Interactive comment

on “One year of attenuation data from a commercial dual-polarized duplex microwave link with concurrent disdrometer, rain gauge and weather observations” by Špačková et al.

The authors present attenuation data from a commercial microwave link (CML) experiment together with multiple observations of disdrometers, rain gauges and weather observations. This data set is unique as there are just a few CML data sets available publicly from which none is supplemented by the presented amount of rainfall and meteorological data. The data is available via zenodo and the internal structure of the data is easy to understand. The manuscript accompanying the data is well written and logically structured. In some points the manuscript and ease of usability of the data could be improved, which are listed in the general comments. Together with some further minor updates of the manuscript I recommend the manuscript for publication in ESSD.

General comments

• Chapter four is not consistent in its content: While 4.1 does only give a summary of past activities on dry/wet classification with the presented data, 4.2 and 4.3 also include open issues with regard to wet antenna attenuation and DSD estimation. In 4.4 the issues from the preceding sub-sections are (partly) repeated with less structure and additional issues are presented. The whole section 4 should be improved by re-structuring its content - for example in sub-sections covering the already conducted applications of the data set and the unsolved issues which could be tackled in the future - in a consistently manner.
• For quick exploration and convenience of the data user a global overview of the data availability and data issues (flags) describe in section 2.5 should be added both to the manuscript and the event calendar view in the html viewer. This would be most important for the CML data and could be done e.g. with an additional figure in the manuscript showing time vs data availability/flags and likewise or with and an additional calendar view in the html viewer. The daily data availability charts are fine, but it takes rather long time to select e.g. the “best” summer month for a certain application. (see Fig. 2 in van Leth et al., 2018)
• An additional kml file of the observation locations in the data publication would add further convenience for the user of the data.

Detailed comments:

p.1 l.2f.: Here you show CML use cases which are not further described in the text. I suggest to either give a short overview of the retrieval of such variables and the potential of the presented data set or to concentrate on rainfall as target variable as you did consistently in the manuscript

p1. l.15: delete “also”

p2. l.20: Use CML attenuation rather than microwave attenuation to be more consistent (e.g. with p.2 l.25).

p.2 l.25ff.: Refer to the availability of open source packages which for CML processing accompanied by small test data sets

(e.g. https://github.com/overeem11/RAINLINK and https://github.com/pycomlink/pycomlink).

p.3 l.57f.: Add citation stating that parameter b is close to 1 for a certain frequency range (e.g from ITU).

p.4 Figure 1: Use different colors for the zooming window and the CML. You also could plot the CML in the un-zoomed map without the disdrometers etc. for a better overview.

p.4 Caption Figure 1: The Figure caption states “the direction of the link” please change to “the path of the CML” and “are less than 6 to 10 km” to “are within 6 to 10 km”.

p.4 l.90: Be more precise about the quantization of the received signal

p.5 l.95: Make clear that even with a shield wet antenna attenuation can potentially be a problem due to dew

p.5 l.108ff: Use plural or singular for rain gauge(s) consistently.

p.6 l.117: Rephrase to “CML attenuation data is available for the period between …”.

p.7 l.119: State again that the reading is instantaneous.

p.9 Figure 5: You could identify the individual rain gauges and disdrometers with small labels. For the caption you could move the data availability etc. to the running text.

p.10 l. 164f: Here as well as in the html viewer it is unclear from which observation the daily cumulative rainfall depth is used. Please provide this information in the manuscript as well as in the html viewer.

p 14. l.213: Delete “last, but not least”.

References

van Leth, T. C., Overeem, A., Leijnse, H., & Uijlenhoet, R. (2018). A measurement campaign to assess sources of error in microwave link rainfall estimation. Atmospheric Measurement Techniques, 11(8), 4645–4669. https://doi.org/10.5194/amt-11-4645-2018

On “One year of attenuation data from a commercial dual-polarized duplex microwave link with concurrent disdrometer, rain gauge and weather observations” by Spačková et al.

In this paper, the author presents a dataset of microwave signal attenuation of a 38-GHz dual-polarized commercial microwave link (CML). The dataset was collected in the context of a one-year-long field campaign and is supplemented with the data of five laser optical disdrometers and three standard tipping bucket rain gauges under the path of the microwave link. In addition, further meteorological data from five synoptic weather stations is given. 

General Comment

As a whole would rate the datapaper as good and support to publish it with some revisions.

The dataset by itself is definitely unique and can serve as a basis for further studies on different issues dealing with quantitative precipitation estimation with the use of CML. On the one hand it is the availability at all, since CML data is often not openly a not available. On the other hand, it is the extend of the data. This holds true for the long-time span of the dataset (one year) of the CML data and also for the collocated data. Next to standard rain gauge and further (synoptic) meteorological data (temperature, humidity, wind and pressure), information of the real drop size distributions is given via the output of five laser optical disdrometers. To sum it up: with this dataset further detailed studies in the evolving field of QPE using CML are made possible.

The presentation of the dataset is generally good. The structure is clear and the data is easy to access. It is possible to find the data one is looking for, to download the data and to work with the data. The data format and the internal structure within the specific datasets is straight. Further on an overview is given by a HTML viewer, which is also easy to access and leads – via the color coding of the different dates based on the precipitation amount - to the – for the majority of the users - more interesting days with higher precipitation amounts.

Nevertheless, there are some – in my opinion – important points, which need a deeper revision.

The two most important point deals with (a) the metadata of the different sensors and stations and (b) a deeper discussion of the data quality. The metadata being one of the most important parts of a dataset, since they enables the future user to get a bit more insights to the situation the data was gathered or – to put it in other words – to comprehend the whole story a bit more. “Metadata are especially important for elements which are particularly sensitive to exposure, such as precipitation, wind and temperature” (Burt, 2012). In this paper and dataset, the metadata, as well as the discussion of the different data quality issues come – in my opinion – a bit very short. I couldn’t find detailed information about the different measurement sites. Generally: For every site an own somehow detailed metadata-sheet (as recommended by the WMO and is the best practice in meteorological measurements) is needed, which should – next to a description of the site and the exposure, the specific instruments in use- also include a couple of photos and a more detailed map. I also would recommend a maintenance plan with information what was done during maintenance.

In addition - regarding the different senors / instruments:

For the disdrometers it would be important to know to which direction each was aligned to, since – because of the construction of the Parsivel laser optical disdrometers precipitation coming from the side is a bit more undisturbed. A short discussion of each disdrometer site with respect to possible wind influences would come handy as well as information of a change of the surrounding (especially trees and bushes) over the course of the year.

For the rain gauges a short discussion on possible wind influences and possible change of the surroundings is also recommendend warmly.

For the CML the information about the height above the ground is missing and of high importance for the use of the data for further studies.

For the airport weather stations and meteo swiss weather stations a short description (list) of the different devices (sensor type, manufacturer, model) is needed in order to generally reproduce the measurements and gain information about the measuring principle. For the airport weather stations it is also important to know, whether the international measuring standards are met.

The second point which needs a more detailed description, and is so far a bit short.

 This point goes a little bit hand in hand with the more detailed metadata and is partly accounted for, when one puts down more detailed metadata (for example a short discussion of the exposure of each site). But one further point is the need to clearly explain, how often the instruments were maintained and what was done during maintenance. For the disdrometers it is for example interesting to know, how possible cob webs were dealt with. For the rain gauge it is important to know, whether wind protection shields were used, or what was done during snow.

Burt, S., 2012: The weather observer's handbook, Cambridge University Press, Cambridge, 444 p.

Detailed comments:

p.1 l.4: … CML of 1.85 km at a high temporal resolution  --- CML of 1.85 km path length at a high …

p.1 l.6:   delete ‘effortlessly’

p.1. l. 7: satisfactory to what? Be a bit more precise. Especially state, whether it is reliable.

p.3. l.66: Formula (4): Give a reference, where the formula comes from. There are some a little bit different moments of the DSD for the rain rate circulating in the literature.

1. 3 l. 73: Five laser optical disdrometers.

p.3 l. 76. ‘intergrated’, this verb seems in this context a bit not best expression. Just writ, that three tipping bucket rain gauges were used.

p.4. Figure 1: Give an explanation to the abbreviations: P and RG.

p.4 l. 86 ff: Section 2.3. use a subsection for each instrument and – as stated – give a link to the specific metadata sheets

p.4 l. 87 ff: CML- as already written – the height above ground is a needed information.

5. 5. l. 96. – laser optical disdrometer

p.6. l. 1. Table 1:  I think it is just the location of the sites, which is stated, not the characteristics

6. 6. l.111 ff ‘located 6 to 10 km from the experimental site’ Be more precise, give the exact distances and directions.
7. 6. l. 116 ff think this is a good place to state, that the tie stamp is UTC (not just in the Appendix).
8. 7. l. 121 ff: in the Parsivel dataset there are 20 variables. Some (esp. “Precipitation code 4680", "Precipitation code 4677") need more explanation and in the “read me file” no units are given. I think this should be changed.

p.8. Figure 3. X axis label / time stamp label should be time of the day and there should be a tie stamp (UTC) – This concerns also the other figures and especially the figures accessed via the html viewer.

8. 8. l. 135 ff: ‘all major weather characteristics’ is absolutely not the right term. Better use for example standard meteorological parameters.
9. 8. l. 137: As stated in the general comment section: The discussion of the data quality is partly a bit to common. What about the standard issues of rain gauge and disdrometer measurements regarding the wind influence and dirt ( and cob webs) and snow.

p.9. l.142 ff. There is somehow a duplication of the description, please check this. I was a bit confused.

11. 11. Table 4. Be a bit more precise in the label of the table.
12. 12. l. 178: has helped develop or could provide – a little bit confusing sentence.

p.12. l.179. font size of subsection caption is bigger that the one of the (main-)section.

-- Section 4 – In my opinion it is a bit to extensive, since it is mostly hypothetic. It is more extensive than the discussion of the data quality and the variable description. This is not the right relation for a data paper. In this way it kind of degrades the more important sections, which are on the other hand a bit short.

Final remarks:

The comments should not hide the fact that the paper is a good data publication. The data are of high scientific interest and the data availability and quality appear to be very good. The paper has also been written in an appealingly clear manner. However, some important points should be changed or expanded in order to provide the best possible basis for future users of the data. What is meant above all is a significant expansion of the metadata and a more detailed discussion of the data quality.