Yang et al. Introduced the structure and performance of an Early Warning System installed for monitoring ice collapses and river blockages in the southeastern Tibetan Plateau. The EWS has successfully warn three collapse-river blockage events and seven small-scale collapses. They also found that the volume and location of the collapse and the percentage of ice content contribute the different velocity of debris flow and magnitude of rive blockages. Such work represents an important contribution to the cryosphere disasters monitoring and warning on the Tibetan Plateau. Although the manuscript provided good descriptions of EWS and their performance, there are a number of issues below that need to be addressed prior to publication.

1. The introduction needs more background on the ground-based early warning systems that have already been installed around the world. The author states that few successful warnings have been reported. Could the authors provide some description about the question of EWS performance? In line 51, the author states that there are no EWS for ice collapse on the Tibetan Plateau. Are specific instruments and structure of EWS needed for ice collapse? More explanation should be added to the manuscript.
2. Section 2 “Study region” and Section 3 “Historical ice-rock collapse around ......” could be merged as Section 2. Study region and historical ice-rock collapses.
3. Figure 7 and the supplementary, the seismic waveform of two abnormal waveform (showed by green arrows) need to be displayed in the supplementary. Are these waveforms typically different with the normal waveform introduced by ice-rock avalanche? Maybe the small-scale rock avalanches without debris flow formation.
4. Line 276-278, the author claimed that both the maximum amplitude and the duration of seismic waveform are useful information for early warning the ice-rock collapse. Did the EWS incorporate this information in the system? In Line 174-175, the warning system seem to only consider the thresholds as the only warning information. More explains should be added for introduce the automatic warning system in the manuscript.
5. Line 295-Line 300. The author provided the values of anomalous glacier thickening on three risking glaciers in the ablation zone. The boundary of this ablation zone should be delineated in Figure 9.
6. The EWS in this manuscript did not include the GNSS for monitoring glacier displacement, which is the direct indicator of glacier abnormal change and the popular way for disaster monitoring (landslide and rock avalanche). Could the GNSS be used on the glacier surface, in particular for the risking glaciers such as the Zelongnong Glacier and the other glacier RGI60-13.01430 for early warning their abnormal dynamic changes?
7. In section 6.2, the author stated that all-weather avalanche radar for real-time monitoring. Do the author know the similar instruments available for EWS?

Minor comments:

1. Line 130. The video of the disaster process is retrieved remotely.
2. Figure 1. Yarlung Tsangpo, Figure 4. Yarlungzangbo River. I suggest consistent use of Yarlung Tsangpo throughout.
3. The quality of Figure 5 should be improved to be more than 300Mdpi. The red arrow was not displayed fully in Figure 5a. and the major tick lines in X and Y axis should be added.
4. Figure 9a, the units is m or m/year?

I had the chance to review the paper ‘Early warning system for ice collapses and river blockages in the Sedongpu Valley, southeastern Tibetan Plateau’. The paper analyses the performance of a three sets early warning system in Sedongpu valley to detect ice rock collapse and alert the communities. Such an analysis is very useful for the emergency and scientific community and while the novelty of the approach is moderate, I recommend this paper for NHESS with major revisions.

Overall, the paper is very interesting, and the data well presented. The strengths of this paper are the EWSs setups which fed the paper and made the analysis relevant. The efficiency of EWS 1, 2 and 3 data is well described. Evidence on the waveform characteristics of a flow is very informative as well as the the dilution of the flow when rocks and ice are mixed. The charts are clear, although I have made some suggestions below. While this paper will find readers from the scientific and the emergency stakeholders, I would recommend to further improve the scientific analysis and take full advantage of this EWSs’ network. The EWSs network introduced here offer a great source of dataset which could be further exploited in this paper. Particularly, EWS1 meteorological records should have been analysed to understand the role of weather parameters on ice-rock detachment (Temperature? Precipitation? Radiation?). EWS1 was established in May 2022, therefore it would be very useful to look at the data records prior to Ice-rock collapses and river blockage on 14 May 2022 and to the repeated ice-rock collapses on 11 August 2022. If there are any anomalies prior to the events, those must be mentioned in the paper to better understand the weather conditions which trigger such hazards. This is the main, if not the only, gap of this paper. Please, see some of my comments in details below.

Scientific Significance: Fair. All statements are supported by clear evidence which contributes to the understanding of these natural hazards. However, weather parameters should be further exploited to explain the weather conditions of such hazards in Sedongpu.

Scientific Quality: Good. The paper is based on exclusive observations of two remote rivers. Some references could be added to the paper to better place this paper in a global context.

Presentation Quality: Good. The data are well described and clear. Again, the weather conditions of these hazards are too concise here and should be further explained.

Line 27: can you provide a figure of the ‘rate of air temperature warming’?

Line 46: ‘Given the short duration of glacier collapse, it is difficult to provide timely

warnings of glacier catastrophes and assess their impacts’, add ‘using remote sensing only’ at the end of this sentence, to emphasize your point about remote sensing’s limitations.

Line 49-51: provide more explanations about the difficulties faced with those EWSs. The reader needs to understand the limitations of EWSs, and how you are going to address these issues in your paper.

In the objective paragraph line 52 to 59, I would suggest being more specific on your objective, rather than mentioning the EWS only. For instance, on what real-time physical measurements the real-time warnings signals are based. Water level and other parameters such as soil moisture or precipitation? We find information about this only in caption of Figure 1.

How your EWS is more transferrable to other regions than other EWS, as stated line 58-59. We need to have a clue of the real novelty of your approach at this point of the paper. All EWS use real-time signals.

Line 57: ‘different types of ice-rock collapse’, define them here in few words.

Line 54: I would, conventionally, keep ‘Second Tibetan Plateau Scientific Expedition and Research Program’ mentioned line 54 to the acknowledgement section.

Line 67: “There were two peaks over 7000 m above sea level (m asl): Mt. Namcha Barwa (7782 m asl) and Mt.Gyala Peri (7294 m asl).” Please change ‘were’ into ‘are’, those peaks still exist.

Line 70: when the summer monsoon is mentioned, I suggest to provide the annual precipitation mean of Sedongpu Valley, and its seasonal distribution (% per season). Moreover, it would be useful to know the share of liquid and solid precipitation. Heavy precipitation, and therefore saturated soil moisture can have an impact on ice-rock collapses.

Section 3 on the history of ice rocks collapses is very informative and interesting. I suggest however to move it to the study area section (section 2) as it is not part of the EWS’ results but rather a review of previous studies on this region. Moreover, information about the consequences (or even the absence of severe destructions) of the cited debris flows on villages in the lower part are needed here, in order to better understand the challenges of this valley. What happened to the villages downstream or the bridge of Gyala after the 22 October 2017 for example, once the water was released?

Line 115: ‘Since 2019, the EWSs have been built around the Sedongpu Valley.’ ‘Along’ the valley would better depict the distribution of EWSs based on Fig 1.

Line 136-137: it would be helpful for the reader to have a brief explanation about the recurrent timing of the small-scale collapses at midnight. Is it due to accumulated solar energy during the day, frost-thaw, or any other factor?

I suppose the meteorological parameters were recorded on an hourly basis. Please precise it in the text.

Line 184: what sensor was used and what resolution?

Figure 3a, 5a, precise the units of the amplitude.

Line 212-213: ‘which agrees well with the delay between the observed ice collapse in October 2019 and its delayed debris flow (Zhao et al., 2022)’, please provide a value (hours) of the time of response between the collapse and the actual water level rise on that day.

Figure 5b, the photo is not visible and deserves a better resolution. I suggest increasing its size below Fig5.

Line 226: ‘EMS1’, EWS1?

In Figure 8 (very informative), it would be better to have the same amplitude axis to better visualize the difference between collapse and seismic waveforms.

In Figure 8, add the Hugonnet and Millan references in the caption.

In the discussion, you may want to go through Tiwari et al., 2022 and compare your approaches using amplitudes. (Potential seismic precursors and surficial dynamics of a deadly Himalayan disaster: an early warning; https://doi.org/10.1038/s41598-022-07491-y). It’s worth adding this reference in the discussion section. See also Cook et al., 2021: ‘Detection and potential early warning of catastrophic flow events with regional seismic networks’, DOI: 10.1126/science.abj122

Review nhess-2023-38

The authors present and impressive EWS setup in the Namche Barwa massif, an area that has been subject to multiple hazard cascades in recent years and decades. They present a combination of sensors at different locations and based on a limited number of months of observations are able to track movements and show the potential of these setups for early warning. Especially useful, as has been shown in other sites, is a geophone that is able to pick up mass movements of different magnitude well before they become the actual risk when damming the main river channel.

The study is well presented and the setup and the insights present a valuable contribution to NHESS. Before recommending it for publication the authors however need to embed their findings into the state of knowledge on EWS in the region, especially in light of their geophone/wavelet analysis angle. Other studies exist and those need to be brought into context to make this not just a description of a successful first test of the setup but a contribution that will be helpful for the scientific community. NHESS requires this to be a novel contribution, and at the moment I miss a discussion of the added value (apart from of course the value the specific EWS has for local stakeholders). While this requires some more work on the Discussion, I believe that seeing the quality of the manuscript up to this stage this should be well possible to do.

I detail my main concerns below, followed by a list of minor comments.

Main comments

L153: What are the ‘various variables'? There is also no discussion on how the meteo data is useful or used. Are they required at all? What role do they play in the EWS chain of communication/analysis?

Figure 4: It is very hard for me to make out changes on these images. In a/b can you indicate flow direction, what the deposit and what water is? In c to f I can not really see change between the two images. Can you indicate what change the reader should perceive? It would also be important in the text to clarify how these images are used in early warning – to validate? Is there an algorithm that traks changes on the images as was the case at Kyagar? Stick to one writing of Yarlung Tsangpo/Yarlungzangbo

Figure 6: As in Figure 4 I find it hard to see anything in c to f. Would it be possible to zoom in for the purpose of the manuscript? You need to tell the reader how these images are useful for interpretation.

L253: Here and then again the Discussion/Conclusion I would expect some critical reflection on potential pitfalls. Why for example did EWS2 not send a warning when the debris flow mass must have reached the location before midnight? Or did it only reach the location that late? You defined a threshold for amplitude based on a few observations, how is this robust? What about false positives or false negatives? The amplitude of the first August 2022 event was larger than the second but you say the second event was larger in volume. What about the smaller events well below 20, would your system then not record them at all and send no potential warning? How does expert judgment of these data and images work, who interprets it and who takes final responsibility for the judgement made? You also introduce a ‘three level warning system’ based on different amplitudes (L264) but then never explain where this comes from or how it is used. This needs to be clarified.

In Figure 8 and the text above you very briefly touch upon this topic but I think this needs to be further expanded. What about the noise before the rock collapse, not visible before a rupture, could this be exploited as a prewarning signal?

L282: The Discussion misses two crucial points:

1. a discussion of the presented approach keeping in mind similar approaches, one of which you mentioned but did not discuss (Maurer et al. 2020) and one that didn’t occur at all (Cook et al. 2021). Both these studies have discussed the potential especially of wavelets in the region for similar events and it is crucial that for a scientific publication the previous literature is looked into and put into context for a solid conclusion on a way forward.
2. A discussion of the institutional challenges. The authors are aware of the challenges within China (e.g. Kyagar) but also with other countries of the region (e.g. downstream Nepal) to make such systems work. How is this setup managed locally for success? Are recommendations understood and transported on? What stakeholders need to be included. I realize that this is not the main part of the manuscript and also doesn’t need to take up a part in the Results but should surface in the Discussion. Otherwise many of the technical observations often become unsustainable. Wanting a radar is nice, but so expensive that it is hardly ever feasible and you need to chain of command at the end of a signal too to turn such data into actual actionable information.

General/minor comments

L32: remove ‘giant’, not a relevant descriptor

L45: ‘before a glacier collapse’

L45f: ‘…data is subject to favourable weather conditions and revisit cycles.’

L46: ‘a glacier collapse’ – also state here what you mean by short, what is the time frame to be expected from known cases.

L47: ‘glacier catastrophes’ is a new term you now use while earlier you used ‘glacier-related disasters’ or just specifically ‘glacier collapse’. I would stick to one and only use another if strictly necessary to describe something else.

L49: remove ‘regimes’

L67: ‘are’ not ‘were’

L128/L151: As you are here describing a setup, details for all sensors are crucial. What kind of thermal and why a thermal camera - refer to the Table where appropriate.

L180: ‘is briefly described’

Figure 3b: Legend missing making clear what is level and what is precipitation

L262: Spelling is generally ‘Medog’ in local Tibetan or if you use Chinese placenames ‘Motuo’

References

Cook, Kristen L., Rajesh Rekapalli, Michael Dietze, Marco Pilz, Simone Cesca, N. Purnachandra Rao, D. Srinagesh, et al. 2021. “Detection and Potential Early Warning of Catastrophic Flow Events with Regional Seismic Networks.” Science 374 (6563): 87–92. https://doi.org/10.1126/science.abj1227.

Maurer, J. M., J. M. Schaefer, J. B. Russell, S. Rupper, N. Wangdi, A. E. Putnam, and N. Young. 2020. “Seismic Observations, Numerical Modeling, and Geomorphic Analysis of a Glacier Lake Outburst Flood in the Himalayas.” Science Advances 6 (38): eaba3645. https://doi.org/10.1126/sciadv.aba3645.