Runout Zone Research Articles

Parameter identification by deep learning of a material model for granular media

Classical physical modeling with associated numerical simulation (model-based), and prognostic methods based on the analysis of large amounts of data (data-driven) are the two most common methods used for the mapping of complex physical processes. In recent years, the efficient combination of these approaches has become increasingly important. Continuum mechanics in the core consists of conservation equations that-in addition to the always-necessary specification of the process conditions-can be supplemented by phenomenological material models. The latter are an idealized image of the specific material behavior that can be determined experimentally, empirically, and based on a wealth of expert knowledge. The more complex the material, the more difficult the calibration is. This situation forms the starting point for this work’s hybrid data-driven and model-based approach for mapping a complex physical process in continuum mechanics. Specifically, we use data generated from a classical physical model by the MESHFREE software (MESHFREE Team in Fraunhofer ITWM & SCAI: MESHFREE. https://www.meshfree.eu, 2023) to train a Principal Component Analysis-based neural network (PCA-NN) for the task of parameter identification of the material model parameters. The obtained results highlight the potential of deep-learning-based hybrid models for determining parameters, which are the key to characterizing materials occurring naturally such as sand, soil, mud, or snow. The motivation for our research is the simulation of the interaction of vehicles with sand. However, the applicability of the presented methodology is not limited to this industrial use case. In geosciences, when predicting the runout zones of landslides or avalanches and evaluating corresponding protective measures, the parameterization of the respective material model is essential.

Forest damage and its braking effect on the extreme snow avalanche on Mt. Nodanishoji, Japan, in 2021

ABSTRACT An extreme snow avalanche occurred on Mt. Nodanishoji, Japan, in 2021. This avalanche was the second largest ever documented in Japan and destroyed many trees and structures. Despite the increasing interest in avalanche disaster mitigation effects of forests, there are limited opportunities to obtain actual data sets on avalanches and forests. In this study, we reconstructed the avalanche velocity based on the cedar forest damage and simulated the avalanche flow using the dynamical run-out model Titan2D to reveal the avalanche movement and ascertain the braking effect of the forest on the avalanche. We successfully simulated the whole movement of the extreme avalanche with a horizontal runout distance of 2,800 m from the starting zone at an altitude of 1,700 m to the run-out zone considering the effect of the forest by increasing the frictional parameters in the model. Whereas in the case of no forest, the avalanche spread beyond its actual reach, considering the forest, the run-out distance decreased and was consistent with the observed reach. Comparing the results with and without forests, we ascertained the distinct braking effect of the forest on the extreme avalanche and quantify it in terms of a locally increased friction angle.

The effect of location and geometric properties of landslides caused by rainstorms and earthquakes

AbstractLandslides triggered by rainstorms and earthquakes are prominent geological hazards that exhibit distinctive spatial and morphological characteristics due to diverse instability mechanisms. However, studies on differences between the two types of landslides remain limited. In this study, we explored differences in location and geometric properties between rainstorm‐induced landslides in Qinzhou, Longchuan and Fukuoka and earthquake‐induced landslides in Lushan, Iburi and Kaikōura. We normalized the location of landslides across the slope and quantified the landslide polygons using four geometric properties. Findings revealed that both location and geometric properties are specific to landslide type and differ between them. Earthquake‐induced landslides are more common near the ridge of a slope, while rainstorm‐induced landslides are more frequent in the valley or near streams. The quantitative analysis of geometric properties showed that earthquake‐induced landslides are generally larger and have a more compact, rounded and less complex shape. The two landslide types present different hazards, particularly in their runout zones, where dispersion of materials occurs. Insights from our quantitative approach serve as a critical foundation for informed decision‐making in emergency scenarios and contribute to enhancing landslide hazard management.

IMPACT OF LEARNING SET AND SAMPLING FOR SNOW AVALANCHE SUSCEPTIBILITY MAPPING WITH RANDOM FOREST

Abstract. Snow avalanche refers to rapid snow mass movement under the influence of gravity and is a frequently observed natural phenomenon in mountainous regions. An area affected by a snow avalanche consists of starting, track (or transition) and runout (or deposition) zones, which have different geomorphologic characteristics that need to be considered in hazard modelling. This study attempted to analyze the impact of the different zones in producing snow avalanche susceptibility maps (ASMs) with data-driven methods. The random forest (RF) method was applied to the data from Gross Spannort Mountain region (Switzerland) for this purpose. Avalanche inventories from two dates were manually delineated to separate the three zones and two RF models were trained with learning datasets; i.e. Inventory-A which includes all snow avalanche zones (original inventory), and Inventory-B with starting and track zones. The conditioning factors were defined based on the literature, data availability and study area characteristics. The trained models (Model-A with Inventory-A and Model-B with Inventory-B) were evaluated with the area under the receiver operating characteristic (ROC), precision, recall and F1 score. The results show that Model-A has AUC of 0.97, precision of 0.89, recall of 0.95 and F1 score of 0.92 and the Model-B has AUC of 0.98, precision of 0.90, recall of 0.94 and F1 score of 0.92 that indicate high prediction performances for both cases. Furthermore, feature importance values were calculated by the Mean Decrease in Impurity (MDI) method, and elevation, aspect and valley depth were found the most influencing conditioning factors.

Can big data and random forests improve avalanche runout estimation compared to simple linear regression?

Accurate prediction of snow avalanche runout-distances in a deterministic sense remains a challenge due to the complexity of all the physical properties involved. Therefore, in many locations including Norway, it has been common practice to define the runout distance using the angle from the starting point to the end of the runout zone (α-angle). We use a large dataset of avalanche events from Switzerland (N = 18,737) acquired using optical satellites to calculate the α-angle for each avalanche. The α-angles in our dataset are normally distributed with a mean of 33° and a standard deviation of 6.1°, which provides additional understanding and insights into α-angle distribution. Using a feature importance module in the Random Forest framework, we found the most important topographic parameter for predicting α-angles to be the average gradient from the release area to the β-point. Despite the large dataset and a modern machine learning (ML) method, we found the simple linear regression model to yield a higher performance than our ML attempts. This means that it is better to use a simple linear regression in an operational context.

Analysis of the Rockfall Phenomena Contributing to the Evolution of a Pocket Beach Area Using Traditional and Remotely Acquired Data (Lo Zingaro Nature Reserve, Southern Italy)

The coastal domain of central western Sicily is characterized by the presence of rocky coasts, which mainly consist of pocket beaches situated between bedrock headlands that constitute ecological niches of great touristic and economic value. In this peculiar morphodynamic system, the sedimentary contributions are mainly derived from the rockfall that affects the back of nearby cliffs or the sediment supply of small streams that flow into it. In this study, we investigated the geomorphological processes and related landforms that contribute to the evolution of a pocket beach area located in a coastal sector of NW Sicily Island. The cliffs in this are affected by several rockfalls, and deposits from these rockfalls also add to the rate of sedimentary contribution. The analysis was conducted through the application of traditional approaches and contemporary methods that have previously been used to forecast the collection of input data in the field, often under difficult conditions due to the accessibility of the sites, and which have been supported by UAV surveys. Through the analysis of the digital models of terrain and orthophotos, geometrical and multitemporal analyses of landforms were carried out. A dedicated software was utilized for the detection of rockfall runout zones and block trajectories and for defining the automatic extraction of rock mass discontinuities. The data were compared with those derived from traditional geomechanical surveys. The availability of the existing and acquired remote sensing data proved essential for this study for both defining the reference geological model and for performing the site-specific analysis of rockfall.

Conventional data-driven landslide susceptibility models may only tell us half of the story: Potential underestimation of landslide impact areas depending on the modeling design

Data-driven landslide susceptibility models (LSM) have been widely used to spatially predict areas likely affected by landslides. Even though such approaches are used to estimate the spatial likelihood of landslide occurrence, they are frequently trained with observations that rather represent landslide release zone conditions without explicitly considering the potential landslide downslope propagation into flat terrain. Landslide runout models, instead, are designed to route the downslope propagation of masses from given release zones but are seldom coupled with landslide release susceptibility models. This research compares the prediction pattern of three conceptually different models: (i) a model trained on landslide release observations only (Model R), (ii) a model trained on landslide body observations only (Model B), and, (iii) a model that explicitly couples landslide-release susceptibility and the subsequent runout zones (Model R + R). The aim is to explore and interpret the implications of differences in the susceptibility classes and associated model performance metrics. In this context, the interpretation focuses on the results observed for gentle terrain that is located downslope of the prevalent steeper hillsides. These partly built-up areas (i.e., buildings, linear infrastructures) are frequently within the reach of open-hillslope debris flows and, therefore of particular interest from a hazard and risk viewpoint.The study area for this research is the 54 km2 catchment Córrego Dantas, situated in the mountainous region of Rio de Janeiro, SE Brazil. After a heavy rainfall event in 2011, an inventory composed of 313 landslides was mapped and used as input for this work. First, a landslide release susceptibility model (Model R) based on an established data-driven method (Generalized Additive model, GAM) was created to explore the spatial likelihood of landslide release. Model R was then combined with r.randomwalk to compute the propensity of downslope regions to be affected by the released landslides, resulting in an integrated landslide impact susceptibility score (Model R + R). In order to achieve this, a constrained top-down random walk approach was used. Furthermore, a GAM model built upon observations from the entire landslide body (Model B) was used for comparison. Results reveal a potential underestimation of the landslide threat in flatter areas. This tendency is most pronounced for the release-trained model (Model R). The runout impact susceptibility scores produced by the random walk (Model R + R) were observed to better represent the actual impact threat of the whole geomorphological phenomena. From a statistical viewpoint, also the simpler model (Model B) exhibited a satisfactory performance to predict the full landslide area that includes release and deposit zones. The outcomes of this research could be of particular interest to modelers aiming to model landslide susceptibility of regions characterized by long-runout hillslope debris flows and for studies focusing on landslide impact and risk.

The Potential of UAV-Acquired Photogrammetric and LiDAR-Point Clouds for Obtaining Rock Dimensions as Input Parameters for Modeling Rockfall Runout Zones

Rockfalls present a significant hazard to human activities; therefore, their identification and knowledge about potential spatial impacts are important in planning protection measures to reduce rockfall risk. Remote sensing with unmanned aerial vehicles (UAVs) has allowed for the accurate observation of slopes that are susceptible to rockfall activity via various methods and sensors with which it is possible to digitally collect information about the rockfall activity and spatial distributions. In this work, a three-dimensional (3D) reconstruction of rock deposits (width, length, and height) and their volumes are addressed, and the results are used in a rockfall trajectory simulation. Due to the availability of different sensors on the UAV, the aim was also to observe the possible differences in the dimension estimations between photogrammetric and LiDAR (light detection and ranging) point clouds, besides the most traditional method where rock deposit dimensions are measured on the field using a measuring tape. The motivation for reconstructing rock dimensions and volumes was solely for obtaining input parameters into a rockfall model. In order to study the differences between rock-measuring methods, rock dimensions were used as input parameters in a rockfall model, and additionally, modeling results such as propagation probability, maximum kinetic energies, and maximum passing heights were compared. The results show that there are no statistically significant differences between the measurement method with respect to rock dimensions and volumes and when modeling the propagation probability and maximum passing heights. On the other hand, large differences are present with maximum kinetic energies where LiDAR point cloud measurements achieved statistically significantly different results from the other two measurements. With this approach, an automated collection and measurement process of rock deposits is possible without the need for exposure to a risk of rockfall during fieldwork.

Avalanche Hazards Around Kedarnath Temple, Mandakini River Valley, Uttarakhand – A Case Study

Abstract The Shri Kedarnath town is located in the western extremity of the Central Himalaya (30°44'6.7" N; 79°04'1"E) in the Mandakini River valley. This has a total catchment area of ~67 km2 (up to Rambara), out of which 23% area is covered by glaciers. In September 2022, avalanches were observed repeatedly in the mountains to the north of Kedarnath town. Based on aerial survey and ground observations of the area around Kedarnath, we found that the avalanches initiated ~6 km (Aerial distance) uphill of the Kedarnath Temple from the peaks of the Mahalaya Parvat (5970 m asl) of the Companion Glacier adjacent to the Chorabari Glacier. The Companion Glacier does not have a well-defined accumulation zone but receives accumulation mostly from the avalanches. The run-out and deposition zones of the avalanches are located ~5 km uphill of the Kedarnath Temple. The avalanche activity thus poses no threat to the Kedarnath Temple and township. In addition, fifteen avalanche sites (A1-A15) have been identified based on slope morphology, grass clearing, the presence of snow drifts, and debris cones in the area. We observed that the Temple of Shri Kedarnath Complex is located at a reasonably safe place as far as avalanche hazards are concerned.

LANDSLIDE HAZARD LEVEL IN PASURUAN REGENCY

Pasuruan Regency is an area with potential for landslides, particularly along the Arjuno-Welirang Slope and the Bromo Tengger Semeru National Park (TNBTS). Geographical conditions, high slopes, and andosols and litosols that are easily eroded, further exacerbated by human activities such as land-use changes, lead to increasingly high landslide potentials. The TNBTS Area and Arjuna-Welirang Mountains are priority destinations for national tourism. Therefore, efforts for disaster risk reduction are imperative. Assessment of landslide hazard levels is necessary as a basis for landslide mitigation. The analysis used to assess landslide susceptibility was the overlay of two maps: a map of landslide-potential zones and a map of potential runout zones produced from the model by Horton et al. (2008) employing the flow direction algorithm. Analysis results indicate that landslide hazards in Pasuruan Regency mostly comprise the low hazard (29,924.73 hectares), followed by high landslide hazard (17,421.39 hectares), and moderate landslide hazard (2,161.8 hectares). The high hazard level is found in Tosari District due to being located close to Mount Bromo, whereas low hazard levels are mostly seen in lowland areas.

SNOW AVALANCHE SUSCEPTIBILITY MAPPING FOR DAVOS, SWITZERLAND

Abstract. Snow avalanches are among destructive hazards occurring in mountainous regions and spatial distribution (susceptibility) of their occurrences needs to be considered for spatial planning and disaster risk mitigation efforts. The susceptibility assessment is the first step in avalanche disaster management and can be carried out using high resolution geospatial data and machine learning (ML) algorithms. In this study, we have assessed the snow avalanche susceptibility in Davos, Switzerland using an inventory delineated on satellite imagery in a previous study. The conditioning factors used for the avalanche susceptibility assessment include elevation, slope, plan curvature, profile curvature, aspect, topographic position index, topographic ruggedness index, topographic wetness index, land use and land cover, lithology, distance to road, and distance to the river. Two ML algorithms, the logistic regression (LR) and the random forest (RF), were comparatively assessed using validation data split from the training data (30/70). The prediction performances of both models were assessed based on the area under the receiver operating characteristic curve (ROC-AUC) value. Although the AUC value obtained from the LR method was relatively low (0.74), the value obtained from the RF (0.96) demonstrated high performance and usability of this approach. The results indicate that the RF method can successfully produce an avalanche susceptibility map for the region, although potential improvements may be possible by investigating various input features and ML algorithms as well as by classifying the starting and runout zones of the avalanche data separately. Furthermore, the accuracy is expected to increase by using a larger training dataset.

A data-driven method for predicting debris-flow runout zones by integrating multivariate adaptive regression splines and Akaike information criterion

A data-driven method for predicting debris-flow runout zones by integrating multivariate adaptive regression splines and Akaike information criterion

A new GPU-accelerated coupled discrete element and depth-averaged model for simulation of flow-like landslides

Flow-like landslides are a common type of natural hazards that may impose a great risk to people and their properties. Different models have been reported for simulating flow-like landslides, all of which possess limitations due to the underlying assumptions and simplifications. Harnessing the advantages of two types of prevailing modelling approaches, a new coupled model is developed which adopts a discrete element method (DEM) model to simulate the complex collapsing process in the source area and a depth-averaged model (DAM) to predict the predominantly convective movement in the runout and deposition zones. The coupled model is finally implemented on the NVIDIA CUDA programming platform to achieve GPU high-performance computing. Two laboratory tests are considered to validate the model and a field-scale landslide event is simulated to verify its applicability in real-world conditions. Satisfactory results confirm that the coupled model is able to reproduce the dynamic process of real-world flow-like landslides.

Potassium chloride wells used as quick-clay landslide mitigation: installation procedures, cost–benefit analysis, and recommendations for design

Retrogressive development of landslides in highly sensitive clays (quick clays) may extend several hundred metres upslope from an initial landslide, and liquified slide-debris may impact buildings or infrastructure in the run-out zone. By installing wells filled with potassium chloride (KCl) in quick clays, the salt migrates into the surrounding clay and increases its remolded shear strength, reducing its sensitivity. The salt-stabilized, nonquick clay volume may act as a barrier preventing backward retrogression, thereby contributing to reducing both the area susceptible for being involved in a quick-clay landslide, and the area of the run-out zone. Installation procedures and design guidelines for salt stabilization are examined herein. Installation procedures generating temporary, very small excess pore-water pressures were tested at National Geo-Test Site Tiller–Flotten. Although the benefit-to-cost ratios related to these procedures are small compared to conventional landslide mitigation measures, reducing the installation costs to less than 30 USD per m well and increasing the center distance between the wells may justify salt-stabilization as a landslide mitigation measure. This paper describes experience from testing safe installation procedures, evaluations of cost–benefit and environmental impact, and proposed design guidelines, introducing KCl as an alternative to conventional landslide mitigation-measures in slopes with highly sensitive quick-clay deposits.

The scale effect in extreme snow avalanche runout distance

The scale effect in avalanche runout distance refers to the question of whether paths with larger scale such as total vertical drop may have longer runout distance than shorter paths. Here the effect is analyzed using 475 values of extreme runout distance measured from five mountain ranges. The results show that within each range there is such a scale effect. However, the statistical significance of the scale effect reduces sharply as the median scale parameter, such as total vertical drop, increases. Results from maximum speed data of avalanches also show that maximum speed increases with path scale. The reduction of significance of the scale effect with the path scale combined with the speed data analysis suggests that the terrain just upslope of the runout zone may be crucial in determining the runout distance. The overall results should be of interest to those attempting to model avalanche runout using either empirical statistical methods or avalanche dynamics.

A Simple Method of Mapping Landslides Runout Zones Considering Kinematic Uncertainties

Landslides can be triggered by natural and human activities, threatening the safety of buildings and infrastructures. Mapping potential landslide runout zones are critical for regional risk evaluation. Although remote sensing technology has been widely used to discover unstable areas, an entire landslide runout zone is difficult to identify using these techniques alone. Some simplified methods based on empirical models are used to simulate full-scale movements, but these methods do not consider the kinematic uncertainties caused by random particle collisions in practice. In this paper, we develop a semi-empirical landslide dynamics method considering kinematic uncertainties to solve this problem. The uncertainties caused by the microtopography and anisotropy of the material are expressed by the diffusion angle. Monte Carlo (MC) simulations are adopted to calculate the probability of each cell. Compared with the existing Flow-R model, this method can more accurately and effectively estimate runout zones of the Yigong landslide where random particle collisions are intense. Combining the D-InSAR technique, we evaluate the runout zones in the Jinsha River from June 2019 to December 2020. This result shows that the method is of great significance in early warning and risk mitigation, especially in remote areas. The source area of the landslide and DEM resolution together affect the number of MC simulations required. A landslide with a larger volume requires a larger diffusion angle and more MC simulations.

Landslide Susceptibility Mapping in the Vrancea-Buzău Seismic Region, Southeast Romania

This study presents the results of a landslide susceptibility analysis applied to the Vrancea-Buzău seismogenic region in the Carpathian Mountains, Romania. The target area is affected by a large diversity of landslide processes. Slopes are made-up of various types of rocks, climatic conditions can be classified as wet, and the area is a seismically active one. All this contributes to the observed high landslide hazard. The paper analyses the spatial component of the landslide hazard affecting the target area, the regional landslide susceptibility. First, an existing landslide inventory was completed to cover a wider area for the landslide susceptibility analysis. Second, two types of methods are applied, a purely statistical technique, based on correlations between landslide occurrence and local conditions, as well as the simplified spatial process-based Newmark Displacement analysis. Landslide susceptibility maps have been produced by applying both methods, the second one also allowing us to simulate different scenarios, based on various soil saturation rates and seismic inputs. Furthermore, landslide susceptibility was computed both for the landslide source and runout zones—the first providing information about areas where landslides are preferentially triggered and the second indicating where landslides preferentially move along the slope and accumulate. The analysis showed that any of the different methods applied produces reliable maps of landslide susceptibility. However, uncertainties were also outlined as validation is insufficient, especially in the northern area, where only a few landslides could be mapped due to the intense vegetation cover.

Multiscale investigation on fatigue properties and damage of a 3D braided SiC/SiC + PyC/SiC composites in the full stress range at 1300 °C

Monotonic tensile and fatigue tests of a SiC/SiC composites were conducted at 1300 °C in the full stress range. The macroscopic behaviors were studied based on the strain data. The mesoscopic morphology was observed by X-ray computed tomography, and the microanalysis was conducted using SEM, EDS and XRD. Besides, the interfacial debonding strength (IDS) were measured by nano-indenter. The results reveal that the fatigue behaviors can be divided into three zones. The inelastic strains accumulation and stiffness reduction can be observed in all three zones due to matrix cracking, interface damage, and failure of fibers. The fatigue life is long in the run-out zone because the maximum stress is lower than the proportional limit stress (PLS). In the stress-insensitive zone, the fracture depends on high-temperature and oxidation effects. The failure in the stress-sensitive zone is dominated by the fiber strength. The interface behaviors greatly affect the fatigue life above the PLS.

A coupled discrete element and depth-averaged model for dynamic simulation of flow-like landslides

Flow-like landslides commonly happen in mountainous areas and may threaten people’s lives, damage their properties, and create negative impact on the environment. Computer modelling has become an effective tool to support landslide risk assessment and management. Models based on discrete element method (DEM) can capture micro-mechanical behaviour of soils, simulate large deformation and have been widely used for landslide simulations. However, these models are computationally too demanding for large-scale applications. On the other hand, depth-averaged models (DAM) have been well reported for simulation of flow-like landslides over large spatial domains due to its relatively high computational efficiency. To combine the advantages of both types of modelling approaches, this paper develops a novel landslide model by coupling a DEM model with a DAM for landslide simulation, in which the DEM component is employed to better simulate the complex landslide dynamics in the source area and the DAM is adopted to predict the predominantly convective movement in the runout and deposition zone. Finally, the new coupled landslide model is validated against several test cases, including a field-scale event. Satisfactory results have been obtained, demonstrating that the coupled model is able to reproduce the dynamic process of flow-like landslides.

Rockfall susceptibility and runout in the Valley of the Kings

The UNESCO world heritage site Valley of the Kings or Wadi el-Moluk (وادي الملوك) near Luxor, Egypt, hosts unique burial places of Egyptian kings and royals from the New Kingdom (c. 1539–1075 BCE) and attracts about 0.5 to 2 million tourists per year. Very steep to subvertical cliffs of Thebes Limestone surround the Valley of the Kings. The rock mass is cut by frequent joints and faults making the cliff walls prone to rockfalls. However, only few rockfall debris are found in the valley, likely due to natural remobilisation by flood events and artificial clearings and excavation works that rendered the natural debris cover over the millennia. This work focuses on rockfall susceptibility and runout and makes use of new high-resolution landscape surface models utilising terrestrial laser scanning. We investigated rockfall release areas by exploring rock mass fractures at 23 cliff segments and analysed the kinematics of potential rockfalls. Furthermore, we estimated potential rockfall deposition areas with CONEFALL supported by nine numerical simulations of single rockfall events using Rockyfor3D. We found that nearly 4500 m2 (26%) of the public walking paths and 24 out of 64 tomb entrance areas locate within potential rockfall runout zones.