Reconstructing the History of Glacial Lake Outburst Floods (GLOF) in the Kanchenjunga Conservation Area, East Nepal: An Interdisciplinary Approach

Round 1

Reviewer 1 Report

This manuscript elucidates the history of flood events in the Kanchenjunga region, Nepal, by combining interviewing of nine local people with satellite and field-based identification of past GLOF events. In addition, numerical modelling is used to estimate wave amplitude and simulate discharge time series of the 1980 Nangama flood.

 

The study provides some new knowledge on historical GLOFs in this region, but several revisions are needed before the manuscript has the quality to be published.

 

Major issues:

1. The text suffers from being repetitive and needs restructuring and shortening. I have made some comments below to help the authors in this process, but I recommend that the authors carefully read through the manuscript to remove superfluous text. The results will appear clearer to the readers, if the text was more brief but comprehensive.
2. The description of the modelling must also be much more concise and scientifically rigorous with respect to assumptions, input parameters, uncertainties and validation. The BASEMENT and Heller-Hager should be clearly presented in the Method section. What assumptions are made? What are the input parameters? How are the models calibrated and validated? What are the uncertainties of the DEM, the lake-bed topography scenarios, and the volume estimate? This part of the study needs to be clearly described in the Method section before readers are convinced about the reliability of the model results.
3. I miss the wider perspective of the study in the Discussion section. The authors present a nice overview of the regional (Nepalese) GLOF history, but the eight GLOF events are not placed in an international context. A characteristic feature of the examined GLOF events is that they lack the traditional GLOF cyclicity, which is related to ice-dammed lakes. This basically tells us that the GLOFs are triggered by moraine breaching (as suggested by the authors) or extraordinary runoff due to snow and ice melting. At least the possibility of the latter trigger mechanism should be discussed. For instance, the internationally best-described GLOF events are from an ice-dammed lake in West Greenland, where the drainage volume is up to 39 x 10^6 m^3 (Russell et al., 2011). However, the largest flood event in this basin was not caused by a GLOF but by extraordinary snow and ice melting (Mikkelsen et al., 2016). Also, a comparison to other sites worldwide where GLOF events have been triggered by landslides or ice avalanching would be appreciated.

 

Minor issues:

L21: Inform the reader what Pabuk Khola and Chheche Pokhari are. I wondered whether Chheche Pokhari was the name of a village or an alluvial fan until I saw Figure 9.

 

L30-43: The Introduction is thin. What is the scientific novelty? Why is this research important? What are the wider perspectives? Have oral accounts been combined with remote sensing and field studies in GLOF studies before? What are the research questions?

 

L49: What is meant by “1,494 mm of Taplejung´s average 1,985 mm/year”?

 

L50: What is the distance between the meteorological station at Taplejung and Nangama Pokhari (or Lelep)?

 

L55: Where is Terai?

 

L61-111: All this is completely irrelevant for this study. It must be deleted.

 

L96: Byers et al. (2020) is not mentioned in the reference list and most likely irrelevant to this study.

 

L111: Do not reference “Byers et al. (forthcoming)” – it is none-existing literature that cannot be checked by the reviewers and the editor.

 

L136: Delete the third “the” in this sentence.

 

L136-137: I recommend that point (g) and the corresponding text in the Results section (L232-238) are deleted. The nine respondents’ guesses about potential future GLOFs have no scientific value or novelty and are irrelevant for the conclusions of the study.

 

L145: Why is a “pers. comm.” to Cook (2020) needed here? How was this dating conducted?

 

L148: What Landsat platforms were used?

 

L155 and elsewhere: Delete the “see:” in front of some references throughout the manuscript. It is not needed and the text should be self-explanatory so that readers do not have to read through several other papers to understand what is meant in this paper.

 

L166-171: These two sentences do not belong in a Results section. Delete them.

 

L173: Be consistent in the naming of the “1980 Nangama GLOF”. It varies throughout the text, in Table 3 and in the caption of Figure 8.

 

L173: I guess that it should be “map symbol 3”, not map symbol 2.

 

L202: Delete “created”.

 

L211-213: This will be another good place to inform the readers what Pabuk Khola and Chheche Pokhari are. Is Pabuk Khola on Figure 1?

 

L216 and elsewhere: There seems to be double spacing in front of some sentences in the manuscript. Please correct this.

 

L229: Lack of spacing between the two brackets.

 

L231: What is Chhochenphu Himal? I cannot find it on Figure 1.

 

L232-238: Delete this non-scientific guesswork.

 

L253-254: No need for a new paragraph here. Both paragraphs concern the Olangchun Gola flood.

 

L256: Insert spacing between “the” and “1963”.

 

L260: Change “the fact” to “its occurrence” or something similar.

 

L269: What is an englacial conduit flood? In order to be a GLOF, the water must, by definition, derive from a glacial lake. If not, it is not a GLOF. I recommend that in the Introduction section readers are informed about the classification of different flood events in glacierized basins and about various potential trigger mechanisms for GLOF events. This will be a good basis for a discussion about trigger mechanisms in the Discussion section.

 

L271-273: This sentence should probably be moved to L276.

 

L281: In my version of the manuscript, the last part of this sentence is missing.

 

L287-332: All this text belong in the Method section, not in a Results section. It also needs to be shortened.

 

L349: “… because of contemporary warming trends and accelerations in glacial melt”. This statement must be supported by one or more references.

 

L350-352: Be more specific. How fast have the number of non-glacial lakes increased (number per year)? How many lakes have transformed?

 

L364: I need information about the location and elevation of the lake spillway before the moraine was breached in 1980? Maybe change Figure 9 to a close-up of the Nangama Lake and the hanging glacier, where the pre-1980 spillway is showed by an arrow.

 

L381-402: Insert “Figure” or replace “image” with “Figure” in front of 8x in these paragraphs.

 

L388: Typo in Pokhari.

 

L413: Again, I have no clue on what an “englacial conduit flood” is. I guess that it is a flood causes by an extraordinary amount of meltwater, but readers need to know about the applied terminology and definitions.

 

L417-503: All this text and Table 5 belong in the Method section. The text is repetitive and needs to be shortened.

 

L465: Is this avalanche volume validated by a similar loss of hanging ice in satellite imagery from before and after 1980?

 

L478: Why the scaling factor of 1.1?

 

L491: What is the estimated uncertainty of using the grain size distribution of Worni et al. (2013)?

 

L512: What is this sediment transport model? Explain it in the Method section.

 

L538: Insert “flood” after Tama Pohari.

 

L540: It is not clear to me how the model predicts whether or not Chheche Pokhari will be dammed. Explain in the Method section what the threshold value for this assessment is.

 

L563-618: There is a lot of repetition in the Discussion (e.g., L563-566). It is not really a scientific discussion, but more a summary with a few qualitative comparisons to nearby GLOFs. I recommend that the Discussion is rewritten in a more scientifically rigorous fashion with discussions of characteristics of regional GLOF and flood events, and potential trigger mechanisms. There should also be a discussion about the wider perspectives of the interdisciplinary approach, the modelling efforts and the frequency of GLOFs in Himalaya.

 

Figure 3: When is the photo taken?

 

Figure 4: When is the photo taken?

 

Figure 14: The layout of Figure 14 is not good in my version of the manuscript. Where did the water go before the moraine breach?

 

Figure 17 caption: Replace “where some believe could trigger a new GLOF” with one or more literature references.

 

Table 2: Is it possible available information on the time of year for GLOF events?

 

Table 3: Delete the “0” in total area.

 

Table 4: Fix the layout issue with respect to the Basin row.

 

Footnotes: Delete all footnotes. They are scientifically irrelevant for the study.

 

Author Response

Open Review

(x) I would not like to sign my review report 
( ) I would like to sign my review report 

English language and style

( ) Extensive editing of English language and style required 
( ) Moderate English changes required 
(x) English language and style are fine/minor spell check required 
( ) I don't feel qualified to judge about the English language and style 

 

	Yes	Can be improved	Must be improved	Not applicable
Does the introduction provide sufficient background and include all relevant references?	( )	( )	(x)	( )
Is the research design appropriate?	( )	(x)	( )	( )
Are the methods adequately described?	( )	( )	(x)	( )
Are the results clearly presented?	( )	( )	(x)	( )
Are the conclusions supported by the results?	( )	(x)	( )	( )

Comments and Suggestions for Authors

This manuscript elucidates the history of flood events in the Kanchenjunga region, Nepal, by combining interviewing of nine local people with satellite and field-based identification of past GLOF events. In addition, numerical modelling is used to estimate wave amplitude and simulate discharge time series of the 1980 Nangama flood.

 

The study provides some new knowledge on historical GLOFs in this region, but several revisions are needed before the manuscript has the quality to be published.

 

Major issues:

1. The text suffers from being repetitive and needs restructuring and shortening. I have made some comments below to help the authors in this process, but I recommend that the authors carefully read through the manuscript to remove superfluous text. The results will appear clearer to the readers, if the text was more brief but comprehensive.
2. The description of the modelling must also be much more concise and scientifically rigorous with respect to assumptions, input parameters, uncertainties and validation. The BASEMENT and Heller-Hager should be clearly presented in the Method section. What assumptions are made? What are the input parameters? How are the models calibrated and validated? What are the uncertainties of the DEM, the lake-bed topography scenarios, and the volume estimate? This part of the study needs to be clearly described in the Method section before readers are convinced about the reliability of the model results.

 

 

The authors would like to thank the reviewer for his/her constructive and supportive comments. We agree that the modeling process could be more explicitly described. The introduction to the modeling section has been substantially revised, and a new figure demonstrating the calibration and modeling process (Figure 5) has been added. To clarify the modeling process, input parameters, and avalanche volume estimates, the following text has been updated (italics):

 

Numeric Modelling of the 1980 Nangama GLOF

L177--Because the Nangama GLOF—triggered by an avalanche that entered the lake and generated a large impulse wave—was among the largest of the GLOF events to have occurred in the region, a numerical simulation of the flood was performed in order to develop estimates related to triggering mechanisms, wave amplitude, and total flood volume. The Basic Simulation Environment for Computation of Environmental Flow and Natural Hazard Simulation (BASEMENT) model (Vetsch et al. 2017) was used for this purpose, incorporating a range of direct measurements of the Nangama terminal moraine and breach characteristics that included feature height, width, altitude, high water marks, glacier terminus condition, and new potential flood triggers. BASEMENT is a hydrodynamic model based on the 2-D shallow water equations with additional capabilities for sediment transport and erosion modeling (Vetsch et al., 2017). Validation of BASEMENT’s wave generation and propagation procedure was achieved by additionally employing an empirical model (the “Heller-Hager model;” Heller et al., 2009), which uses basic morphological attributes of the lake and surrounding slopes to characterize avalanche-induced impulse waves in lakes. BASEMENT’s capabilities make it ideal for GLOF modeling (Worni et al., 2014), and a combination of BASEMENT simulations with calibration via the Heller-Hager model has been carried out on a variety of GLOFs, both for historic reconstruction and to evaluate future hazard (Somos-Valenzuela et al., 2016; Byers et al., 2018; Lala et al., 2018). The model process chain followed the following steps: (1) using the Heller-Hager model, estimate the mass of the avalanche given the wave height and morphology of the lake; (2) iteratively set inflow conditions for the BASEMENT model to characterize the transfer of energy from the avalanche to the wave, until the wave amplitude matches that of the Heller-Hager model; (3) continue BASEMENT model downstream under various sediment transport models to characterize moraine erosion and downstream discharge. Figure 5 depicts this process chain in a brief flow chart.

 

Regarding uncertainty in the lake depth and bathymetry, the following has been updated at Line 216: “Lake depth was estimated from an empirical analysis of area-depth-volume relationships developed by Cook and Quincey (2015); given a lake level of 4,894 masl, the pre-GLOF area of the lake was approximately 1.1 km² and the average depth was estimated at 38 m. While this is a rough estimate (± 10 m for a lake of Nangama’s size), the Heller-Hager model produced very similar wave heights (8.0 m versus 8.3 m amplitude for a depth of 28 m or 48 m, respectively) and thus 38 m was deemed acceptable for modeling purposes. Further uncertainty in the lake depth was considered by estimating bathymetry in two ways…”

               

Regarding DEM uncertainty, the following has been altered: “A digital elevation model (DEM) of the region was taken from the High Mountain Asia 8-meter DEM Mosaics Derived from Optical Imagery, Version 1 (Shean, 2017), which has an absolute horizontal and vertical accuracy of <0.5 m.”



1. I miss the wider perspective of the study in the Discussion section. The authors present a nice overview of the regional (Nepalese) GLOF history, but the eight GLOF events are not placed in an international context. A characteristic feature of the examined GLOF events is that they lack the traditional GLOF cyclicity, which is related to ice-dammed lakes. This basically tells us that the GLOFs are triggered by moraine breaching (as suggested by the authors) or extraordinary runoff due to snow and ice melting. At least the possibility of the latter trigger mechanism should be discussed. For instance, the internationally best-described GLOF events are from an ice-dammed lake in West Greenland, where the drainage volume is up to 39 x 10^6 m^3 (Russell et al., 2011). However, the largest flood event in this basin was not caused by a GLOF but by extraordinary snow and ice melting (Mikkelsen et al., 2016). Also, a comparison to other sites worldwide where GLOF events have been triggered by landslides or ice avalanching would be appreciated.

L567-643—this was an excellent point and has since been addressed. The summary/repetition has been removed, and replaced with an actual discussion of (a) strengths of an oral testimony component, (b) englacial conduit floods, (c) brief review of contemporary GLOFs in Nepal, where most of those now documented for the Kanchenjunga region are absent, (d) GLOF triggering mechanisms, including the more recent (and comparatively little known and studied) phenomena of permafrost thaw and its increasing links to catastrophic landslide-initiated GLOFs, and (e) recent concerns about lateral moraine landslides triggering a GLOF at Nangama, with citations from other regions of the world. We feel that this section is now an actual discussion of additional and important elements that surfaced during the course of the field work and data analysis. In particular, based upon recent lectures given by the PI in Nepal, the phenomenon of permafrost thaw is still little known among scientists and field practitioners there, which the current paper could help to reverse.

Minor issues

L21: Inform the reader what Pabuk Khola and Chheche Pokhari are. I wondered whether Chheche Pokhari was the name of a village or an alluvial fan until I saw Figure 9.

 L21—This sentence has been revised for greater clarification as follows: “Debris from the flood dammed the Pabuk Khola (river) 2 km below the lake to form what is today known as Chheche Pokhari (lake).”

L30-43: The Introduction is thin. What is the scientific novelty? Why is this research important? What are the wider perspectives? Have oral accounts been combined with remote sensing and field studies in GLOF studies before? What are the research questions?

This is an excellent point and observation, which we have addressed by inserting the following in what are now lines 39-69:

(L31) On 28 June, 1980, the Rising Nepal reported that a glacial lake outburst flood (GLOF) had occurred in the Tamor valley of eastern Nepal (Rising Nepal, 28 June 1980) (Figure 1). The estimated date of the flood was cited as occurring on or about 23 June, 1980 (Watanabe et al. 1998). Flood damage included “….all the houses in Olangchung Gola village” (Rising Nepal 1980), four bridges, 10 human fatalities (Khanal 1996), and within the downstream settlements of Lunthun, Siwa, and Dobhan (Watanabe et al. 1998). The source of the flood was later determined to be the Nangama glacial lake in the Kanchenjunga[i] region of east Nepal, about 8 km south of the border with Tibet (Watanabe et al. 1998; Figure 2).

GLOFs are highly destructive events, caused when stored lake water is suddenly released by a triggering event that can include dam failure, snow and/or debris avalanches, and heavy rainfall (Richardson and Reynolds 2000). While there is now a copious Himalayan (ICIMOD 2011) as well as international literature concerned with GLOFs (Carrivick and Tweed 2016), no detailed scientific studies are available for the Kanchenjunga region  of eastern Nepal.  In absence of such information, we undertook an integrated, interdisciplinary field investigation of the upper Tamor and Kanchenjunga basecamp regions between April and May 2019. Specific research questions included:

 

1. What were the likely causes and triggers contributing to the 1980 Nangama GLOF?
2. What was the damage caused by the 1980 Nangama GLOF when compared to reports from the press release of the same year?
3. What other GLOFs may have occurred in the Kanchenjunga region during contemporary and/or historical times, and how can they be identified and verified?, and
4. What role can interdisciplinary studies play in the advancement of knowledge concerning GLOFs, their causes, impacts, and mitigation opportunities?

 

Methods included oral history interviews, field measurements, remote sensing analyses, literature reviews, and flood modeling. The April-May 2019 field phase focused primarily upon interviews and field measurements, followed by the laboratory-based remote sensing and modeling work of September-December 2019.

It should be noted that GLOF studies incorporating interdisciplinary approaches (e.g., Sherpa et al. 2019; Byers et al. 2014; Byers et al. 2017) are extremely rare when compared to those based entirely upon physical science approaches, in spite of the former method’s acknowledged importance (Watanabe et al2016). The present research project thus presented an opportunity to further test the strength and viability of interdisciplinary methods in the clarification of GLOF history in a remote and understudied region of Nepal.  

.  

L49: What is meant by “1,494 mm of Taplejung´s average 1,985 mm/year”?

On lines 71-78, this has now been clarified as follows:

For example, the village of Taplejung (100 km south of Nangama glacial lake)  averages 1,985 mm/year, with 1,494 mm of this total falling during the monsoon (Climate-data.org 2020).

L50: What is the distance between the meteorological station at Taplejung and Nangama Pokhari (or Lelep)?

 On lines 70-75, this has been included as follows:

A monsoon climate dominates weather patterns, with most of the annual precipitation falling between June and September. For example, the village of Taplejung (100 km south of Nangama glacial lake)  averages 1,985 mm/year, with 1,494 mm of this total falling during the monsoon) (Climate-data.org 2020). Taplejung’s mean annual temperature is 15.4 °C, and while no weather data for Nangama or other glacial lakes exists at present, the mean annual temperature of the higher elevation village of Ghunsa (3,407 m, Figure 1) has been recorded by data loggers as 4.9 ^oC (Watanabe et al. 2006).

L55: Where is Terai?

 Line 79—this has been clarified as follows:

Altitudes range from 40 meters above sea level (masl) in the southern plains (known as the Terai region) to the summit of Kanchenjunga at 8, 856 masl to the north within a relative short air distance of approximately 150 km.

L61-111: All this is completely irrelevant for this study. It must be deleted.

L61-111—This entire section has been deleted as per the Reviewer’s request. However, for the time being we have retained this information as an Endnote, primarily because the Kanchenjunga region is so little known in terms of the contemporary problems of wildlife, poaching, road building, and tourism. The information will be of tremendous interest to the paper’s audience, especially the Government of Nepal, which is now planning to promote the Kanchenjunga region in an effort to attract adventure tourists to post-COVID 19 Nepal.

L96: Byers et al. (2020) is not mentioned in the reference list and most likely irrelevant to this study.

L96--This recently published work is now part of the proposed endnote, and has been entered into the reference list.

L111: Do not reference “Byers et al. (forthcoming)” – it is none-existing literature that cannot be checked by the reviewers and the editor.

 L111—this reference has been deleted.

L136: Delete the third “the” in this sentence.

L136—we were unable to find the third “the” in L136 but assume that it has been taken care of through spell check and proof reading.

L136-137: I recommend that point (g) and the corresponding text in the Results section (L232-238) are deleted. The nine respondents’ guesses about potential future GLOFs have no scientific value or novelty and are irrelevant for the conclusions of the study.

L136-137—Point g, or “(g) the likelihood of future floods occurring in the region,” was an emerging theme and persistently salient in the interviews conducted with local residents. Although the sample size might be small, local residents’ perspectives on the past flood events and the likelihood of repeating similar events in the future are highly valuable, especially to understand how those residents are preparing themselves with appropriate adaption and mitigation measures. Their knowledge of the local ecology, its resources, and catastrophic events is highly sophisticated and can complement the scientific understanding of glacial lakes and GLOFs. Given the interdisciplinary nature of the paper, we feel that point “g” should be retained with appropriate revision to reflect the complementary nature of scientific and indigenous knowledge. Likewise, we feel that the informant responses provided in L232-238 are important as well, but have revised this section to read as follows:

“Most respondents agreed that other floods could be expected to come in the future, although it was frequently added that it is “better not to talk about such things.” People’s reluctance to discuss negative phenomena (e.g., floods, death, misfortune) is common among Buddhist-Bón communities throughout the Himalayas, as this is viewed as representing an invitation to their actual occurrence (Sherpa 2014). Other informants, however, expressed little concern about the possibility of future floods.”

 L145: Why is a “pers. comm.” to Cook (2020) needed here? How was this dating conducted?

L145—this ‘pers. comm.’ reference has been removed, as references to the tree ring analyses conducted at the Columbia Tree Ring Laboratory are now contained in the proposed Endnote regarding tourism impacts on alpine ecosystems.

L148: What Landsat platforms were used?

L120: This sentence has been revised to the following:  

A remote sensing analysis using CORONA and Landsat MSS and Landsat 5 data platforms was conducted after the field phase to determine if flood-related landscape features could either support or challenge the oral reports of contemporary and historical flood occurrence.

L155 and elsewhere: Delete the “see:” in front of some references throughout the manuscript. It is not needed and the text should be self-explanatory so that readers do not have to read through several other papers to understand what is meant in this paper.

 L155 and elsewhere: all references to “see” have been deleted, with the exception of the paragraph starting at line 299 in the track changes manuscript, as shown below, since the Benn et al. 2017 paper is in addition to those contained in Rounce et al. 2917:

This informant also reported that another small flood of the Ghunsa river had occurred in “either August or September” of 2011, the approximate time that a deadly earthquake (6.9 magnitude) centered in the Kanchenjunga region occurred that killed an estimated 111 people (BBC News 2011). The source of the flood is unknown, but may have been an englacial conduit flood from the Ramdam or other glacier to the northeast, similar to those reported for the Lhotse glacier in Khumbu for the years 2015 and 2016 (Rounce et al. 2017). As opposed to GLOFs, usually triggered by catastrophic ice avalanches and resultant dam breaching, englacial conduit floods “…develop on debris-covered glaciers…through cut-and-closure mechanisms with meltwater streams, the exploitation of high permeability areas that provide alternative pathways to the impermeable glacier ice, and through hydrofracturing processes (Rounce et al. 2017, citing Gulley and Ben 2007; Benn et al. 2009; Gulley et al. 2009a; Gulley et al. 2009b; see also: Benn et al. 2017). In general, however, little concern was expressed by informants regarding future GLOFs in the region, although the Nupchu Pokhari above Khangpachen was singled out as a potentially dangerous glacial lake (Figure 1).

 

L166-171: These two sentences do not belong in a Results section. Delete them.

The Methods and Results sections have been significantly modified as per the Reviewer’s later suggestion, such that these two sentences are no longer a problem.

L173: Be consistent in the naming of the “1980 Nangama GLOF”. It varies throughout the text, in Table 3 and in the caption of Figure 8.

L173 and elsewhere—A search and replace was conducted on all variations of Nangama so that it is now consistent throughout the text

L173: I guess that it should be “map symbol 3”, not map symbol 2.

 L190—this has now been corrected as follows:

Figure 1 and Table 3 presents a list of GLOFs, date of occurrence, source, and location as recorded by the oral testimony component. In addition to the 1980 Nangama GLOF (map symbol 3), participants identified an additional five GLOFs that had occurred in the region since 1921 (Map symbols 1, 2, 4, 5, and 8). Map symbols 6 and 7 refer to two additional flood events that were identified by the remote sensing component, but not mentioned by informants.

 

L202: Delete “created”.

 L 125—the second ‘created’ has been deleted

L211-213: This will be another good place to inform the readers what Pabuk Khola and Chheche Pokhari are. Is Pabuk Khola on Figure 1?

L133--this sentence has been clarified as follows:

Floodwaters periodically lowered and became clear of debris, only to surge again sometime later, indicating the pulse-like nature reported for GLOFs elsewhere (Byers et al. 2019). Chheche Pokhari, located at the foot of the Nangama valley, was reported to have been formed from the blockage of the Pabuk Khola by flood sediments and debris (Figure 1).

 

L216 and elsewhere: There seems to be double spacing in front of some sentences in the manuscript. Please correct this.

 L216 and elsewhere—a search and replace was performed and 30 double spaces were replaced with single.

L229: Lack of spacing between the two brackets.

 L256—this sentence has been corrected as follows:

The flood was also linked by religious leaders to the fact that “modern people have become wicked, so bad things happen,” a frequently heard correlation in highland Nepal that links greed, the erosion of traditional practices, and lack of respect for gods, deities, and spirits with negative consequences, both physical (e.g., earthquakes, floods) and social (e.g., death, suffering) (Dawson and Uhrig (1984); Bjønness (1986);Sherry and Curtis (2017)).

L231: What is Chhochenphu Himal? I cannot find it on Figure 1.

L86—Figure 1 has been revised and now contains Chhochenphu Himal

L232-238: Delete this non-scientific guesswork

L2353—as in previous social science sections within the paper, we feel that the information presented here is important and of use to the reader’s overall understanding of the range of issues involved with GLOFs in the Himalayas. However, we have revised the sentence as follows; 

Respondents linked the cause of the 1963 and 1968 Olangchun Gola floods (map symbols 5 and 4) to icefall from Chhochenphu Himal (6,260 m), which fell into Tiptola glacial lake and caused a surge wave that then breached the terminal moraine. Most respondents agreed that other floods could be expected to come in the future, although it was frequently added that it is “better not to talk about such things.” People’s reluctance to discuss negative phenomena (e.g., floods, death, misfortune) is common among Buddhist-Bón communities throughout the Himalayas, as this is viewed as representing an invitation to their actual occurrence (Sherpa 2014). Other informants, however, expressed little concern about the possibility of future floods.

 

 

L253-254: No need for a new paragraph here. Both paragraphs concern the Olangchun Gola flood. –

L376—The sentence break has been deleted, and “Other informants…” is now part of the preceding paragraph

 

L256: Insert spacing between “the” and “1963”.

L363—spacing has been inserted

L260: Change “the fact” to “its occurrence” or something similar.

 L182—‘the fact’ has been changed to ‘its occurrence’

L269: What is an englacial conduit flood? In order to be a GLOF, the water must, by definition, derive from a glacial lake. If not, it is not a GLOF. I recommend that in the Introduction section readers are informed about the classification of different flood events in glacierized basins and about various potential trigger mechanisms for GLOF events. This will be a good basis for a discussion about trigger mechanisms in the Discussion section

 Englacial conduit floods are now explained in the paragraph starting with L388 as follows:

This informant also reported that another small flood of the Ghunsa river had occurred in “either August or September” of 2011, the approximate time that a deadly earthquake (6.9 magnitude) centered in the Kanchenjunga region occurred that killed an estimated 111 people (BBC News 2011). The source of the flood is unknown, but may have been an englacial conduit flood from the Ramdam or other glacier to the northeast, similar to those reported for the Lhotse glacier in Khumbu for the years 2015 and 2016 (Rounce et al. 2017). As opposed to GLOFs, usually triggered by catastrophic ice avalanches and resultant dam breaching, englacial conduit floods “…develop on debris-covered glaciers…through cut-and-closure mechanisms with meltwater streams, the exploitation of high permeability areas that provide alternative pathways to the impermeable glacier ice, and through hydrofracturing processes (Rounce et al. 2017, citing Gulley and Ben 2007; Benn et al. 2009; Gulley et al. 2009a; and Gulley et al. 2009b; see also: Benn et al. 2017). In general, however, little concern was…………..

 

L271-273: This sentence should probably be moved to L276.

 L198—this sentence has been edited to fit more closely within the paragraph as follows, and so has not been moved:

In general, however, little concern was expressed by informants regarding future GLOFs in the region, although the Nupchu Pokhari above Khangpachen was singled out as a potentially dangerous glacial lake (Figure 1).

 

L281: In my version of the manuscript, the last part of this sentence is missing

L111—the sentence has been corrected as follows:

The pattern reflects that of glacial lake development and GLOF frequency found elsewhere in the Himalayas, usually linked to contemporary changes in local regional, and global climate (Haritashya et al. 2018).

 

L287-332: All this text belong in the Method section, not in a Results section. It also needs to be shortened

L287-332—All text in this section has been shortened and moved to the Methods section. Likewise the numeric modeling methods sections, previously under Results, have now been moved to Methods.

L349: “… because of contemporary warming trends and accelerations in glacial melt”. This statement must be supported by one or more references

L129—this sentence has been modified as per the following:

Between 1964 and 2018 the number of glacial lakes >5,000 m² in size increased from 31 to 157, with total lake area increasing from 1.74 km² 5.74 km² during the same period. While both are increasing in number and area, glacier-fed lakes are doing so more rapidly than the non-glacier-fed lakes because of contemporary warming trends and accelerations in glacial melt (Benn et al. 2017; Haritashya et al. 2018).

 L350-352: Be more specific. How fast have the number of non-glacial lakes increased (number per year)? How

L128--The following has been added to provide greater specificity:

Figure 10 illustrates several trends over time in glacier-fed and non-glacier-fed lake development and expansion. Between 1964 and 2018 the number of glacial lakes >5,000 m² in size increased from 31 to 157, with total lake area increasing from 1.74 km² 5.74 km² during the same period.

 

L364: I need information about the location and elevation of the lake spillway before the moraine was breached in 1980? Maybe change Figure 9 to a close-up of the Nangama Lake and the hanging glacier, where the pre-1980 spillway is showed by an arrow.

Line 373: Given the lack of DEMs before 1980, the morphology of the terminal moraine and spillway was interpolated from areas on each side of the floodplain that avoided scouring (see Lines 467-470). The following has been added to Line 453: “Judging from remaining parts of the moraine and spillway that were not affected by the GLOF, the terminal moraine rose above the bottom of the spillway (~4820 m.a.s.l.) by approximately 80 m and over a distance of 130 m, for a slope of approximately 30°.”

L381-402: Insert “Figure” or replace “image” with “Figure” in front of 8x in these paragraphs

L460-470—‘Figure’ been placed in from of all 8x designations; ‘image’ has been removed

 

L388: Typo in Pokhari.

Entire mss--the spelling of ‘Pokhari’ has been corrected through search and replace

L413: Again, I have no clue on what an “englacial conduit flood” is. I guess that it is a flood causes by an extraordinary amount of meltwater, but readers need to know about the applied terminology and definitions

 Please see the explanations provided previously and above

L417-503: All this text and Table 5 belong in the Method section. The text is repetitive and needs to be shortened

As mentioned, the Methods and Results sections for Remote sensing and Numeric Modeling have been completely revised as per the Reviewer’s suggestions.

L465: Is this avalanche volume validated by a similar loss of hanging ice in satellite imagery from before and after 1980?

Co-author Lala reports that he did not have access to pre-1980 DEMs or satellite imagery. However he has updated the text to provide more clarity re: how the Heller-Hager model estimates avalanche volume.

L478: Why the scaling factor of 1.1?

       L245--This factor is the ratio of Nangama’s avalanche volume to that of Imja Tsho’s. The text has been altered: Because BASEMENT only accepts water as inflow, this difference in density was considered by simply increasing the Imja Tsho entry rate by a factor of 1.25, in addition to another scaling factor of 1.1 to account for the ratio of the Nangama avalanche’s volume to that of Imja Tsho’s (Figure 8).

L491: What is the estimated uncertainty of using the grain size distribution of Worni et al. (2013)?

 Worni et al. (2013) provided two grain size distributions from various sites in India. A previous GLOF modeling study in Nepal (Byers et al., 2018) found the two distributions to have little effect on moraine erosion; thus we elected to use only a single distribution which corresponded to two lakes at a similar altitude to that of Nangama. The text has been updated: “…the grain size distribution was instead taken from an inventory of glacial lakes in the Indian Himalaya (Worni et al., 2013; Table 5) that had performed well in recreating previous Nepal GLOFs. Despite uncertainty in the actual grain size distribution, a similar GLOF modeling study in the Barun Valley (Byers et al., 2018) found little difference in simulated moraine erosion between the grain size distributions listed in Worni et al. (2013).”

L512: What is this sediment transport model? Explain it in the Method section.

 The sediment transport model has now been more clearly explained in the text: “BASEMENT characterizes sediment transport using the Meyer-Peter and Müller (MPM) equations (Vetsch et al., 2017), which estimate suspended and bed load fluxes by calculating shear stresses within the flow through a modified Shields parameter (Shields, 1936). Soil mechanics parameters (i.e. critical slope angles for dry, wetted, and deposited sediment) required by the MPM equations were estimated from a number of studies in the Nepal Himalaya and elsewhere based on both soil samples and visual observations of slopes (see Lala et al., 2018).”

L538: Insert “flood” after Tama Pohari.

 L545—this insertion has been made as follows:

For comparison, this amount was nearly twice that of the famous 1985 Langmoche “flash flood” in Khumbu (Vuichard and Zimmerman 1987), but half that of the 1998 Tama Pokhari in the Hinku valley of Makalu-Barun National Park and Buffer Zone (Lamsal et al. 2015).

L540: It is not clear to me how the model predicts whether or not Chheche Pokhari will be dammed. Explain in the Method section what the threshold value for this assessment is.

There were no assumptions regarding the damming of Chheche Pokhari; once the BASEMENT model was calibrated for wave heights within the lake, it was allowed to run downstream. The simulations invariably resulted in flooding of the [dry] Chhechke Pokhari basin followed by the deposition of sediment between this flooded area and the riverbed, thereby damming and creating Chhechke Pokhari. The text (Line 577) has been clarified to reflect this: “All three scenarios resulted in the creation of Chheche Pokhari (e.g., Figure 17) through flooding of its basin and subsequent damming via sediment deposition, granting some level of confidence to the results.”

L563-618: There is a lot of repetition in the Discussion (e.g., L563-566). It is not really a scientific discussion, but more a summary with a few qualitative comparisons to nearby GLOFs. I recommend that the Discussion is rewritten in a more scientifically rigorous fashion with discussions of characteristics of regional GLOF and flood events, and potential trigger mechanisms. There should also be a discussion about the wider perspectives of the interdisciplinary approach, the modelling efforts and the frequency of GLOFs in Himalaya.

Please see the responses above that address these issues, as well as the revisions in the Discussion section.

 Figure 3: When is the photo taken?

May 2019 was the date of the photo and has been inserted into the caption

 Figure 4: When is the photo taken?

May 2019 was the date of the photo and has been inserted into the caption

Figure 14: The layout of Figure 14 is not good in my version of the manuscript. Where did the water go before the moraine breach?

Figure 15 (since re-numbered)--The figure has been reformatted to fully depict the legend. Note that the initial lake/water level is below the initial moraine level; there was no outlet to the lake.

Figure 17 caption: Replace “where some believe could trigger a new GLOF” with one or more literature references.

L627— This sentence has been removed from the caption, and references provided on L627: “Such phenomena have been documented for various glaciers in the Peruvian Andes (Emmer and Vilímek (2014); Klimeš et al. (2016)), and suggest that regular monitoring of Nangama, Tiptola, and other glacial lakes in the region be conducted.”

Table 2: Is it possible available information on the time of year for GLOF events?

It is unfortunately not possible to obtain precise information re: the timing of the GLOF events shown in Table 2. However, nearly all GLOF events on record have occurred during the summer (monsoon) period, and the following was added to the caption:

L318…….Most of the post-1960 GLOF events were described as occurring during the monsoon season, although exact dates are unknown, except for Nangama.

 Table 3: Delete the “0” in total area.

‘0’ has been deleted in the table

 Table 4: Fix the layout issue with respect to the Basin row.

 The layout issue with respect to the Basin row has been corrected.

Footnotes: Delete all footnotes. They are scientifically irrelevant for the study.

All footnotes have been deleted. However, as discussed previously, we are proposing that they be included as Endnotes or Supplementary Material, given the scarcity of information for the Kanchenjunga region.

 

Submission Date

02 June 2020

Date of this review

08 Jun 2020

 

 

Reviewer 2 Report

In an interesting and well-written paper, Byers et al. present an integration of interdisciplinary approaches to document episodes of glacial lake outburst floods in eastern Nepal. Drawing upon evidence based on oral histories, remote sensing, and field observations, the authors corroborate the occurrence of several known flood events and suggest the possibility of two previously unknown or unrecognized events. The authors also use a numerical model to reconstruct possible peak discharges and flood volumes associated with a recent outburst flood in the region.

In doing so, the paper will principally be of interest to those concerned with the mitigation of the hazards posed by outburst floods. However, their inventory of “contemporary” glacial lakes - especially the apparent apparent increase in formation owing to increased melt in response to warming climate - together with their geomorphic observations will broaden the paper’s appeal to glaciologists, climate scientists, and researchers interested in landscape evolution in mountainous terrain. Moreover, as pointed out by the authors, this paper demonstrates the value of anecdotal evidence (i.e. local knowledge) in augmenting “scientific” or more “rigorous” studies of natural events.

Briefly, the problem investigated by the authors is well-defined, and they make clear its significance in terms of consequences for human settlements (i.e. flood hazards) and for an ecologically sensitive region. With some minor exceptions of the modeling (see below), sufficient details are provided for the methodologies used, and the results are clearly presented. Their conclusions are supported by the results.

I personally found the cultural, socio-economic, and ecological aspects of the paper illuminating.

It is my recommendation that the paper be accepted with minor changes and/or corrections. Importantly, I emphasize that although my ratings of the scientific soundness and reader interest are average, they are in reality above average (not an option) but not yet high owing respectively to some of vagaries in the description of, or uncertainties involved in the modeling and the geographic location being very specific.   

 

General comments:

My overarching, albeit minor concern is that the numerical model (BASEMENT) and modeling is inadequately described. Perhaps this is beyond the scope of the journal’s audience. Nevertheless, sufficient details should be given in order to understand the nature of the model without having to go to the original source (Vetsch et al., 2017). The few details given in lines 417-430 do not, for example neither explicitly indicate that the model apparently (?) requires a trigger event (e.g. avalanche?), nor that it provides an estimate of the resulting seiche wave height (see below). I would have appreciated a brief paragraph that systematically explained the required input(s) and the steps leading to the reconstruction of flood hydrographs - e.g. avalanche/mass movement generates a seiche wave with a specific wave height, if the wave tops the moraine dam erosion of a breach channel is determined using [specific] sediment erosion and transport equations, discharge is calculated using [specific] equations and are contingent on the evolution of breach cross-section, etc. In this way the reader will have a better overall sequence of processes and understanding of the modeling scheme. I note that some - but not all - of this can be gleaned later in the paper, however the paper would benefit from a concise overview of the model and modeling procedure (a flow chart?) earlier in the manuscript. (And just in case, I did not have access to any supplementary information wherein the modeling detail might appear.)      

Specific comments and/or suggestions

Line 55 - “southern Terai.” Where is this? It is not on the accompanying map (Fig. 1). A few words might serve to indicate where this is for those of us not intimately familiar with the regional geography.

Lines 144-45 - What is the significance of the dendrochronology data? In an endnote there is some discussion of the ecological impact of harvesting, but I did not see the connection to the present study. Did I miss something here?

Lines 172-3 - I think “the 1980 Nangama GLOF (map symbol 2)” should be “the 1980 Nangma GLOF (map symbol 3 [three]).”

Line 187 - Here and elsewhere, the Nangama GLOF is now referred to as the Yangma GLOF and the names are used somewhat inconsistently; cf. lines 36, 239, 357 for example. I appreciate that locations and features can be referred to differently by different cultures and local peoples, but this can be acknowledged and for the sake of clarity for the reader one name should be used consistently thereafter.

Line 269 - Here and elsewhere, the term “englacial conduit flood” is used. I have some familiarity with ice-dam failure and resulting GLOFs and have not yet come across this term. At first I thought this to be akin to subglacial tunnel enlargement and ice dam failure. Apparently (via the Rounce et al. reference) this refers to floods resulting from englacial conduit enlargement and subsequent rapid drainage of supraglacial (and subglacial?) lakes. The authors might briefly define this term for the benefit of readers, like me, that are less familiar with this mode of catastrophic water discharge.

Line 288 - “In the next section...” Shouldn’t this be “this section” or the “current section?”

Lines 313-15 - What about the vertical (elevation) uncertainties? Wouldn’t they be important in determining lake surface (and other) elevations, hence lake volumes?

Line 325 - Perhaps make clear or explicit that the 5000 m² threshold is a minimum.

Lines 333-4 - “...and trend assessment component.” I do think these tables provide or reveal a trend (which is interesting) but there’s no “component” or quantification on the tables themselves; i.e. the reader must infer the trend. Rather the trend is shown on Fig. 6.

      Relatedly, on these tables, are the areas given the sum of all lakes for a given year and sub-basin? This should be made clear.

Line 362 - “this paper’s pre-GLOF estimate [of lake area] of 1.1 km².” Perhaps the authors might point out that this value is derived subsequently (specifically line 452). I initially thought this value was presented earlier in the paper.

Line 425 - Shouldn’t a reference be provided for the Heller-Hagar empirical model?

Lines 450-52 - The authors might be more forthcoming in the use of the Cook and Quincey (2015) empirical relationship between lake area and water depth as that paper shows lake area is a poor predictor of depth. This does not mean that relationship shouldn’t be used, but the uncertainties associated with depth can be - if I am interpreting the Cook and Quincey graph correctly- on the order of 10s of meters for a lake of the size considered in the current paper.

Line 459 - Unless I’ve missed something here as well, there was no previous mention that an avalanche was the trigger for the seiche wave so this line reads rather matter-of-fact. This also addresses the criticism above regarding the succession of events resulting in, and modeling of the GLOF.

Line 466 - I did not have the supplemental material, but I presume more details regarding the Heller-Hagar method would provide some indication of the nature of this approach?

Line 474 - It is unclear to me why “ice and snow avalanches” would have a density of 1000 kg m^-3, i.e. of water. Glacial ice is ~900 and snow less. Is there an assumption that debris is contained in the ice and snow? Also, the Schneider et al. reference is missing in the references section.

Lines 505-6 - These lines also underscore the need for more details of the modeling approach used as suggested above. This is maybe the first explicit statement indicating that the BASEMENT model provides an estimate of wave heights.

References

As noted above, the Schneider et al. 2014 reference is missing.

Tables

Also as noted above, in Tables 3 and 4 it should be made clear that the areas given for the specific sub-basins are the cumulative area represented by n number of lakes in each basin (assuming this is correct!).

Figures

On Fig. 14, an explanation should be provided for the different colored lines. I presume each is associated with a specific scenario as per the caption. This figure also suffered from formatting issues; axis labels are printed over axis values.

 

Author Response

Open Review

(x) I would not like to sign my review report 
( ) I would like to sign my review report 

English language and style

( ) Extensive editing of English language and style required 
( ) Moderate English changes required 
(x) English language and style are fine/minor spell check required 
( ) I don't feel qualified to judge about the English language and style 

 

	Yes	Can be improved	Must be improved	Not applicable
Does the introduction provide sufficient background and include all relevant references?	(x)	( )	( )	( )
Is the research design appropriate?	(x)	( )	( )	( )
Are the methods adequately described?	( )	(x)	( )	( )
Are the results clearly presented?	(x)	( )	( )	( )
Are the conclusions supported by the results?	(x)	( )	( )	( )

Comments and Suggestions for Authors

In an interesting and well-written paper, Byers et al. present an integration of interdisciplinary approaches to document episodes of glacial lake outburst floods in eastern Nepal. Drawing upon evidence based on oral histories, remote sensing, and field observations, the authors corroborate the occurrence of several known flood events and suggest the possibility of two previously unknown or unrecognized events. The authors also use a numerical model to reconstruct possible peak discharges and flood volumes associated with a recent outburst flood in the region.

In doing so, the paper will principally be of interest to those concerned with the mitigation of the hazards posed by outburst floods. However, their inventory of “contemporary” glacial lakes - especially the apparent apparent increase in formation owing to increased melt in response to warming climate - together with their geomorphic observations will broaden the paper’s appeal to glaciologists, climate scientists, and researchers interested in landscape evolution in mountainous terrain. Moreover, as pointed out by the authors, this paper demonstrates the value of anecdotal evidence (i.e. local knowledge) in augmenting “scientific” or more “rigorous” studies of natural events.

Briefly, the problem investigated by the authors is well-defined, and they make clear its significance in terms of consequences for human settlements (i.e. flood hazards) and for an ecologically sensitive region. With some minor exceptions of the modeling (see below), sufficient details are provided for the methodologies used, and the results are clearly presented. Their conclusions are supported by the results.

I personally found the cultural, socio-economic, and ecological aspects of the paper illuminating.

It is my recommendation that the paper be accepted with minor changes and/or corrections. Importantly, I emphasize that although my ratings of the scientific soundness and reader interest are average, they are in reality above average (not an option) but not yet high owing respectively to some of vagaries in the description of, or uncertainties involved in the modeling and the geographic location being very specific.   

General comments:

My overarching, albeit minor concern is that the numerical model (BASEMENT) and modeling is inadequately described. Perhaps this is beyond the scope of the journal’s audience. Nevertheless, sufficient details should be given in order to understand the nature of the model without having to go to the original source (Vetsch et al., 2017). The few details given in lines 417-430 do not, for example neither explicitly indicate that the model apparently (?) requires a trigger event (e.g. avalanche?), nor that it provides an estimate of the resulting seiche wave height (see below). I would have appreciated a brief paragraph that systematically explained the required input(s) and the steps leading to the reconstruction of flood hydrographs - e.g. avalanche/mass movement generates a seiche wave with a specific wave height, if the wave tops the moraine dam erosion of a breach channel is determined using [specific] sediment erosion and transport equations, discharge is calculated using [specific] equations and are contingent on the evolution of breach cross-section, etc. In this way the reader will have a better overall sequence of processes and understanding of the modeling scheme. I note that some - but not all - of this can be gleaned later in the paper, however the paper would benefit from a concise overview of the model and modeling procedure (a flow chart?) earlier in the manuscript. (And just in case, I did not have access to any supplementary information wherein the modeling detail might appear.)      

The authors would like to thank the reviewer for his/her constructive and supportive comments, and agree that the process chain within the BASEMENT numerical model should be more clearly described. The section “Numeric Modeling of the 1980 Nangama GLOF” has thus been substantially rewritten as follows to clarify many of these issues:

 

Line 178: Because the Nangama GLOF—triggered by an avalanche that entered the lake and generated a large impulse wave—was among the largest of the GLOF events to have occurred in the region, a numerical simulation of the flood was performed in order to develop estimates related to triggering mechanisms, wave amplitude, and total flood volume. The Basic Simulation Environment for Computation of Environmental Flow and Natural Hazard Simulation (BASEMENT) model (Vetsch et al. 2017) was used for this purpose, incorporating a range of direct measurements of the Nangama terminal moraine and breach characteristics that included feature height, width, altitude, high water marks, glacier terminus condition, and new potential flood triggers. BASEMENT is a hydrodynamic model based on the 2-D shallow water equations with additional capabilities for sediment transport and erosion modeling (Vetsch et al., 2017). Validation of BASEMENT’s wave generation and propagation procedure was achieved by additionally employing an empirical model (the “Heller-Hager model;” Heller et al., 2009), which uses basic morphological attributes of the lake and surrounding slopes to characterize avalanche-induced impulse waves in lakes. BASEMENT’s capabilities make it ideal for GLOF modeling (Worni et al., 2014), and a combination of BASEMENT simulations with calibration via the Heller-Hager model has been carried out on a variety of GLOFs, both for historic reconstruction and to evaluate future hazard (Somos-Valenzuela et al., 2016; Byers et al., 2018; Lala et al., 2018). The model process chain followed the following steps: (1) using the Heller-Hager model, estimate the mass of the avalanche given the wave height and morphology of the lake; (2) iteratively set inflow conditions for the BASEMENT model to characterize the transfer of energy from the avalanche to the wave, until the wave amplitude matches that of the Heller-Hager model; (3) continue BASEMENT model downstream under various sediment transport models to characterize moraine erosion and downstream discharge. Figure 5 depicts this process chain in a brief flow chart.

 

Figure 5. Flow chart of calibration and modeling process

 

Additionally, a flowchart of the modeling process (Figure 8) has been added to improve clarity.

Specific comments and/or suggestions

Line 55 - “southern Terai.” Where is this? It is not on the accompanying map (Fig. 1). A few words might serve to indicate where this is for those of us not intimately familiar with the regional geography.

Line 80—the Terai has been more fully explained in the revised sentence, as follows: “Altitudes range from 40 meters above sea level (masl) in the southern plains (known as the Terai region) to the summit of Kanchenjunga at 8, 856 masl…..”

Lines 144-45 - What is the significance of the dendrochronology data? In an endnote there is some discussion of the ecological impact of harvesting, but I did not see the connection to the present study. Did I miss something here?

This reference has been removed and placed in a proposed Endnotes section as per the instructions of Reviewer 1, where it now reads: “The average age and diameter of four samples collected at Lhonak was 52 years at 5.08 cm (2 inches)(Cook 2019, pers. comm.).” The impacts of contemporary adventure tourism upon Himalayan and Andean alpine ecosystems happens to be one of the specialty areas of PI Byers, with previous published studies conducted in the Everest (Nepal), Makalu-Barun, Tibet, and Peru. Data on conditions in the Kanchenjunga region were collected as a tangential component to the GLOF study, since no information on alpine impacts is known to exist. Like endnote iv concerned with the impacts of road building in the region, the alpine impacts endnote was included as some of the first observational as well as quantitative data to emerge from the Kanchenjunga region, and thus of potential interest to a range of disciplines and management concerns. Please see the references contained in this Endnote for further information related to the importance of alpine ecosystem and tourism impacts studies in the sustainable management of the Himalayan alpine zone.

 

 

Lines 172-3 - I think “the 1980 Nangama GLOF (map symbol 2)” should be “the 1980 Nangma GLOF (map symbol 3 [three]).”

L291—the map symbol number has been corrected as follows: “In addition to the 1980 Nangama GLOF (map symbol 3), participants identified an additional five GLOFs that had occurred in the region since 1921 (Map symbols 1, 2, 4, 5, and 8)….”

Line 187 - Here and elsewhere, the Nangama GLOF is now referred to as the Yangma GLOF and the names are used somewhat inconsistently; cf. lines 36, 239, 357 for example. I appreciate that locations and features can be referred to differently by different cultures and local peoples, but this can be acknowledged and for the sake of clarity for the reader one name should be used consistently thereafter.

 “Nangama” has replaced “Yangma” whenever the text is referring to the GLOF throughout the manuscript. Where “Yangma” refers to the village of Yangma, this has been clarified by using “Yangma village.”

Line 269 - Here and elsewhere, the term “englacial conduit flood” is used. I have some familiarity with ice-dam failure and resulting GLOFs and have not yet come across this term. At first I thought this to be akin to subglacial tunnel enlargement and ice dam failure. Apparently (via the Rounce et al. reference) this refers to floods resulting from englacial conduit enlargement and subsequent rapid drainage of supraglacial (and subglacial?) lakes. The authors might briefly define this term for the benefit of readers, like me, that are less familiar with this mode of catastrophic water discharge.

Beginning at line 391, an expanded definition of englacial conduit floods is presented as follows:

 

“The source of the flood is unknown, but may have been an englacial conduit flood from the Ramdam or other glacier to the northeast, similar to those reported for the Lhotse glacier in Khumbu for the years 2015 and 2016 (Rounce et al. 2017). As opposed to GLOFs, usually triggered by catastrophic ice avalanches and resultant dam breaching, englacial conduit floods “…develop on debris-covered glaciers…through cut-and-closure mechanisms with meltwater streams, the exploitation of high permeability areas that provide alternative pathways to the impermeable glacier ice, and through hydrofracturing processes (Rounce et al. 2017, citing Gulley and Ben 2007; Benn et al. 2009; Gulley et al. 2009a; and Gulley et al. 2009b; see also: Benn et al. 2017). In general, however, little concern was….”

 

 

All additional references as cited in the above paragraph have been entered into the paper’s reference section.

 

Line 288 - “In the next section...” Shouldn’t this be “this section” or the “current section?”

Line 114: “In the next section…” has been changed to “In this section,…..”

 

Lines 313-15 - What about the vertical (elevation) uncertainties? Wouldn’t they be important in determining lake surface (and other) elevations, hence lake volumes?

Here, optical images were used to determine the surface area of the lakes, while digital elevation model for the respective inventory years are not available. Hence, we don’t have uncertainty associated with the elevation. We are limited to surface area of the lake and not volume of the lakes. There are empirical techniques to estimate the volume of the lakes, which is very rough estimation and proportional to the surface area. Therefore, we did not apply those technique for estimating the volume of the all glacial lakes, this can be used in further study related with GLOF modeling, hazard mapping and others. No changes to the manuscript have been made.

Line 325 - Perhaps make clear or explicit that the 5000 m² threshold is a minimum.

We used a common threshold of 5,000 m² to prepare a consistent decadal inventory from 1964 to 2018 since satellite images used were from different sensor and different spatial resolution, which is easy to compare the inventory and trend between the years. In order to give more detail for the recent year, we prepared an inventory of the lakes using Sentinel-2 images of 10 m spatial resolution, where we used a threshold of 500 m².

Lines 333-4 - “...and trend assessment component.” I do think these tables provide or reveal a trend (which is interesting) but there’s no “component” or quantification on the tables themselves; i.e. the reader must infer the trend. Rather the trend is shown on Fig. 6.

Relatedly, on these tables, are the areas given the sum of all lakes for a given year and sub-basin? This should be made clear

L442: We have replaced table 4 with a new table that shows the trend of changes in number and area of the glacial lakes between different periods as follows: The description for table 4 also revised as follows:

 

The number and area of the lakes with size greater than 5,000 m² have grown significantly in the past century, while the largest number and area increase was observed between 1975 and 1987 (Table 4).

 

Table 4. Changes in number and area of the glacial lakes with size >5,000 m² between different periods.

Period	1964-1975		1975-1987		1987-2000		2000-2010		2010-2018		1964-2018
Basins	No.	Area	No.	Area	No.	Area	No.	Area	No.	Area	No.	Area
Tamor  (at Olangchun Gola)	1	-0.2	17	0.63	1	0.05	1	0.04	10	0.12	30	0.64
Yangma	2	0.56	8	0.47	7	0.24	4	0.26	2	0.05	23	1.58
Tamor (at Lelep)	3	0.07	1	0.06	0	0.01	2	0.02	0	0.04	6	0.2
Ghunsa	7	0.38	26	0.34	0	0.11	8	0.17	9	0.07	50	1.07
Simbuwa	-2	0.14	18	0.32	-1	-0.03	5	0.04	-3	0.02	17	0.49
Total	11	0.96	70	1.83	7	0.37	20	0.53	18	0.31	126	4

 

Line 362 - “this paper’s pre-GLOF estimate [of lake area] of 1.1 km².” Perhaps the authors might point out that this value is derived subsequently (specifically line 452). I initially thought this value was presented earlier in the paper

Since the 1.1 km2 estimate is based on the DEM and mainly relates to the numeric modeling, I would keep it where it is simply include a reference to it (e.g. “(see details of the numeric modeling scheme below)”) as suggested by the reviewer.

Line 425 - Shouldn’t a reference be provided for the Heller-Hager empirical model?

Line 764: the Heller-Hager reference has been added at line 779 as: Heller, V, Hager, W, Minor HE (2009) Landslide generated impulse waves in reservoirs: basics and computation. Laboratory of Hydraulics, Hydrology, and Glaciology, ETH Zürich 172 pp

               

Lines 450-52 - The authors might be more forthcoming in the use of the Cook and Quincey (2015) empirical relationship between lake area and water depth as that paper shows lake area is a poor predictor of depth. This does not mean that relationship shouldn’t be used, but the uncertainties associated with depth can be - if I am interpreting the Cook and Quincey graph correctly- on the order of 10s of meters for a lake of the size considered in the current paper.

Agreed; although this is a rough estimate, lake depth does not considerably alter the characteristics of the overtopping wave and thus would likely not affect downstream flood peaks. The following has been added to Line 459: “While this is a rough estimate (± 10 m for a lake of Nangama’s size), the Heller-Hager model produced very similar wave heights (8.0 m versus 8.3 m amplitude for a depth of 28 m or 48 m, respectively) and thus 38 m was deemed acceptable for modeling purposes. Further uncertainty in the lake depth was considered by estimating bathymetry in two ways…”

Line 459 - Unless I’ve missed something here as well, there was no previous mention that an avalanche was the trigger for the seiche wave so this line reads rather matter-of-fact. This also addresses the criticism above regarding the succession of events resulting in, and modeling of the GLOF.

Line 178—Reference to the avalanche as a triggering mechanism has been entered earlier in the paper on line 178, specifically as: “The 1980 Nangama GLOF—triggered by an avalanche that entered the lake and generated a large impulse wave—resulted in a considerable loss of life, infrastructure, and property downstream……

Line 466 - I did not have the supplemental material, but I presume more details regarding the Heller-Hagar method would provide some indication of the nature of this approach?

Line 187: While we retained the Heller-Hager calculations (a large number of empirical equations) in supplemental material for brevity, the following has been added when introducing the model: “Validation of BASEMENT’s wave generation and propagation procedure was achieved by additionally employing an empirical model (the “Heller-Hager model;” Heller et al., 2009), which uses basic morphological attributes of the lake and surrounding slopes to characterize avalanche-induced impulse waves in lakes.”

Line 474 - It is unclear to me why “ice and snow avalanches” would have a density of 1000 kg m^-3, i.e. of water. Glacial ice is ~900 and snow less. Is there an assumption that debris is contained in the ice and snow? Also, the Schneider et al. reference is missing in the references section.

Line 243--Yes, we assume here that some debris is contained in the avalanche. The manuscript has been corrected to “ice and snow-dominated avalanches.” The Schneider et al. reference has also been added to the reference section.

Lines 505-6 - These lines also underscore the need for more details of the modeling approach used as suggested above. This is maybe the first explicit statement indicating that the BASEMENT model provides an estimate of wave heights.

Again, we apologize for the lack of clarity and have updated the manuscript to explicitly mention the process chain (see Lines 412-437), as stated in the responses above.

References

As noted above, the Schneider et al. 2014 reference is missing.

                The reference has been added as in the previous response.

Tables

Also as noted above, in Tables 3 and 4 it should be made clear that the areas given for the specific sub-basins are the cumulative area represented by n number of lakes in each basin (assuming this is correct!).

A new table has been prepared to show the change in number and area of the lakes between different periods.

The area next to name of the sub-basins in table 3 are the respective sub-basin’s area not the cumulative area of the lakes. To make this clearer, the column name has been changed to sub-basin area (km²instead of simply area (km²)

Figures

On Fig. 14, an explanation should be provided for the different colored lines. I presume each is associated with a specific scenario as per the caption. This figure also suffered from formatting issues; axis labels are printed over axis values.

Figure 14 (not corrected to Figure 15) has now been corrected to include a legend, which was not visible in the original manuscript because of formatting problems.

 

 

Submission Date

02 June 2020

Date of this review

11 Jun 2020 15:24:31

 

Round 2

Reviewer 1 Report

In my opinion, the authors have done a solid job in addressing my comments. When reading the revised version of the manuscript, I just noted a few minor issues that the authors may like to address before a final acceptance.

 

Minor issues:

L21: This sentence could probably be rewritten in a more elegant way to avoid the brackets around “(river)” and “(lake)”.

 

L43: Check the spacing between words in this sentence.

 

L63: It should be “Watanabe et al. 2016”.

 

L71-73: If possible, it will be good to include the elevation of the Taplejung meteorological station. Precipitation changes significantly with elevation (in addition to other variables) in mountainous areas such as the Himalayas.

 

L79: I think that it is more common to write “m asl” with spacing between m and asl.

 

L80: Typo in the altitude of the Kanchenjunga summit. I guess that it should be 8,586 m asl.

 

L85: The endnote should be “ii”, not “iii”. Also, correct this in the endnotes.

 

L85: In my opinion, the text that has been removed to the Endnotes section does not belong in a scientific paper, although I understand and sympathize with the authors’ argument that some readers, including the local government, may find it interesting. I would recommend that the authors use this text in a popular science article about GLOFs in the Kangchenjunga region, maybe to a Nepalese popular science journal or online science site. I think that the text fits better there and receive a higher impact on the non-scientific community.

 

L127: “… CORONA and Landsat MSS and Landsat 5 data platforms …” - this sentence is unclear as MSS instruments have been on several Landsat platforms. Maybe rephrase the sentence to “… CORONA and Landsat 4 and 5 …” – if data is from the Landsat 4 and 5.

 

L279: Consider deleting “see” in this sentence.

 

L397-402: GLOFs are not usually triggered by ice avalanches – this is a rare phenomenon compared to GLOFs triggered by sudden drainage of meltwater from ice-dammed lakes. Also, it is still not clear to me where the water for “englacial conduit floods” comes from. My guess would be that the water origins primarily from a supraglacial sources such as rapid melting of snow in the spring, supraglacial lakes in the debris-covered part of the glacier or an extreme melting event during the summer. The cited text just describe potential englacial pathways of meltwater in general, so there is no need to cite this text. Maybe instead of using the term “englacial conduit floods”, the authors could change the text to reflect that the likely source of the flood (L395) is supraglacial meltwater, as there are no indications of drained ice-dammed lakes or ice avalanches.

Author Response

Open Review

(x) I would not like to sign my review report 
( ) I would like to sign my review report 

English language and style

( ) Extensive editing of English language and style required 
( ) Moderate English changes required 
(x) English language and style are fine/minor spell check required 
( ) I don't feel qualified to judge about the English language and style 

 

	Yes	Can be improved	Must be improved	Not applicable
Does the introduction provide sufficient background and include all relevant references?	(x)	( )	( )	( )
Is the research design appropriate?	(x)	( )	( )	( )
Are the methods adequately described?	(x)	( )	( )	( )
Are the results clearly presented?	(x)	( )	( )	( )
Are the conclusions supported by the results?	(x)	( )	( )	( )

Comments and Suggestions for Authors

In my opinion, the authors have done a solid job in addressing my comments. When reading the revised version of the manuscript, I just noted a few minor issues that the authors may like to address before a final acceptance.

Again, the authors would like to express their sincere appreciation for the reviewer’s comments and suggestions on the first manuscript, which we feel has greatly improved as a result.

It the spaces below we will address the remaining minor issues identified by the reviewer.

Minor issues:

L21: This sentence could probably be rewritten in a more elegant way to avoid the brackets around “(river)” and “(lake)”.

L21—the brackets have been removed from (river) and (lake) so that the sentence now reads more smoothly as “Debris from the flood dammed the Pabuk Khola river 2 km below the lake to form what is today known as Chheche Pokhari lake.” This is technically repetitive, as “khola” in Nepali means “river,” and “pokhari” means lake, but for the non-English readers of the paper the revised sentence should make for a more flowing read.

L43: Check the spacing between words in this sentence.

L43—the spacing between the words has been corrected.There were a number of similar sentence justifications in the original MPDI manuscript, with origins unknown (probably from deletions elsewhere) which we corrected as we reviewed the previous version.

L63: It should be “Watanabe et al. 2016”.

The citation has been changed to “Watanabe et al. 2016.”

 

L71-73: If possible, it will be good to include the elevation of the Taplejung meteorological station. Precipitation changes significantly with elevation (in addition to other variables) in mountainous areas such as the Himalayas.

L71-73—The elevation of Taplejung is 1, 441 m. Since “m” is a universally understood symbol for “meters,” we suggest that we delete the “m asl” further down and simply use “m” starting with Taplejung’s altitude. Thus, sentence 71 now reads, “For example, the village of Taplejung (1441 m, and 100 km south of Nangama glacial lake) averages…………

The following sentence now reads as: “…..). Altitudes range from 40 m in the southern plains (known as the Terai region) to the summit of Kanchenjunga at 8, 856 m to the north within a relatively short air distance of 150 km.

 

L79: I think that it is more common to write “m asl” with spacing between m and asl.

Please see the above response. If the reviewer prefers “….meters above sea level (m asl)” then we would be happy to make the adjustment

 

L80: Typo in the altitude of the Kanchenjunga summit. I guess that it should be 8,586 m asl.

L80—this sentence has now been changed to “…to the summit of Kanchenjunga at 8,856 m to the north, within a relatively short air distance of 150 km” as per the discussion under comment L71-73

 

L85: The endnote should be “ii”, not “iii”. Also, correct this in the endnotes.

L85—Please see below 

L85: In my opinion, the text that has been removed to the Endnotes section does not belong in a scientific paper, although I understand and sympathize with the authors’ argument that some readers, including the local government, may find it interesting. I would recommend that the authors use this text in a popular science article about GLOFs in the Kangchenjunga region, maybe to a Nepalese popular science journal or online science site. I think that the text fits better there and receive a higher impact on the non-scientific community.

L85—both endnotes i and ii have been removed. The spelling of Kanchenjunga as used in this paper is the same as that used by the Department of National Parks and Wildlife Conservation, trekking maps, and other scientific papers, such that an explanation of the spelling used here is probably not needed as well. Likewise, we agree with the reviewer that the detail provided in endnote ii is extraneous to this particular paper, and have thus removed it.

L127: “… CORONA and Landsat MSS and Landsat 5 data platforms …” - this sentence is unclear as MSS instruments have been on several Landsat platforms. Maybe rephrase the sentence to “… CORONA and Landsat 4 and 5 …” – if data is from the Landsat 4 and 5.

L127—this sentence has been corrected to “….CORONA and Landsat 2 and 5 data platforms....."

L279: Consider deleting “see” in this sentence.

 L278—“see” has been removed from the reference citation.

L397-402: GLOFs are not usually triggered by ice avalanches – this is a rare phenomenon compared to GLOFs triggered by sudden drainage of meltwater from ice-dammed lakes. Also, it is still not clear to me where the water for “englacial conduit floods” comes from. My guess would be that the water origins primarily from a supraglacial sources such as rapid melting of snow in the spring, supraglacial lakes in the debris-covered part of the glacier or an extreme melting event during the summer. The cited text just describe potential englacial pathways of meltwater in general, so there is no need to cite this text. Maybe instead of using the term “englacial conduit floods”, the authors could change the text to reflect that the likely source of the flood (L395) is supraglacial meltwater, as there are no indications of drained ice-dammed lakes or ice avalanches.

L397-402—the reviewer is correct in that a more technically accurate description is “glacier outburst flood,” with flood waters originating from surficial meltwater ponds as well as sub-surface, water-filled conduits or caves within the glacier. Himalayan GLOFs, however, are in fact usually triggered by ice/snow avalanches leading to morainal collapse, which we document with appropriate references. The paragraphs have been re-written as follows:

“This informant also reported that another small flood of the Ghunsa river had occurred in “either August or September” of 2011, the approximate time that a deadly earthquake (6.9 magnitude) centered in the Kanchenjunga region occurred that killed an estimated 111 people (BBC News 2011). The source of the flood is unknown, but may have been a glacier outburst flood from the Ramdam or other glacier to the northeast, similar to those reported for the Lhotse glacier in Khumbu for the years 2015 (Byers et al. 2017) and videoed in 2016 (Rounce et al. 2017). As opposed to GLOFs, which in the Himalayas are usually triggered by catastrophic ice avalanches and resultant moraine dam breaching and collapse (Richardson and Reynolds 2000; Falatkova 2016), water sources from glacier outburst floods originate within the glacier itself, i.e., from its inner system of englacial conduits and supraglacial ponds that are often inter-connected (Rounce et al. 2017; Benn et al. 2017; Benn 2007; Benn et al. 2009; Gulley et al. 2009a; Gulley et al. 2009b).  The sudden and rapid drainage of a large meltwater pond can trigger the release of water stored within conduits and other surficial ponds located further down the glacier, collectively forming a potentially dangerous flood downstream (see: https://www.youtube.com/watch?v=UM0UnoDGEAc).

 

In general, however, little concern was expressed by informants regarding future GLOFs in the region. The one exception was the Nupchu Pokhari above Khangpachen (Figure 1), singled out by informants as a growing and potentially dangerous glacial lake that should be avoided (Figure 1).

Furthermore, references to “englacial conduit flood” have been replaced throughout the remainder of the manuscript with “glacier outburst flood” where needed or appropriate, such as at L523, L537, and L620.

 

Submission Date

02 June 2020

Date of this review

24 Jun 2020 15:23:03