You will see from the reports that while Referee 1 is now satisfied with the paper Referee 2 is not. I believe that Referee 2 has thought very carefully about this paper and I take their view seriously.

My own interpretation of Referee 2's concerns is that the method you choose to incorporate the effect of a mean flow has quite significant disadvantages (and indeed Referee 1 pointed out some of these disadvantages in their report on the first version of the paper). Referee 2 is particularly concerned about the fact that you are restricted to considering the transient behaviour -- you cannot consider how the statistically steady state has been changed by the mean-flow effect. In the end Referee 2 accepts that it is unrealistic to expect you to carry out a completely new set of simulations and also that it would be unreasonable simply to block any publication resulting from the simulations that you have. However Referee 2 is uncomfortable -- reasonably in my view -- that too much is being claimed for an investigation which has some significant limitations.

My own recommendation, having looked at the paper quite carefully myself, is that you are more candid about whether the technique you use in these simulations is to be recommended to be used in future. (Referee 2 makes the comment, for example, that because what you simulate are features that propagate in the x-direction, the whole advantage of the approach you choose is lost.) You should also included a few sentences that make it absolutely clear what ingredients would be present in a straightforward (and realisable) configuration with a mean flow (which would for example, include vertical shear) and what is present in the configuration that you investigate -- and what are the differences between the two. To me the abstract and the title emphasise 'mean flow' and it is only in one phrase in the abstract that you say that actually the mean flow is implemented only by adjusting the surface fluxes.

A particular point that concerns the Referee is that there is great emphasis in the Introduction on the relevance of your investigation to the MJO. Clearly there are differing views on the whole question of relevance of convective aggregation to the MJO, but the referee makes the point that, firstly any mean flow present as the active phase of the MJO develops is likely to have vertical shear and secondly, the scales that you simulate are far smaller than the overall scale of the MJO. Of course the MJO is a multi scale phenomenon, but several steps would be required to go from the modifying effect that you are finding at the mesoscale, which is in any case limited by the constraints of the experiment, to the overall implications for the MJO. Again I think that it would be much better if you 'dialled down' the emphasis on the MJO in the introduction which, to be honest, doesn't really seem justified by the your eventual results and conclusions.

Please consider Referee 2's comments carefully and provide responses plus a revised version of the paper. I very much hope then to be able to accept the paper without going back to Referee 2 -- but I will need to see evidence serious consideration of Referee 2's comments.

You have made further adjustments to the paper in response to the comments one Referee 2 and I think that the fair option is now to proceed very quickly to accept the paper for publication in ACP.

On the basis of reading through the latest version of the paper myself I make the comments below -- please consider making further modest changes in response to those comments and provide a final version of the paper -- which I will then be pleased to accept.

abstract: 'aggregation' is mentioned in the first sentence of the abstract, but the term is not used again in the rest of the abstract. A reader might wonder whether the first sentence has any relevance to the rest of the abstract. I think that you are interested in the effect of a mean flow on the aggregated structure -- saying something like 'propagation of aggregated structures in convection' in the second sentence would help the reader see the link.

abstract l9: 'retards the propagation of the convection' would be clearer than 'retards the convection'. Incidentally this last line of the abstract seems an important conclusion, but '5% of the mean wind' may be misleading -- you don't actually know whether the effect is a fixed percentage of the mean wind (as the mean wind varies).

l109: 'mechanism denial experiments' -- but in fact there is only one experiment.

l145:
'In the case of WISHE, the convective cluster moves against the mean wind (e.g.,urel <0ms−1).' -- what you actually mean here I think, is 'If the hypotheticated effect of WISHE was realised then the convective cluster would move against the mean wind'

l249: 'Even in the simulation with the dynamic feedback removed and the WISHE-induced asymmetry in surface fluxes preserved, the effect on the propagation of convective clusters is small.' -- why 'Even'? Isn't the point that in this case the effect is sustained, but weak by some measure?


Two further comments -- which you may or may not wish to take into account in producing the final version of the paper. I am essentially setting out my own understanding of the paper -- if I have misunderstood something then you may wish to consider whether other readers may also have similar misunderstandings -- and alter the text a bit to avoid this. (Of you may consider that my misunderstanding unlikely to be shared by most readers.)

1) My own interpretation of your findings is that, whilst the effect of the mean flow might according to some have a significant effect on the propagation because of the associated difference in surface fluxes upstream and downstream of the aggregated system. You are intending to identify this effect by considering u_rel and finding that it is negative (i.e. in the opposite sense to the imposed mean flow. What you find is that u_rel is certainly negative, but a large part of the explanation for this is that the upstream/downstream different cannot be sustained close to the surface inding is that this effect cannot be sustained because the momentum exchange with the surface limits the upstream/downstream difference -- the low-level mean flow is quickly reduced in magnitude. Of course this does not rule out a sustained effect in a different system where there is active dynamical driving of the low-level flow.


2) Personally every time I looked at this paper I had difficulty relearning what system it was exactly that you were studying. I think that you are considering a system in which, from the point of view of an observer fixed relative to the Earth's surface, at t=0 the air velocity at all levels is u_b, and you choose to observe/simulate the system whilst moving at speed u_b.

When the effect of the surface is transferred to the atmosphere above by a momentum flux, the velocity in the atmosphere naturally reduces towards that of the surface, i.e. in your frame of reference the velocity in the atmosphere naturally adjusts towards the value -u_b. When there is no momentum flux from the surface then the velocity in the atmosphere remains at u_b, i.e. in your frame of reference it remains at zero.

If you could make a simple statement saying something like that then it would help a lot -- the reader would not have to worry about which physical effects were included in your formulation -- equations (1) and (2) -- and which were not.

Thank you for making further changes to your paper in response to my comments. I am now pleased to accept the paper for publication in ACP.