generally, there are a few typos here and there that I didn’t bother pointing out, but an additional round of proofreading before re-submission would be beneficial