Even though the reviewers seem quite critical they also both acknowledge the value of the paper. Your response seem to address all the issues raised well and will lead to a paper that is interesting and also fits well within the scope of the journal (and of the special issue in particular).

I think your revisions makes good sense. Given that the reviewers were quite critical I will send it to them again for review.

Both reviewers have noted substantial improvements in the revised manuscript. However, they still have some comments I would like you to address.


I copy the comments from one of the reviewers below. Apparently his review was overdue by a few hours so he emailed them to me directly.

The authors have made a comprehensive revision of the initial submission resulting in a much more readable manuscript. As such I think the manuscript is close to a stage where it can be published, but I still have a number of issues/questions/suggestions:

Page 2, line 25: There might be a strong practical reason why engineers prefer to focus on peak flow rather than time, as flow magnitude determines hydraulic design?

Page 7, top. Maybe spell out that the hard work of collecting and quality controlling the data was done by USGS staff

Eq. (2) I am not familiar with the 'inf' notation. Please explain

Page 9, line 16: Please explain what ' a rough indication of the left tail properties' really means.

Page 11, line 12: A method from can 2004 hardly be referred to as a classic

Page 12, line 8: 'of the log-scale of ln(\lambda):'

Figure 2: I think the two graphs looks quite similar, i.e. full of black lines. Maybe chose, say, one year to better bring out the differences

Section 3.5: In general, I think this section needs more grounding in previous studies to better emphasise which findings are novel and which once are already well-known and even expected. Every hydrological model I know scales flow magnitude with basin area - anything else would be very surprising. There are several studies linking flow variation to basin areas as well, notably CV of flood events (e.g. Blöschl and Sivapalan, 1997). Basin control on skewness is generally more difficult to identify due to sampling variability of skewness indicators. With regards to urbanization, most models assume increasing urbanisation result in increased runoff. Notably Kjeldsen (2010) reported that this is not always the case, and sometime flood magnitude is less than expected in an urban basin. Kjeldsen (2010) also investigated the effect of urbanization on higher order moments (CV and skewness) and also reported a decrease in skewness for more urban basins.

Page 24, 2. paragraph: The attribution of results to specific features of the urban environment (dams, detension ponds, etc) is speculative as this has not been investigated in the study

Page 24, line 26-27: It would be more convincing if 'contrary to finding of many studies' could be made more factual with references to which findings and what studies.

References
Blöschl, G., and M. Sivapalan (1997), Process controls on regional flood frequency: Coefficient of variation and basin scale, Water Resour. Res., 33(12), 2967–2980

Kjeldsen, T.R., 2010. Modelling the impact of urbanization on flood frequency relationships in the UK. Hydrology Research, 41(5), pp.391-405.

Thank you for a thorough revision of your paper. I think it is publishable as-is now.