Please see the attached PDF.

Duran et al propose a method for separating out the contributions to effective radiative forcing from aerosol-cloud interactions (ERFaci) into the Twomey effect and adjustments due to liquid water path and cloud fraction change. This is achieved using an extension to the set of standard monthly diagnostics from CMIP experiments, and is a valuable addition to the list of methods currently available such as the rough and ready attempt I made at this separation in 2020. I hope that many modelling groups will implement the MODIS simulator to allow this decomposition to be calculated in several models in CMIP7. As the authors point out, the individual terms of the decomposition may lend themselves better to observational constraints than the overall ERFaci, and if well-founded, can contribute important knowledge towards narrowing uncertainty in warming projections. In RFMIP we have requested that it be added to the data request. It's a neat method, and the paper is well written – one of those rare papers where it’s difficult to find any fault, which is why this review is comparatively short.

My only real comment is that the method doesn’t quite decompose linearly. Summing the components from each model in Table 1 underestimates the ERFaci for the three non-MIROC models by more than 10%. I suggest that the residual term is given the same level of prominence in this table and the text. In the opinion of the authors, is this a non-linearity due to second order cross-terms, or could there be other aspects of the ERFaci that we don’t or can’t evaluate?

Minor comments:

Line 32: just clarify that this is the instantaneous forcing due to aerosol-cloud interactions.

Lines 128-132: could you clarify? >70% liquid means classified as liquid, <30% liquid means classified as ice, and 30-70% liquid means undetermined. Did you also mean 0.3 to 0.7 optical thickness in line 131, or is this a coincidence?