akthough the legend has been updated as indicated in the author's response, die figure (now) 3:still features

the misleading description: Mean EVI (Area under the curve)

I am approaching this manuscript as a fresh reviewer. The work seems to have had a tricky passage through the discussion and review process to date. Although the comments from the two original reviewers were largely positive,the revision of the first version (https://essd.copernicus.org/preprints/essd-2019-198/) seems to have not been accepted.  The current version has been under discussion since September 2021! 

Overall, my impression is broadly positive, and I see no major flaws. I appreciate the effort by the authors to provide basic online visualization capabilities, and confirm that I could download the dataset and understand the contents of the various sub-directories (i.e. the metadata provided is generally sufficient).

However, numerous language improvements are still required. More substantively, I was slightly disappointed to that the authors did not provide the code they used to generate their datasets; this hinders the reproducibility and transferability of their work to a certain extent. If at all possible, even though the regulations of ESSD may not require it, I would ask them to consider providing the code so that others could apply the methods to other areas most efficiently. The authors should also ensure that high resolution figures are uploaded; in the present version, some aspects of the figures are hard to read. Finally, I am not sure if the detailed analysis and interpretation of the datasets that are presented as “results” is necessary / fits within the scope of a data descriptor; such material could arguably constitute or be expanded into a full novel research article, which simply cites the data descriptor.

In summary, I recommend publication subject to these remarks, as well as the more specific (generally minor) comments in the attached PDF, being addressed.