As both the authors and reviewers make clear, this manuscript marks the start of something, not a final product. Such steps need to be documented so that future progress can be made. My impression is that the authors have done the best they can in their present circumstances to describe the new coupled modelling system and code developments, and so, on these grounds I intend to accept the paper, even though it could be argued that some aspects of the paper fall short of the ideals expressed in the GMD webpages.

Minor revisions.
I was given access to Wavewatch but can only download the distributions. The good news is that model version 4.18 is accessible; but the bad news is that I have not been given access to the development tree so can't see the information pertinent to table c6. So - it seems that the information provided in the manuscript is not sufficient to attain the correct level of access.

Thanks for adding the version numbers for the various code on github. There remains an issue with OASIS3-MCT. All I can do is download the latest version. Is there a way of obtaining a particular versions?

Weblinks are ephemeral, and at GMD we prefer information to be made more permanently accessible with a DOI. However, given the "step along the way" nature of this manuscript, and the hope and expectation that it will be built upon in a relatively short period of time, I will permit some weblinks. However, please check them all. I found that several weblinks in the current manuscript already resolve to versions of 404!!! What clearer indication of this ephemeral nature of weblinks is needed!? I suggest you consider whether any of the information can be made more permanently accessible. Failing that, please provide sufficient information related to the link such that authors can use search engines to find the relevant data in the case where the weblink fails in future. Please provide a revised manuscript in which all links work! (Note that one of the links in the caption to table c6 resolves to a "no permission" warning)

There has been some to-and-fro with the lead author over email, in regard to the availability, or not, of the WAVEWATCH code. The situation was quite confused but I think we do now understand the situation. It is true that the official WAVEWATCH releases are accessible after registration (and a month's delay!), but this does not give the new user, "developer" status, so they cannot access the svn repository and obtain the precise version of the code used in this paper. However, the appropriate version of the code is on the MetOffice's own repository, so is accessible to collaborators of the UKC2 project. The manuscript needs to be updated to reflect this situation.

It is a bit disappointing that the code is not more generally accessible. In future it would be good to see the MetOffice fall more in to line with institutions world-wide and work towards making some versions of their code publicly accessible, such that the whole UKC code snapshot that, for this paper, is only available on the MetOffice repository, could be uploaded to a public repository with a DOI.