Review: Exploratory analysis of the annual risk to life from debris flows - January 2024

Overview

I thank the authors for their submission to NHESS. They present an interesting topic, which inverts typical hazards estimates to evaluate how a range of debris flow return intervals might influence risk to human life. 

In essence, the authors perform a sensitivity analysis for a range of unknown debris flow return intervals to identify catchments where annual risk may warrant further investigation. The key contribution of this work is identification of a “window of non-recognition”, or a range of debris flow frequencies in a catchment where it is likely that no debris flows have been observed given the length of settlement, but where risk to human life is still elevated above a desirable threshold. This is an important conclusion and aligns with hazards practitioners who respond to natural disasters which seem to take residents by surprise. As the authors note in the conclusion, complacency may prevent emergency managers or residents from anticipating potential hazards.  

However, this key finding is currently vague, partly because it is difficult to estimate risk parameters and partly because the current manuscript has limited/unclear presentation of the specific return intervals that result in unacceptable risk. Are they intended to be global/generalizable estimates, or regionally specific? Fully characterizing hazard and risk for decision-making would require additional process-based investigation or empirical approaches that use local/regional data. I encourage the authors to describe how their model results could be used to support actionable strategies for prioritizing further risk-reduction efforts.

I suggest several major modifications that may improve the weight of the conclusions:

1. Significant re-framing to better reflect the conclusions. This could involve a more specific paper title, updates to the abstract, introduction, and conclusions that better reflect the contributions of their work. Specifically, I would suggest that the authors re-align the work with the primary contribution as described in the first paragraph of this review.
2. Deeper grounding in real-world processes. This could be accomplished through improved basis in the literature and in consideration of both physical and social processes. Most importantly, the authors mention field evidence in the conclusions, but do not introduce study areas, present field observation, or describe data collection methodologies. These are critical for understanding the scientific contribution as well as its ability to be generalized or extrapolated to other study regions.
3. Expanded/restructured introduction and methods section that present existing knowledge and the research question (introduction) and the author’s novel approach (methods).
4. Reconsideration for some of the assumptions presented as justification of the work and for the values assigned to risk calculations. Some examples are described in the line comments below.
5. Finally, the authors describe a single, universal “window of non-recognition,” which reflects the generic or best estimate parameter values for risk. However, considering that risk estimates and the window of observation relies on settlement periods, risk tolerance, and physical processes (probability of avulsion), I am not convinced that this value would apply to broad areas. Instead, I suggest that the authors present their window with a methodology framework, investigate a specific study area, and/or evaluate a wider range of parameter values for each risk parameter. 

Writing and Organization

The writing style is clear and easy to follow. As noted in the overview, I suggest some improvements to the organization of the introduction and methods that better separates the two sections. A few specific recommendations are described in the line comments, but in general I recommend separating background contextual information in the introduction, with specific methodology descriptions for the author’s work in the methods. Currently, the methods section describes the research problem, which would be better suited in an introduction section.

Line comments (organized by line number)

• 1. Consider a more descriptive title which highlights specific analysis or finding of your work.
• 24. “No history of debris flows,” is too vague. Do you mean geologic history? Oral history? Written history? Consider describing your study area, as settlement history is also highly variable around the world. Expanding urban areas and limited records absolutely result in low public awareness of debris flow hazards, but consider acknowledging that human settlement in New Zealand is relatively short (hence the challenge the authors’ research question) densely inhabited areas in Eurasia may have hundreds of years of detailed written records, and oral histories for many indigenous peoples describe landslides or other debris flows over millennia.
• 28-34. Please provide citations to the literature so that readers can seek additional context on debris flow processes, sediment pathways, and debris flow hazards in New Zealand.
• 35. This statement poorly reflects the pioneering work of many geoscientists and engineers in landslide and debris flow research in New Zealand. Consider providing context for what is known about debris flows in NZ and globally and then highlight the specific gaps that persist in debris flow knowledge, which may include things like regional hazards mapping or public preparedness.
• Section 2.1. Much of this section would be better suited to the introduction.
• 58. Table 1 does not summarize ARIs for different catchments. Please add the summary table and revise your in-text citation.
• 62. It is true that field evidence & topographic analysis can be costly to collect and process. However, I would argue that these are the best tools for developing specific hazards understanding & precise risk estimates. Your approach for risk assessment should not replace process-based assessment, but may be useful in prioritizing communities/residences for improved outreach & risk awareness.
• 68-70. Consider “catchment gradient is associated with debris flow occurrence,” and reference other specific topographically based tools for landslide susceptibility (E.g., Montgomery et al., 1994; Dietrich et al., 2001).
• 70-71. “Most common” is hard to justify. I recommend instead describing the merits of the Melton Ratio (E.g. a simple metric that identifies catchments likely to produce debris flows, which is easily calculated for many catchments over large areas, even where topographic resolution is poor or computation is limited. Also it sounds like these values have already been calculated for large areas of NZ).
• 87-89. I understand that this need for further investigation is one of the primary justifications for your work–specifically, in helping prioritize catchments for further investigations? Is that correct? If so, consider making that link more explicitly.
• 96-97. I understand that you you need to determine threshold risk values, but consider providing more context on how and why risk tolerance may vary, and why these values (10-3-10-4) are appropriate according to Taig. et al. Here or in the discussion, you may want to acknowledge that “unacceptable” risk reflects the values & tolerances of individuals and communities.
• EQ 2. Are these standard abbreviations? I found them hard to follow (where does the “H” come from? Is PS:H a ratio? E stands for exposure?) Consider using simpler abbreviations or adding some description to help readers keep track of which parameter is which.
• 132-134 / 147. I’m not sure I agree with this estimate. In my experience, very few debris flows impact the entire debris flow fan. While it is challenging to estimate the probability of avulsion, are there any estimates in the literature which might describe the distribution of areas, as a proportion of total fan area, that occur during a debris flow? De Haas and others, or works by C Scheidl, DM Staley, or D Rickenmann may be useful places to look for an estimate.
• 165-167. It may be worth adding a section in the discussion on how further investigation could be used to refine generic estimates. Sediment volume calculations, for example, could be used to improve estimates of debris flow area and deposit depths for exposure calculations. Consider expanding on this statement and adding appropriate support from the literature. 
• 177. What do you mean by permanent materials?
• 183-184. These values are lower than I would expect. Can you provide more detail about the data used to make these estimates? Considering that your estimates are intended to describe worst-case scenarios where whole debris flow fans are inundated, 0.1 seems far too low. 
• 208. I don’t agree with this assumption, and assuming the worst-case scenario conflicts with your goal of best-estimate risk calculation.
• Figure 1. What is the value shown on the y-axis?
• Figure 2. Can you provide a clearer description of the populations shown in this figure? Are these real-world catchments or monte carlo-type simulations?
• 357. The danger of complacency seems like an important application of this work. Consider adding discussion of how your model or method could be used to inform hazards communication and improve/prioritize risk awareness in at-risk communities.
• 390. No field evidence is presented here. Please add description and summary of field evidence that you use to draw conclusions. 

Some Literature Recommendations

Consider reviewing these and other works to better support your manuscript.

Bertrand, M., Liébault, F., & Piégay, H. (2013). Debris-flow susceptibility of upland catchments. Natural Hazards, 67, 497–511. 

de Haas, T., Densmore, A.L., Stoffel, M., Suwa, H., Imaizumi, F., Ballesteros-Cánovas, J.A., Wasklewicz, T. (2018). Avulsions and the spatio-temporal evolution of debris-flow fans. Earth-Science Reviews, 177, 53-75. https://doi.org/10.1016/j.earscirev.2017.11.007 

Dietrich, W. E., Bellugi, D., & Real de Asua, R. (2001). Validation of the Shallow Landslide Model, SHALSTAB, for Forest Management. Land Use and Watersheds: Human Influence on Hydrology and Geomorphology in Urban and Forest Areas, 2, 195–227. https://doi.org/10.1029/ws002p0195, 

Montgomery, D. R., & Dietrich, W. E. E. (1994). A physically based model for the hydrologic control on shallow landsliding. Water Resources Research, 30(4), 1153–1171. https://doi.org/10.1029/2005WR004369

Nirupama, N. (2012). Risk and vulnerability assessment: a comprehensive approach. International Journal of Disaster Resilience in the Built Environment. 3(2) 103-114. https://doi.org/10.1108/17595901211245189

Pierson, T.C. (1980) Erosion and deposition by debris flows at Mt Thomas, North Canterbury, New Zealand. Earth Surface Processes and Landforms, 5(3) 227-247. https://doi.org/10.1002/esp.3760050302

Scheidel, C., Rickenmann, D., (2009). Empirical prediction of debris-flow mobility and deposition on fans. Earth Surface Processes and Landforms, 35(2) 157-173. https://doi.org/10.1002/esp.1897

See supplement