Thank you for your thorough replies to the reviewer and editor comments, which required major revisions of the initial paper.

I will look into the revised version of the paper and make a decision how to proceed.

Dear authors

Thank you very much for your careful revisions adressing many of the reviewer comments and improving the clarity of the manuscript considerably.

Reading the manuscript, the reviewer comments, and your replies, I think the paper has clearly advanced providing a unique long-term reconstruction of global biome changes from both data and the biome model BIOME4. This reconstruction will be of immense use for future studies and clearly justifies publication in CP. In line with referee #2, I am still worried that no quantitative estimate can be given on the similarity of model and data based biome reconstructions, which makes an uncertainty estimate (for example for NPP etc.) difficult. However, I admit that deriving such a statistically robust measure within the scope of this paper may be difficult or even impossible considering that uncertainties in the forcing of the climate models, in the climate models themselves and in the BIOME model exist.

While the paper clearly advances our knowledge on terrestrial vegetation changes over the last 120,000 years, I feel strongly that the further discussion of the results on carbon storage and d13C changes in the ocean (Section 4.3 & 4.4.) are walking on too thin ice given the large uncertainties in the biome reconstruction but also in the limitations of your approach:

- the BIOME model cannot quantify changes in inactive/inert carbon pools such as permafrost, which may dominate the carbon storage changes. In fact, your Fig. 7 in the replies to referee #1 shows that including such pools provides a completely different answer and thus your total numbers on terrestrial carbon storage changes (in Gt C) cannot be regarded robust. It would be a bad thing if other scientists used these numbers from your modeling approach for further studies
- your mass balance approach for isotopes in the ocean does not include isotope dilution effects on this long time scale that come about by volcanic, weathering as well as sediment fluxes. Accordingly, your mass balance approach overestimates the d13C changes in the ocean.
- it may be questioned that the turnover times of different biomes are constant over time, impacting your terrestrial carbon storage estimate
- the results are obviously highly dependent on the forcing data (some of the CO2 data used are lower than latest, improved reconstructions. This has a considerable effect through CO2 fertilization in your model) and the model version/setup used and the uncertainty of the reconstructed carbon storage is of the same size as the total changes. Taking the additional uncertainties in your approach to calculate terrestrial carbon storage and d13C in the ocean into account none of these numbers are known to a sufficient degree.

Accordingly, my editoral suggestion is to remove section 4.3 and 4.4. from the manuscript, before final publication in CP. You could shortly comment on the implications of your results on terrestrial carbon storage and d13C in the ocean in the Discussion but any such statement would have to be qualified by stating that the current uncertainties in the approach do not allow to reconstruct robust numbers in terr. carbon storage and d13C in the ocean. This should also be stated in the Abstract.

Specific comments: Please see also some minor corrections and some additional comments in the annotated manuscript attached to my editor's comment's.
In contrast to what I suggested before, I would ask you to move your new figure 2 into the supplement, where it nicely complements the data.

I hope you will understand my concerns and I am looking forward to the revised version of the manuscript.

All the best

Hubertus Fischer
(editor Climate of the Past)

Dear author's

Thank you for responding very positively to my last comments/editor's review and for revising the manuscript according to my recommendations. I am happy to say that it is now ready for publication in Climate of the Past. A few typos can be corrected during the copy editing process.

Thank you for all your work and for publishing in Climate of the Past.

Hubertus Fischer
(CP editor)