|Due to the limited time for this review round, I could only take a look at the revisions of my comments (and not the ones from all reviewers). Since the paper has been modified substantially, there is the possibility that my comments may contradict the ones from the other reviewers.|
Most of my major concerns from the first round of revisions have been addressed. There are however still two issues that remain
- Clustering: I acknowledge your line of argument that the actual number of clusters is not the main focus when making a point that the scaling can be classified based in the inflections points. And maybe you can justify 6 clusters, but when looking a figure 8 (in the revised manuscript), categories 1 and 2 or 4 and 5 look very similar to me. So why not go with 4 categories instead? The interpretation of the results in section 3.5 is based on those 6 categories but this interpretation may turn out quite different when fewer categories are chosen (more probably won’t make any sense). And if a different number of clusters leads to different results and subsequently to a different interpretation that would an issue that needs to be addressed. If not, then it would support the choice of six classes.
Did you look into this and could you comment on this? Maybe a supplement or appendix showing results for different clusters would be an option?
I still don't understand why you are going into the imputation bridge part. You are definitely correct that an interpolation of NaN does not make any sense in the case of rainfall. But you could justify this simply by stating that there could be e.g. extremes that occurred during a gap (and thus remain undetected) or that an interpolation between two values before and after a gap would lead to false information because you do not know what happened in between. But what is the message you a trying to convey in Figure 2b? A map with maximum values between gaps does not really make sense to me because it is not stated how many gaps there are in total and how long these gaps are. Some statistics about the number and length of these gaps would be more meaningful from my point of view.
Minor comments (line numbers referring to the manuscript with track changes)
First of all, the manuscript should be proofread again. There are still numerous errors, especially in the changed passages. Furthermore, for the sake of readability, some sentences should be formulated in a more direct and compact manner.
- avoid e.g. “spatio-temporal” vs. “spatio temporal”
- try to avoid double brackets like “(Fig. X (a))” or similar
- on more than one occasion you write that “pixels experience” something. This sound strange and I’m not sure if this is correct English. The same goes for e.g. “the entire of Germany”, “the whole of Germany” and similar.
Numbers: use “,” for separating thousands, i.e. 1,000 etc. This is inconsistent throughout the manuscript, including the figures!
Regarding the use of “world record”: I made a mistake in my first review, I wanted to say that this term should NOT be used, sorry about this. In my opinion it sounds more like yellow press than scientific paper. Maybe something like “global precipitation extremes” is better...
Specific corrections (without claim for completeness)
Abstract: I suggest you write years, hour and day instead of yrs,h, and d here
- l. 1 investigated (the rest is past tense as well)
- l. 19 “urban and non-urban flash floods” → are there any others? Just say “flash floods”
- full stop missing the same line
- l. 21 “needing”? Sounds strange, rephrase
- l. 22. “Obstacles to identifying and investigating extremes” → sounds strange, rephrase.
- l. 23ff. makes no sense, “50% obwhich observed extremes”(?), “even more”(?) This is not clear, be more specific!
- l. 30 from may point of view, attenuation is also a major reason for uncertainty because this may reduce extremes
- l. 34 this is a 1:1 quote from the AMS Glossary of Meteorology. Mark this accordingly (or reformulate) and update the source to American Meteorological Society, 2020: Climatology. Glossary of Meteorology, http://glossary.ametsoc.org/wiki/
Figure 1: I suppose you removed the triangles because another reviewer suggested this. I my opinion, the first version with the triangles was better….
l. 82 change to “extremes found”
l. 85 “...a reprocessed...” delete “carefully”
l. 101f. change to “One QPE from German radar data RADOLAN (German: RADar OnLine Aneichung)” and add a citation
l. 111 “two formats” You mean “two versions”, I suppose the format is the same
l.117 due → because of
l. 118 “have been”
l. 123: “values (compared below) compared to RADOLAN” repetition, rephrase
l.125 I still think that the details about files sizes are not really relevant and this sentence can be deleted. If you decide to keep it nevertheless: “making” sounds wrong in this context
l. 131 “Some time series” → you mean “data in some areas”?
l. 136 “...red spots could mean...” sounds strange, rephrase
Caption Figure 2 “needs”
l. 144 “than the missing of any extreme values.” sounds strange
l. 149ff difficult to understand, reformulate
l. 164 “built” → developed?
l. 171 it is called “k-means”, sometimes the K is also capitalized, please check and use this consistently
l. 198: “xx minutes” (?)
l. 199f “Hesse state” vs. “Bavaria” → stick to one convention. Furthermore, what the purpose of mentioning that this has not been documented in the news?
l. 205 What are “maximum locations”? Do you mean the locations of maxima?
Table 1 is a good idea, but much too detailed. Reduce to lat/lon to 2 trailing digits, precipitation to max 1, use CET and 24h Format for times, Location is WGS84 not WG84
l. 209ff consider deleting brackets with the ranks, is does not enhance the readability. It’s sufficient to mention this in Figure 4
l. 213 “3-regime-form” ->?
l. 215 “Figure 5 shows the location of the high quantile rainfall” → delete the numbers for the sake of readability
l. 216 reformulate this sentence, this is not correct
l. 218-220 same as l. 216, hard to understand, rephrase
Legend figure 4: “392,128 in total” or “n=392,128”
l. 230 please note that the red and yellow spots are difficult to see in a printed version
l. 233 similarly located → rephrase
l. 241 why don’t you mention that the reference is a single power law right here?
l. 245 what does “attached to the maximum rainfall event” mean? Rephrase!
l. 254 “The k–mean clustering algorithm successfully classified the depth–duration relationship into six categories” -> Of course, because you defined 6 classes, what’s successful about this?
l. 257 “grid points” not elements
l. 257ff. Why to you capitalize “Category”?
l. 261 “beginning of the curve”
l. 265 smaller pixels → I assume that your pixel size is constant
l. 267 “Category 1’s” possessive s in combination with a number looks strange
l. 268 repetition of “location”, rephrase!
Caption Fig. 8: Rephrase to “Resulting clusters of the maximum depth-duration relationships”
Figure 9 “??” in caption, missing reference?