The manuscript by Greta Hoe Wells and colleagues submitted to the NHESSD focuses on the evolution and outburst hazard of the proglacial lake Fjalsjökull. While the study aims to address an important topic and the rationale is clear, it does not address it with appropriate state-of-the-art methods and approaches and reads like a preliminary assessment rather than a research article. Firstly, the linear extrapolation of glacier retreat/lake growth rates over the next eight decades regardless of climate change scenarios (RCPs), lake-glacier interactions, etc. is trivial. Secondly, the approach used to look at potential GLOF triggers (deriving mass movement prone areas from DEM analysis) is adequate for a landslide susceptibility assessment of a mountain range or a country, but not for a detailed case study of one lake. Instead, a widely used and widely accepted slope stability/deformation analysis (based on field data, based on SAR data or both) could be used to map potential release zones and estimate potential landslide volumes. Thirdly, the generation, propagation, attenuation and impacts of potential mass movement/displacement wave and a GLOF lack rigorous analysis and require the use of modelling approaches. Further processing of this manuscript is recommended only if substantial revisions and additional analyzes are carried out.

The manuscript “Proglacial lake evolution and outburst flood hazard at Fjallsjökull glacier, southeast Iceland” offers valuable insights regarding the effect of the glacier retreat on hazards related to rockfall/rock avalanche provoking a GLOF that could have a catastrophic impact on the region. The manuscript is generally well-written and organised. However, the simplifications of the glacier retreat and rock avalanche models are questionable and need closer examination to ensure the robustness of the conclusions (see specific comments). I recommend the manuscript to be reconsidered after major revisions, following which the manuscript has the potential to make a significant contribution to the field of glacier-related natural hazards.

Specific comments:

• To estimate a reasonable glacier retreat, the 2000-2021 average retreat rate was taken (l. 139-141 and 243) and kept constant over the next century. However, based on projected temperature scenarios, other studies showed that a linear glacier retreat is somewhat optimistic (e.g., Bosson et al., 2023). Similarly, performing a model run for a more pessimistic temperature curve (like SSP5-8.5) would be interesting, where the glacier retreat would keep accelerating in the coming decades. Then, it would be possible to give a range of glacier retreat dates for the different locations studied instead of a single one (and ranges in Table 1, too).
• While the effect of calving is mentioned at l.39 and 323-332, it does not influence the projection of glacier retreat in Figure 5, despite the strong change in bed depth for the N part of the glacier tongue. A faster melt in the N due to the over-deepened basin (similar to what happened in the S in recent years) would have consequences in the timing of deglaciation for two of the three identified potential rockfall source areas and could, therefore, be significant. I suggest considering this when estimating the glacier's future extent.
• At l. 173-175, the H/L ratio is defined as connecting “the highest zone point and the lowest deposit point along estimated flow path”. Accordingly, the Fahrböschung angle is defined classically. However, later in the manuscript and Figure 4, the lowest deposit point is replaced by the lake shore. Then, this ratio and angle change over time as the lake expands. In this case, it represents the maximum angle to reach the lake, not the Fahrböschung angle. I suggest modifying the terminology accordingly.
• A more detailed description of the geology and structures at the identified source zones is necessary.

Technical comments:

• The sentence at l. 38-42 could be split in two.
• l. 86: up to... -> a maximum depth of ... (same clarification needed at l. 190 and l. 192)
• Figure 1: remove the black outline around figures for all figures. In 1C, a capital N is probably to be removed. The figures could be the same width as the text. I would consider having a map instead of the orthophoto in 1C. Could Figure 2 be slightly extended to replace 1C?
• l. 113: Vertical and horizontal direction.
• l. 114: Are the multibeam readings corrected with the surface temperature or temperature profiles? The temperature data could be nice to have in the supplementary material.
• l. 119: The vertical uncertainty for the multibeam sonar survey could be added.
• Figure 2: the colours of the successive glacier extents should follow a continuous colour scale for better readability (some exist for colour-blinded, too). The sea should be blue to avoid being confused with a light grey glacier. The coordinate system used should be in the caption.
• l. 148-158: the assumption that sedimentation in the lake was negligible over the last 70 y should be made clear here already. The assumption is then discussed at l. 296-300 as a potential source of error.
• l. 150: “Manually” can be omitted here if other datasets have been digitised manually, too.
• Figure 3: Continuous colour scale for the lake extent would help readability.
• l. 174-175: The lowest point of the mass movement deposit can be at the bottom of the lake, not necessarily at the lake shore. Could you reformulate, for example, writing that H is taken from the lake surface instead of the lowest deposit?
• Figure 4: The bottom of the figure could be extended to show the contact of the bedrock below the lake and bedrock-glacier as well as glacier-lake contacts. The figure could be modified so the glacier does not stop the rockfall before it reaches the lake.
• Figure 5 should be improved following the specific comment above.
• l. 241: Is it possible to have a lake level 15 m below sea level? Or should the calculation start at 0 m instead?
• l. 250-252: can we make an assumption based on the orthophoto regarding the geology and main structural features? They are, in general, essential to understand the landslide hazard. If possible, the geology at the source zones should be better described.
• Discussion: Reworking the discussion so that the main outputs and hazard scenarios for rock avalanches in the lake appear before the study’s main sources of uncertainty could strengthen the discussion.
• l. 305-306: It would be interesting to discuss the possible evolution of the lake in case a GLOF happens in the coming years. Would a significant erosional event mean a lower lake level and intrusion of warmer seawater at each tide?
• In Figure 8, a travel distance of 0 m is estimated in 2120 between Miðaftanstindur and the lake. However, the polygon is not directly touching the lake. Similarly, for Eyðnatindur, could the rockfall/rock avalanche area extend below the current glacier surface?

Comments by Wilfried Haeberli

on

Proglacial lake evolution and outburst flood hazard at Fjallsjökull glacier, southeast Iceland

Paper submitted to Natural Hazards and Earth System Sciences by

G.H. Wells, Þ. Sæmundsson, F. Pálsson, G. Aðalgeirsdóttir, , E. Magnússon, R. Hermanns and S. Guðmundsson

General

The submitted paper describes and discusses the development in time of a proglacial lake in southern Iceland and of related hazards from potential impact/flood waves produced by landslides into the lake and threatening tourist and traffic infrastructure. The text is clearly written, its structure is logical, the illustrations are fine and reference to the scientific literature on the topic is rich. The general level of the study is comparable to a first, relatively rough assessment of climate change impacts and related changes in hazard conditions. As such a preparatory stage for more refined practical analysis, it is fine and corresponds to international guidelines (GAPHAZ 2017, Allen et al. 2022). As a scientific paper, however, it would benefit from more critical reflection and more advanced use and/or discussion of modern quantitative approaches. Such more quantitative approaches should at least be mentioned and discussed in view of their possible applicability. The following comments indicate recent examples of corresponding possibilities and references.

Comparable analyses are, for instance, Sattar et al. (2023) or Gantayat et al. (2024b). Reference could be made to such recent case studies and the involved quantitative approaches, simplifications and potentials/needs for further improvement. Concerning retreat rates of calving fronts, quantitative relations between calving speeds and water depths had been developed already decades ago (Brown et al. 1982; cf. the discussion in Vieli 2021). Such simple, application-oriented approaches should at least be mentioned or – better – as far as possible be applied. The empirical-quantitative approach concerning rock avalanche propagation by Cathala et al. (2024) could also be mentioned and applied. Modeling the propagation of potential impact waves has become quite common practice and could at least be discussed in view of the involved factors and uncertainties (see the recent overview provided by Rinzin et al. 2024).

Estimating lake volumes using volume-area self-correlations had earlier been quite common but is problematic and must either (better!) be avoided or at least be critically discussed. Statistical volume-area correlation relates a mathematical product (volume = area times depth) with one of the factors (area), from which this product had been calculated. This is not a generally accepted scientific approach but violates the fundamental basics of statistical regression (the independence of the variables to be compared) and can even be seen as a statistical data manipulation, which creates misleading results (unrealistic correlation coefficients, nice-looking double-logarithmic graphs with seemingly suppressed scatter). Huggel et al. (2002 as mentioned in the text) use a statistically correct depth-area relation and only transform it into a volume-area self-correlation for intercomparison. Cook and Quincy (2015 as mentioned in the text) state that “Empirical volume–area relationships can also give a misleading impression of the predictability of lake volumes because lake volume is dependent on area (Wang et al., 2012; Haeberli, 2015). Hence, higher degrees of correlation between lake area and volume often mask the complexity of lake basin morphometry”. Volume-area self-correlations can easily be transformed back into correct depth-area regressions by dividing the volumes through the related areas and correlating the so-obtained original depth values with the related areas. Adequate statistical treatment is, for instance, provided by Muñoz et al. (2020). A new/better empirical approach is described in Gantayat et al. (2024a).

The authors rightly state that their “study provides input data for risk assessments and mitigation strategies at Fjallsjökull; a template for investigating this hazard at other proglacial lakes in Iceland; and field data to advance understanding of overdeepenings and lake–terminus interactions in proglacial lakes worldwide”. Improving the scientific merit of their interesting submitted paper is certainly possible but needs additional reflection about the present state of knowledge in the field.

Some more technical comments are contained in the annotated file.

References:

Allen, S., Frey, H., Haeberli, W., Huggel, C., Chiarle, M. and Geertsema, M. (2022): Assessment principles for glacier and permafrost hazards in mountain regions. Oxford Research Encyclopedia of Natural Hazard Science. doi.org/10.1093/acrefore/9780199389407.013.356

Brown, C.S., Meier, F. and Post, A. (1982): Calving speed of Alaska tidewater glaciers, with application to Columbia Glacier. US Geologiocal Surbey Professional Paper 1258-C.

Cathala, M., Magnin, F., Ravanel, L., Dorren, L., Zuanon, N., Berger, F., Bourrier, F. and Deline, P. (2024): Mapping release and propagation areas of permafrost-related rock slope failures in the French Alps: A new methodological approach at regional scale. Geomorphology 448, 109032. doi.org/10.1016/j.geomorph.2023.109032

GAPHAZ (2017): Assessment of Glacier and Permafrost Hazards in Mountain Regions – Technical Guidance Document. Prepared by Allen, S., Frey, H., Huggel, C., Bründl, M., Chiarle, M., Clague, J.J., Cochachin, A., Cook, S., Deline, P., Geertsema, M., Giardino, M., Haeberli, W., Kääb, A., Kargel, J.,Klimes, J., Krautblatter, M., McArdell, B., Mergili, M., Petrakov, D., Portocarrero, C., Reynolds, J. and Schneider, D. Standing Group on Glacier and Permafrost Hazards in Mountains (GAPHAZ) of the International Association of Cryospheric Sciences (IACS) and the International Permafrost Association (IPA). Zurich, Switzerland / Lima, Peru, 72 pp.

Gantayat, P., Sattar, A. Haritashya, U.K., Watson, C.S. and Kargel, J.S. (2024a): Bayesian Approach to Estimate Proglacial Lake Volume (BE‐GLAV). Earth and Space Science, 11, e2024EA003542. doi.org/10.1029/2024EA003542

Gantayat, P., Sattar, A., Haritashya, U.K, Ramsankaran, R., Kargel, J.S. (2024): Evolution of the Lower Barun lake and its exposure to potential mass movement slopes in the Nepal Himalaya. Science of the Total Environment, 949, 175028. doi.org/10.1016/j.scitotenv.2024.175028

Muñoz, R., Huggel, C., Frey, H. Cochachin, A. and Haeberli, W. (2020): Glacial lake depth and volume estimation based on a large bathymetric dataset from the Cordillera Blanca, Peru. Earth Surface Processes and Landforms 45, 1510-1527. doi:10.1002/esp.4826

Rinzin, S., Dunning, S., Carr, R.J., Sattar, A. and Mergili, M. (2024): Exploring implications of input parameter uncertainties on GLOF modelling results using the state-of-the-art modelling code, r.avaflow.doi.org/10.5194/egusphere-2024-1819

Sattar, A., Allen, S., Mergili, M., Haeberli, W., Frey, H., Kulkarni, A.V., Haritashya, U.K., Huggel, C., Goswami, A. and Ramsankaran, R.A.A.J. (2023): Modelling potential glacial lake outburst flood process chains and effects from artificial lake-level lowering at Gepang Gath Lake, Indian Himalaya. Journal of Geophysical Research – Earth Surface. doi:10.1029/2022JF006826.  

Vieli, A. (2021): Retreat instability of tidewater glaciers and marine ice sheets. In: Haeberli, W., Whiteman, C. (Eds.), Snow and Ice-Related Hazards, Risks, and Disasters. Elsevier, 671–706.