Search

The task of landslide recognition focuses on extracting the location and extent of landslides over large areas, providing ample data support for subsequent landslide research. This study explores the use of UAV and deep learning technologies to achieve robust landslide recognition in a more rational, simpler, and faster manner. Specifically, the widely successful DeepLabV3+ model was used as a blueprint and a dual-encoder design was introduced to reconstruct a novel semantic segmentation model consisting of Encoder1, Encoder2, Mixer and Decoder modules. This model, named DeepLab for Landslide (DeepLab4LS), considers topographic information as a supplement to DeepLabV3+, and is expected to improve the efficiency of landslide recognition by extracting shape information from relative elevation, slope, and hillshade. Additionally, a novel loss function term—Positive Enhanced loss (PE loss)—was incorporated into the training of DeepLab4LS, significantly enhancing its ability to understand positive samples. DeepLab4LS was then applied to a UAV dataset of Baihetan reservoir, where comparative tests demonstrated its high performance in landslide recognition tasks. We found that DeepLab4LS has a stronger inference capability for landslides with less distinct boundary information, and delineates landslide boundaries more precisely. More specifically, in terms of evaluation metrics, DeepLab4LS achieved a mean intersection over union (mIoU) of 76.0% on the validation set, which is a substantial 5.5 percentage point improvement over DeepLabV3+. Moreover, the study also validated the rationale behind the dual-encoder design and the introduction of PE loss through ablation experiments. Overall, this research presents a robust semantic segmentation model for landslide recognition that considers both optical and topographic semantics of landslides, emulating the recognition pathways of human experts, and is highly suitable for landslide recognition based on UAV datasets. [ABSTRACT FROM AUTHOR]

Fast and accurate assessment of slope topography and geological information and understanding of slope deformation evolution are of great significance in emergency investigation and monitoring of landslides. In this study, using the geometric information contained in terrestrial laser scanning (TLS) point clouds, an automatic discontinuity identification algorithm and a quantitative overall displacement evaluation method were developed. The automatic discontinuity identification algorithm is based on a k-means fuzzy cluster, allowing fast visual identification of any discontinuity from the acquired point cloud data to obtain accurate geometric information including occurrence, spacing, and continuity. It can provide adequate assessment of potentially unstable blocks in landslides during emergency disposal. The quantitative overall displacement evaluation method characterizes the overall displacement trend of local areas based on statistics from the M3C2 algorithm results. It can be used to quantify the deformation/displacement of relatively independent blocks (or local areas of slopes) and provides technical support for multi-temporal monitoring of potentially unstable blocks during emergency landslide disposal. These methods were applied to the Baige landslide during emergency disposal in 2019. Based on a comprehensive geological survey and monitoring results, the overall remnant slope was in a relatively stable state after the Baige landslide. However, the back edge of the slope was somewhat broken, and there were two large blocks with a relatively high risk of instability. The results show that these methods can provide effective technical support for emergency investigation, monitoring, and risk assessment of large-scale landslides. [ABSTRACT FROM AUTHOR]

Landslides induced by rainfall pose a notable threat to post-earthquake reconstruction in mountainous areas. Taking Qingliu landslide as an example, this paper reveals the potential failure mechanism and range of the landslide mass through field investigation and monitoring data, and further uses the improved computational fluid dynamics method to simulate and predict the disaster range after landslide destabilization. The results show that the landslide was initially deformed and cracked under the influence of strong earthquakes and formed seepage channels for water infiltration. The circulation of surface recharge of groundwater in rainy season and long-distance recharge of groundwater from melting ice and snow in winter season leads to the continuous softening and reduction of sliding resistance in the sliding zone. Under the action of this change chain, the cracks along the back edge and boundary of the landslide continue to develop and expand and finally will lead to landslide collapse failure. The equivalent fluid method can better simulate the movement process and accumulation process of landslide, since the Qingliu landslide is a soil landslide with fluid characteristics in the process of landslide movement. The results indicate that the maximum speed of the mass flow could reach 49 m/s, which could destroy the infrastructures along the sliding path, besides, both of the two landslides could bury Xiameng town below and block the Mengtun River to form a landslide-dammed lake. The research results could provide a reference for further comprehensive treatment of landslides and the formulation of emergency plans. [ABSTRACT FROM AUTHOR]

Numerous Quaternary deposits are existed in the mountainous areas of Southwest China, especially in the transition zone between the Qinghai-Tibet Plateau and the Sichuan Basin, where strong tectonic movements and frequent climatic changes increase the potential landslides. The possible deformation and failure process of potential landslides and their impacts on the surrounding environment are important research topics. Field investigation and monitoring indicate that the Qingliu landslide in Xiameng town, Li County, Sichuan Province, China has been continuously deforming since August 2020. The deformation zone has a maximum deformation depth of approximately 18.9 m, a total area of 54,628 m2, and a volume of 34.0×104 m3, which seriously threatens infrastructure projects and dwellings. As a result, understanding the Qingliu landslide evolution process, assessing the hazard risk, and planning disaster prevention measures are of great significance for reducing disaster loss. In this study, the mass movement process and hazard risk of the Qingliu landslide are evaluated, and the effects of different prevention measures are compared and discussed. By using the depth-integrated method, the mass movement of the Qingliu landslide is analyzed. The numerical simulation results indicate that the maximum velocity of the Qingliu landslide is approximately 37.5 m/s, and the duration of the landslide is approximately 90s. The simulated landslide can eventually form a deposited mass with a maximum deposit thickness of 19.4 m and an area of approximately 60,168.3 m2, thereby blocking the river and burying dwellings. Furthermore, a risk assessment of the Qingliu landslide under different forms of protection measures is also produced and discussed by considering the hazard level and economic vulnerability level of the affected area. Setting three layers of anti-slide piles on the deformation zone to reduce the hazard risk of the Qingliu landslide is a better choice. Our results may be useful for planning prevention measures and improving disaster emergency response systems. [ABSTRACT FROM AUTHOR]

Landslides are serious natural disasters frequently occurring in the mountainous areas along the Sichuan–Tibet railway. Landslides are often induced by rainfall and greatly threaten people's lives and property. Via the use of GIS technology, a probabilistic analysis model was proposed for rainfall-induced shallow landslides along the above route. The process of the proposed probabilistic analysis model could be divided into three parts: (i) an infiltration hydrogeological model, (ii) a slope stability model, and (iii) probabilistic analysis. In this model, soil cohesion and internal friction angle were regarded as random parameters. The Green–Ampt model was used to dynamically describe the soil infiltration process during rainfall. Combined with the infinite slope stability model, the failure probability in the study area was analyzed. The model was applied to simulate the Bayi catchment landslide event of August 18, 2010, to evaluate its reliability. Through comparison of the simulation results to landslide occurrence locations, it was determined that the model achieves a satisfactory prediction performance. In addition, compared to deterministic analysis methods and the r.slope model, the proposed probabilistic analysis model achieved a satisfactory evaluation performance. The presented physics-based probabilistic analysis model could provide important theoretical support for hazard prevention in regard to rainfall-induced shallow landslides in mountainous areas. [ABSTRACT FROM AUTHOR]

The Hongshiyan rock avalanche is a remarkable landslide disaster with approximately volume of 12 × 106 m3, triggered by the 2014 Ms. 6.5 Ludian earthquake in Yunnan Province, China. This study conducted a comprehensive analysis by the model based on discrete element (DEM-based) numerical simulation to understand the transport process and mechanism for this rock avalanche. The simulation results showed that the transport process of the rock avalanche depends on the input seismic duration and motion. The average velocity of the rock avalanche sharply increases to peak value of 27 m/s and then gradually decreases to zero, and 64% and 36% of the total energy are dissipated by collision and friction, respectively. In this process, the progressions from simple disintegration along pre-existing discontinuities to fragmentation that creates new fracture surface are documented, and gradual increase of the fragmentation degree over time results in the decrease of fragment size and the formation of well-graded and narrower-interval gradation. This fragmentation evolution creates a conductive condition to the development of internal shear, and is closely associated with the dense flow regime that dominates the main body of the rock avalanche but presents discontinuous distribution along the flow thickness direction. In addition, further analyzing the simulated results indicates that more likely effects of fragmentation on mobility of rock avalanches depend on fragmentation-induced special flow structure, which makes a rock avalanche in a flow state with lower friction and lower energy consumption. [ABSTRACT FROM AUTHOR]

Landslides are one of the major factors threatening human life and social development. Monitoring and early warning of landslides are important parts of disaster prevention and control. However, due to complex terrain conditions, landslide monitoring in mountainous watersheds and high mountain valleys usually faces issues of high difficulty, high danger, poor data integrity, and low data accuracy compared with plains or hilly areas. Based on the fusion of terrestrial laser scanning (TLS) and unmanned aerial vehicle (UAV) photogrammetry, this article proposes a landslide deformation monitoring method that can be applied to complex terrain conditions where there are access restrictions and blind areas. The method is based on TLS to increase the number and range of ground control points (GCPs) of UAVs within the visual range and then uses assumed control points (ACPs) to reconstruct the UAV model and calculate landslide displacement in invisible areas. In addition, this article proposes and verifies a formula based on an allometricl equation to estimate the error between the calculated and real displacements based on this method. Then, we applied this method to a real landslide case and successfully realized landslide monitoring under the conditions of access limitation and blind areas. The proposed method enables landslide monitoring to overcome terrain restrictions, and it makes it possible to perform large-scale and high-precision ground deformation measurements with limited working ranges, which can be an effective tool for landslide monitoring in emergency situations and long-term risk evaluation. [ABSTRACT FROM AUTHOR]

Occurrence of a reservoir landslide and its potential secondary hazards near a dam can result in significant losses and casualties, such as those that resulted from the Vajont landslide. In this study, a cataclinal rock slope in the Maoergai reservoir was taken as a case to study the characteristics of the gravitational deformation process and to analyse the potential threat. The stability of the rock slope was analysed using the limit equilibrium method, and the potential landslide movement and subsequent waves were simulated. Results indicated that lithology, geological structure, reservoir water-level changes and artificial activities all play an important role in large deformations of rock slopes that are characterized by a combination of bending–toppling and, principally, shear slip. Pre-calculations of potential threats indicated that the impact of a landslide wave would be greater at dead water levels than at the normal water level and could result in the blockage of the inlet to the water-diversion structure on the opposite right bank. These findings provide implications for the control of reservoir rock slopes: (i) serious attention should be paid to the influence of water on rock strength in the early investigation of geological disasters; and (ii) infiltration must be prevented during water-level rise. Thematic collection: This article is part of the Role of water in destabilizing slopes collection available at: https://www.lyellcollection.org/cc/Role-of-water-in-destabilizing-slopes [ABSTRACT FROM AUTHOR]

The intrusion of debris avalanches into reservoirs frequently occurs in mountainous regions and is a complex solid-fluid interaction. A coupled discrete element method (DEM) and smooth particle hydrodynamics (SPH) method is developed to forecast the entire process of disaster chains including landslide motion, wave generation, and wave propagation on Kaiding slopes at the Houziyan Reservoir, Sichuan Province, China. The sliding masses are treated as granular assemblies with collisions instead of rigid solid. The simulation results indicate that the maximum velocity of debris avalanches when swarming into the water surface is beyond 30 m/s, and the maximum wave elevation on the watercourse and the maximum run-up wave elevation generated by landslides reach 1867.4 m and 1883.0 m, respectively, and the minimum wave trough is 1822.1 m occurring near opposite bank of landslide. Furthermore, the range of landslide-generated waves is estimated to be approximately a propagation distance of 800.0 m along the axis of the river. The coupled DEM-SPH simulation method provides important inspiration for hazard prevention on reservoir landslides. [ABSTRACT FROM AUTHOR]

The 2019 Shuicheng landslide was a typical landslide in that failure occurred in the slope deposits and there was a close correlation with the antecedent rainfall. The initiation mechanism was analysed using field investigations, laboratory tests and numerical simulations and the characteristics of mass movement were quantitatively assessed. The numerical results indicated that the infiltration of rainfall decreased the shear resistance of the slope and increased the pore water pressure and soil weight, resulting in a large deformation of the slope and a gradual deterioration in its stability. After saturation, as the pore water pressure coefficient (ru) approached 0.3, the overall safety factor decreased to 1 and failure eventually occurred. Quantitative mass assessment showed that the unique terrain aggravated the erosion and entrainment effects and that the inclusion of loose material and water promoted the movement of the debris avalanche by changing it into a fluidized state, resulting in a much larger landslide volume with a long runout distance. As a result, the volume of the failed mass, which was calculated by an unmanned aerial vehicle and terrestrial laser scanning modelling techniques to be c. 4.7 × 105 m3 in the source area, almost quadrupled to >2.0 × 106 m3. [ABSTRACT FROM AUTHOR]

Frequent rockfall events at the Huashiya cliff near the dam site of Jinsha Hydropower Station poses great threats to construction projects along the Jinsha River. In order to protect the construction area, a comprehensive analysis of rockfall hazard was carried out including field survey, numerical simulation, quantitative assessment and mitigation strategy. High resolution Digital Terrain Model (DTM) was created by Terrestrial Laser Scanning (TLS), then used for the identification of rockfall sources and simulation of rockfall run-out behaviours. Approximately 60 rockfall simulations were conducted to analyze the interaction between rockfall block and protection nets. Quantitative assessment of rockfall hazard was then made to visualize the distribution and hazard of rockfall. The difference between the 60 rockfall simulations was quantified through Rockfall Interception Ratio (RIR) and the Energy Ratio of Overflying Rockfall (EROR). By integrating RIR, EROR and cost, the performance of protection measures were quantitatively assessed. Then, an optimization strategy was proposed to improve the cost-efficiency of the protection system. At the end of this paper, some problems that can be encountered when creating DTM and hazard maps were discussed, and some advices for designing passive protection system were introduced. [ABSTRACT FROM AUTHOR]

Multiple reactivation of a large landslide might dam river course at the same site in different periods, which would exert long-term and considerable influence on society and environment. In this paper, we consider such successive damming episodes as an integrated system and take the 2018 Baige event due to successive landslides as an example to explore the hazard characteristics of successive landslide-dammed lakes. Based on the river valley conditions at dam site before the first and second blockage, the height of dams formed by landslides with different volumes is estimated, and then the stability and peak discharge of these landslide dams are evaluated. The calculation results show that the hazard of the dammed lake triggered by subsequent slide event in the successive landslides could increase significantly. The cause of such hazard increasement of the subsequent landslide dam is also discussed. The experience obtained in this case study can facilitate better understanding the hazard characteristics of similar landslide-dammed lakes, as well as planning engineered controls in advance in order to decease the hazard of dammed lake formed by the potential subsequent landslide. [ABSTRACT FROM AUTHOR]

Hai-bo Li, Yue-ren Xu, Jia-wen Zhou, Xie-kang Wang and Jie Dou contributed equally to this work. Because of these local emergency responses, 11 survivors were successfully rescued from the landslide-affected area shortly after the occurrence of the landslide. The Shuicheng landslide is a typical example of a heavy rainfall-induced long-runout landslide, which was also affected by some human activities. [Extracted from the article]

The formation of the landslide dam is believed to be case-specific and closely related to the dynamics of failure mass movement and river bed topography. This study focuses on two massive landslides, which blocked the Jinsha River twice on October 11 and November 3, 2018, respectively. The unique feature of these two sequential events was that the remnant of the 1st landslide dam body after break had a significant influence on the formation and shape of the 2nd landslide dam. Multiple methods were used to investigate landslides triggering factors and formation mechanism of these two landslide dams, including traditional field investigation, satellite remote sensing, and unmanned aerial vehicle (UAV) 3D image technology. After long-term creeping deformation under gravitational and hydrological effects, the first failure eventually occurred due to the short-term heavy rainfall in early October 2018. With a high speed along the steep slide bed, the failure mass of the 1st landslide rushed towards the opposite (left) bank of the Jinsha River, while the failure mass of 2nd landslide mainly accumulated at the toe of the slope (right bank) as a result of obstruction from the remaining body of the 1st landslide dam after overflowing. Another important observation is the scraping and entrainment effect on the 2nd landslide; consequently, the volume of the failure mass, which was about 2.0 × 106 m3 at source area based on UAV data, increased significantly to more than 6.0 × 106 m3. This substantially enlarged the volume of the new landslide dam with an unexpected height; thus, overflowing was not possible without manual excavation. An additional investigation implies that the Baige remnant slope is likely to fail again with considerable failure mass. These analyses presented in the paper indicate that multiple occurrences of landslides can significantly increase the risks associated with the compound landslide dams, considering also the volume amplification effect. Thus, timely measures should be carried out to reduce the unexpected devastating consequences prior to possible re-occurrence of landslide. [ABSTRACT FROM AUTHOR]

Reinforcing a landslide dam and converting a landslide-dammed lake to a hydraulically engineered lake is a sound means to address this natural disaster. The Hongshiyan landslide-dammed lake reconstruction project provides an excellent example. However, the stability of the remnant slope is crucial to the reconstruction of this project. It is essential to analyse the formation failure mechanisms of the Hongshiyan landslide and evaluate the stability of the remnant slope. Combined with field investigations and unmanned aerial vehicle (UAV) 3D image technologies, the failure mechanisms of the Hongshiyan landslide and the stability of the remnant slope were qualitatively studied and discussed. The dynamic process and failure mechanism of the Hongshiyan landslide are significantly different to conventional landslides. The dynamic process of the Hongshiyan landslide can be divided into three stages: time-dependent deformation stage, earthquake-induced failure stage and an unloading recovery stage. The failure mechanism can be summarised as follows: tension–crush–shattering–sliding. The stability conditions of the remnant slope are worse than those of conventional landslides under the same conditions. Toppling and small collapse are possibly occurring at the back scarp of the remnant slope because of the steep slope gradient, well-developed tension fractures and frequent occurrence of aftershocks and rainstorms. Based on the density, opening degree, porosity and connectivity of the cracks, as well as instability risk probabilities, the rock mass of the back scarp of the remnant slope can be divided into three zones: the seismically damaged zone, the unloading damaged zone and the stable zone. To guarantee the safety of the remnant slope and reduce secondary earthquake or rainstorm disasters, corresponding comprehensive treatment measures must be taken to ensure long-term stability. [ABSTRACT FROM AUTHOR]

Landslides falling into water bodies can generate impulsive waves. The process of wave generation induced by far subaerial granular landslide has been investigated through series physical experiments. The slide Froude number has been quantified considered the parameters of slide mass and water such as grain size, water depth and impact angle. The effect of impact angle, grain size and water depth on the variation of generated wave amplitude is 0.11%, 1.52% and 2.89% respectively when every 1% change in the value of the parameter. On the three governing parameters the water depth is the most sensitive to the wave amplitude, secondly the grain size and final for the impact angle. The intruding pattern of granular debris, the landslide generated wave types and maximum wave amplitude have been systematically discussed. The degree of impact interaction between the slides and water under different conditions is characterized by the value of the slide Froude number, and the empirical relationships linking the slide Froude number with the produced wave amplitude and wave classifications have been conducted based on experimental data, which is useful to illustrate the triggering mechanisms and dynamic process of impact landslide waves on high debris avalanches. • Landslide-generated wave experiments from subaerial granular debris were conducted in different conditions. • Degree of impact interaction between slides and water can be characterized by value of slide Froude number. • Wave classification and maximum amplitude of landslide generated wave were quantized based on slide Froude number. [ABSTRACT FROM AUTHOR]

The breaching of landslide dams is a severe geological disaster, and the resulting breaching floods significantly threaten people in downstream areas. In this study, a large-scale physical model was constructed in a natural river channel using unconsolidated dam material with a wide grain size distribution. Terrestrial laser scanning was used to capture three-dimensional geometric information of the landslide dam during the breaching process. The results showed that the dominant erosion pattern during dam breaching was surface progressive erosion. The breaching process could be divided into three stages based on different erosion characteristics: headward erosion, rapid erosion, and attenuated erosion. Longitudinal and lateral evolution models were proposed based on the field test results. The longitudinal evolution model describes the erosion at the dam crest and the sedimentation at the dam toe during the breaching process by defining the erosion and rotation points. The lateral evolution model revealed different lateral erosion characteristics at different cross-sections. Both the longitudinal and lateral evolution models were validated using two dam break events of the Baige landslide. The evolution models proposed in this study provide a scientific reference for numerical simulations of dam breaching and the prevention and mitigation of landslide dams. • A large scale field model test was conducted in a natural river to reveal the breaching process of landslide dam. • The whole breaching process was divided into three stages: headward erosion, rapid erosion, and attenuated erosion. • A longitudinal and a lateral evolution models were proposed based on the field test results. [ABSTRACT FROM AUTHOR]

• Dam breach and flood routing processes of landslide dams are studied by physical and numerical models. • Numerical simulations indicated that the evolution of breach morphology was influenced by hydrodynamics. • Channel terrain controls the spatial and temporal distribution of the flow velocity. • Channel topography and flood hydraulics are essential factors affecting deposition distribution. • Upstream inflow significantly affects landslide dam breach and flood propagation processes. Landslide dams are common hazards in mountainous regions, and outburst floods following dam breaches significantly impact people and infrastructure along rivers. The breaching processes associated with landslide dams and the interactions of outburst floods with channel terrain are very complex. The lack of event-based observations hampers the understanding of the mechanisms underlying these processes. Here, we combine a large-scale field experiment conducted on a natural river channel with a 3D numerical simulation. The numerical model was calibrated by a flume test and validated by experimental data from the large-scale physical model, demonstrating that it can accurately simulate the breaching process of landslide dams. The results show that the channel terrain controls the spatial and temporal distribution of the flow velocity and depth of the breach flood. The area of high flow velocity is mainly located in the block-confined narrow channel. In addition, outburst floods with high sediment concentrations led to extensive deposition in the downstream channels. The deposition thickness generally decreases in the flow direction and is affected by the channel topography. The measured changes in the particle size distribution of bed sediment before and after the dam breach support our findings. Moreover, the effect of upstream inflow on breach discharge was also investigated. It was found that the increasing inflow rate significantly accelerated the breaching process and increased the peak discharge. Our study enhances the understanding of the breach mechanism and reveals the connection between the hydrodynamic processes and geomorphic effects of breach floods, which can provide insight into the risk assessment of landslide dam hazards. [ABSTRACT FROM AUTHOR]