Dear authors,

I have iterated your manuscript with one of the reviewers who asked me to be able to comment on the revised version again. The reviewer finds that the manuscript has improved significantly and that many questions have been clarified. However, a few minor issues remain, notably the aspect of hourly averaging not being sufficiently well motivated as well as a few places where the text needs (further) revision and streamlining with the answers given to the reviewer.

The manuscript is thus on a good track, but before I can recommend publication, these minor issues need to be resolved.

Best regards
Dominik Brunner

Dear authors,

Thank you for the new revisions and responses to the referee. The manuscript has further improved and I am happy to let you know that I accept the manuscript for publication. However, as you will see below, I still have a rather large number of further comments that, although being all minor, will have to be addressed before the manuscript can be published.

1. Title: The new title of the method "Adaptive selection of diurnal least variation" sounds less correct to me than the previous title. Even though I am not a native English speaker, I am quite sure that the adjective "least" can not be used in this way. A variation can be small but not less, and thus it can be smallest but not least. Actually, I don't think that anything was wrong with the previous title " Adaptive Selection of Diurnal Minimum Variation", but rather that the method was later called "Adaptive Diurnal Minimum Variation" which sounds like the "variation" would be "adaptive" rather than the "selection". I would propose to call the method "Adapative Selection of Diurnal Minimum Variation (ASDMV)" or, closer to the present name, "Adaptive Diurnal minimum Variation Selection (ADVS)".

2. Please note that it is standard practice to present validated rather than raw data, and therefore the emphasis on "validated data" in this manuscript is quite confusing. Furthermore, I would refer to such data as "calibrated and quality controlled" rather than "validated", since validation typically requires comparison against independent data.
"Data selection", on the other hand, is a very general term that is not necessarily associated with baseline data selection. I am, therefore, not particularly happy with the usage of the terms "validated data" and "selected data" in this manuscript as they are prone to confusion, but since the terms are introduced at the beginning, it is probably ok but certainly not ideal.

I have added all my other suggestions (about 150 in total) as annotations to the attached PDF document, which will hopefully help improve the clarity of the manuscript. Please incorporate these suggestions as much as possible, though you don't need to answer them one by one or reiterate with me. However, please contact me if something is unclear or if you disagree with a particular suggestion.

Best regards
Dominik Brunner