Comment on hess-2021-499

In this manuscript the climate sensitivities of runoff (catchment, glacierized and nonglacierized parts) in two contrasting glacierized Himalayan catchments are analyzed. The obtained climate sensitivities are then used to derive a measure for the standard deviation of streamflow and to make projections of how streamflow will change under climate change. A hydrological model combined with a glacier melt model is used to simulate streamflow timeseries for the two catchments. Instead of deriving the streamflow variability and the future changes directly from the model output, the model simulations are used to get the climate sensitivities of summer runoff to annual precipitation and summer temperature. This step is likely needed because the hydrological model only includes a static glacier. The study comes to the (not very novel) conclusion that glacierized parts of the catchment are sensitive to temperature and that the nonglacierized parts are sensitive to changes in precipitation. It also suggests that climate sensitivities can be used to estimate magnitude and timing of peak water, but it is unclear how climate sensitivities should be derived for ungauged catchments and what the advantages are of not directly estimating peak water from (glacio)-hydrological models. Overall, I think that, despite the interesting topic of climate sensitivity of glacierized catchments which is suitable for HESS, the study does not do a good job in addressing a clear research gap. Besides, I have some doubts about some parts of the methods, there are some unclear descriptions and there is a lack of discussion on the simplifications, interpretation of the findings and the implications. Please find below my major and minor comments.

In this manuscript the climate sensitivities of runoff (catchment, glacierized and nonglacierized parts) in two contrasting glacierized Himalayan catchments are analyzed. The obtained climate sensitivities are then used to derive a measure for the standard deviation of streamflow and to make projections of how streamflow will change under climate change. A hydrological model combined with a glacier melt model is used to simulate streamflow timeseries for the two catchments. Instead of deriving the streamflow variability and the future changes directly from the model output, the model simulations are used to get the climate sensitivities of summer runoff to annual precipitation and summer temperature. This step is likely needed because the hydrological model only includes a static glacier. The study comes to the (not very novel) conclusion that glacierized parts of the catchment are sensitive to temperature and that the nonglacierized parts are sensitive to changes in precipitation. It also suggests that climate sensitivities can be used to estimate magnitude and timing of peak water, but it is unclear how climate sensitivities should be derived for ungauged catchments and what the advantages are of not directly estimating peak water from (glacio)-hydrological models.
Overall, I think that, despite the interesting topic of climate sensitivity of glacierized catchments which is suitable for HESS, the study does not do a good job in addressing a clear research gap. Besides, I have some doubts about some parts of the methods, there are some unclear descriptions and there is a lack of discussion on the simplifications, interpretation of the findings and the implications. Please find below my major and minor comments.

Novelty and research gap
The introduction of the manuscript is very minimal, it touches on a few topics but does not show how this study fits in between previous studies and it does not clearly explain the research gap and what the study aims to achieve and why. The sentence 'Due to a lack of long-term data…. may still be lacking' does not do justice to all the studies that exist on streamflow and its projections of Himalayan catchments. Here I would expect to read what climate sensitivities can add to the existing (modelling) studies. Then in the second part of the introduction, the explanation of how climate sensitivities are related to long-term changes and the glacier compensation effect are very unclear. What to do with the sentence 'Climate-sensitivity based predictions for future changes in runoff are reliable ………..over the calibration period'? How does that match with the peak water exercise in the manuscript? The relation between climate sensitivity and glacier compensation effect is also not clear and requires more explanation. In the last sentence 'We also attempt to do this and that (glacier compensation effect and peak water)' I miss reasoning on why these attempts are needed.

Methodology
The workflow in this manuscript is not completely clear to me. The aim of the study is to assess climate sensitivities, because those can help to understand the variability and changes of streamflow. Since there is only limited streamflow data available, timeseries of streamflow are simulated with the VIC model. However, in theory, such models can also provide information on variability and change, so as a reader I need some argumentation why climate sensitivities are a useful alternative route, especially when there are no or only few streamflow observations available.
The simulations of streamflow are crucial here for the derivation of climate sensitivities, and I am surprised by the similar sensitivities of the two catchments, while their precipitation seasonality and mass balance type are so different. How is snowmelt simulated in the VIC model? Is there a different parameter for snow and glacier melt? If there is snow falling on the summer accumulation type glacier, is melt then also reduced in the model (albedo effect)? Have you tested if there is a difference in summer runoff sensitivity to summer precipitation in the non-glacierized parts? How is ET modelled? This should be important for the non-glacierized runoff sensitivity to temperature. Regarding the parameter sensitivity tests, were the optimized DDF and ap parameters fixed? Low parameter sensitivity may suggest that the model is not very suitable to model the system. Also, summer runoff may not be the optimal variable to test with, as timing of melt and snow/rain ratio will be important to model right to extract the sensitivities in a meaningful way. Could Qg and Qr for the 40 year of simulations be easily plotted, and compared with other modelling studies?
Climate sensitivities are derived for catchment runoff, glacier runoff and non-glacierized runoff. There is a formula given (eq4) for how to derive catchment runoff sensitivity from the glacier and non-glacier runoff, but, if I am right, it is not used for the results. Has this been tested for?
In Eq. 8, the changes in runoff due to changes in glacier cover are estimated by the recent difference in runoff from the glaciers and the non-glacierized parts. This, however, neglects the process of usually increasing precipitation with elevation. For large changes of glacier cover this may become quite relevant. Also, assuming 'the recent ratio of winter to summer runoff remain unchanged' contradicts many previous studies of increased winter flow and decreased summer flow. If these assumptions need to be made, I wonder how the results could be used, as many of the models do actually include these kinds of feedbacks.
Throughout the results section, the climate sensitivities are presented as mm change per change in degree C or per change of mm of precipitation. Based on these outcomes, some sensitivities are regarded as zero. However, these results are misleading if they are not communicated in how much T and P varies per year. In general, it would be helpful, I think, to communicate them in percentage from the mean flow, and also present all of them in an overview table.

Unclear descriptions
Throughout the manuscript there are quite some words missing or misspelled (please carefully check!), and unclear descriptions or presentation (see also list below). For example, units are missing in equations, and there is a mix of units in m and in mm.
The bias correction methods description is very unclear. Apparently, temperature is corrected based on station data, but precipitation not and instead is corrected via a calibration parameter. Why is that? How is the meteorological input data used in VIC? Are T and P lapse rates used? If so, how are they obtained? For the VIC modelling, how does the coupling between the glacierized and non-glacierized parts work? Is there snow redistribution from the non-glacierized parts onto the glacier? And does glacier melt contribute to baseflow?
How are the mass balances calculated? Per catchment or per glacier? And how are they compared with available data, per glacier or per catchment? How is the glacier runoff modelled in a similar way to Huss and Hock? Are the same sensitivities as for the nonglacierized parts used for the parts that get de-glacierized?
In general, there are a lot of references to supplementary material, and I would suggest to better describe some of these in the main text.

Lack of discussion
Section 4.9 is quite a deception to read. Basically, it summarizes the methods to obtain streamflow simulations. The approach used in this manuscript is very theoretical, and at least in the discussion section I think a translation again to the glacio-hydrological processes is needed (e.g. compensation of the glacierized and non-glacierized runoff parts, connecting precipitation importance for mass balance to changes in summer streamflow, describing why temperature is not relevant for non-glacierized parts, interaction of P and T processes).
Also, as mentioned before, it would be good to show what can be learned from these derived sensitivities, how can this approach be implemented to derive sensitivities in other catchments, or how do these results give a different perspective from what we already know?
And last but not least, I think a comparison with other climate sensitivities (also outside the Himalayas) is needed (e.g. He. 2021, Engelhardt et al., 2017, Moore et al., 2020, and some reasoning why there where no differences found between the two catchments (summer acc. types are thought to be very sensitive to temperature) and/or the differences in peak water timing in the two catchments, and the large differences in temperature sensitivities found in the studies that you cite in section 4.5