On May 28, 2025, a high-altitude and long-runout ice-rock avalanche disaster occurred at the Birch Glacier in the Alps of the Valais region in southern Switzerland. This incident completely devastated the downstream towns of Blatten and Ried, leading to the emergency evacuation of over 300 individuals, with one person reported missing. This study presents a systematic investigation into the developmental characteristics, evolutionary processes, and disaster dynamics of the “5•28” Birch high-altitude and long-runout ice-rock avalanche, utilizing multi-temporal satellite remote sensing images, UAV data collected pre- and post-disaster, landquake signal, and on-site video footage. Preliminary results indicate that the Nesthorn Peak, located at a relative altitude of approximately 300 meters on the south side of the upper Birch Glacier, frequently experienced rockfalls driven by a combination of global climate warming and freeze-thaw cycles. While the accumulated debris on the glacier surface suppressed glacial ablation, it enhanced plastic flow, intensified bulging at the glacier front, and promoted the expansion of ice crevasses. Remote sensing interpretation revealed that the glacier area has expanded by approximately 44% over the past decade, with the glacier tongue advancing about 110 meters. During the disaster, around 3 million cubic meters of wedge-shaped sliding mass experienced high-altitude instability, continually impacting the lower Birch Glacier at a velocity of about 36 m/s. This triggered a total instability involving approximately 6 million cubic meters of glacial material and its covered debris, which subsequently transformed into a rapidly moving ice-rock avalanche that surged out of the valley at an average speed of 64 m/s, accumulating upon collision with the opposite mountainside. Such high-altitude and long-runout geological disasters, characterized by ice-rock compositions and developed in high-mountains area, are widely distributed throughout the Himalayan orogenic belt in China, posing serious threats to the geological safety of major engineering projects. This research may provide useful references for disaster prevention and mitigation strategies.

Currently, the technology of large language models is evolving rapidly and accelerating its integration in geological disaster prevention and control. It has been expanding the application scenarios and breaking the limitations in data analysis and complex modeling capabilities as well as innovating the traditional research paradigm. To further promote new breakthroughs in AI technologies in the intelligent prevention and control of geological disasters, this article reviews the evolution characteristics of large language model technology and the application scenarios in multiple fields, and also discusses the key technologies including small sample learning, multimodal data fusion, lightweight model and transfer application, as well as expert knowledge embedding and human-computer collaboration, which are also the main ideas and research focus directions for achieving intelligent identification of geological disaster hazards. The article also proposes an "AI + geological disasters" research framework, technical ideas and typical application scenarios based on core elements including "application scenarios, key issues, mechanism of action, data modalities, sample characteristics, model development, expert knowledge, and human-computer collaboration". This highlights the important value of AI technology in geological disasters research in solving the dealing with multi-dimensional, multi-scale, nonlinear and complex relationship modeling problems. The purpose of this article is to promote AI technologies to integrate into geological disaster prevention and control work at a deeper level, from data, models, and knowledge, and also better leverage AI technology to promote the development of disaster prevention and mitigation towards a greater precision and intelligence.

On July 23, 2019, a catastrophic landslide occurred in Jichang Town, Shuicheng District, Liupanshui City, Guizhou Province, resulting in the tragic loss of 43 lives, with 9 people still missing, and causing direct economic losses estimated at approximately 103 million Yuan. Following the landslide, significant debris and an unstable slope continued to threaten the nearby residences in the Pingzhai and Pingxing groups. Through comprehensive on-site investigations, this study identifies and analyzes both the geological environment and contributing factors of the Shuicheng landslide, providing an in-depth understanding of its triggering mechanisms. For the first time, a thorough stability analysis of the post-landslide residual deposits and the adjacent unstable slope under varying precipitation conditions is presented. The results indicate that extreme rainfall may trigger secondary landslide within the extensive debris deposits from the Shuicheng landslide. Additionally, we performed a dynamic behavior analysis of the unstable slope using rheological models and depth-averaged approaches. The findings provide essential insights into the dynamical mechanisms driving secondary landslide disasters and are crutial for safeguarding the lives and properties of residents at the slope toe.

Ground collapse due to mining activities is a prevalent issue in underground coal mining processes. Without timely monitoring and control, it can adversely affect the surrounding structures and the environment. Addressing the challenges of traditional subsidence monitoring methods in the mining areas with uneven hilly terrain, this study focuses on the 20314, 20313, and 20312 working faces within the 203 panel of Gaojialiang coal mine area, Inner Mongolia. It employs 12 images of Sentinel-1 radar from April 2018 to December 2020 processed using the small baseline subset differential interferometry InSAR (SBAS-InSAR) technique to derive average displacement velocities and temporal subsidence data in the study area. The study analyzes the dynamic characteristics of subsidence in the area. The results show that the overall subsidence rate is higher in the northern part of the study area compared to the south, with the maximum subsidence rate of approximately −17.2 mm/a observed in the northern third of the 20313 working face. The subsidence pattern generally progresses from south to north and from west to east, corresponding to the actual mining sequence. Major subsidence areas are concentrated in the northern portions of the 20314 and 20313 working faces, with maximum subsidence reaching about −106 mm. The application shows that SBAS-InSAR technology has effective results and significantly technical advantages in monitoring land subsidence in hilly mining areas, thereby providing certain method support for land subsidence monitoring in mining areas.

China is one of the countries most frequently affected by landslide disasters in the world, making landslide susceptibility assessment crucial for effective disaster prevention and mitigation. Due to variations in the adaptability of different machine learning models in different regions, in order to better carry out landslide disaster prevention and control work in Badong County, Hubei Province, ten influencing factors including slope gradient, slope direction, curvature, degree of undulation, stratigraphy, overburden, NDVI, road density, water system density, and slope structure were selected. 4 different models, including logistic regression, support vector machine, multilayer perceptron, and random forest, were used for landslide susceptibility evaluation. Three evaluation methods were used to assess the accuracy of the model: receiver operating characteristic curves, mean square error, determination coefficient, and the ratio of landslide to study area. The experimental results show that there are differences among the models in different evaluation methods. Overall, the RF model exhibits the highest accuracy and generates more reasonable susceptibility zoning maps. The susceptibility distribution maps generated by the four models are similar, with high and very high susceptibility areas predominantly located in the southern riverside area. Areas near Guandukou Town and Jiaojiawan Village along the southwest coast exhibit relatively high susceptibility. The assessment results can provide reference for landslide control in Badong County.

Traditional information value models for evaluating geological hazard susceptibility typically involve simply summing the information values of various evaluation factors, without considering the differences in weight among these factors. This can affect the scientific rigor and rationality of susceptibility zoning to some extent. To address this issue, this paper takes Shidian County of Yunnan Province as an example and introduces the random forest model to calculate the weights of each evaluation factor. After constructing an appropriate evaluation index system, the information value and weight of each factor are calculated individually, followed by a weighted summation. According to the equal interval classification method, the study area is then divided into four susceptibility levels--extremely high, high, medium, and low. To verify the accuracy of the model, the latest geological hazard hidden points identified through detailed investigations and risk assessments over the past 3 a were overlaid with the susceptibility zones. The accuracy was analyzed through hazard point density analysis and ROC curve comparison. Based on the comparison of research results, after introducing the random forest weighting, the density of extremely high-risk hidden hazard points increased from 1.754 to 1.926, and the AUC value improved from 0.809 to 0.847. The research results indicate that introducing random forest for weighting in a single information quantity model can effectively reflects the weight differences among factors, enhancing the precision of geological disaster susceptibility zoning. This method shows higher accuracy in practical applications.

Sedongpu gully, situated in the Yarlung Zangbo Grand Canyon in Xizang, is an area of frequent massive river-damming landslides that threaten the geological safety of border towns and major engineering projects. This study focused on a detailed analysis of two river-damming events that occurred on April 15 and May 14, 2024. The disaster formation processes, main causes, and developing trends were widely analyzed. From the methods of water level monitoring, ground motion monitoring, investigating by helicopter, and survey by high-altitude and Unmanned Aerial Vehicles, the formation and development processes of these river-damming events were identified and analyzed. It was found that the disaster body slid along the gully reached 8 minutes and the river-damming lasted over 10 hours. The second disaster was relatively more serious because the landslide-dammed lake had not completely burst, which significantly aggravated the damming of the main channel of the Yarlung Zangbo River. Their causes were examined from topographical and geological conditions, seismic factors, and climate factors. It was determined that the Sedongpu gully has large height differences, fractured rock structures, and a rich source of loose deposits, which provides favorable conditions for disaster formation; the temperature rising during the alternation of the spring and summer leads to the acceleration of glacier melting and the enhancement of hydrodynamic effect, triggering the occurrence of river-damming disaster chain. Furthermore, it is believed that Sedongpu gully has entered an active period of river-damming disaster chain, based on the interpretation of the comprehensive remote sensing images. Our findings revealed that the major river-damming hazards that occurred in 2018 had caused significant changes to the river morphology of the Yarlung Zangbo River; subsequent large-scale river-damming events resulted in more and more clogged up with the river channel, which increased the risk of forming a giant landslide dam. Finally, this paper provided some suggestions for addressing the issues related to river damming, rising water, outburst flooding, monitoring, early warning, and disaster reduction measures for the high-altitude and long runout disasters in the Sedongpu gully.

The Shouling temple in Luhuo County, Sichuan faces an urgent situation due to the construction of the Future Hall. A five-level excavation slope with a height of 21.8 m and a slope angle of 55° to 75° was formed on the northwest side of the site, leading to local damage at the toe of the slope. Field investigation and numerical simulation analysis indicate that the slope formed a circular sliding surface within the silty clay layer, with stress concentration at the toe of the slope. Under rainfall conditions, the bulk density of the soil increases, while the cohesion and internal friction angle decrease sharply. The maximum horizontal displacement can reach 3.4 m, resulting in a pushing-type landslide and slope instability. The study area is located in a high-intensity seismic zone. The vibration loads generated by earthquakes and large-scale machinery operations will have serious adverse effects on the slope. Numerical simulation analysis shows that under seismic conditions, the slope forms a pushing-type landslide, causing shear failure of the foundation soil with an impact depth of approximately 5 m. Considering the stability requirements of the slope and the damage to the foundation soil under seismic conditions, reinforcement measures such as anti-slide piles with grid anchoring were adopted to strengthen the slope, which have been applied to this project with good governance effectiveness. The analysis method and control measures can provide reference experience for the treatment of similar artificial excavation slope projects.

Numerical simulation is commonly used to address large deformation geological disasters such as collapses, landslides, and debris flows. Accurately and efficiently simulating these issues has always been a challenge. The material point method (MPM), as emerging numerical method, overcomes the grid distortion problems of traditional numerical methods such as the finite element method (FEM) and finite difference method (FDM) when simulating large deformations, and has been successfully applied in the large deformation analysis of geological disasters. In order to understand the research progress of MPM in the large deformation simulation of geological disasters, this paper briefly introduces the basic principles of MPM based on current research. It also summarizes the application of MPM in simulating large deformations of geological disasters such as landslide, debris flow, and ground fracture, highlighting the latest research progress. Furthermore, it identifies issues in existing MPM research, such as accuracy, computational efficiency, and coupling of multi-physics fields, and discusses future trends in MPM development withinengineering geology.

The Tanbu limestone mine in Guangzhou has been in operation for over 30 years, resulting in the formation of a series of terraced slopes, with the highest reaching approximately 195 m. Between 2012 and 2022, the mining area experienced 26 incidents of slope and dangerous rock body collapses. Recent investigations have identified 16 remaining dangerous rock bodies on these excavated slopes. The risk of slope collapse in this open-pit mine poses a significant threat to ongoing mining operations. Analyzing the collapse behavior of these dangerous rock masses is of great significance for developing effective preventive measures for the collapse of dangerous rock mass in mines. This study focuses on the WY3 and WY11 dangerous rock mass, using Rockfall numerical simulation software. A 1000 kg rockfall was simulated, released from the same starting point 50 times to calculate its fall trajectory, bounce height, kinetic energy, and other related behaviors. The results indicate that over 84% of the falling rocks reached the foot of the slope, with the bounce heights ranging between 5 and 15 m and total kinetic energy between 302.3 and 399.2 kJ. The simulation results closely matched the actual historical paths observed in the study area. The numerical simulation results of these rockfall behaviors based on the Rockfall software provide valuable data to support the identification of rockfall risk zones and the selection of appropriate protective measures. This study also offers a specific case reference for the analysis of rockfall behavior of dangerous rock masses in large open-pit mine slopes.

With the intensification of global climate change, extreme rainfall events have become increasingly frequent, leading to recurrent rainfall-triggered landslides and causing significant casualties and economic losses. With the context of climate change, this study systematically reviews the research progress on advancements in probabilistic risk assessment of rainfall-triggered landslides, focusing on three key aspects: (1) slope reliability assessment under rainfall conditions considering climate change; (2) vulnerability assessment of slopes considering the uncertainty of rainfall patterns; and (3) rainfall-induced landslide hazard assessment based on machine learning methods. On this basis, this study further analyzes the multidimensional challenges faced by rainfall-triggered landslide risk assessment under climate change, including uncertainties associated with climate change, the lack of high spatio-temporal resolution geological and meteorological data, and the adaptability of models across different regions. Finally, from the perspectives of detailed geological surveys, multi-factor disaster gestation mechanisms, this study looks towards future research directions for enhancing resilience in rainfall-induced landslide disaster prevention, from landslide mechanisms under multiple factors, to resilience-based risk assessment. This study aims to provide theoretical support and methodological references for the disaster prevention and mitigation work of rainfall-triggered landslides, promoting the scientific, systematic, and refined development of landslide risk management.

Extreme rainfall is often accompanied by mass geological disasters, which seriously endangers the safety of people 's lives and property in prone areas and affects the healthy development of the economy and society. Summarizing and analyzing the time-space distribution characteristics of geological disasters due to extreme rainfall and the effectiveness of early warning is of great significance for improving the comprehensive defense ability against geological disasters. Taking the sudden geological disasters caused by “23•7” heavy rainfall in 2023 as the research object, based on the refined precipitation data from the Beijing sudden geological disaster monitoring and early warning system, the time-space distribution characteristics of “23•7” heavy rainfall and the development and distribution characteristics of geological disasters were analyzed, and the early warning effect of geological disasters was discussed. The results show that the“23•7”heavy rainfall has the characteristics such as a large total amount, strong rainfall, long duration and wide range, and the disasters due to extreme rainfall have the characteristics of group occurrence. The multi-dimensional early warning of geological disaster classification has achieved remarkable results and has achieved the goal of zero casualties due to geological disasters under extreme weather conditions. The research results can provide a reference for actively preventing and scientifically responding to extreme rainfall geological disasters.

This research explores the integration of machine learning in assessing landslide susceptibility, scrutinizing the selection of non-landslide samples. Taking Wenchuan County, Lixian County, and Maoxian County in Sichuan Province as the study areas, 7 evaluation factors were considered, including slope, aspect, elevation, distance to the water system, distance to the fault, lithology, and land use. Non-landslide samples were randomly selected from the lower and extremely low susceptibility zones divided by the information value model (I), weight of evidence model(WOE), coefficient of determination model (CF), and frequency ratio model(FR), as well as form the buffer zones (B) and the entire region (G). These samples were then analyzed using a support vector machine (SVM) model. The results showed that the AUC values for I-SVM, WOE-SVM, CF-SVM, and FR-SVM were 0.9804, 0.9726, 0.9368, and 0.8451, respectively, which were superior to the AUC values of B-SVM (0.7869) and G-SVM (0.7389). This highlight the effectiveness of using mathematical-statistical models for the selection of non-landslide samples, with particular emphasis on the accuracy of the information value model. This study offers a novel approach to selecting non-landslide samples, significantly enhancing predictive accuracy in landslide susceptibility assessments.

Debris flow is a high-concentration, heterogeneous, multiphase flow typically triggered by intense rainfall or snowmelt. Its complex formation and movement processes make accurate susceptibility assessment vital for disaster monitoring and mitigation. Traditional methods often fall short in predictive accuracy, leading to a growing adoption of machine learning algorithms in this field in recent years. This study proposes a debris flow susceptibility assessment model, SPY-RF, which integrates the random forest (RF) algorithm with the spy technique (SPY), using the upper Minjiang River Basin as a case study. The SPY method addresses the common issue of class imbalance by generating high-quality pseudo-negative samples from unlabeled data, thereby enhancing the model’s classification performance. A total of fourteen assessment factors, including gully density, lithology, area, and others, were selected based on geological disaster data and remote sensing imagery to construct a comprehensive debris flow dataset. The SPY technique was utilized to optimize the negative sample selection process, which was then combined with the RF model to evaluate susceptibility. The findings indicate that the SPY-RF model outperforms the traditional RF model, achieving an AUC of 0.98 compared to 0.93. The predicted distribution of extremely high susceptibility areas aligns closely with the current debris flow points, indicating that the SPY-RF model predicts debris flow susceptibility with greater accuracy and stability. Additionally, the model also successfully identifies debris flow occurrences in low-risk and extremely low-risk susceptibility areas. The quality of negative samples was greatly increased by using SPY technology in terms of negative sample acquisition and filtering techniques, which raised the prediction accuracy and dependability of the model. The proposed SPY-RF model serves as a useful guidance for managing the risk of debris flows in the upper Minjiang River Basin.

Geological disasters of ground collapse in mining area have a significant impact on infrastructure, such as roads, pipelines, residents’ lives and property, and safety production in mining area. Yingcheng gypsum mine in Hubei Province has a mining history of nearly 400 years, and long-term underground mining has formed a large range of goaf and ground collapse. Based on the systematic collection of pre-mining data and supplementary survey, the types and distribution rules of ground collapse were analyzed by multi-factor comprehensive and geological method. Based on the theory of "three zones", the cause mechanism of ground collapse of old hole type and mining cavity type were studied. The research shows that the ground collapse of gypsum mine is mainly small to medium in Yingcheng, and the geological hazards of ground collapse are divided into mining cavity type and old hole type. The caving type of ground collapse mainly includes two types: pillar breakage type and bending settlement type. Pillar breakage type is mainly the pillar and roof collapse caused by room and pillar mining, and bending settlement type is mainly the roof collapse caused by insufficient filling rate of longwall filling method mining. The main controlling factors of goaf-type surface collapse are the goaf filling condition and the ratio of depth to thickness. When the ratio of depth to thickness is less than 60, the surface collapse occurs more often, and with the increase of the ratio of depth to thickness, the ground collapse gradually decreases. The degree of subsidence deformation in the old hole type depends on whether the old hole is connected with the large-scale goaf and whether it is filled with water. The research results have guiding significance for gypsum mine risk management, safety assessment, monitoring and early warning system construction.

The geological hazards of submarine landslides can cause serious damage to infrastructure such as offshore wind power, submarine optical cables, and marine platforms, posing a serious challenge to the major strategic task of building a maritime power and ensuring the geological safety of marine engineering. The article systematically reviews the research process of submarine landslide turbidity current geological hazards, summarizes the dynamic characteristics of submarine landslide-turbidity flow chain, dynamic erosion types, mechanisms of triggering, evolution, migration, erosion and sedimentation, theoretical models of erosion, and the influence of complex landforms such as uplift, canyons, and basins. A novel dynamic erosion approach is put forward of submarine landslide-turbidity flow chain, including quantitative, multiphase, whole process, erosion flow-state transformation. Finally, in view of the development of major projects such as offshore wind power, marine resource development, marine transportation, and marine engineering equipment, the geological model and identification technology are discussed of the erosion-prone structure of submarine landslide landslide-turbidity flow chain, as well as the composite, overlapping, and heterogeneous dynamic erosion mechanic model of the disaster chain, and the issues of prevention and control of boundary layer dynamic erosion.

Geological disaster investigations enable timely detection of hazards, issuance of early warnings, and prevention of loss of life and property. To address the challenges of high risk and low efficiency of high steep slopes investigation, this study proposes a method of three-dimensional reconstruction and structural plane identification of high steep slope based on UAV close-range photogrammetry. Using Zengziyan in Nanchuan, Chongqing as a case study, the process begins with acquiring high-definition aerial photographs through UAV close-range and supplemental route photogrammetry. The SFM-MVS algorithm is utilized to construct detailed 3D models and point clouds. An adaptive KNN algorithm is introduced to enhance the coplanarity detection passing rate in point clouds, while optimal planar equations are fitted using the least squares method. Point cloud clustering is achieved using a genetic annealing fuzzy C algorithm. Finally, according to the point cloud covariance matrix eigenvalues and eigenvectors, the point cloud plane parameters and normal vectors are inverted, and the structural surface identification and structural surface yield parameters extraction are completed. The results indicate a 99.6% passing rate for point cloud coplanarity detection, with a maximum deviation in identified orientation parameters of only 4.82°. This research provide insights for rapid acquisition of geological information, stability evaluation, and disaster prevention and mitigation for high steep slopes.

Ground fissure disasters pose significant threats to the safety of residential areas and urban-rural planning in localized areas of Beijing. To analyze the impact of ground fissure activity on the dynamic response characteristics, patterns, and scope of affected sites, this study focuses on the Songzhuang ground fissure in Tongzhou District, a typical development area. Based on the field investigations, the study captured the development characteristics and influence scope of ground fissures and obtain key parameters such as the predominant frequency of the site through microtremor testing. Three dynamic analysis methods were applied to investigate the dynamic response characteristics and patterns of the ground fissure site. The results show that there is no obvious relationship between the excellence frequency and excellence period of the ground fissure site in Songzhuang, Beijing and the location of the measurement points. The ground fissure site exhibits a pronounced “hanging wall effect”, with the Fourier superior frequency, the excellent cycle of the response spectra, and the peak intensity of Arias located in the hanging wall of the ground fissure generally exceeding those on the footwall. Additionally, the amplitude of the ground pulsation test is greater than that of the footwall in the area nearer to the ground fissure. The results of this study can provide valuable references for planning, construction, and seismic fortification of structures in the Songzhuang ground fissure development area.

This paper presents a high fill slope solution for the steep and karst-developed terrain at the southern end, west side of Wulong Airport in Chongqing, where a combination of high embankment and ultra-high counterweight retaining wall was adopted. The foundation of the gravity retaining wall includes three karst formations, covering more than 45% of the total area. These karst areas are fully filled and have a maximum depth exceeding 30 m, making them typical examples of special and complex foundations. To address challenges such as strong non-uniformity, low bearing capacity of the karst foundation, and instability of slopes and retaining walls in karstic foundations, a solution involving the excavation of karst infill to a certain depth and backfilling with concrete was adopted. Through theoretical calculations, this paper comprehensively analyzes the failure modes, stability, stress, and deformation of the high slope and retaining wall with varying concrete replacement depths, ultimately determining a suitable replacement depth. The research results show that the adoption of a certain depth of replacement can effectively improve the non-uniformity of the karst foundation, reduce stress concentration effect on the retaining wall, decrease the deformation of the retaining wall and high embankment, and significantly enhance the stability of the retaining wall and slope. Field monitoring indicates that the horizontal and vertical displacements of the high retaining wall and high slope after construction are both less than 4mm, with deformation curves converging, demonstrating good stability of the slope and retaining wall. The research findings have significant reference value for the planning, design, and construction of high-fill slope projects in complex mountainous areas.