Review of the manuscript Reanalysis of multi-year high-resolution X-band weather radar

observations in Hamburg from Burgemeister et al.

The authors have compiled quite an impressive time series of a X-band weather radar in Hamburg with a high sptio-temporal resolution which fits the scope of ESSD well. They describe the need for such a dataset for applications in meteorological and hydrological applications and put their dataset in perspective to studies using such data. The description of the radar is exhaustive. They provide detailed information on the removal of noise, the alignment of the radar, clutter removal, attenuation correction, and rainfall retrieval. Finally, they list the use cases of the data and point out potential limitations. The overall quality of the manuscript and its figure is high, albeit in some cases more precise information could be given and the abstract would benefit from a major revision. The dataset itself is already at a good level, but user-friendliness could be further increased by added cf compliances, a reduction of the number of files, and the removal of numerical instability in the time dimension. The biggest issue is that WDCC requires a login and tracks downloads which is not compliant with ESSDs policy. After solving these issues I’d recommend the publication of this manuscript.

General comments on data

• Data are under restricted access: WDCC requires a login and downloading persons are tracked. This does not comply with ESSDs policy https://www.earth-system-science-data.net/policies/data_policy.html 
• Downloading more than one single data file was not possible for me. I could neither download a part of the dataset via the interface nor a complete monthly file.
• Having files stored as daily or even monthly files would reduce the amount of files drastically. In the cases of maintenance, you could split the daily file into two to avoid adding the dim time to additional variables.
• Data could be more cf compliant as the editor already pointed out in his assessment.
• In the individual files I could download I found numerical instabilities on the dimension time e.g.  ‘2016-05-01T09:12:00.000000256’,  or '2016-05-01T09:02:29.999999744'. Rounding/reindexing the dim time to the full minute would remove this and errors e.g. in the selection or aggregation of data could be avoided.

General comments on the manuscript:

• In general, the manuscript could be more precise. Try to avoid general statements like “more or less” (L200) or “in rare cases” (L371),  and be more consistent with abbreviations (e.g. L211 “LAWR radar” vs L212 “LAWR HHG”). A bit picky from my side but still improving consistency is settling on either “rain rate” or “rainfall rate” (“rainfall intensity” is also used once).
• L1-9 This part of the abstract is not well-structured and somewhat imprecise. It is unclear what is part of the manuscript and what is part of previous studies and only after reading the respective sections in the manuscript it becomes clear which parts are done here. Please restructure the abstract. You could start with the second part (L.10-15), which is more concise, and expand it to include all sections included in the manuscript. 
• L133 A table or graphic describing the number/percentage of clutter points each filter removed, and hence having an estimate on the number of measured vs interpolated pixel values could add more perspective on the importance of the clutter filters and the quality of the raw LAWR observations. 
• L231 Correct attenuation. Similar to a quantitative estimate of clutter you can also provide estimates of stable and unstable attenuation correct in section 3.5. This can then be used e.g. in L371 to give a number instead of the expression “in rare cases”.
• Fig. 4,5, and 6 You could consider using a colormap that has more variation for the colorbar, e.g. viridis or turbo
• Fig. 9 You could add an animation of the event as an animated gif into the supplements of the manuscript.
• Table 4 It is not entirely clear to me if the days if all “rare days” of maintenance are listed as logged in Table 4. If yes, please indicate that the days between periods here are the maintenance days or otherwise provide this information elsewhere.

Technical comments on the manuscript

L2: “large domains and small temporal scales” is very unspecific, please modify the sentence

L14: change “will” to “can”

L22 Add “networks” after “Rain gauges provide” to make the sentence correct

L67 “The measurements refine the observations of..” This and several similar statements e.g. in L.4 sound like the refinement is done operationally

L.121 “continues”

L135 artificial: is this the correct wording?

L156 defines abbreviations TDBZ and SPIN on first use

L156 be more precise by stating that you use two variants of a spike and two variants of a ring filter

L182ff Please elaborate on the choice of N and W for the four filters 

L194 Give some information on the Kriging method 

L200 more or less can be more precisely described by L214

L221 being instead of is

L228 remove is

L238 is the iterative scheme from Krämer and Verworn, 2008 or HHitschfeld and Bordan (1954)? This sentence is not entirely clear to me

L254 change introduces to past tense

L262 either write “from the level 1 data set” or “from level data”

L263 same as 262

L279 benefit might be the wrong word here, suggestion: “but would perform better compare to” could be

L301 be more precise about common volumes

L346 sentence structure

The manuscript presents a new precipitation data set from high-resolution X-band radar measurements over the city of Hamburg between 2013 and 2021. It gives a good description of the radar itself and the data processing chain from raw reflectivity measurements to quality controlled rainfall estimates. An overview of previous studies that have already proven the usefulness of the dataset is also presented as well as recommendation for future applications.

The manuscript fits in the scope of ESSD, it is well written and clearly structured. The described data set is unique and allows for a variety of applications and studies in the field of small scale (extrem) precipitation and hydrology.  However, accessing the data is a bit complicated and the large amount of files can be challenging. I recommend publishing the manuscript after taking the following suggestions and comments into account:

• In the abstract there is a lot of information on former studies. In my opinion the abstract should be more about the current manuscript and less about other studies. Please consider rewriting the abstract with more focus on this manuscript.
• In section 2.1 the author state that the elevation angle was adjusted several times during the measuring period. Does that influence the homogeneity of the data set in any kind of way?
• Section 3.1: A flow chart or maybe some equations would be helpful to understand the method for the noise removal

Minor comments:

• 1, L.12: discuss à discusses
• 5, L.121: continues à continuous
• 8, L.183-184: The last sentence of the paragraph should be moved to beginning of the description of the ring and spike filters
• 9, L.209: Do the authors really many “e.g.” or do they mean “i.e.”?
• 9, L.210: sufficient à sufficiently
• 12, L.252: Should’nt it be 6.91 x 10^-5 for alpha instead of 6.91?
• 13, Fig.4: Why is the upper boundary of the uncertainty outside of actual range of the measurements?
• 15, L.324: reproduces à reproduce
• 17, L.336: continuos à continuous
• 19, L.380: open-acess à open-access

Dataset:

• Getting the data is bit difficult, because an account for WDCC is required. The users have to give information about their name, affiliation, which data they want to access and what they are planning to do with the data. It took about a day until my access was granted.
• The large amount of files (one file per hour) is a bit overwhelming. I wasn’t able to download a full month of data because of the size of the file. Downloading a part of the dataset was also quite a challenge. The user can choose a range of the dataset between “start record” and “end record”. This works fine as long as the first number of “end record” is lower than the first number of “start record”, e.g. start record = 1, end record = 9 works. But start record = 5 and end record = 20 gives an error, because end record starts with 2 which lower than 5. There seems to be a bug in the system.
• The netcdf-datafile itself is easy to handle, but could be more cf-compliant as the first reviewer and the editor already said.