Dear Erin and Co-Authors,

We have now received two positive reviews for your manuscript. I am grateful for your responses to the reviewers and your detailing of how you will address their minor revisions. Both reviewers note that this is a significant step forward in Pliocene Data-Model comparison and recommend your paper for publishing. I agree with their sentiments and look forward to receiving a revised version of the manuscript taking into account the changes noted by the reviewers.

Regarding the comments from Reviewer 2 in terms of the impact of using different baseline datasets from which to create an anomaly (NOAA v WOA) I take your point about the limited impact and your preference to maintain the oceanographic causes for Data-model discrepancies as the key focus in the manuscript. However, please consider adding the analysis that you have done comparing the two baseline datasets to the Supplementary Information to demonstrate that this issue has been taken into account.

Best wishes,
Aisling

Dear Erin and Co-authors,

Thank you very much for your re-submission which incorporated all of the changes to the reviewers comments and my own.

Upon final reading of the manuscript I noticed 2 areas, that I feel need addressing before the manuscript can be published. These are minor and the second I think is a result of some changes you made to the manuscript in the most recent submission.

1. Please reference the underlying SST data in Figure 2. At present, the figure caption only makes reference to the the PlioVAR sites.

2. One line 282 of the revised submission, you state that: "Our proxy-based mean global SST is larger than most PlioMIP models (Figure 3); because air temperature increases are larger over the land than over the ocean (Haywood et al...)"...

Can you specify which 'proxy-based mean global SST you are refering to? Proxy data (all)?

Also, the statement regardig temperature increases over land being larger than over the ocean seems out of place here in reference to Figure 3, which I assume is depicting a global mean SST for the models as well as the data. Whilst the statement is true, I don't think it is a justification of the differences between models and data in Figure 3. Please can you have a look at lines 282-285 and consider rephrasing them.

Many thanks,
Aisling

The authors have now addressed all comments by the reviewers and the minor comments I suggested for the final version of the manuscript. I am very pleased to see this contribution ready for publciation now.