The paper presents a statistical analysis of trends in streamflow in western Austria, pointing to differences in high altitude and low altitude stations. Mostly positive trends in annual runoff at high altitudes are attributed to more glacier melt and mostly negative trends in annual runoff at low altitudes are attributed to increased losses due to higher temperatures. Analyzing daily flows the authors conclude that the main drivers of alpine streamflow changes are increased glacial melt and earlier snowmelt.

The topic of the paper is timely and interesting. There have been a number of recent papers published on this topic in the European Alps and elsewhere. This also presents a challenge, because different methodologies and datasets are used, stress is placed on different aspects, some conclusions cannot be generalized. In this sense the authors have chosen their path and executed it in a series of two papers. The problem is, as the authors correctly state, a problem of attribution rather than trend detection. In this sense, their work also does not prove the causality of streamflow trends, rather finds some relations between intuitive variables which should explain it. Indeed, although data-gleaned evidence of change is the most reliable, causality can be convincing only if data are combined with some hydrological process understanding/modelling including uncertainty. In this light data studies such as this are a valuable contribution.

The paper was reviewed by 4 referees, all of whom are experts in the topic covered by the paper. The reviews were constructive and quite critical, 3 of the 4 suggested major revisions of the paper. I recommend therefore to the authors a major revision of their work which will address very clearly the concerns of the referees, in line with the posted author’s responses. Three of the referees have requested to review the revised paper, so there will be another fast review round of the final revised submission.

I would especially like to highlight the following key points for the revision that the referees raised.

1/ Ref 1 states that the presented trend assessment does not provide a confirmation of the causal physical processes. He correctly advises more careful interpretations in the paper to the meaning of the results. Ref 3 asks for more details on the most important methods applied. Overall practically all of the referees found parts of the work cumbersome to read and understand (I concur with them) and suggested many improvements to the readability of the paper in their detailed reviews. Please pay very close attention to these in your revision. Don’t worry about keeping the manuscript short - focus on making it understandable.

2/ It is also a good recommendation to prepare a new Discussion section in your paper where you engage closer with the literature that is directly relevant to your results as well (Refs 1 and 2). Do not present the results from your earlier paper Korman et al. (2014) anymore, but be very careful and explicit in specifying what is the novelty in your HESS paper. I have read the preprints and your explanation of the content and aim of this work and conclude that overlap with the HESS paper is there, but is not excessive. However duplication of the presented results in both papers is unacceptable.

3/ Several of the referees raised the fact that there is other existing research you may have missed, and also correct your interpretations of past studies (especially Refs 2&4). Check that throughout the paper the interpretations you ascribe to past studies are accurate. This is very important.

4/ I would like you to explain more the issues of trends or shifts in your analysis, and how you separate them. This is more than a terminological issue (e.g. see Dery et al., 2009), a simple shift in time in a hydrograph will inadvertently lead to (possibly very strong) rising and dropping “trends” in time. These may have nothing to do with an overall increase or decrease in runoff, and can affect your interpretation and causality analysis. Can you please provide a better explanation of this fact in your revised text where appropriate?

5/ You state on several occasions that yours is a high resolution study as opposed to annual, seasonal, monthly sums. In principle I agree, since you are directly using daily data. However the 30-day moving average operator in reality goes back to monthly sums (Ref 2). How does this averaging window impact the results? In your response you state that it reduces fluctuations, I suggest you engage more with the meaning of the averaging, especially in identifying dates of change and dQ/dt terms which are dependent on it, even if you have not explicitly tested other averaging windows.

6/ Please explain better the estimation of the minimum detectable trend (Ref 3).

7/ To the point of attributing melt to ice/snow and the point raised by Ref 4 that in spring glacier melt cannot dominate runoff, please engaged in the revised text more with this issue.

8/ Several of the referees raised issues about regulation effects on the streamflow data (Refs 1&2). I understand of course that regulation cannot be completely discounted, as so many rivers are to some degree regulated, but please engage with this topic more in the paper. Also connected to station choice are nested basins which are not independent (Refs 2&4) and you have explained your choice in the response. In the regression analysis you use climate, snow and streamflow gauging station data. It is not clear how the assignment of climate and snow stations to each streamflow gauging station was made.

9/ The hourly analysis is not well connected with the rest of the paper (Ref 4). Consider how important this part is to the message of your paper.

Peter Molnar

Dear Authors,

I will not wait any longer for the third reviewer to submit their review of your revised manuscript. I am in agreement with Referee #1 that a minor revision of the text is still necessary to make the paper stronger and clearer. I ask you to please take action on the items that the Referee #1 indicates in his review (I copy it below). There will only be an Editor Review after you have completed this last revision.

Best regards, Peter Molnar

===
REFEREE #1

I appreciate the effort given into the revision. Currently I have only one main and a couple of minor comments on the manuscript. The main comment relates to the fact that authors tried to accept/include many of the reviewer's comments, which however does not result always in a coherent and clear story. I would suggest to have an additional look on the story and condense the text a bit in order to highlight more the important findings and clarify the storyline. For example, some parts of the Introduction section are not very informative/relevant with respect to the main objective of the paper - with a consequence that the overall story does not clearly converges to a clear message about the missing gaps in the literature and motivation to the research questions investigated in the manuscript. For example p.4, l3-4 is a typical example of an "empty sentence". Or paragraph p.4 l.10-16 which starts with introducing other approaches, but lines 14-16 refer to declining snowpacks. This makes the manuscript (story) somehow difficult to read.


Minor comments

1) Abstract: trends on daily basis. I would suggest to be more specific here, e.g. by saying filtered daily data. (the same applies for the results or discussion).

2) Section 2 - Data: Please consider to switch the Data with Methods. I would say that the main messages are connected more with applied (new) methodology than with uniqueness of applied data. So I would suggest to present methodology first and data after. The main comment applies also for the Data section, which is rather unsorted. (E.g one paragraph refers to streamflow gauges, next to the basin characteristics, then the period used and then again streamflow).

3) Methods. From the headings, it is not clear what is the difference between 3.1 and 3.2, both refer to trend detection. Please consider to condense some parts. E.g. Section 3.1.2 is too long and does not refer only to the methodology applied in the manuscript. (the section referring to global precipitation estimation is confusing here.) or paragraph on p.10. l.20-24 could be condensed (if needed). Some parts of the methodology are likely more related to the discussion (p.11, l.18-24).

4) Table 1. Please consider to provide also maximum elevation to support the interpretations about lower and higher catchments.

5) Figure 8. I like to see a real hydrographs. Please be more specific and indicate which years the hydrographs represent.
===

Dear Authors,

I appreciate your effort to shorten the text. I think you could have gone further in that effort and been more concrete in your conclusions, but overall I judge that your work provides a coherent and informative picture of the streamflow change in western Austria especially insofar as the confounding effects of altitude, temperature and snow-related processes are concerned. I hope your paper will be considered as a valuable contribution to our Austrian colleagues working in these areas. There are some minor grammar mistakes and typos in the text. Please fix those at the proofing stage.

Best regards, Peter Molnar