In 1968 a road was constructed along the coast on the western side of the Tröllaskagi peninsula in central north Iceland. The road, which until 2010 was the only whole year road to the town of Siglufjörður, crosses three large landslides, the Hraun landslide in the south, the Þúfnavellir landslide and the Tjarnardalir landslide in the north, in an area named Almenningar. Since its construction extensive damages have occurred on the road often causing hazardous conditions.

In 1977 the Icelandic Road and Coastal Administration began to monitor the deformation. In the beginning the measurements were achieved with several years intervals, but over the last decades yearly measurements have been carried out. In the year 2022, nine GNSS stations were installed along the road and a rain gauge, giving us for the first time the possibility of 24/7 monitoring on the displacements and connect the movement to weather variations, such as temperature variation and precipitation.

The dataset, which spans now over 47 years, gives us a unique opportunity to correlate the displacements on the road to external factors. Written source of deformation in the Almenningar area dates back to 1916 and since then more than 50 movement events have been listed affecting the road.

These measurements show the deformation along the road, but recent studies using “feature tracking” and InSAR show us that the whole landslide masses show signs of movement.

Our studies show that the highest movement rate takes place along the frontal parts of the landslide masses and that the movement is strongly related to both weather variations, e.g. precipitation, snowmelt and coastal erosion.

How to cite: Sæmundsson, Þ. and Geirsson, H.: Interaction between weather variations and large scale displacements along the Siglufjarðarvegur road in the Almenningar area, in central North Iceland, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8775, https://doi.org/10.5194/egusphere-egu25-8775, 2025.

The reactivation of earthflows in fine-grained geological media represents a complex phenomenon characterised by transitions from prolonged dormant phases to sudden accelerations.  While dormant stages typically exhibit slow movements (less than 1 m/year), critical rainfall conditions may trigger rapid surges in which the landslide mass can attain velocities up to several meters per hour within a limited time frame. Despite the extensive literature on the subject, the mechanisms and dynamics underlying this peculiar behaviour remain incompletely understood, largely due to challenges in acquiring direct field data that accurately capture these episodic events.

This study presents field data documenting the October 2024 reactivation of the large, dormant Ca’ di Sotto earthflow, located in the Northern Apennines (Italy) within the municipality of San Benedetto Val di Sambro. During the initial stages of reactivation, adverse weather conditions, including persistent fog and rainfall, severely hindered direct visual observation and aerial monitoring of the landslide's evolution. To overcome these challenges, a GNSS-based monitoring system was promptly deployed, comprising 31 evenly distributed periodic measuring points (surveyed daily) as well as three dual-frequency permanent GNSS stations.

GNSS data revealed an exceptionally rapid reactivation of the Ca’ di Sotto earthflow. The initial failure quickly propagated from the source area through the entire 2-km-long landslide body within a few days irreversibly compromising the functionality of a water bypass system built at the toe of the earthflow along the Sambro Stream after a previous reactivation in 1994. The failure of this bypass caused a critical water level rise in an upstream impoundment that had formed during the 1994 event.

In the following weeks, as precipitations significantly subsided, the landslide mass progressively decelerated, transitioning from peak velocities of 100 – 150 m/day recorded during the initial phase to rates of a few cm/day. At this stage, the monitoring system was enhanced with periodic drone surveys and a robotic total station, providing hourly measurements with millimetric precision across 24 regularly distributed monitoring prisms. Particularly, two transverse prism arrays were strategically installed at different elevations to serve as early warning systems for potential future reactivations.

Additionally, emergency hydraulic risk assessments were conducted, examining plausible scenarios of river blockage, impoundment water level fluctuations and management with contingency water pumping systems. These scenarios were evaluated considering ad hoc impoundment characteristic curves and hydrographs derived for design rainfall events, following the standardised NRCS (Natural Resources Conservation Service) unit hydrograph methodology.

How to cite: Zuccarini, A., Ciccarese, G., Dal Seno, N., Bartola, M., Rani, R., Zamboni, L., Caputo, G., Carboni, R., Fantini, A., Monti, L., and Berti, M.: Field monitoring of the recent reactivation of the large dormant Ca’ di Sotto earthflow, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6962, https://doi.org/10.5194/egusphere-egu25-6962, 2025.

In Nepal, efforts to establish landslide-triggering rainfall thresholds across multiple scales are well underway, aiming to enhance the effectiveness of landslide early warning systems (LEWS). These thresholds are being developed at regional, provincial, municipal, and single-slope levels, supporting landslide prediction across diverse geographic and administrative contexts.

In the vast and landslide-prone Himalayan region, establishing a regional rainfall threshold is crucial. Analysis of historical data from 1951–2006 covering 677 landslides identified a threshold relationship between rainfall intensity and duration, revealing that daily precipitation exceeding 144 mm significantly increases landslide risk. This regional threshold, developed by Dahal and Hasegawa (2008), serves as a valuable basis for early warnings across the Nepal Himalaya, providing essential risk management information for large-scale events.

Landslides in Nepal are frequently triggered by high-intensity rainfall, seismic activity, and hillside modification. A study using satellite rainfall data and landslide records from 2011 to 2022 developed a provincial-level threshold for Bagmati Province. The threshold equation at a 5% non-exceedance probability indicates that even low rainfall levels can trigger landslides due to geological weaknesses, particularly following the 2015 Gorkha Earthquake. This provincial threshold has been validated through real-time analysis, establishing a robust foundation for LEWS of Bagmati Province.

At the municipal level, rainfall thresholds have been developed for Helambu and Panchpokhari Thangpal municipalities. Using inventory data, intensity-duration threshold equations were established with high accuracy, showing strong predictive capability for landslides in these areas. The correlation between landslide susceptibility and terrain features underscores the importance of localized LEWS and community awareness initiatives to improve response during intense rainfall events.

On a single-slope scale, a physically based model was used to establish a landslide threshold for a slope on the Narayangadh-Mugling road. Here, variations in pore water pressure were analyzed under different rainfall return periods, revealing that increased topographic hollow size amplifies pore water pressure, which elevates landslide risk. The slope at Nau Kilo, with an extremely low safety factor, is highly susceptible to collapse during heavy rainfall, underscoring the need for targeted monitoring and stabilization at high-risk sites.

Developing these multilevel rainfall thresholds, tailored to Nepal’s diverse landscapes, provides essential tools for advancing LEWS and reducing landslide impacts on vulnerable communities. Enhancing rain gauge density, ensuring consistent landslide data management, and refining thresholds continuously will further improve prediction accuracy, offering valuable insights for disaster preparedness and community risk reduction across landslide-prone areas of Nepal.

How to cite: Dahal, R. K.: Multi-Scale Rainfall Thresholds for Landslide Prediction: Advancing Early Warning Systems in Diverse Landscapes of Nepal, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-276, https://doi.org/10.5194/egusphere-egu25-276, 2025.

Early detection and monitoring of rock slope instabilities are critical due to their sudden onset and significant risks to life and infrastructure. Understanding the factors controlling the dynamic evolution of rock slopes towards catastrophic failure remains a major challenge as mechanisms driving the failure occur at depths inaccessible to surface-based measurement techniques. 

Once rock bridge failures grow and coalesce to a continuous failure plane under (sub)critical stress, a rockslide enters the mobilisation phase. From then, it creeps or slides until it evacuates the source area. For many hillslope instabilities, it is unclear how the interplay between internal mechanisms and external, often meteorological drivers governs the time to collapse and the extent of structural damage during displacement.

In Brienz/Brinzauls, Switzerland, near-field seismic data from a network of geophones and broadband sensors captured precursory signals originating on or within the active “Insel” compartment of a large landslide complex, as it accelerated from 50 mm/day in late April to over 5000 mm/day, before its collapse on 15 June 2023. During prolonged mobilisation, we analyse the link between precipitation and internal mechanisms and assess how these internal processes independently drive the unstable rock mass to catastrophic collapse in the absence of external meteorological forces.

We apply a supervised XGBoost machine learning model based on seismic features to detect and classify surface rockfall events and sub-surface micro-earthquake events (internal rock bridge failures and basal stick-slip) from continuous seismic time series. 

Initial increases in surface and sub-surface event rates were rainfall-driven, with sub-surface event spikes lagging behind surface events due to progressive water infiltration into the landslide mass. After rainfall ends, surface event rates decrease earlier than sub-surface event rates as water drains from the landslide mass. Rocksliding transitioned to a phase of internal control, leading to the nonlinear evolution of surface and sub-surface events until the main collapse, in the absence of rainfall. After the transition, subsurface activity accelerated without a corresponding change in rockfall activity. Rockfall activity from the "Insel" increased after a 9-day lag, likely driven by the upward propagation of stress imbalances caused by an enhanced rate of basal sliding. A continuous decrease in sub-surface events per unit slip indicates rate-weakening behaviour at the sliding surface with slip progressively eroding asperities, reducing frictional resistance. In this context, the disintegration of rock fragments along the sliding surface generates transient families of repeating seismic events characterized by near-identical waveforms.

Our observations underline the critical role of dynamic roughness evolution at the sliding interface in governing rock mass mobilisation, with the transition from meteorologically driven sliding to internally controlled acceleration predominantly reflected in basal stick-slip and internal cracking, rather than surface rockfall activity. This highlights the need for spatially extensive monitoring of rock-internal processes to understand the non-linear dynamics of large slope instabilities during failure preparation, beyond precipitation-based models.

How to cite: Dash, S., Dietze, M., Zhou, Q., Makus, P., Walter, F., Fulde, M., Turowski, J., and Hovius, N.: Seismic precursors reveal the role of internal processes in driving mobilisation of the 15th June 2023 Brienz/Brinzauls Rockslide, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7312, https://doi.org/10.5194/egusphere-egu25-7312, 2025.

Understanding particle fragmentation and its resulting particle-size distribution is crucial for interpreting shear zone behavior in geological processes like faulting and landslides, especially under high-stress conditions. This study uses the 3-D fractal dimension (D3) to measure particle-size distribution and potential self-similarity. While previous models predict D3 values around 2.58 or 3.0, field data show significant variation. We conducted rotary shear experiments to investigate how D3 evolves with shear displacement under different normal stresses, velocities, and mineral compositions. Our results show that D3 increases monotonically with shear displacement, converging to an ultimate value highly dependent on mineral composition, but much less affected by normal stress and shear velocity. A modified large-strain model incorporating size-dependent grain-breakage probability is proposed, which explains the divergence of D3 from previous predictions. This model highlights the complexity of particle fragmentation in dense grain flows and provides a possible explanation for the high but variable D3 observed in natural shear zones. Further, we acknowledge that additional mechanisms, such as abrasion and grinding, can further contribute to particle size reduction. This study offers valuable insights into the dynamics of particle fragmentation in geological shear zones.

How to cite: Ge, Y., Hu, W., and Li, Y.: Fractal Dimension Evolution in Dense Granular Flows: Insights from Rotary Shear Experiments, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3981, https://doi.org/10.5194/egusphere-egu25-3981, 2025.

Determining the shear-velocity dependence of dry granular friction can provide insight into the controlling variables in a dry granular friction law. Scattered laboratory results suggest that granular friction is greatly affected by shear-velocity (v), but shear experiments over the large range of naturally occurring shear-velocities are lacking. Herein we examined the shear velocity dependence of dry friction for three granular materials, quartz sand, glass beads and fluorspar, across nine orders of magnitude of shear velocity (10^-8 m/s - 2 m/s). Within this range, granular friction exhibited four regimes, following a broad approximate "m" shape including two velocity-strengthening and two velocity-weakening regimes, and we discussed the possible physical mechanisms of each regime. This shear velocity dependence appeared to be universal for all particle types, shapes, sizes, and for all normal stresses over the tested range. We also found that ultra-high frequency vibration as grain surfaces were scoured by micro-chips formed by spalling at high shear velocities, creating ~20 µm diameter impact pits on particle surfaces. This study provided laboratory laws of a friction-velocity (μ-v) model for granular materials.

How to cite: hu, W.: Variation in granular frictional resistance across nine orders of magnitude in shear velocity, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4620, https://doi.org/10.5194/egusphere-egu25-4620, 2025.

The importance of erosion processes in influencing the long-distance travel of geophysical mass movements (such as debris flows, rock and snow avalanches, and landslides) is well recognized. However, numerical modeling of these processes remains difficult and is frequently overlooked. Typically, researchers have neglected entrainment or employed empirical models, where the entrainment parameters must be back-calculated to achieve the observed erosion volume and runout. Instead, in this work, we use a two-phase depth-resolved model, within an elasto-plastic framework, utilizing a dilatant Mohr-Coulomb constitutive model based on Terzaghi's effective stress principle. This model effectively captures large deformations and the interactions between solid and liquid phases in water-saturated soils subjected to overriding granular flows. Consequently, it naturally simulates bed liquefaction—the transition of initially solid soil into a liquid-like state—when overridden by debris material. The simulations reveal that the initial characteristics of the bed material, such as its permeability and consolidation degree, are crucial in influencing pore pressure generation and dissipation, degree of bed material mobilization and flow travel distance, consistent with observations from natural events. The study highlights the need to consider ground hydrological and geotechnical properties when predicting landslide hazards while also offering a detailed quantitative analysis of how bed mechanical properties influence the potential for liquefaction. Since bed material properties can potentially be measured through laboratory and field tests, the two-phase depth-resolved model has the capabilities to predictively simulate real events.

How to cite: Vicari, H., Tran, Q.-A., Metzsch, M., and Gaume, J.: Two-phase depth-resolved numerical model captures debris flows entraining water-saturated sediments, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6888, https://doi.org/10.5194/egusphere-egu25-6888, 2025.

Water plays a crucial role in the initiation and runout patterns of most landslides. Variations in the degree of saturation and porewater pressure influence the strength of landslide materials, thereby determining the final runout for a given topography. These properties can vary within a single landslide body, leading to different movement patterns and mobility across different sections, with associated implications for the landslide hazard and risk. This complexity poses challenges for numerical modelling aimed at accurately predicting landslide runout.

In this study, we used Material Point Method—a hybrid Lagrangian-Eulerian method—coupled with mixture theory (Tampubolon et al., 2017) to simulate elastoplastic deformation and runout behaviour of the landslide body. To better capture the evolution of movement patterns with minimal manual constraints and to enhance the accuracy of runout predictions, we integrated an excess pore pressure generation curve (Wang, 1999) into the computational workflow. This allowed us to simulate the excess pore pressure induced by the negative dilatancy of the solid phase under conditions of rapid motion or low permeability. The integration of this mechanism captures the effects of dilatancy, which arise from compaction and grain crushing in the sliding zone during the runout process. We show that by accounting for this localised material strength loss and the pore pressure dissipation, the evolution of landslide movement and landslide runout may be more accurately simulated.

The model was validated against a two-dimensional cross-sectional slope failure scenario with varying permeability conditions. Subsequently, it was applied to two typical multi-pattern landslide cases: a giant loess landslide on the Qinghai-Tibet Plateau and another one in London Clay on the northern shore of the Isle of Sheppey. The initial state of the slope was reconstructed based on pre-landslide digital elevation model data, while the groundwater variations, driven by either rainfall or tidal influences, were modelled as the triggering factors. This approach effectively captures the localised pore pressure effects, thereby improving the accuracy of runout distance and area predictions. We expect our model to be broadly applicable to improve runout simulation and associated hazard assessment for a broad range of hydrologically modulated landslides.

How to cite: Chen, Y., Van Wyk de Vries, M., and Wang, F.: Improved landslide runout prediction by integrating the pore pressure response to dilatancy, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12359, https://doi.org/10.5194/egusphere-egu25-12359, 2025.

Erosion can tremendously amplify the volume and destructive potential of mass flows with spectacularly increased mobility. However, the mechanism and consequences of erosion and entrainment of such flows are still not well understood as these processes are inherently complex due to the composition of the flow as well as the erodible bed material and their physical properties. Erosion rate, erosion velocity, and momentum production are the key factors essentially controlling all the processes associated with erosive mass transport. Here, we present experimental results on the dynamics of impact-induced mobility of erosive mass flows. Experiments are conducted at the Laboratory Nepnova – Innovation Flows in Kathmandu using some native Nepalese food grains as well as geological granular materials. As we focus on erosion in the inclined channel, transition, and run-out zone, we determine how the flow and the bed conditions control the erosion rate, erosion velocity, and momentum production. This includes the change in volume, composition, and physical properties of the released mass and the erodible bed and its slope. We establish some quantitative functional relationships among the erosion rate, the erosion velocity, and the mobility of the mass transport aiming at providing a foundation for developing predictive models and innovative strategies for erosion control and mitigation from landslide hazard.    

How to cite: Tiwari, C. N., Dangol, B. R., Kattel, P., Kafle, J., and Pudasaini, S. P.: The dynamics of impact-induced erosive mass flow mobility, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14875, https://doi.org/10.5194/egusphere-egu25-14875, 2025.

Earth dams, either natural or developed as part of mining operations (tailing dams) are prone to failure. In particular, recent studies show that tailing dams have a worldwide failure rate close to one collapse per year [1].

In this work we present the developments done in the monitoring and risk assessment for dams; including sensor technology, real-time numerical modelling and safety factor calculation. The recent surge in the availability of sensors allows enhancing the data that can be gathered to monitor the mechanical and hydraulic state of the dams. Numerical models can be used to enrich the local information collected by the sensors (e.g. piezometers, inclinometers) and provide the current physical state of the dam.

For monitoring purposes, numerical models are only useful if they provide results fast enough to react to an unsafe state. The results presented include the works of [2] and [3], where model order reduction techniques are applied in the context of data assimilation to learn about the state of dams. A transient nonlinear hydro-mechanical model describing the groundwater flow in unsaturated soil conditions is solved using Reduced Basis method. Hyper-reduction techniques (DEIM, LDEM) are tested and show time gains up to 1/100 with respect to standard finite element methods.

REFERENCES

[1] Clarkson, Luke, and David Williams. "Critical review of tailings dam monitoring best practice.International Journal of Mining, Reclamation and Environment, 34.2: 119-148, doi:10.1080/17480930.2019.1625172, 2020.

[2] Nasika C., P. Díez, P. Gerard, T.J. Massart and S. Zlotnik. Towards real time assessment of earthfill dams via Model Order Reduction. Finite Elements in Analysis & Design, Vol. 199, 103666, doi:10.1016/j.finel.2021.103666, 2022.

[3] Nasika C., P. Díez, P. Gerard, T.J. Massart and S. Zlotnik. Discrete Empirical Interpolation for hyper-reduction of hydro-mechanical problems in groundwater flow through soil. Journal for Numerical and Analytical Methods in Geomechanics, doi:10.1002/nag.3487, 2022.

How to cite: Zlotnik, S., García-González, A., Díez, P., and Massart, T.:  Monitoring and real time risk analysis of earth dams, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11221, https://doi.org/10.5194/egusphere-egu25-11221, 2025.