Thank you for addressing the comments by the referees. I feel a couple of comments require somewhat further addressing. Besides the already suggested changes, I would like to see the following treated.

1. In particular the notion of assessing uncertainty has not been addressed to my satisfaction. The authors say a number of times that it is difficult to assess uncertainty, or validate, and that using model output is justifiable because the focus of the paper is the method, rather than results. However, for a proper justification that the method is valid, an assessment of its uncertainty and/or validation against some real data (synthetic events against actual events from observations as suggested by Jongman) is required. The lowest hanging fruit to do this to my mind, is to benchmark the method against a set of observations, and perform assessment of the robustness of the method by testing a number of assumptions (for instance the GPD quantile threshold). This will require some additional work.

2. Compare against work by previous authors as suggested by Jongman. Even if this is qualitatively done, for instance by comparing a spatial plot of outcoming correlations of events against correlations found by Jongman et al., 2015 over a certain area would give a reasonable idea if the methods show the same spatial correlation patterns. I leave it to the authors to find a method to compare.

3. The comment on the used hydrologic/hydraulic model by Jongman needs to be better elaborated upon. They claim that "no hydrological/hydraulic model has been used for this study", but they do depend strongly on results by Alfieri et al. 2014, that are generated using a hydrology/hydraulics model cascade. The authors should explain what this entails, and discuss if and how the schematization can result in correlations that have no relation with reality but are a result of the schematization. Further "we do not have data of reality" may also not be true. There will certainly be basins in the study region where daily discharge data can easily be obtained by the authors.

4. the missing fields author contributions, competing interests and disclaimer have to be treated.

After revision, I will ask the reviewers for further review.

A few small points were left to improve. These are minor revisions after which I recommend the paper to be published.

1. The conclusions and the new section 5.3 in my feeling still do not make a strong case for the study. It should be better explained why a 10,000-year synthetic series of very limited coverage is still preferable to a 25-year discharge series for whole of Europe. Currently section 5.3 mostly lists issues with applying the model, rather than making a positive case.
2. Figure 2 still misses both a legend and a scale, including explanation what variable is shown in the background and what is the data source. Similarly other figures miss legends and/or scale. In Figure 7 the positioning of panels doesn’t seem to conform with the figure caption – please check.
3. The authors changed the confusing “large-scale” in the title to “continental-scale”, but expressions like “large-scale”, “The larger the spatial scale” or “large spatial scales” etc., was retained throughout the paper. Please make sure this is corrected in the text.
4. “River discharge waves generally propagate through the network in downstream direction”. I guess it should be “always” instead of “generally”. (P6L11)