General comments

This manuscript accompanies a radon-222 dataset (derived from measuring radon's decay product) from eight German monitoring stations with measurements dating back to 2014.  Compared with earlier publications, this release handles the effect of very humid conditions on the measurement by flagging out periods when humidity exceeds a particular threshold.

With this additional quality control measure, the data are ready for immediate use in subsequent analyses.  In my opinion, because radon-222 is an important natural tracer, and because of the other measurements available from these ICOS stations, this data set is very likely to be used in a range of studies.

The method for determining the humidity threshold is appropriate, the reasoning behind it is transparent, and the one-filter measurement technique is well described in previous publications, which are appropriately cited.

Regarding the linked data set, the data is of high quality, well formatted and well described.  Apart from the queries below, which should be simple to address, I consider that the data set will be reused productively in the future and recommend the manuscript for publication.

Specific comments

I have three minor suggestions; two related to the humidity threshold and one observation about the data itself.

First, I am uncertain about whether there is a single humidity threshold, applied to all stations, or if the humidity threshold is different for each station (“We have therefore developed relative humidity (RH) flagging thresholds for the individual stations…”, line 59).  Elsewhere, including the dataset landing page, it is implied that data is flagged as passing manual QC only when RH<98%, which is to say a single threshold of 98% is used across all sites.  If this is the case, an unambiguous statement around line 225 (conclusions) and in the abstract is recommended; if the threshold is station-dependent then the threshold (as used during QC of the published data) should be included in Table 1.

Second, it seems rather likely that a particular use might require a different humidity threshold.  This would be extremely straightforward if the humidity values were included in the data files, or acceptably straightforward if links to the meteorological data were included in this paper (along with instructions about which humidity sensor to use to replicate the published threshold, as there are likely to be many at each site).

Third, there is a period of data at the beginning of the Schauinsland (SSL) record, from February 2014, which is anomalously high compared with the rest of the record even though it is flagged “O” (Manual QC passed).  Since this is at the start of the record, and there is a break in monitoring before ‘normal’ measurements resume, it seems worth double-checking the classification (or making a note in the paper about what may have caused this – if it is thought to be non-instrumental).

Technical comments

Line 13: “..about 98% RH…”, if the threshold of 98% was used uniformly across all sites then add a comment here.

L 17: “…flat areas…”  I read this as implying that the mountain sites are not useful (at least, not ‘analysis-ready’), even when humidity is low.  I’m not sure that this is intended, based on the rest of the paper. In any case, the abstract should provide concrete guidance, to avoid the misuse of this data set, by linking these recommendations to how a new user could get started.  For instance, a statement like, “Measurements flagged as passing quality control from the stations GAT, STE, LIN, JUE, and KIT meet these criteria whereas other measurements should be treated with more care”.

L19: I think that typical style for isotopes, when the element name is written out in full, is the hyphenated form (Radon-222)

L20: “…as gaseous constituent…” -> “as a gas”

L51: “function” -> “functions”

Dataset

A typo in data headers (“depent” -> “dependent”):  Disequilibrium: specifies the sampling height-depent factor between calculated atmospheric 222Rn activity concentration in air and measured 214Po activity concentration in air

There is a column called “QualityId” – not defined in the data file headers (if this is of no use to the end-user, it could be described in the headers as “Heidelberg University internal use only”)

There is a header describing the Data Format as Version 1.0.  Is there a link to this format, for example is it standardised across the ICOS network?  If so, is there any sample code in popular analysis languages (R, Python) which would read the data and apply the QC flags?  If sample code like this does exist, it could be linked from the data files or from the ESSD paper.  This is not necessary, as the data file is in a simple plain-text format, but some users may benefit from some demonstration code and therefore be more likely to access the data and use it correctly.

Please see attached PDF