the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 15 Apr 2026
Modelling climate-induced instability of ice-rich permafrost slopes
Abstract. Climate-induced slope failures in ice-rich permafrost environments typically manifest in surficial materials as newly initiated retrogressive thaw slumps (RTS) and active layer detachment failures (ALDF). Instability is linked to high pore water pressures developed during thaw of ice-rich layers that reduce the factor of safety below unity. Here we develop a module within the CryoGrid community model that uses meteorological inputs and soil geotechnical characteristics to simulate ice segregation and thaw consolidation, and predicts potential instability at all levels within the soil column through time using an infinite slope analysis. The analysis is expanded spatially using clustering of slope gradients and aspect. The model was tested using a multi-decadal database of RTS initiation for an area of 2300 km2 on Banks Island, Canada. Results showed that the Thawing Slope Stability Index (TSSI), based on the severity and duration of slope instability, was correlated with years in which tens to hundreds of RTS were initiated. These years were characterized by high summer air temperatures and high incoming short-wave radiation which led to a deepening of the active layer in the model, melting of ice-rich layers, and increased pore water pressures. Furthermore, the newly initiated RTS were concentrated on slopes with the highest TSSI values. The newly developed modelling scheme represents a significant step towards evaluating the stability of ice-rich permafrost slopes for different land use and climate change scenarios.
- Preprint
(46024 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 13 Jun 2026)
- RC1: 'Comment on egusphere-2026-916', Ian Shirley, 23 May 2026 reply
-
RC2: 'Comment on egusphere-2026-916', Sebastian Uhlemann, 29 May 2026
reply
General Response
This paper introduces a physically-based slope stability index, integrated within the geomechanical module of the CryoGrid community model, to evaluate climate-induced slope instability in ice-rich permafrost environments.
The implementation of this geomechanical scheme represents a significant step forward in modeling and understanding the impacts of climate change on permafrost slope failures. By moving beyond empirical or statistical approaches, the integration of these geotechnical formulations into a model that dynamically simulates permafrost thermal and hydrological regimes provides a valuable tool for assessing landscape-scale vulnerabilities. A particular strength of this work is the development of the Thawing Slope Stability Index (TSSI) and its application to bridge process-based modeling with regional-scale observations on Banks Island, Canada.
However, while the model successfully captures broad susceptibility trends, there are still several limitations and mechanistic gaps that the manuscript should highlight more thoroughly. Specifically, the manuscript would benefit from a more explicit discussion regarding the physical limitations of the model setup such as the treatment of subsurface lateral fluxes, bulk density changes during thaw, and the influence of variable snow distribution. Some of those points are discussed in the Limitations section, but they should be discussed upfront, not as an afterthought. Furthermore, clarifying the distinction between the simulated active layer detachment failures (ALDF) and the retrogressive thaw slumps (RTS) used for validation early in the text will better contextualize the model's capabilities and its current constraints.
Overall, the paper is well-written, timely, and makes an important contribution to the field of permafrost geohazards. Please see some more detailed comments below for further refinement.
Detailed Comments
1. Conceptual and Terminology Clarifications
- The methodology section should explicitly highlight that the developed model accounts for the failure process of ALDFs. Because ALDFs frequently initiate RTSs, validating the model against an RTS database is perfectly valid, but this specific limitation and linkage should be clearly established early on.
- The TSSI should not be explicitly termed a "likelihood," as its mathematical definition represents a cumulative duration and magnitude of risk rather than a strict statistical probability.
2. Missing Physical Mechanisms
- The manuscript asserts that elevated pore pressures are a primary triggering condition, yet the simulations do not show an obvious correlation between unstable conditions and precipitation. In many field settings, elevated landslide occurrence is closely tied to wet summers; this discrepancy and the limited direct effect of summer rainfall on the simulated slope stability require further explanation. This suggests there is still missing understanding on why and when these slopes fail.
- While the model details how the internal friction angle and cohesion adapt to ice content and temperature, it is unclear if the model accounts for the change in normal stress. The phase change from ice to water inherently increases the bulk density of the material, which should alter the normal stress. You refer to this in the discussion of the limitations, but there it is rather brief, and I think this should be discussed in the model setup to be upfront with potential limitations.
- The discussion of "convective fluxes" needs clarification on whether it refers strictly to density-driven fluxes or if it includes advective fluxes. On hillslopes, a relatively thin advective layer can result in significant lateral downslope flow. Furthermore, if the model incorporates lateral fluid flow, it should be clarified whether this simultaneously drives lateral heat flow. You also mention this in the limitations section, but again, I think that this should be highlighted earlier on, and then the impacts of not considering advective flow should be discussed in the discussion section.
3. Spatial and Topographic Controls
- The model implementation should address whether the natural variability of snow distribution is reflected. Thicker snow accumulation in topographic depressions versus thinner snow on uplands creates thermal conditions that could increase ground temperatures at the toe of a slope, potentially reducing shear strength and increasing failure risk. Whether or not this is reflected in the model remained unclear to me.
- Because the results are described based on topographical clusters, it would be highly beneficial to include a figure showing the simulated responses (either permafrost conditions or slope stability) separated by these specific clusters. This would clearly illustrate which "types" of slopes are most vulnerable and which remain stable.
- While the results suggest that high susceptibility occurs across all cardinal directions, there appears to be a higher prevalence for easterly to southerly facing slopes. This nuance should be addressed, as it aligns with the expected impact of increased solar irradiation.
You have developed an impressive tool that truly advances the state of the art in permafrost modeling. I hope this feedback is helpful as you finalize this excellent paper.
Citation: https://doi.org/10.5194/egusphere-2026-916-RC2
Model code and software
CryoGrid source code for "Modelling climate-induced instability of ice-rich permafrost slopes" Juditha Aga https://doi.org/10.5281/zenodo.18492187
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 258 | 187 | 18 | 463 | 13 | 19 |
- HTML: 258
- PDF: 187
- XML: 18
- Total: 463
- BibTeX: 13
- EndNote: 19
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This paper introduces a slope stability index within the recently developed geomechanical module of the CryoGrid community model to evaluate slope instability associated with thawing ice-rich permafrost terrain.
The geomechanical scheme is based on established geotechnical formulations and is appropriately implemented. Its coupling to CryoGrid, which mechanistically simulates permafrost thermal and hydrological dynamics, represents an important advance for physically based prediction of thaw-driven slope instability under climate change. The work is timely given the expected increase in permafrost slope failures and their implications for infrastructure, landscape evolution, and carbon-climate feedbacks.
A particular strength of the paper is the attempt to bridge process-based permafrost modeling with regional-scale observations of slope failure occurrence using the proposed Thawing Slope Stability Index (TSSI). At Banks Island, Canada, the modeled slope stability index compares favorably with long-term records of retrogressive thaw slump (RTS) initiation. I do think, however, that the paper would benefit from a more explicit and nuanced discussion of this comparison, given that the model represents active layer detachment failures (ALDF) rather than the observed RTS failures.
The paper is very well written and clearly organized, and will make an important contribution to the field. Please see some more detailed comments below.
1. Comparison between modeled ALDF and observed RTS initiation
The main conceptual issue I have with this manuscript concerns the comparison between the modeled instability and the observational dataset used for validation. The implemented mechanics represent ALDF-style instability, since thaw consolidation and elevated pore water pressures drive reduced shear strength at the permafrost table. The observational dataset, on the other hand, consists of RTS initiation, which involves additional thermo-erosional and geomorphic drivers. I therefore think the manuscript should state much more explicitly throughout (including in the abstract/discussion) that the model is simulating ALDF-style instability rather than RTS initiation itself.
I do think the comparison to RTS observations is meaningful, particularly given the strong association between widespread RTS initiation and simulated ALDF, but the manuscript should spend more time discussing why this association exists. For example, ALDF may directly contribute to RTS initiation through exposure of massive ice and subsequent thermoerosional degradation of the exposed headwall. This idea is somewhat implicit in the manuscript, but is never explicitly discussed. At the same time, the manuscript overstates the degree to which RTS initiation can be attributed specifically to ALDF, e.g. “with warm summers causing the majority of RTS in the fieldsite, their initiation is likely connected to the scar zones of ALDF, rather than slope-undercuttings or blockfalls ” (L185). I do not think this conclusion is yet fully justified mechanistically. Warm summers could plausibly enhance RTS initiation in different ways, including via thaw-driven ALDF, but also via warming accelerated fluvial or coastal thermoerosion. Indeed, the authors observe that RTS are strongly concentrated along rivers, lakes, and coastlines (~L340), which suggests that undercutting mechanisms may still play an important role. Some of the agreement between model and observations may therefore reflect the fact that the meteorological conditions that increase susceptibility to ALDF also increase susceptibility to RTS in these environments, even where the precise triggering mechanisms differ.
Overall, I think the manuscript would be strengthened by framing the model more explicitly as a simulation of ALDF-style slope instability that may contribute to the observed RTS initiation through exposure of massive ice, while also recognizing that ALDF and RTS initiation may simply be promoted by similar meteorological conditions. This would help avoid implying that RTS initiation itself is being directly simulated.
2. Aspect controls on TSSI
In section 3.3.2, you discuss the somewhat surprising result that variation in aspect has relatively little influence on simulated TSSI. This is an interesting and potentially important result, given the strong topographic control on incoming shortwave radiation that is often assumed to influence thaw-driven slope instability. However, I do not think the implications of this result are fully explored, and the discussion around ~L365-370 somewhat overstates the role of aspect relative to the presented results. I think this discussion should be revised to better reflect the weak modeled aspect sensitivity shown in Fig. 6.
Since this weak aspect dependence is somewhat unexpected, I also think it would be useful to include at least a simple comparison of observed RTS occurrence versus aspect class. If observed RTS occurrence likewise shows weak aspect dependence, this would substantially strengthen the interpretation that landscape-scale variability in thaw-driven slope instability at this site is controlled more strongly by factors such as slope angle and ground ice conditions than by aspect-driven differences in radiation loading. On the other hand, if the observations exhibit a stronger aspect dependence than the model, this could point to an important model limitation.
Minor comments