Dear authors

Thank you for submitting this manuscript to HESS. Based on the review comments and on my own examination of the manuscript, I find that your manuscript requires substantial revisions to address the issues that have been raised through the review process.

While the paper contains interesting results, I believe that some of the review comments have not been fully addressed in your response. I strongly recommend that your address them fully in your revised manuscript which will be sent out for rereview. Please also write a reply to each of the editorial comments below:

Figure 3 needs to be changed as it is misleading. As it is now it suggests that the flood frequency has increased immensely in the past millennium. More likely this is an artefact of the data collection. The figure needs to be redrawn and the caption completed to make it crystal clear that the main pattern shown is observation bias.

Information is missing in the paper which needs to be supplied in the revision. For example, information needs to be given on what rivers individual floods refer to (eg. by colouring the floods shown by the river, giving percentages for each decade etc). It also needs to be stated whether the raw data used refer to individual locations or river reaches. How was a flood wave along a river counted, multiple times or once, and what was the threshold to decide among these alternatives?

In a similar vein, a much more detailed and accurate description of the sources is needed. Please include exact references to the main sources (rather than simply the source type). As you are probably aware, scientific papers need to be repeatable, and without accurate information it will not be able for the reader to trace where the information came from which, however, is a requirement for publication.

A more technical point is what is referred to as a “non critical approach”. Full justification is needed about how realistic this approach is and why it is preferred over the more usual critical approach, in particular the gain in accuracy over the critical approach.

The flood rich periods, apparently, are derived from fitting a polynomial to the flood frequencies per 31 years. This looks like an arbitrary method and needs justification. Why a polynomial? Why 31 years? The fracture point analysis is based on a similar averaging window so does not provide much extra information.

Finally, the English needs improvement.

In highlighting these points, it is not my intention to discount other elements of the reviews.

Günter Blöschl, Editor

Dear Dr. Böhm,
many thanks for revising the paper which improved its quality.
The reviewers have suggested a number of minor changes which I would ask you to implement.
Please submit a revised manuscript.
Best wishes
Günter Blöschl

Dear authors
thanks for the changes. There are still embarrassing typos in Fig. 3. This should read "annum", i.e. double n.
Please correct and resubmit.
Best wishes
Günter Blöschl

Thanks for the additional changes
Best wishes
Günter Blöschl