Dear Marianne and co-authors,

Thank you for submitting a revised manuscript and I wish to apologise for the delay in reading your revised manuscript.

On reading your responses to the reviewers' comments, there are a couple of minor points, where I feel that there might have been a slight misunderstanding and so, I ask you for further clarification below. Otherwise, on the whole, I felt that you had satisfactorily addressed the reviewers' comments and as soon as you've addressed the two minor points below, I will recommend that your paper be accepted and published in Geoscientific Model Development.

Regards,
Fiona O'Connor

Minor comments:

1. In Sect 2.3 Please report the length of each time slice simulation and how this impacts the signal to noise in resulting RF estimates. Are you running 20-30 years for each time slice?

Each simulation is one year, with six months spin up. Text has been clarified and now reads: “All simulations are one year with six months spin-up. “

I think the referee is concerned about the potential of poor signal-to-noise affecting the RFari calculations. It should be made clear that the meteorology of the simulations is identical and that the RFari is being calculated offline and so, signal to noise with such short simulations isn’t an issue.

2. In Sect 3.3 please describe clearly either explicitly or through appropriate references how you calculated the RFari. Have you calculated an effective radiative forcing or used the more traditional radiative forcing metric? It is currently unclear. Also, how are the individual species RFari determined - through species only runs or is the RT code able to output this?. Please provide this detail in the manuscript.

Regarding the first half of this comment: The radiative forcing calculations described in detail in Sect. 2.4 of the paper, including a definition of RFari, i.e., forcing due to aerosol-radiation interactions, not including any rapid adjustments as for the effective RF. To clarify, we refer to back to the description in Sect. 3.3. The species-specific RFari is obtained by individual runs, where the concentration of the respective species is set to the pre-industrial level. This has been clarified in the text: “The RFari of individual aerosols is obtained by separate simulations, where the concentration of the respective species is set to the pre-industrial level.”

Here, I don’t think that you’ve fully addressed the referee’s comments. In particular, although RFari is clearly defined as the radiative forcing due to aerosol-radiation interactions, it is not clear whether it is being calculated at the tropopause or at the top-of-the atmosphere, does it include stratospheric adjustment etc..? Can you please make this clear?