Although you have made a significant effort in addressing the comments/questions from the reviewers it is clear there is a significant amount of confusion regarding many aspects of the paper. These include whether models are used to generate datasets versus those that are used to validate, and the underlying algorithms used to generate the LULC-based focing datasets from the pollen proxy data. Given that this is a GMD and not a discipline-specific journal, I think it would make sense for you to pass through the article again with a mind to this, and also clarifying the nomenclature accordingly. Furthermore, uncertainties in the protocol will also be subject to significant uncertainties and more insight into these in the paper would be useful.

A further comment on code and datasets. GMD endeavors to make paper submission code and datasets used to facilitate the research as available as possible. Would it be possible to make a limited subset of the data that was used to make the figures available on a code/data serving service (e.g., zenodo). In a perfect world it would be possible to recreate the figures with the provided code and data.

It was notable that there were a few grammatical and spelling errors in the revisions that need to be collected also.

Thank you for making the datasets available publicly. It would be ideal to make them available from just a few places with a couple of doi, but if the multiple Zenodo dois are what is needed then I believe the paper can be accepted for publication. Thanks!