Dear authors,
thanks for answering to the reviewers comments.
None of the reviewers asked to see the manuscript again, so I will be the one who will assess these minor revisions. Looking forward to receiving the revised manuscript.
Regards
Kerstin

Dear authors,

thanks for the revisions. I found several of them satisfying, but when reading the paper in detail now, I found there are crucial remaining issues related to the reviewer comments, which I do not see addressed sufficiently in the replies and added material. Also there are number of smaller editorial aspects relating to figures etc. that need to be improved, before I can make a final decision. I will keep the option open to appoint another reviewer again to advise.

Referee 1 in points 2 and 3 raised questions about detailed land cover and about the potential influence of snow accumulation over the years. Referee 2 found that streamflow is not enveloped in the simulation and the reviewers lacked some important background material. I think the comments are very important for the interpretation of the results and are not sufficiently elucidated and not transparently addressed. And they might be related.

The added short description and the new figure with some example time series of SWE are both not sufficient. In particular, it is not proven whether so-called snow towers (e.g. see Freudiger et al. 2017 in WIRES Water for a review) build up only in the accumulation area of glaciers and nowhere else (in the past but in particular also in the future runs). The manuscript needs to be more informative and transparent about spatial and long-term cumulative, quantitative, information about cryosphere water balance terms.
The main aim is climate change assessment and hence long-term changes. Hence, this long term change needs to be the benchmark for the model simulation in the past. In such an environment long-term accumulation of snow and the correct glacier evolution (such as retreat) are interdependent and will be primary determinants of long-term changes to the water balance (and thus changes in the hydropower potential which is cited as the underlying motivation).

Indeed, now after learning from the revised text that there is considerable glacier area in the modelled catchment, I need to request more modelling details a) how snow redistribution is handled in WaSIM and b) how the glacier(s?) and glacier change is/are treated within the model. Please also add exactly how much the glacierized area is in the here-shown modelled catchment, not the entire Clutha (as the new sentence says now) and in particular how much the streamflow contribution from glacier mass balance changes is. A correct representation of the long-term glacier change will be the best proof for a generally correct snow model as well!

The reviewers pointed out unexplained discrepancies in the observed and modelled hydrograph dynamics. The sentence on parameter uncertainly added is not a sufficient address to that concern.

May these differences and somewhat systematic errors be due also to the glacier melt modelling? The runoff generation not well represented needs to be better assessed and clearly explained and discussed - or transparently ruled out.

For the SWE-comparison of the two model variants, it would be necessary to be more transparent on how snow accumulation and melt are treated also within the glacier mass balance model component (normally, models convert snow to firn and hence 'extract' snow on glacier but not off-glacier from the water balance, snow may be differently redistributed or melt differently parameterized on or off-glacier ,etc.... - there are many potential differences that may affect the entire budget and hence the catchment-wide comparison of SWE and runoff).

If the ice melt contribution to generated runoff is too small to be relevant, the following comments may perhaps be disregarded. But if glacier ice melt is a substantial contribution (which the reader doesn't learn in the current manuscript), the necessary information on modelling must be added to the manuscript to make the study results credible and usable.

The Reviewers criticized the lack of description of the calibration procedure and the quality of the reference period simulation against which to assess future changes. Here, the dissertation by Jobst 2017 is cited, but if these details are relevant, and they are, and as this is not peer reviewed literature, the details need to be described here instead. Otherwise, it cannot be evaluated whether the quality of the model allows the conclusions drawn. Please include the calibration strategy and multi-criteria data information as it is necessary to assess how well the correct runoff generation processes are described and runoff is not generated for the wrong reason. As we all know, it is easy to model very seasonal glacial hydrographs because of the possible compensation of accumulation and melt by parameters of the hydrological model and NS efficiencies have limited value (Schaefli and Gupta, HP - Do Nash Values have value?).

Necessary information to be added also includes what happened to the glaciers between 1992 and 2012. What does that change entail for the calibrated parameters and resulting over-/underestimations in the validation period (see reviewer comments)?
Was this change also the reason for the (to me very unusual) choice of a short recent calibration period and long past validation period? From my experience in such environments long model spin ups are required and calibration over phases of wet/cold and warm/dry phases necessary. Short calibration periods may lead to biased storage fillings and thus biased results. The motivation and procedure of these choices also need to be more transparently described as well - reader cannot be referred to read a thesis to learn this important methodological reasoning.
In addition to the question how well was the long-term glacier change in the past was modelled (see former comment that this is the best evaluation of a good snow model), it is also important to learn what happened to the glacier in the future simulations as the loss or gain will be reflected in the water balance. A future hydrograph without that information is not useful.

In summary, the manuscript does not yet go far enough into modelling reasons and therefore my expectation is that it will also not have the impact it perhaps deserves. The conclusions need to go beyond the general warmer-less snow paradigm to make an interesting contribution to international science in the field. Therefore, I need to ask for further revisions regarding more transparency of the modelling details and more in-depth analyses of the reasons for the difference and range in changes founds.


Further Editorial comments

Figure 1: Please make this figure more specific to this study! Legend for the red polygon is missing. What is this? Is it relevant for this study at all? Otherwise please delete this polygon. 'snow cal' is not a nice or useful abbreviation here - please write out what is meant with that in this introductory figure. The order of a, b, c is very confusing/unusual. Please switch graphs or a, b,c to go from upper left to lower right as is usually the custom. In fact, a) is really not needed and anyway very dark colors hard to see anything. Instead, it would be more useful to give more landcover details of the modelled catchment, such as glaciers, forest vs grassland, etc. Most of what is in the map c) is not modelled and investigated in the study anyway (reservoirs and lakes only mentioned once as overall motivation).

Figure 7: colors in the legend are not consistent with colors of shaded model ranges (orange needs to be paler, black should be gray). Please adjust.

Figures 7-9 Axis labels are too small and too close to the axes. There is enough space to make them considerably larger and keep a bit of distance from the axes.

Figure 10a is not possible to understand without any indication what is on the x axis - what are the different ranges? The colors repeat, so it is not at all clear to me (and likely other readers) what the difference among these bars is.
10b-d. I suggest that these are sufficiently different to warrant a separate Figure 11 and then larger panels. At this size, it is hard to see differences clearly. Also, perhpas since the provided figure resolution is not acceptable. I assume that will be much better in the final production, but still, I think these are too small.

Table 1: give units where applicable

Table 3: why is SWE given in km3 and not also in mm similar to all other terms. This is not only highly unusual, it also doesn't allow comparison. Please change.

Dear authors,
thanks for the clarifications and careful suggestions for revisions, which now make the case and boundaries much clearer to readers. It will be good to see the most important clarifications on snow and ice contributions included, but I agree that it should not overwhelm the final paper and take away attention from the main point.
Regards
Kerstin