Dear authors,

Thank you for submitting a revised version of the manuscript. I consider that it is now essentially ready to be accepted for publication.

However, I would like to expand upon one of the comments by the reviewers first. Specifically, one of the reviewers raises the issue of conservation of mass and tracers. This issue is of fundamental importance, particularly when using ocean models to conduct very long simulations (such as for palaeoclimates). In your response, you state that this issue has not been studied.

I consider that this issue needs to be acknowledged in some form in the manuscript. At an absolute bare minimum, it should be discussed in the text. However, I would like you to present results if at all possible. Presumably it is possible to assess any conservation errors in the simulations that you ran? Even if you cannot calculate conservation errors directly, it is presumably possible to compare the evolution of global-mean salinity (etc.) between the simulations with full precision and reduced precision?

I suspect, because of your accuracy tests, that the reduced precision does not introduce significant conservation errors. However, it would strengthen your manuscript to confirm this by quantifying the magnitude of any errors in conserving mass or tracers.

Otherwise, I would appreciate it if you could address the following minor comments:

* Page 3, lines 10-1; page 4, lines 19-21; page 12, lines 6-9: One or more words appear to be missing from these sentences.

* Page 7, lines 31-32: Could you please check the wording of this sentence carefully?

* Page 8, lines 12-14: Please provide a reference to support this statement.

* Page 13, lines 5 and 8: I think "significant" should be "significand" in both cases.

Kind regards,
Steven Phipps
Handling Topical Editor

Dear authors,

Thank you for submitting a revised version of your manuscript. I am now happy to accept it for publication in Geoscientific Model Development, subject to you addressing the following two technical comments:

1. Could you please add panel numbers (a and b) to the new Figure 1, and improve the labelling of the y-axes? The axis labels should, I believe, read "Global heat content (10^25 J)" and "Global salt content (10^18 g)", rather than having these numbers appear separately on the figure. I realise that the global salt content has been offset by 4.731e22 g, but it might be clearer to state this in the caption.

2. Could the new Figure S1 be added to the manuscript as an Appendix B? It is undesirable to have a supplement that contains only one figure.

Kind regards,
Steven Phipps
Handling Topical Editor