|The authors have responded comprehensively to my comments and have made many good modifications to the text based on them and those of the other reviewers. As a result the contents of the manuscript is improved. Having said that, some of the main issues I highlighted in my first review remain. The manuscript remains very long, it constantly mixes results and discussion, and it makes what is in my view inappropriate use of supplementary material. As all of these issues are about style, rather than substance, I feel the manuscript can be published as is, even though a better manuscript could be produced.|
1) Length - In my first review I had noted that the manuscript was very long with 26 pages and 28 figures. It is now 24 pages and 27 figures. The two page saving is from moving the model description into the supplementary material. So in essence, nothing has changed. I still think that by improving the writing and focussing on the main argument, the manuscript could be significantly shorter, but if the journal is happy with the length, I am not inclined to push harder.
2) Supplementary Figures - The manuscript makes what is in my view inappropriate use of supplementary material. I believe examples for an appropriate use is showing NH figures when the paper is focussing on the SH. It’s also appropriate to show a model mean figure in the main text and show every model’s figure in the supplement. I do not believe it is sensible for major arguments in the paper to rely on supplementary figures. Here are the first dozen or so figures referenced in the paper in order of appearance: S1, S2, S2, S3, S3, 1, S3, S4, 2, 2, S6, S7. From this it is evident that the supplementary figures carry the entire first half of the results section and a must be anything but supplementary. I think many of them are used to hide the true length of the manuscript, which is still caused by a lack of focus.
3) Mixing of discussion and result - The authors continue to mix discussion and results as they walk through the results section. That is their choice, but I still feel it gets into the way of the argument. The fact that so many supplementary figures are cited early on is related to this, as it is usually the discussions that are often supported by the supplementary figures. I still feel a clean separation of presenting results and discussing them in a separate section will both shorten and sharpen the manuscript.
As mentioned above, while I feel the manuscript could be further improved, most of the improvements are on style. It can therefore be published as is unless the authors wish to make further improvements.