the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 19 Aug 2026
National-scale debris-flow hazard indication modelling for Swiss railway infrastructure
Abstract. Debris flows are among the most destructive alpine hazards, threatening settlements and critical infrastructure with average annual damages in Switzerland of CHF 321 million. Hazard indication modelling of debris flows, among other hazards, offers the possibility to identify hazard hot spots at a regional and national scale. Additionally it lays the foundation for future scenario modelling. We present the Large-Scale Hazard Indication Modelling (LSHIM) framework for debris flows as commissioned by the Swiss federal railway (SBB) with guiding value return periods of 30, 100 and 300 years. For disposition, we combine high-resolution digital terrain models, climatic and geologic layers, and statistical modelling with a Random Forest approach. Disposition is cross-validated against the Swiss event database (StorMe) with an average accuracy of 81.4 % and application results in 16,756 possible release areas. Dynamic runout is calculated with a tailored version of RAMMS::Debrisflow, where we define hydrograph volume from MeteoSchweiz-modelled extreme precipitation. Landscape erosivity is assessed by geology and geomorphology. Results indicate 74, 99 and 110 km of railway affected by debris flow hazard for a 30, 100 and 300 year return period, respectively. The highest intensity class shows the highest relative increase under greater return periods. This approach enables a large scale hazard indication assessment based on high-quality input data. The results are a valuable base for further planning and investigations that incorporate climate change. Limitations in the methodology lie in the static definition of release area and deterministic parameterization.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Natural Hazards and Earth System Sciences.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(2536 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-4604', Anonymous Referee #1, 15 Sep 2026
General CommentsThis manuscript presents a comprehensive national-scale hazard indication modeling framework for debris flows along the Swiss federal railway (SBB) infrastructure, integrating a statistical Random Forest susceptibility model with dynamic runout simulations via RAMMS::LSHIM-Debrisflow. The overall framework is methodologically sound, rigorously validated against the StorMe event database, and provides actionable insights for natural hazard risk management on critical infrastructure.I recommend publication after the authors address the following minor points.Specific Comments
- Rationale for the 750 m PRA Spacing: The authors mention placing Potential Release Areas (PRAs) along susceptible channels at a fixed interval of 750 m. While adopting a static disposition is well-justified for computational efficiency across a national domain, the rationale behind selecting 750 m specifically is not detailed. Please add a brief explanation or citation justifying this specific interval (e.g., balancing computational burden with typical initiation reach lengths).
- Justification Regarding Short-Duration Rainfall Intensities (5–10 min): The hydrograph volume parameterization uses 1-hour extreme precipitation. However, debris flow initiation and peak discharge in steep mountain channels are frequently governed by short-duration, high-intensity convective rainfall bursts (e.g., 5-minute or 10-minute peak intensity). While the authors state that 1-hour precipitation approximates concentration time and fits data availability, please add a discussion in the limitations section on why sub-hourly intensities were excluded and how this choice might impact modeled peak discharge.
- Glaciers as Hazard-Amplifying Factors in Future Work: In the current hydrological parameterization, glacier accumulation areas are modeled as buffering zones that reduce runoff response (using an adjusted runoff coefficient of 0.0). Although the introduction and discussion acknowledge cryospheric hazards such as glacial lake outbursts and permafrost degradation, the current runout model treats glaciated terrain primarily as a runoff-attenuating factor. Please expand the future work section to explicitly discuss perspectives on incorporating glacial processes as potential hazard/risk amplifiers (e.g., GLOFs, thermokarst/glacier bed destabilization, or cascading ice-sediment surges).
Citation: https://doi.org/10.5194/egusphere-2026-4604-RC1 -
RC2: 'Comment on egusphere-2026-4604', Anonymous Referee #2, 15 Sep 2026
General comments
This manuscript introduces a first national-scale debris-flow hazard indication using an LSHIM framework, applied to the Swiss railway network (SBB). The workflow is rigorous and generally well documented, and the contribution is genuine. Additionally, the use of the StorMe inventory for cross-validation is a particular strength. However, the manuscript does not fully clarify many significant choices, such as parameter values or thresholds. Therefore, I recommend publication after the authors address the points below.
Major comments:
- Selection of key parameters for the study is introduced without further detail or sufficient explanation/citation. Some examples: i) Buffer of 2000 m defining the Bilanz-gebiete; ii) The fixed interval of 750 m for the PRAs; iii) Runoff coefficients for the accumulation and ablation zones; iv) Threshold value of P df > 50%. Please briefly explain or cite accordingly.
- Consistency between the parametrization of mu and xi in the Voellmy model: The friction coefficient mu is spatially distributed from fan data, but xi is kept constant at the national scale. Since RAMMS depends on both parameters jointly, please explain why a constant value and the reason behind this specific value.
- Please clarify in the discussion section why the output is framed as a hazard indication rather than a hazard map. I follow the intent to not introduce legal constraints directly, but this should be explicitly stated.
- L259-L265: Deriving mu spatially from data sourced is a good idea. However, two clarifications are needed: i) Which criterion was applied with “after visual inspection of results”?, and ii) Why the 5th percentile and not the 10th or the 25th?
- Use of the 10 km moving window: I follow the intent, but the choice of value should be stated as well as the authors should acknowledge that this smoothing trades local variability of mu for larger spatial assessments.
- L275: The decision criteria used to define the erosion categories are not given, and it is not clear how the erosion parametrization was implemented in RAMMS nor the obtained values presented in Tables A4/A5.
- Figure 6: A national-scale visualization of the flow-height is understandably difficult. However, the authors should motivate the specific examples shown, i.e., if they were chosen to represent contrasting geological, geomorphological or other environmental settings.
Minor comments:
- L276: define Altschutt/Jungschutt for the debris cover in the text, using the definitions in Table 1.
- Section 6.2: If the Illgraben catchment is used as a representative, well-known site, state this before Figures 4 and 5.
- Section 6 partially interprets the results. Consider renaming it “Results and interpretation”.
- L340-L341: this would strengthen the outcome of this study: “Since then these approaches have been applied in large-scale modeling of rockfall (Bründl et al., 2025) and debris flow (this paper)” reviewed in “[…]. The results of this study show that a national application to debris-flow hazard can also be achieved and provide general indications for further detailed studies”.
Technical corrections:
- References are sometimes ordered chronologically and sometimes alphabetically; adopt one criterion and keep it consistent.
- Equations should be referred to by number in the text, e.g., (1).
- Ensure all variables in the equations are defined in the text before or after each equation.
- L260 typo: “which is in turn is approximated”.
- Add subfigure labels (e.g., Figure 5a-c) and refer to them in the text - Figures 6 and 8 would benefit.
- L285: Cite the reference work used to justify omitting the flow height values.
- Apply abbreviations consistently after the first definition, eg. PRA in L347.
- Figure 7: Align labels for the three areas (Emmental, Tessin and Wallis) with Figure 6 (Emmental, Ticino and Valais).
I hope these comments can help the authors refine the already good quality of the manuscript.
Citation: https://doi.org/10.5194/egusphere-2026-4604-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 177 | 92 | 72 | 341 | 61 | 60 |
- HTML: 177
- PDF: 92
- XML: 72
- Total: 341
- BibTeX: 61
- EndNote: 60
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
