Accurate assessment of seismic landslide susceptibility is crucial for disaster prevention and emergency decision-making.Although machine learning methods have been widely applied in this field,they exhibit a strong d...Accurate assessment of seismic landslide susceptibility is crucial for disaster prevention and emergency decision-making.Although machine learning methods have been widely applied in this field,they exhibit a strong dependence on large quantities of highquality samples,resulting in significantly low prediction accuracy of existing studies under data-scarce or crossregional prediction scenarios,which fail to meet practical application requirements.To address this issue,this study proposes an intelligent prediction model integrating transfer learning and a sampling optimization strategy,aiming to enhance the accuracy and applicability of seismic landslide susceptibility assessment.The model first improves the sample collection method through the sampling optimization strategy to enhance the precision and representativeness of training samples.This not only ensures the accuracy of origin area training but also further strengthens the model's predictive ability in the target area.Subsequently,it incorporates Transfer Component Analysis(TCA)to overcome the differences in environmental characteristics between the origin area and target area,and couples TCA with the Light GBM algorithm to construct the TCA-Light GBM model,realizing the assessment of seismic landslide susceptibility in sample-free areas.Validated through case studies of the Jiuzhaigou and Luding earthquakes,the results demonstrate that the proposed TCALight GBM transfer learning method exhibits excellent applicability in seismic landslide susceptibility prediction.After optimization with the TCA algorithm,the model's prediction performance in the target domain is significantly improved,with the AUC value increasing from 0.719 to 0.827,representing an increase of approximately 15.02%.This indicates that TCA technology can effectively alleviate the feature distribution discrepancy between the source domain and target domain,enhancing the model's generalization ability.The method is particularly suitable for scenarios with data scarcity and cross-regional prediction and can provide reliable technical support for the emergency response and risk prevention and control of seismic hazards.展开更多
As a critical ecological barrier in China,the Qinling Mountains see their ecological functions significantly impaired by frequent shallow landslides.However,existing research on the distribution characteristics and dr...As a critical ecological barrier in China,the Qinling Mountains see their ecological functions significantly impaired by frequent shallow landslides.However,existing research on the distribution characteristics and driving mechanisms of such landslides remains relatively limited.To address this knowledge gap,the present study integrated data analysis,field investigations,and remote sensing interpretation to construct a landslide database for the core area of the Qinling Mountains,and systematically analyzed the spatial patterns,development characteristics,and environmental driving factors of shallow landslides.The results reveal that shallow landslides are predominantly small-to-medium in scale,concentrated in regions with an altitude of 800–1000 m and a slope gradient of approximately 30°,with a distinct tendency to develop on sunny(southfacing)slopes.The occurrence frequency of these landslides exhibits a significant positive correlation with the soil moisture content of the weathered layer and the degree of groundwater enrichment in the study area.Specifically,these landslides are mainly developed in bedrock fissure water zones and karst fissure water zones with favorable water-bearing capacity,indicating that rainfall and surface hydrological processes are the key triggering factors for shallow landslides.Notably,vegetation exerts a mediating role in the"vegetation-hydrology-landslide"system:shallow landslides occur most frequently in areas with artificial or shrub-grass vegetation,peaking at a moderate coverage of 50%–60%.This peak suggests that vegetation within this range is ineffective at regulating soil moisture,while the interaction between specific vegetation types and hydrological enrichment further exacerbates landslide risk.By prioritizing the weights of vegetation and hydrological factors,we enhanced the information quantity model,which significantly improved its performance and increased the AUC value to 0.83.These findings confirm the pivotal roles of vegetation and hydrological factors,thereby providing a robust scientific basis for targeted landslide prevention and control in this region.展开更多
The investigation of the Akchour landslide(AKL)demands precise examination on a local scale,which necessitates field surveys that are often hindered by the landslide's steep and extensive nature of the landslide(1...The investigation of the Akchour landslide(AKL)demands precise examination on a local scale,which necessitates field surveys that are often hindered by the landslide's steep and extensive nature of the landslide(1100 m×400 m,ΔZ of 300 m).Digital Elevation Models(DEMs)are among the key datasets used to achieve this objective.A comparative study between freely available DEMs such as Shuttel Radar Topography Mission(SRTM)(30 m×30 m)and Phased Array type L-band Synthetic Aperture Radar(PALSAR)(12.5 m×12.5 m),alongside those generated by unmanned aerial vehicles(UAVs)demonstrates their significant potential for both geomorphological and geomorphometric analysis.Indeed,scaling issues can lead to the oversight of crucial geological elements.Aerial photos at a 1/20000 scale,previously utilized for anaglyph,provide a broad overview but lack detailed information.To address this limitation,we employed the UAV to capture high-resolution aerial views(with a ground resolution of 17 cm).This approach enabled exploration of inaccessible areas,photogrammetry for orthophotos,and the generation of precise DEM supported geomorphological studies.The orthophoto allowed for detailed visual assessment,while the DEM facilitated geomorphological study.The dynamic behaviors within the landslide.Furthermore,the former irrigation network likely exacerbates the situation.Fractures delineating an unstable area are prominent along the main scarp suggesting the possibility of further sliding.This UAV-mapping revealed three distinct zones with varying based approach significantly enhances our understanding of the AKL,surpassing the limitations of traditional methods and providing critical insights into its morphology and potential risks.展开更多
The assessment of landslide susceptibility triggered by earthquakes serves as a fundamental basis for effective emergency response and post-disaster reconstruction efforts.However,current predictive models often face ...The assessment of landslide susceptibility triggered by earthquakes serves as a fundamental basis for effective emergency response and post-disaster reconstruction efforts.However,current predictive models often face limitations in accuracy,with the prediction rates of most models ranging from 80%to 90%.This study introduces a new hybrid machine learning framework,termed the Subtractive Clustering Method-based Adaptive Neural Network Fuzzy Inference System(SCM-ANFIS),and evaluates its performance in the Wenchuan earthquake region.This region features distinctive geology(e.g.,Longmenshan Fault-governed complex tectonics)and abundant fundamental data;additionally,the 2008 Wenchuan Earthquake provides a pertinent case for earthquakeinduced landslide model evaluation.Based on a literature review and correlation analysis,this study systematically identified 12 key influencing factors that collectively characterize the region's high landslide susceptibility,shaped by intense seismic activity,complex terrain,and fragmented rock masses.Positive and negative samples were extracted as target variables through buffer sampling to calculate earthquake-induced landslide susceptibility.The susceptibility zoning map was then calibrated and generated by incorporating the regional landslide area percentage.The study concludes the following:(1)Compared to traditional machine learning approaches,the model demonstrates strong performance and stability,achieving a prediction accuracy of 98.5%.Approximately 97.89%of historically documented landslides in the Wenchuan region were located within areas identified as having high susceptibility,which aligns well with observed spatial distributions.(2)Increase in the buffer distance contributes to enhance prediction accuracy while a larger sample size improves model stability.(3)The model exhibits superior performance and possesses scalability for application in other regions,such as Jiuzhaigou and Luding.(4)Nonetheless,limitations remain regarding uncertainty,sample composition,algorithmic design,and practical implementation.Future research should focus on improving data precision and optimizing algorithmic frameworks.Overall,this study provides valuable support for landslide susceptibility assessments and contributes essential data for disaster risk mitigation efforts.展开更多
Debris flows have increased in frequency within the arid Daheba Basin on the northeastern Tibetan Plateau,but their sediment sources remain poorly quantified.Using high-resolution UAV-derived DEMs from 51 small catchm...Debris flows have increased in frequency within the arid Daheba Basin on the northeastern Tibetan Plateau,but their sediment sources remain poorly quantified.Using high-resolution UAV-derived DEMs from 51 small catchments,this study evaluates the relative contributions of landslide-derived and channel-derived sediment in controlling debris-flow fan magnitude,and quantifies sediment supply during the 2023 rainy season using DEM differencing.A total of 766 landslides occurred predominantly on slopes of 40°-50°and southeast-southwest aspects,generating 36.17×10^(4)m^(3)of material.Gully heads exhibit exceptionally lower landscape dissection thresholds compared with loess and Quaternary regions in China,indicating high susceptibility to failure under intensified runoff.The results show that Landslide area-volume scaling exponent(b)varies with hillslope geometry(K_(u)):b>1.3 for K_(u)<8 and generally b<1.3 for K_(u)>8,indicating more complete scar evacuation upslope and partial erosion downslope.Despite the abundance of landslides,their contribution to debris flow fan magnitude is minor(<25%),with channel debris dominating(>75%).DEM differencing of a small catchment before and after the 2023 rainy season further reveals that sediment supply originates primarily from the main channel(60.6%)and tributaries(23.3%),with smaller contributions from channel banks(6.8%)and channel heads(9.2%).Tributaries exhibit the greatest mean erosion depth(4.2 m),exceeding that of the main channel(3.8 m).These findings demonstrate that debris-flow material supply in the Daheba Basin is transport-limited and controlled mainly by fluvial entrainment rather than slope failures.Climatic warming and wetting may enhance slope instability,but sediment mobilization is dominantly governed by runoff-driven channel erosion.This study underscores the importance of prioritizing channel sediment dynamics in debris flow hazards assessments for arid regions of the Tibetan plateau.展开更多
Landslides pose a formidable natural hazard across the Qinghai-Tibet Plateau(QTP),endangering both ecosystems and human life.Identifying the driving factors behind landslides and accurately assessing susceptibility ar...Landslides pose a formidable natural hazard across the Qinghai-Tibet Plateau(QTP),endangering both ecosystems and human life.Identifying the driving factors behind landslides and accurately assessing susceptibility are key to mitigating disaster risk.This study integrated multi-source historical landslide data with 15 predictive factors and used several machine learning models—Random Forest(RF),Gradient Boosting Regression Trees(GBRT),Extreme Gradient Boosting(XGBoost),and Categorical Boosting(CatBoost)—to generate susceptibility maps.The Shapley additive explanation(SHAP)method was applied to quantify factor importance and explore their nonlinear effects.The results showed that:(1)CatBoost was the best-performing model(CA=0.938,AUC=0.980)in assessing landslide susceptibility,with altitude emerging as the most significant factor,followed by distance to roads and earthquake sites,precipitation,and slope;(2)the SHAP method revealed critical nonlinear thresholds,demonstrating that historical landslides were concentrated at mid-altitudes(1400-4000 m)and decreased markedly above 4000 m,with a parallel reduction in probability beyond 700 m from roads;and(3)landslide-prone areas,comprising 13%of the QTP,were concentrated in the southeastern and northeastern parts of the plateau.By integrating machine learning and SHAP analysis,this study revealed landslide hazard-prone areas and their driving factors,providing insights to support disaster management strategies and sustainable regional planning.展开更多
The Phewa watershed is under constant landslide threat because of its complicated topography,climate,and biology.The floral structure of landslide-prone areas possesses a significant impact on determining the ecologic...The Phewa watershed is under constant landslide threat because of its complicated topography,climate,and biology.The floral structure of landslide-prone areas possesses a significant impact on determining the ecological processes involved in slope stabilization.Plant roots,for example,serve as physical anchors in the soil,enhancing slope stability.Therefore,this study aims to determine appropriate plant species that can enhance soil stability in Phewa Watershed by examining their floral structure in landslide areas.Floral diversity was assessed throughout field visits.Ten of the 46 landslides were selected with 15 plots based on aspect,watershed zones,and normalized difference vegetation index(NDVI)value.Six plant species were selected to evaluate root traits,uprooting force,and cellulosic testing based on their Important Value Index(IVI)value,native characteristics,and regeneration.The uprooting force was calculated using a‘winch’with a force transducer,while the root characteristics were measured manually and using‘ImageJ software’.Results show that 319 species from 92 families are registered in the buffer zones and landslide scars,and the NDVI suggest that vegetation covers more than 49%of the landslide areas.The floral composition of the landslides in the 15 plots contains 140 species from 52 families,with Poaceae dominating.In six plant species,the Ochiai index suggests a significant level of association.The uprooting force is correlated to the root diameter and number but is insignificant in terms of root length and area.Saccharum spontaneum is the best option for landslide stability based on uprooting force(882.63±245.175)N,cellulose content(67.038±4.766)%and root number characteristic(69.333±24.338)whereas Themeda arundinacea is preferred due to its root diameter traits(0.054±0.022)cm.Finally,it emphasizes the significance of selecting key species in lowering the risk of landslides,strengthening soil stability,and building resilient ecosystems in susceptible watershed areas.展开更多
Landslides pose a significant threat in the mountainous regions of Nepal.Landslide susceptibility maps are commonly used to identify potential landslide zones by statistically analyzing geological,topographical,and hy...Landslides pose a significant threat in the mountainous regions of Nepal.Landslide susceptibility maps are commonly used to identify potential landslide zones by statistically analyzing geological,topographical,and hydrological factors,assuming that similar conditions may trigger future failures.While such maps provide valuable insights into landslide-triggering conditions,they are limited in assessing risk to settlements and infrastructure located downslope or in valley bottoms.This study integrates machine learning based landslide susceptibility with numerical runout modeling to provide a comprehensive landslide hazard assessment in the Bhotekoshi watershed,overcoming the limitations of traditional models that focus solely on statistical susceptibility.To conduct the susceptibility analysis,a total of 439 landslides were mapped from 2012 to 2021 using satellite images.Of these,70%were used for training two machine learning(ML)models:random forest and Xtreme Gradient Boosting(XGBoost),and the remaining 30%were used for validation.Among the two ML models,Random Forest model demonstrated slightly superior performance,achieving higher predictive accuracy.After the machine learning susceptibility analysis,the study transitions into a regional-scale landslide runout analysis.First,a back analysis of the past landslide event was conducted to fine-tune the model parameters(internal angle of friction and basal friction angle)and validate performance of the runout model.Following the back analysis,the regional-scale numerical modeling of landslide runout was conducted by designating areas classified as the highest susceptibility class in the Random Forest susceptibility map as potential release zones.This approach allows for a detailed examination of landslide propagation and potential impacts along the downslope settlements and infrastructures.The analysis clearly demonstrates that integrating both machine learning and numerical runout methods significantly increases the estimated exposure of population,buildings,and roads within the very high hazard class compared to relying solely on susceptibility methods.Specifically,population exposure rises from 360 to 7743,buildings increase from 97 to 2771,and road exposure expands from 41 to 251 km.This result highlights the significant risk of underestimating exposure in the analyses that solely rely on landslide susceptibility models.Integration of susceptibility and runout analysis improves landslide risk assessment,aiding in land-use planning and disaster mitigation strategies.展开更多
Landslide dams often undergo seepage due to poor particle gradation and loose structure,yet most existing studies focus solely on overtopping-induced breaching mechanisms,neglecting the potential influence of pre-brea...Landslide dams often undergo seepage due to poor particle gradation and loose structure,yet most existing studies focus solely on overtopping-induced breaching mechanisms,neglecting the potential influence of pre-breaching seepage.Seepage may alter the dam's erodibility,structural stability,and material composition,thereby affecting the overtopping breaching process.Through flume experiments,this study investigates the breaching mechanisms of cohesionless landslide dams with different gradations within the same particle size range under coupled seepage-overtopping conditions.The results demonstrate that pre-breaching seepage significantly impacts breaching dynamics.Within a specific particle size range,compared to pure overtopping,seepage reduces downstream slope stability,increases material erodibility,shortens breaching duration,amplifies peak discharge,and advances the timing of peak flow.As the median particle size(D_(50))increases,the amplification effect of seepage on peak discharge initially increases then decreases,the advancement of peak flow timing diminishes,and the breach erosion rate declines.When D_(50)is sufficiently large,seepage has negligible effects on breach development.For smaller D_(50),seepage markedly accelerates breach widening and deepening.Furthermore,coupled seepage-overtopping extends the downstream deposition area and exacerbates channel erosion due to differences in sediment sorting.These findings highlight the critical role of seepage in landslide dam breaching,providing a scientific basis for hazard prevention and mitigation.展开更多
Shallow landslides are strongly controlled by near-surface lithological variability,yet conventional geological maps are often too generalized to support accurate susceptibility assessment in complex terrains.This rev...Shallow landslides are strongly controlled by near-surface lithological variability,yet conventional geological maps are often too generalized to support accurate susceptibility assessment in complex terrains.This review synthesizes recent advances in remote sensing–based lithological mapping and evaluates their integration into landslide susceptibility modeling.Evidence from the literature indicates that remote sensing-derived lithological products,particularly those incorporating mineralogical information and higher spatial resolution,consistently outperform traditional geological maps in improving model accuracy and spatial detail,especially in heterogeneous environments.However,key challenges remain,including scale mismatches between surface observations and subsurface controls,limited ground validation,uncertainty propagation,and restricted model transferability across regions.The review identifies multi-sensor data fusion and explainable machine learning as the most promising directions for advancing lithological discrimination and model reliability.Future progress depends on integrating remote sensing with process-based understanding,improving validation strategies,and standardizing uncertainty reporting.These developments are essential for enabling more robust,scalable,and operationally relevant landslide susceptibility assessments in complex terrains.Lastly,we describe the directions of research that focus on multi-sensor fusion,explainable machine learning,UAV(Unmanned Aerial Vehicle)-enabled validation,and standardized uncertainty reporting that can help articulate landslide susceptibility assessment,making them even more robust and operationally significant.展开更多
Reservoir-induced landslides in China's Three Gorges Reservoir area are prone to tensile cracks due to the influenceof their own weight and fluctuationsin water levels.The presence of cracks indicates that the ten...Reservoir-induced landslides in China's Three Gorges Reservoir area are prone to tensile cracks due to the influenceof their own weight and fluctuationsin water levels.The presence of cracks indicates that the tensile stress in the area has exceeded the tensile strength of the soil,leading to local instability.To explore the impact of tensile failure behavior on the stability and failure modes of reservoir landslides,the Huangtupo Riverside Slump#1 is taken as a case study.By considering local tensile failure,potential tensile cracks are incorporated into the analysis via the limit equilibrium method and reliability theory.The reliability of landslides under different tensile failure scenarios is quantified.Strain-softening characteristics of the soil are combined to further analyze the failure transmission path of the landslide.Finally,these potential failure modes were validated through physical model tests.The results show that cracks developing at rear positions reduce the stability of the slope and increase the probability of instability.During the destruction process,retrogressive failures with multiple sliding surfaces are likely to occur.However,tensile failure at the forefront reduces the likelihood of an individual slide mass descending.Progressive failure results in both regular and skip transmission patterns.Additionally,cracks and water level changes can also lead to shifts in the positions of the most dangerous blocks.Therefore,in practical landslide analysis and prevention,it is necessary to consider local tensile damage and identify potential tensile crack locations in advance to optimize prevention measures and accurately evaluate landslide risk.展开更多
Global climate change is intensifying the impact of slope hazards,particularly rainfall-induced landslide hazards(RILH),on mountain road networks(MRNs).However,effective quantitative models for dynamically assessing M...Global climate change is intensifying the impact of slope hazards,particularly rainfall-induced landslide hazards(RILH),on mountain road networks(MRNs).However,effective quantitative models for dynamically assessing MRNs vulnerability under RILH disturbances are still lacking.To bridge this gap,this study develops a Cascading Failure Model for Rainfall-Induced Landslide Hazard(CFM-RILH).Validation via a case study of the GarzêTibetan Autonomous Prefecture Road Network(GTPRNs)reveals key characteristics of MRNs system vulnerability under RILH disturbances:(1)Under the disturbance effects of RILH,the vulnerability of the MRNs system follows a nonlinear phase transition law that intensifies with increasing disturbance intensity,exhibiting a distinct critical threshold.When the disturbance intensity exceeds this threshold,the system undergoes a global cascading failure phenomenon analogous to an“avalanche.”(2)Under RILH disturbances,the robustness of the MRNs system possesses a distinct safety boundary.Exceeding this boundary not only fails to improve hazard resistance but instead substantially elevates the risk of large-scale cascading failure.(3)Increasing network redundancy may be considered one of the primary engineering measures for enhancing MRNs resilience against such disturbances.Based on these findings,we propose a“Two-Stage Emergency Response and Hierarchical Fortification”strategy specifically to improve the resilience of GTPRNs impacted by RILH.The CFM-RILH model provides an effective tool for assessing road network vulnerability under such hazards.Furthermore,its modeling framework can also inform vulnerability assessment and resilience strategy development for road networks affected by other types of slope hazards.展开更多
Automatic segmentation of landslides from remote sensing imagery is challenging because traditional machine learning and early CNN-based models often fail to generalize across heterogeneous landscapes,where segmentati...Automatic segmentation of landslides from remote sensing imagery is challenging because traditional machine learning and early CNN-based models often fail to generalize across heterogeneous landscapes,where segmentation maps contain sparse and fragmented landslide regions under diverse geographical conditions.To address these issues,we propose a lightweight dual-stream siamese deep learning framework that integrates optical and topographical data fusion with an adaptive decoder,guided multimodal fusion,and deep supervision.The framework is built upon the synergistic combination of cross-attention,gated fusion,and sub-pixel upsampling within a unified dual-stream architecture specifically optimized for landslide segmentation,enabling efficient context modeling and robust feature exchange between modalities.The decoder captures long-range context at deeper levels using lightweight cross-attention and refines spatial details at shallower levels through attention-gated skip fusion,enabling precise boundary delineation and fewer false positives.The gated fusion further enhances multimodal integration of optical and topographical cues,and the deep supervision stabilizes training and improves generalization.Moreover,to mitigate checkerboard artifacts,a learnable sub-pixel upsampling is devised to replace the traditional transposed convolution.Despite its compact design with fewer parameters,the model consistently outperforms state-of-the-art baselines.Experiments on two benchmark datasets,Landslide4Sense and Bijie,confirm the effectiveness of the framework.On the Bijie dataset,it achieves an F1-score of 0.9110 and an intersection over union(IoU)of 0.8839.These results highlight its potential for accurate large-scale landslide inventory mapping and real-time disaster response.The implementation is publicly available at https://github.com/mishaown/DiGATe-UNet-LandSlide-Segmentation(accessed on 3 November 2025).展开更多
Water storage in the Three Gorges Reservoir in China has increased the regional microseismicity.Bedding-rock landslides,one of the most common slope structures in the Three Gorges Reservoir,are highly prone to sliding...Water storage in the Three Gorges Reservoir in China has increased the regional microseismicity.Bedding-rock landslides,one of the most common slope structures in the Three Gorges Reservoir,are highly prone to sliding under seismic loading.Existing research primarily focuses on the stability of bedding rock landslides under strong earthquakes,while studies on the cumulative damage and long-term stability of bedding rock landslides under high-frequency microseismicity remain immature.In this study,we considered bedding rock landslides under high-frequency microseismicity in the Three Gorges Reservoir area as the research subject and equivalent microseismicity as pre-peak cyclic loading.First,we analyzed the shear strength deterioration of rock mass structural planes under pre-peak cyclic loading conditions and found that the deformation and failure of structural planes involve contact and damage effects.The shear strength of the rock mass structural planes under pre-peak cyclic loading conditions is affected by the confining pressure,loading rate,loading amplitude,and number of loading cycles.Among these factors,the shear strength of the structural planes was the most sensitive to the number of loading cycles.As the number of cycles increased,the rock mass structural planes underwent three stages:stress adjustment(increase in shear strength),fatigue damage(gradual decrease in shear strength),and structural failure(rapid decrease in shear strength).The stability of bedding rock landslides under high-frequency microseismicity was analyzed,revealing that the stability of bedding rock landslides under high-frequency microseismicity can be divided into three stages:short-term enhancement,gradual degradation,and rapid deterioration,exhibiting characteristics of gradual and sudden changes.展开更多
Landslide susceptibility mapping(LSM)is an essential tool for mitigating the escalating global risk of landslides.However,challenges such as the heterogeneity of different landslide triggers,extensive engineering acti...Landslide susceptibility mapping(LSM)is an essential tool for mitigating the escalating global risk of landslides.However,challenges such as the heterogeneity of different landslide triggers,extensive engineering activities exacerbated reactivation,and the interpretability of data-driven models have hindered the practical application of LSM.This work proposes a novel framework for enhancing LSM considering different triggers for accumulation and rock landslides,leveraging interpretable machine learning and Multi-temporal Interferometric Synthetic Aperture Radar(MT-InSAR)technology.Initially,a refined fieldinvestigation was conducted to delineate the accumulation and rock area according to landslide types,leading to the identificationof relevant contributing factors.Deformation along the slope was then combined with time-series analysis to derive a landslide activity level(AL)index to recognize the likelihood of reactivation or dormancy.The SHapley Additive exPlanation(SHAP)technique facilitated the interpretation of factors and the identificationof determinants in high susceptibility areas.The results indicate that random forest(RF)outperformed other models in both accumulation and rock areas.Key factors including thickness and weak intercalation were identifiedfor accumulation and rock landslides.The introduction of AL substantially enhanced the predictive capability of the LSM and outperformed models that neglect movement trends or deformation rates with an average ratio of 81.23%in high susceptibility zones.Besides,the fieldvalidation confirmedthat 83.8%of newly identifiedlandslides were correctly upgraded.Given its efficiencyand operational simplicity,the proposed hybrid model opens new avenues for the feasibility of enhancement in LSM at urban settlements worldwide.展开更多
This study was aimed to prepare landslide susceptibility maps for the Pithoragarh district in Uttarakhand,India,using advanced ensemble models that combined Radial Basis Function Networks(RBFN)with three ensemble lear...This study was aimed to prepare landslide susceptibility maps for the Pithoragarh district in Uttarakhand,India,using advanced ensemble models that combined Radial Basis Function Networks(RBFN)with three ensemble learning techniques:DAGGING(DG),MULTIBOOST(MB),and ADABOOST(AB).This combination resulted in three distinct ensemble models:DG-RBFN,MB-RBFN,and AB-RBFN.Additionally,a traditional weighted method,Information Value(IV),and a benchmark machine learning(ML)model,Multilayer Perceptron Neural Network(MLP),were employed for comparison and validation.The models were developed using ten landslide conditioning factors,which included slope,aspect,elevation,curvature,land cover,geomorphology,overburden depth,lithology,distance to rivers and distance to roads.These factors were instrumental in predicting the output variable,which was the probability of landslide occurrence.Statistical analysis of the models’performance indicated that the DG-RBFN model,with an Area Under ROC Curve(AUC)of 0.931,outperformed the other models.The AB-RBFN model achieved an AUC of 0.929,the MB-RBFN model had an AUC of 0.913,and the MLP model recorded an AUC of 0.926.These results suggest that the advanced ensemble ML model DG-RBFN was more accurate than traditional statistical model,single MLP model,and other ensemble models in preparing trustworthy landslide susceptibility maps,thereby enhancing land use planning and decision-making.展开更多
Landslide susceptibility mapping(LSM)plays a crucial role in assessing geological risks.The current LSM techniques face a significant challenge in achieving accurate results due to uncertainties associated with region...Landslide susceptibility mapping(LSM)plays a crucial role in assessing geological risks.The current LSM techniques face a significant challenge in achieving accurate results due to uncertainties associated with regional-scale geotechnical parameters.To explore rainfall-induced LSM,this study proposes a hybrid model that combines the physically-based probabilistic model(PPM)with convolutional neural network(CNN).The PPM is capable of effectively capturing the spatial distribution of landslides by incorporating the probability of failure(POF)considering the slope stability mechanism under rainfall conditions.This significantly characterizes the variation of POF caused by parameter uncertainties.CNN was used as a binary classifier to capture the spatial and channel correlation between landslide conditioning factors and the probability of landslide occurrence.OpenCV image enhancement technique was utilized to extract non-landslide points based on the POF of landslides.The proposed model comprehensively considers physical mechanics when selecting non-landslide samples,effectively filtering out samples that do not adhere to physical principles and reduce the risk of overfitting.The results indicate that the proposed PPM-CNN hybrid model presents a higher prediction accuracy,with an area under the curve(AUC)value of 0.85 based on the landslide case of the Niangniangba area of Gansu Province,China compared with the individual CNN model(AUC=0.61)and the PPM(AUC=0.74).This model can also consider the statistical correlation and non-normal probability distributions of model parameters.These results offer practical guidance for future research on rainfall-induced LSM at the regional scale.展开更多
Forecasting landslide deformation is challenging due to influence of various internal and external factors on the occurrence of systemic and localized heterogeneities.Despite the potential to improve landslide predict...Forecasting landslide deformation is challenging due to influence of various internal and external factors on the occurrence of systemic and localized heterogeneities.Despite the potential to improve landslide predictability,deep learning has yet to be sufficiently explored for complex deformation patterns associated with landslides and is inherently opaque.Herein,we developed a holistic landslide deformation forecasting method that considers spatiotemporal correlations of landslide deformation by integrating domain knowledge into interpretable deep learning.By spatially capturing the interconnections between multiple deformations from different observation points,our method contributes to the understanding and forecasting of landslide systematic behavior.By integrating specific domain knowledge relevant to each observation point and merging internal properties with external variables,the local heterogeneity is considered in our method,identifying deformation temporal patterns in different landslide zones.Case studies involving reservoir-induced landslides and creeping landslides demonstrated that our approach(1)enhances the accuracy of landslide deformation forecasting,(2)identifies significant contributing factors and their influence on spatiotemporal deformation characteristics,and(3)demonstrates how identifying these factors and patterns facilitates landslide forecasting.Our research offers a promising and pragmatic pathway toward a deeper understanding and forecasting of complex landslide behaviors.展开更多
Influenced by complex external factors,the displacement-time curve of reservoir landslides demonstrates both short-term and long-term diversity and dynamic complexity.It is difficult for existing methods,including Reg...Influenced by complex external factors,the displacement-time curve of reservoir landslides demonstrates both short-term and long-term diversity and dynamic complexity.It is difficult for existing methods,including Regression models and Neural network models,to perform multi-characteristic coupled displacement prediction because they fail to consider landslide creep characteristics.This paper integrates the creep characteristics of landslides with non-linear intelligent algorithms and proposes a dynamic intelligent landslide displacement prediction method based on a combination of the Biological Growth model(BG),Convolutional Neural Network(CNN),and Long ShortTerm Memory Network(LSTM).This prediction approach improves three different biological growth models,thereby effectively extracting landslide creep characteristic parameters.Simultaneously,it integrates external factors(rainfall and reservoir water level)to construct an internal and external comprehensive dataset for data augmentation,which is input into the improved CNN-LSTM model.Thereafter,harnessing the robust feature extraction capabilities and spatial translation invariance of CNN,the model autonomously captures short-term local fluctuation characteristics of landslide displacement,and combines LSTM's efficient handling of long-term nonlinear temporal data to improve prediction performance.An evaluation of the Liangshuijing landslide in the Three Gorges Reservoir Area indicates that BG-CNN-LSTM exhibits high prediction accuracy,excellent generalization capabilities when dealing with various types of landslides.The research provides an innovative approach to achieving the whole-process,realtime,high-precision displacement predictions for multicharacteristic coupled landslides.展开更多
Landslides significantly threaten lives and infrastructure, especially in seismically active regions. This study conducts a probabilistic analysis of seismic landslide runout behavior, leveraging a large-deformation f...Landslides significantly threaten lives and infrastructure, especially in seismically active regions. This study conducts a probabilistic analysis of seismic landslide runout behavior, leveraging a large-deformation finite-element (LDFE) model that accounts for the three-dimensional (3D) spatial variability and cross-correlation in soil strength — a reflection of natural soils' inherent properties. LDFE model results are validated by comparing them against previous studies, followed by an examination of the effects of univariable, uncorrelated bivariable, and cross-correlated bivariable random fields on landslide runout behavior. The study's findings reveal that integrating variability in both friction angle and cohesion within uncorrelated bivariable random fields markedly influences runout distances when compared with univariable random fields. Moreover, the cross-correlation of soil cohesion and friction angle dramatically affects runout behavior, with positive correlations enlarging and negative correlations reducing runout distances. Transitioning from two-dimensional (2D) to 3D analyses, a more realistic representation of sliding surface, landslide velocity, runout distance and final deposit morphology is achieved. The study highlights that 2D random analyses substantially underestimate the mean value and overestimate the variability of runout distance, underscoring the importance of 3D modeling in accurately predicting landslide behavior. Overall, this work emphasizes the essential role of understanding 3D cross-correlation in soil strength for landslide hazard assessment and mitigation strategies.展开更多
文摘Accurate assessment of seismic landslide susceptibility is crucial for disaster prevention and emergency decision-making.Although machine learning methods have been widely applied in this field,they exhibit a strong dependence on large quantities of highquality samples,resulting in significantly low prediction accuracy of existing studies under data-scarce or crossregional prediction scenarios,which fail to meet practical application requirements.To address this issue,this study proposes an intelligent prediction model integrating transfer learning and a sampling optimization strategy,aiming to enhance the accuracy and applicability of seismic landslide susceptibility assessment.The model first improves the sample collection method through the sampling optimization strategy to enhance the precision and representativeness of training samples.This not only ensures the accuracy of origin area training but also further strengthens the model's predictive ability in the target area.Subsequently,it incorporates Transfer Component Analysis(TCA)to overcome the differences in environmental characteristics between the origin area and target area,and couples TCA with the Light GBM algorithm to construct the TCA-Light GBM model,realizing the assessment of seismic landslide susceptibility in sample-free areas.Validated through case studies of the Jiuzhaigou and Luding earthquakes,the results demonstrate that the proposed TCALight GBM transfer learning method exhibits excellent applicability in seismic landslide susceptibility prediction.After optimization with the TCA algorithm,the model's prediction performance in the target domain is significantly improved,with the AUC value increasing from 0.719 to 0.827,representing an increase of approximately 15.02%.This indicates that TCA technology can effectively alleviate the feature distribution discrepancy between the source domain and target domain,enhancing the model's generalization ability.The method is particularly suitable for scenarios with data scarcity and cross-regional prediction and can provide reliable technical support for the emergency response and risk prevention and control of seismic hazards.
基金supported by the National Key R&D Program of China(No.2024YFF1306502)three Special Programs of the National Natural Science Foundation of China(Nos.42341101,42442045,42307220)the Basic Scientific Research Business Funds of Central Universities(Nos.300102263401,300102265501,300102264103)。
文摘As a critical ecological barrier in China,the Qinling Mountains see their ecological functions significantly impaired by frequent shallow landslides.However,existing research on the distribution characteristics and driving mechanisms of such landslides remains relatively limited.To address this knowledge gap,the present study integrated data analysis,field investigations,and remote sensing interpretation to construct a landslide database for the core area of the Qinling Mountains,and systematically analyzed the spatial patterns,development characteristics,and environmental driving factors of shallow landslides.The results reveal that shallow landslides are predominantly small-to-medium in scale,concentrated in regions with an altitude of 800–1000 m and a slope gradient of approximately 30°,with a distinct tendency to develop on sunny(southfacing)slopes.The occurrence frequency of these landslides exhibits a significant positive correlation with the soil moisture content of the weathered layer and the degree of groundwater enrichment in the study area.Specifically,these landslides are mainly developed in bedrock fissure water zones and karst fissure water zones with favorable water-bearing capacity,indicating that rainfall and surface hydrological processes are the key triggering factors for shallow landslides.Notably,vegetation exerts a mediating role in the"vegetation-hydrology-landslide"system:shallow landslides occur most frequently in areas with artificial or shrub-grass vegetation,peaking at a moderate coverage of 50%–60%.This peak suggests that vegetation within this range is ineffective at regulating soil moisture,while the interaction between specific vegetation types and hydrological enrichment further exacerbates landslide risk.By prioritizing the weights of vegetation and hydrological factors,we enhanced the information quantity model,which significantly improved its performance and increased the AUC value to 0.83.These findings confirm the pivotal roles of vegetation and hydrological factors,thereby providing a robust scientific basis for targeted landslide prevention and control in this region.
文摘The investigation of the Akchour landslide(AKL)demands precise examination on a local scale,which necessitates field surveys that are often hindered by the landslide's steep and extensive nature of the landslide(1100 m×400 m,ΔZ of 300 m).Digital Elevation Models(DEMs)are among the key datasets used to achieve this objective.A comparative study between freely available DEMs such as Shuttel Radar Topography Mission(SRTM)(30 m×30 m)and Phased Array type L-band Synthetic Aperture Radar(PALSAR)(12.5 m×12.5 m),alongside those generated by unmanned aerial vehicles(UAVs)demonstrates their significant potential for both geomorphological and geomorphometric analysis.Indeed,scaling issues can lead to the oversight of crucial geological elements.Aerial photos at a 1/20000 scale,previously utilized for anaglyph,provide a broad overview but lack detailed information.To address this limitation,we employed the UAV to capture high-resolution aerial views(with a ground resolution of 17 cm).This approach enabled exploration of inaccessible areas,photogrammetry for orthophotos,and the generation of precise DEM supported geomorphological studies.The orthophoto allowed for detailed visual assessment,while the DEM facilitated geomorphological study.The dynamic behaviors within the landslide.Furthermore,the former irrigation network likely exacerbates the situation.Fractures delineating an unstable area are prominent along the main scarp suggesting the possibility of further sliding.This UAV-mapping revealed three distinct zones with varying based approach significantly enhances our understanding of the AKL,surpassing the limitations of traditional methods and providing critical insights into its morphology and potential risks.
基金financial support from the National Institute of Natural Hazards,Ministry of Emergency Management of China(Grant No.ZDJ2021-12)。
文摘The assessment of landslide susceptibility triggered by earthquakes serves as a fundamental basis for effective emergency response and post-disaster reconstruction efforts.However,current predictive models often face limitations in accuracy,with the prediction rates of most models ranging from 80%to 90%.This study introduces a new hybrid machine learning framework,termed the Subtractive Clustering Method-based Adaptive Neural Network Fuzzy Inference System(SCM-ANFIS),and evaluates its performance in the Wenchuan earthquake region.This region features distinctive geology(e.g.,Longmenshan Fault-governed complex tectonics)and abundant fundamental data;additionally,the 2008 Wenchuan Earthquake provides a pertinent case for earthquakeinduced landslide model evaluation.Based on a literature review and correlation analysis,this study systematically identified 12 key influencing factors that collectively characterize the region's high landslide susceptibility,shaped by intense seismic activity,complex terrain,and fragmented rock masses.Positive and negative samples were extracted as target variables through buffer sampling to calculate earthquake-induced landslide susceptibility.The susceptibility zoning map was then calibrated and generated by incorporating the regional landslide area percentage.The study concludes the following:(1)Compared to traditional machine learning approaches,the model demonstrates strong performance and stability,achieving a prediction accuracy of 98.5%.Approximately 97.89%of historically documented landslides in the Wenchuan region were located within areas identified as having high susceptibility,which aligns well with observed spatial distributions.(2)Increase in the buffer distance contributes to enhance prediction accuracy while a larger sample size improves model stability.(3)The model exhibits superior performance and possesses scalability for application in other regions,such as Jiuzhaigou and Luding.(4)Nonetheless,limitations remain regarding uncertainty,sample composition,algorithmic design,and practical implementation.Future research should focus on improving data precision and optimizing algorithmic frameworks.Overall,this study provides valuable support for landslide susceptibility assessments and contributes essential data for disaster risk mitigation efforts.
基金supported by the Chengdu University of Information Technology Doctoral Fund‘Study on the Initiation Mechanism of Hydraulic Debris Flow Based on Shields Stress’(KYTZ202275)the Second Tibetan Scientific Expedition and Research Program(Grant No.2019QZKK0902)。
文摘Debris flows have increased in frequency within the arid Daheba Basin on the northeastern Tibetan Plateau,but their sediment sources remain poorly quantified.Using high-resolution UAV-derived DEMs from 51 small catchments,this study evaluates the relative contributions of landslide-derived and channel-derived sediment in controlling debris-flow fan magnitude,and quantifies sediment supply during the 2023 rainy season using DEM differencing.A total of 766 landslides occurred predominantly on slopes of 40°-50°and southeast-southwest aspects,generating 36.17×10^(4)m^(3)of material.Gully heads exhibit exceptionally lower landscape dissection thresholds compared with loess and Quaternary regions in China,indicating high susceptibility to failure under intensified runoff.The results show that Landslide area-volume scaling exponent(b)varies with hillslope geometry(K_(u)):b>1.3 for K_(u)<8 and generally b<1.3 for K_(u)>8,indicating more complete scar evacuation upslope and partial erosion downslope.Despite the abundance of landslides,their contribution to debris flow fan magnitude is minor(<25%),with channel debris dominating(>75%).DEM differencing of a small catchment before and after the 2023 rainy season further reveals that sediment supply originates primarily from the main channel(60.6%)and tributaries(23.3%),with smaller contributions from channel banks(6.8%)and channel heads(9.2%).Tributaries exhibit the greatest mean erosion depth(4.2 m),exceeding that of the main channel(3.8 m).These findings demonstrate that debris-flow material supply in the Daheba Basin is transport-limited and controlled mainly by fluvial entrainment rather than slope failures.Climatic warming and wetting may enhance slope instability,but sediment mobilization is dominantly governed by runoff-driven channel erosion.This study underscores the importance of prioritizing channel sediment dynamics in debris flow hazards assessments for arid regions of the Tibetan plateau.
基金The National Key Research and Development Program of China,No.2023YFC3206601。
文摘Landslides pose a formidable natural hazard across the Qinghai-Tibet Plateau(QTP),endangering both ecosystems and human life.Identifying the driving factors behind landslides and accurately assessing susceptibility are key to mitigating disaster risk.This study integrated multi-source historical landslide data with 15 predictive factors and used several machine learning models—Random Forest(RF),Gradient Boosting Regression Trees(GBRT),Extreme Gradient Boosting(XGBoost),and Categorical Boosting(CatBoost)—to generate susceptibility maps.The Shapley additive explanation(SHAP)method was applied to quantify factor importance and explore their nonlinear effects.The results showed that:(1)CatBoost was the best-performing model(CA=0.938,AUC=0.980)in assessing landslide susceptibility,with altitude emerging as the most significant factor,followed by distance to roads and earthquake sites,precipitation,and slope;(2)the SHAP method revealed critical nonlinear thresholds,demonstrating that historical landslides were concentrated at mid-altitudes(1400-4000 m)and decreased markedly above 4000 m,with a parallel reduction in probability beyond 700 m from roads;and(3)landslide-prone areas,comprising 13%of the QTP,were concentrated in the southeastern and northeastern parts of the plateau.By integrating machine learning and SHAP analysis,this study revealed landslide hazard-prone areas and their driving factors,providing insights to support disaster management strategies and sustainable regional planning.
文摘The Phewa watershed is under constant landslide threat because of its complicated topography,climate,and biology.The floral structure of landslide-prone areas possesses a significant impact on determining the ecological processes involved in slope stabilization.Plant roots,for example,serve as physical anchors in the soil,enhancing slope stability.Therefore,this study aims to determine appropriate plant species that can enhance soil stability in Phewa Watershed by examining their floral structure in landslide areas.Floral diversity was assessed throughout field visits.Ten of the 46 landslides were selected with 15 plots based on aspect,watershed zones,and normalized difference vegetation index(NDVI)value.Six plant species were selected to evaluate root traits,uprooting force,and cellulosic testing based on their Important Value Index(IVI)value,native characteristics,and regeneration.The uprooting force was calculated using a‘winch’with a force transducer,while the root characteristics were measured manually and using‘ImageJ software’.Results show that 319 species from 92 families are registered in the buffer zones and landslide scars,and the NDVI suggest that vegetation covers more than 49%of the landslide areas.The floral composition of the landslides in the 15 plots contains 140 species from 52 families,with Poaceae dominating.In six plant species,the Ochiai index suggests a significant level of association.The uprooting force is correlated to the root diameter and number but is insignificant in terms of root length and area.Saccharum spontaneum is the best option for landslide stability based on uprooting force(882.63±245.175)N,cellulose content(67.038±4.766)%and root number characteristic(69.333±24.338)whereas Themeda arundinacea is preferred due to its root diameter traits(0.054±0.022)cm.Finally,it emphasizes the significance of selecting key species in lowering the risk of landslides,strengthening soil stability,and building resilient ecosystems in susceptible watershed areas.
基金China Scholarship Council(CSC)for providing a fully funded post-graduate study in institute of mountain hazards and environment UCASsupported by the National Natural Science Foundation of China(Grant Nos.42361144880)+3 种基金the Science and Technology Program of Xizang(Grant No.XZ202402ZD0001)the Basic Research Program of Qinghai Province(2024-ZJ-904)the Postdoctoral Fellowship Programs of CPSF(Grant Nos.GZC20232571,2024M753153)the International Cooperation Overseas Platform Project,CAS(Grant No.131C11KYSB20200033).
文摘Landslides pose a significant threat in the mountainous regions of Nepal.Landslide susceptibility maps are commonly used to identify potential landslide zones by statistically analyzing geological,topographical,and hydrological factors,assuming that similar conditions may trigger future failures.While such maps provide valuable insights into landslide-triggering conditions,they are limited in assessing risk to settlements and infrastructure located downslope or in valley bottoms.This study integrates machine learning based landslide susceptibility with numerical runout modeling to provide a comprehensive landslide hazard assessment in the Bhotekoshi watershed,overcoming the limitations of traditional models that focus solely on statistical susceptibility.To conduct the susceptibility analysis,a total of 439 landslides were mapped from 2012 to 2021 using satellite images.Of these,70%were used for training two machine learning(ML)models:random forest and Xtreme Gradient Boosting(XGBoost),and the remaining 30%were used for validation.Among the two ML models,Random Forest model demonstrated slightly superior performance,achieving higher predictive accuracy.After the machine learning susceptibility analysis,the study transitions into a regional-scale landslide runout analysis.First,a back analysis of the past landslide event was conducted to fine-tune the model parameters(internal angle of friction and basal friction angle)and validate performance of the runout model.Following the back analysis,the regional-scale numerical modeling of landslide runout was conducted by designating areas classified as the highest susceptibility class in the Random Forest susceptibility map as potential release zones.This approach allows for a detailed examination of landslide propagation and potential impacts along the downslope settlements and infrastructures.The analysis clearly demonstrates that integrating both machine learning and numerical runout methods significantly increases the estimated exposure of population,buildings,and roads within the very high hazard class compared to relying solely on susceptibility methods.Specifically,population exposure rises from 360 to 7743,buildings increase from 97 to 2771,and road exposure expands from 41 to 251 km.This result highlights the significant risk of underestimating exposure in the analyses that solely rely on landslide susceptibility models.Integration of susceptibility and runout analysis improves landslide risk assessment,aiding in land-use planning and disaster mitigation strategies.
基金support of the National Natural Science Foundation of China(42107189,U20A20111)。
文摘Landslide dams often undergo seepage due to poor particle gradation and loose structure,yet most existing studies focus solely on overtopping-induced breaching mechanisms,neglecting the potential influence of pre-breaching seepage.Seepage may alter the dam's erodibility,structural stability,and material composition,thereby affecting the overtopping breaching process.Through flume experiments,this study investigates the breaching mechanisms of cohesionless landslide dams with different gradations within the same particle size range under coupled seepage-overtopping conditions.The results demonstrate that pre-breaching seepage significantly impacts breaching dynamics.Within a specific particle size range,compared to pure overtopping,seepage reduces downstream slope stability,increases material erodibility,shortens breaching duration,amplifies peak discharge,and advances the timing of peak flow.As the median particle size(D_(50))increases,the amplification effect of seepage on peak discharge initially increases then decreases,the advancement of peak flow timing diminishes,and the breach erosion rate declines.When D_(50)is sufficiently large,seepage has negligible effects on breach development.For smaller D_(50),seepage markedly accelerates breach widening and deepening.Furthermore,coupled seepage-overtopping extends the downstream deposition area and exacerbates channel erosion due to differences in sediment sorting.These findings highlight the critical role of seepage in landslide dam breaching,providing a scientific basis for hazard prevention and mitigation.
文摘Shallow landslides are strongly controlled by near-surface lithological variability,yet conventional geological maps are often too generalized to support accurate susceptibility assessment in complex terrains.This review synthesizes recent advances in remote sensing–based lithological mapping and evaluates their integration into landslide susceptibility modeling.Evidence from the literature indicates that remote sensing-derived lithological products,particularly those incorporating mineralogical information and higher spatial resolution,consistently outperform traditional geological maps in improving model accuracy and spatial detail,especially in heterogeneous environments.However,key challenges remain,including scale mismatches between surface observations and subsurface controls,limited ground validation,uncertainty propagation,and restricted model transferability across regions.The review identifies multi-sensor data fusion and explainable machine learning as the most promising directions for advancing lithological discrimination and model reliability.Future progress depends on integrating remote sensing with process-based understanding,improving validation strategies,and standardizing uncertainty reporting.These developments are essential for enabling more robust,scalable,and operationally relevant landslide susceptibility assessments in complex terrains.Lastly,we describe the directions of research that focus on multi-sensor fusion,explainable machine learning,UAV(Unmanned Aerial Vehicle)-enabled validation,and standardized uncertainty reporting that can help articulate landslide susceptibility assessment,making them even more robust and operationally significant.
基金supported by the Major Program of National Natural Science Foundation of China(Grant No.42090055)the National Key ScientificInstruments and Equipment Development Projects of China(Grant No.41827808)the National Nature Science Foundation of China(Grant No.42207216).
文摘Reservoir-induced landslides in China's Three Gorges Reservoir area are prone to tensile cracks due to the influenceof their own weight and fluctuationsin water levels.The presence of cracks indicates that the tensile stress in the area has exceeded the tensile strength of the soil,leading to local instability.To explore the impact of tensile failure behavior on the stability and failure modes of reservoir landslides,the Huangtupo Riverside Slump#1 is taken as a case study.By considering local tensile failure,potential tensile cracks are incorporated into the analysis via the limit equilibrium method and reliability theory.The reliability of landslides under different tensile failure scenarios is quantified.Strain-softening characteristics of the soil are combined to further analyze the failure transmission path of the landslide.Finally,these potential failure modes were validated through physical model tests.The results show that cracks developing at rear positions reduce the stability of the slope and increase the probability of instability.During the destruction process,retrogressive failures with multiple sliding surfaces are likely to occur.However,tensile failure at the forefront reduces the likelihood of an individual slide mass descending.Progressive failure results in both regular and skip transmission patterns.Additionally,cracks and water level changes can also lead to shifts in the positions of the most dangerous blocks.Therefore,in practical landslide analysis and prevention,it is necessary to consider local tensile damage and identify potential tensile crack locations in advance to optimize prevention measures and accurately evaluate landslide risk.
基金financially supported by the National Key R&D Program of China(2024YFE0111900)The National Natural Science Foundation of China(U2468214,52378370,52278372)+1 种基金The National Ten Thousand Talent Program for Young Top-notch Talents(2022QB04978)The Science and Technology Program of Hebei Province(2023HBQZYCSB004)。
文摘Global climate change is intensifying the impact of slope hazards,particularly rainfall-induced landslide hazards(RILH),on mountain road networks(MRNs).However,effective quantitative models for dynamically assessing MRNs vulnerability under RILH disturbances are still lacking.To bridge this gap,this study develops a Cascading Failure Model for Rainfall-Induced Landslide Hazard(CFM-RILH).Validation via a case study of the GarzêTibetan Autonomous Prefecture Road Network(GTPRNs)reveals key characteristics of MRNs system vulnerability under RILH disturbances:(1)Under the disturbance effects of RILH,the vulnerability of the MRNs system follows a nonlinear phase transition law that intensifies with increasing disturbance intensity,exhibiting a distinct critical threshold.When the disturbance intensity exceeds this threshold,the system undergoes a global cascading failure phenomenon analogous to an“avalanche.”(2)Under RILH disturbances,the robustness of the MRNs system possesses a distinct safety boundary.Exceeding this boundary not only fails to improve hazard resistance but instead substantially elevates the risk of large-scale cascading failure.(3)Increasing network redundancy may be considered one of the primary engineering measures for enhancing MRNs resilience against such disturbances.Based on these findings,we propose a“Two-Stage Emergency Response and Hierarchical Fortification”strategy specifically to improve the resilience of GTPRNs impacted by RILH.The CFM-RILH model provides an effective tool for assessing road network vulnerability under such hazards.Furthermore,its modeling framework can also inform vulnerability assessment and resilience strategy development for road networks affected by other types of slope hazards.
基金funded by the National Natural Science Foundation of China,grant number 62262045the Fundamental Research Funds for the Central Universities,grant number 2023CDJYGRH-YB11the Open Funding of SUGON Industrial Control and Security Center,grant number CUIT-SICSC-2025-03.
文摘Automatic segmentation of landslides from remote sensing imagery is challenging because traditional machine learning and early CNN-based models often fail to generalize across heterogeneous landscapes,where segmentation maps contain sparse and fragmented landslide regions under diverse geographical conditions.To address these issues,we propose a lightweight dual-stream siamese deep learning framework that integrates optical and topographical data fusion with an adaptive decoder,guided multimodal fusion,and deep supervision.The framework is built upon the synergistic combination of cross-attention,gated fusion,and sub-pixel upsampling within a unified dual-stream architecture specifically optimized for landslide segmentation,enabling efficient context modeling and robust feature exchange between modalities.The decoder captures long-range context at deeper levels using lightweight cross-attention and refines spatial details at shallower levels through attention-gated skip fusion,enabling precise boundary delineation and fewer false positives.The gated fusion further enhances multimodal integration of optical and topographical cues,and the deep supervision stabilizes training and improves generalization.Moreover,to mitigate checkerboard artifacts,a learnable sub-pixel upsampling is devised to replace the traditional transposed convolution.Despite its compact design with fewer parameters,the model consistently outperforms state-of-the-art baselines.Experiments on two benchmark datasets,Landslide4Sense and Bijie,confirm the effectiveness of the framework.On the Bijie dataset,it achieves an F1-score of 0.9110 and an intersection over union(IoU)of 0.8839.These results highlight its potential for accurate large-scale landslide inventory mapping and real-time disaster response.The implementation is publicly available at https://github.com/mishaown/DiGATe-UNet-LandSlide-Segmentation(accessed on 3 November 2025).
基金sponsored by the General Program of the National Natural Science Foundation of China(Grant No.42407221)the Open Fund of State Key Laboratory of Geohazard Prevention and Geoenvironment Protection(Grant No.SKLGP2024K009)the Hubei Provincial Natural Science Foun-dation,China(Grant No.2023AFB567).
文摘Water storage in the Three Gorges Reservoir in China has increased the regional microseismicity.Bedding-rock landslides,one of the most common slope structures in the Three Gorges Reservoir,are highly prone to sliding under seismic loading.Existing research primarily focuses on the stability of bedding rock landslides under strong earthquakes,while studies on the cumulative damage and long-term stability of bedding rock landslides under high-frequency microseismicity remain immature.In this study,we considered bedding rock landslides under high-frequency microseismicity in the Three Gorges Reservoir area as the research subject and equivalent microseismicity as pre-peak cyclic loading.First,we analyzed the shear strength deterioration of rock mass structural planes under pre-peak cyclic loading conditions and found that the deformation and failure of structural planes involve contact and damage effects.The shear strength of the rock mass structural planes under pre-peak cyclic loading conditions is affected by the confining pressure,loading rate,loading amplitude,and number of loading cycles.Among these factors,the shear strength of the structural planes was the most sensitive to the number of loading cycles.As the number of cycles increased,the rock mass structural planes underwent three stages:stress adjustment(increase in shear strength),fatigue damage(gradual decrease in shear strength),and structural failure(rapid decrease in shear strength).The stability of bedding rock landslides under high-frequency microseismicity was analyzed,revealing that the stability of bedding rock landslides under high-frequency microseismicity can be divided into three stages:short-term enhancement,gradual degradation,and rapid deterioration,exhibiting characteristics of gradual and sudden changes.
基金supported by the National Key R&D Program of China(Grant No.2023YFC3007201)the National Natural Science Foundation of China(Grant No.42377161)the Opening Fund of Key Laboratory of Geological Survey and Evaluation of Ministry of Education(Grant No.GLAB 2024ZR03).
文摘Landslide susceptibility mapping(LSM)is an essential tool for mitigating the escalating global risk of landslides.However,challenges such as the heterogeneity of different landslide triggers,extensive engineering activities exacerbated reactivation,and the interpretability of data-driven models have hindered the practical application of LSM.This work proposes a novel framework for enhancing LSM considering different triggers for accumulation and rock landslides,leveraging interpretable machine learning and Multi-temporal Interferometric Synthetic Aperture Radar(MT-InSAR)technology.Initially,a refined fieldinvestigation was conducted to delineate the accumulation and rock area according to landslide types,leading to the identificationof relevant contributing factors.Deformation along the slope was then combined with time-series analysis to derive a landslide activity level(AL)index to recognize the likelihood of reactivation or dormancy.The SHapley Additive exPlanation(SHAP)technique facilitated the interpretation of factors and the identificationof determinants in high susceptibility areas.The results indicate that random forest(RF)outperformed other models in both accumulation and rock areas.Key factors including thickness and weak intercalation were identifiedfor accumulation and rock landslides.The introduction of AL substantially enhanced the predictive capability of the LSM and outperformed models that neglect movement trends or deformation rates with an average ratio of 81.23%in high susceptibility zones.Besides,the fieldvalidation confirmedthat 83.8%of newly identifiedlandslides were correctly upgraded.Given its efficiencyand operational simplicity,the proposed hybrid model opens new avenues for the feasibility of enhancement in LSM at urban settlements worldwide.
基金the University of Transport Technology under the project entitled“Application of Machine Learning Algorithms in Landslide Susceptibility Mapping in Mountainous Areas”with grant number DTTD2022-16.
文摘This study was aimed to prepare landslide susceptibility maps for the Pithoragarh district in Uttarakhand,India,using advanced ensemble models that combined Radial Basis Function Networks(RBFN)with three ensemble learning techniques:DAGGING(DG),MULTIBOOST(MB),and ADABOOST(AB).This combination resulted in three distinct ensemble models:DG-RBFN,MB-RBFN,and AB-RBFN.Additionally,a traditional weighted method,Information Value(IV),and a benchmark machine learning(ML)model,Multilayer Perceptron Neural Network(MLP),were employed for comparison and validation.The models were developed using ten landslide conditioning factors,which included slope,aspect,elevation,curvature,land cover,geomorphology,overburden depth,lithology,distance to rivers and distance to roads.These factors were instrumental in predicting the output variable,which was the probability of landslide occurrence.Statistical analysis of the models’performance indicated that the DG-RBFN model,with an Area Under ROC Curve(AUC)of 0.931,outperformed the other models.The AB-RBFN model achieved an AUC of 0.929,the MB-RBFN model had an AUC of 0.913,and the MLP model recorded an AUC of 0.926.These results suggest that the advanced ensemble ML model DG-RBFN was more accurate than traditional statistical model,single MLP model,and other ensemble models in preparing trustworthy landslide susceptibility maps,thereby enhancing land use planning and decision-making.
基金funding support from the National Natural Science Foundation of China(Grant Nos.U22A20594,52079045)Hong-Zhi Cui acknowledges the financial support of the China Scholarship Council(Grant No.CSC:202206710014)for his research at Universitat Politecnica de Catalunya,Barcelona.
文摘Landslide susceptibility mapping(LSM)plays a crucial role in assessing geological risks.The current LSM techniques face a significant challenge in achieving accurate results due to uncertainties associated with regional-scale geotechnical parameters.To explore rainfall-induced LSM,this study proposes a hybrid model that combines the physically-based probabilistic model(PPM)with convolutional neural network(CNN).The PPM is capable of effectively capturing the spatial distribution of landslides by incorporating the probability of failure(POF)considering the slope stability mechanism under rainfall conditions.This significantly characterizes the variation of POF caused by parameter uncertainties.CNN was used as a binary classifier to capture the spatial and channel correlation between landslide conditioning factors and the probability of landslide occurrence.OpenCV image enhancement technique was utilized to extract non-landslide points based on the POF of landslides.The proposed model comprehensively considers physical mechanics when selecting non-landslide samples,effectively filtering out samples that do not adhere to physical principles and reduce the risk of overfitting.The results indicate that the proposed PPM-CNN hybrid model presents a higher prediction accuracy,with an area under the curve(AUC)value of 0.85 based on the landslide case of the Niangniangba area of Gansu Province,China compared with the individual CNN model(AUC=0.61)and the PPM(AUC=0.74).This model can also consider the statistical correlation and non-normal probability distributions of model parameters.These results offer practical guidance for future research on rainfall-induced LSM at the regional scale.
基金supported by the Postdoctoral Fellowship Program of CPSF(Grant No.GZB20230685)the National Science Foundation of China(Grant No.42277161).
文摘Forecasting landslide deformation is challenging due to influence of various internal and external factors on the occurrence of systemic and localized heterogeneities.Despite the potential to improve landslide predictability,deep learning has yet to be sufficiently explored for complex deformation patterns associated with landslides and is inherently opaque.Herein,we developed a holistic landslide deformation forecasting method that considers spatiotemporal correlations of landslide deformation by integrating domain knowledge into interpretable deep learning.By spatially capturing the interconnections between multiple deformations from different observation points,our method contributes to the understanding and forecasting of landslide systematic behavior.By integrating specific domain knowledge relevant to each observation point and merging internal properties with external variables,the local heterogeneity is considered in our method,identifying deformation temporal patterns in different landslide zones.Case studies involving reservoir-induced landslides and creeping landslides demonstrated that our approach(1)enhances the accuracy of landslide deformation forecasting,(2)identifies significant contributing factors and their influence on spatiotemporal deformation characteristics,and(3)demonstrates how identifying these factors and patterns facilitates landslide forecasting.Our research offers a promising and pragmatic pathway toward a deeper understanding and forecasting of complex landslide behaviors.
基金the funding support from the National Natural Science Foundation of China(Grant No.52308340)Chongqing Talent Innovation and Entrepreneurship Demonstration Team Project(Grant No.cstc2024ycjh-bgzxm0012)the Science and Technology Projects supported by China Coal Technology and Engineering Chongqing Design and Research Institute(Group)Co.,Ltd..(Grant No.H20230317)。
文摘Influenced by complex external factors,the displacement-time curve of reservoir landslides demonstrates both short-term and long-term diversity and dynamic complexity.It is difficult for existing methods,including Regression models and Neural network models,to perform multi-characteristic coupled displacement prediction because they fail to consider landslide creep characteristics.This paper integrates the creep characteristics of landslides with non-linear intelligent algorithms and proposes a dynamic intelligent landslide displacement prediction method based on a combination of the Biological Growth model(BG),Convolutional Neural Network(CNN),and Long ShortTerm Memory Network(LSTM).This prediction approach improves three different biological growth models,thereby effectively extracting landslide creep characteristic parameters.Simultaneously,it integrates external factors(rainfall and reservoir water level)to construct an internal and external comprehensive dataset for data augmentation,which is input into the improved CNN-LSTM model.Thereafter,harnessing the robust feature extraction capabilities and spatial translation invariance of CNN,the model autonomously captures short-term local fluctuation characteristics of landslide displacement,and combines LSTM's efficient handling of long-term nonlinear temporal data to improve prediction performance.An evaluation of the Liangshuijing landslide in the Three Gorges Reservoir Area indicates that BG-CNN-LSTM exhibits high prediction accuracy,excellent generalization capabilities when dealing with various types of landslides.The research provides an innovative approach to achieving the whole-process,realtime,high-precision displacement predictions for multicharacteristic coupled landslides.
基金supported by the National Natural Science Foundation of China(Grant No.U22A20596)the Shenzhen Science and Technology Program(Grant No.GJHZ20220913142605010)the Jinan Lead Researcher Project(Grant No.202333051).
文摘Landslides significantly threaten lives and infrastructure, especially in seismically active regions. This study conducts a probabilistic analysis of seismic landslide runout behavior, leveraging a large-deformation finite-element (LDFE) model that accounts for the three-dimensional (3D) spatial variability and cross-correlation in soil strength — a reflection of natural soils' inherent properties. LDFE model results are validated by comparing them against previous studies, followed by an examination of the effects of univariable, uncorrelated bivariable, and cross-correlated bivariable random fields on landslide runout behavior. The study's findings reveal that integrating variability in both friction angle and cohesion within uncorrelated bivariable random fields markedly influences runout distances when compared with univariable random fields. Moreover, the cross-correlation of soil cohesion and friction angle dramatically affects runout behavior, with positive correlations enlarging and negative correlations reducing runout distances. Transitioning from two-dimensional (2D) to 3D analyses, a more realistic representation of sliding surface, landslide velocity, runout distance and final deposit morphology is achieved. The study highlights that 2D random analyses substantially underestimate the mean value and overestimate the variability of runout distance, underscoring the importance of 3D modeling in accurately predicting landslide behavior. Overall, this work emphasizes the essential role of understanding 3D cross-correlation in soil strength for landslide hazard assessment and mitigation strategies.
