Long-term monitoring of active large-scale landslides for non-structural risk mitigation-integrated sensors and web-based platform

Large-scale deep-seated landslides pose a significant threat to human life and infrastructure.Therefore,closely monitoring these landslides is crucial for assessing and mitigating their associated risks.In this paper,the authors introduce the So Lo Mon framework,a comprehensive monitoring system developed for three large-scale landslides in the Autonomous Province of Bolzano,Italy.A web-based platform integrates various monitoring data(GNSS,topographic data,in-place inclinometer),providing a user-friendly interface for visualizing and analyzing the collected data.This facilitates the identification of trends and patterns in landslide behaviour,enabling the triggering of warnings and the implementation of appropriate mitigation measures.The So Lo Mon platform has proven to be an invaluable tool for managing the risks associated with large-scale landslides through non-structural measures and driving countermeasure works design.It serves as a centralized data repository,offering visualization and analysis tools.This information empowers decisionmakers to make informed choices regarding risk mitigation,ultimately ensuring the safety of communities and infrastructures.

Spatial Distribution of Large-scale Landslides Induced by the 5.12 Wenchuan Earthquake

The 5.12 Wenchuan Earthquake in 2008 induced hundreds of large-scale landslides. This paper systematically analyzes 112 large-scale landslides (surface area > 50000 m2), which were identified by interpretation of remote sensing imagery and field investigations. The analysis suggests that the distribution of large-scale landslides is affected by the following four factors: (a) distance effect: 80% of studied large-scale landslides are located within a distance of 5 km from the seismic faults. The farther the distance to the faults, the lower the number of large-scale landslides; (b) locked segment effect: the large-scale landslides are mainly located in five concentration zones closely related with the crossing, staggering and transfer sections between one seismic fault section and the next one, as well as the end of the NE fault section. The zone with the highest concentration was the Hongbai-Chaping segment, where a great number of large-scale landslides including the two largest landslides were located. The second highest concentration of large-scale landslides was observed in the Nanba-Donghekou segment at the end of NE fault, where the Donghekou landslide and the Woqian landslide occurred; (c) Hanging wall effect: about 70% of the large-scale landslides occurred on the hanging wall of the seismic faults; and (d) direction effect: in valleys perpendicular to the seismic faults, the density of large-scale landslides on the slopes facing the seismic wave is obviously higher than that on the slopes dipping in the same direction as the direction of propagation of the seismic wave. Meanwhile, it is found that the sliding and moving directions of large-scale landslides are related to the staggering direction of the faults in each section. In Qingchuan County where the main fault activity was horizontal twisting and staggering, a considerable number of landslides showed the feature of sliding and moving in NE direction which coincides with the staggering direction of the seismic faults.

Insights into some large-scale landslides in southeastern margin of Qinghai-Tibet Plateau

The southeastern margin of Qinghai-Tibet Plateau(SMQTP)is of a typical large landslide-prone area due to intense tectonic activity,deeply incised valleys,high geostress and frequent earthquakes.To gain insights into large landslides in southeastern margin of Qinghai-Tibet Plateau,an area covering 3.34×105 km2 that extends 80e150 km on both sides of the Sichuan-Tibet traffic corridors(G318)was used to examine the spatial distribution and corresponding characteristics of landslides.The results showed that the study area contains at least 629 large landslides that are mainly concentrated on 7 zones(zones IeVII).Zones IeVII are in the southern section of the Longmenshan fault zone(with no large river)and sections with Dadu River,Jinsha River,Lancang River,Nujiang River and Yarlung Zangbo River.There are more landslides in the Jinsha River section(totaling 186 landslides)than the other sections.According to the updated Varnes classification,408 large landslides(64.9%)were recognized and divided into 4 major types,i.e.flows(275 cases),slides(58 cases),topples(44 cases)and slope deformations(31 cases).Flows,which consist of rock avalanches and iceerock avalanches,are the most common landslide type.Large landslide triggers(178 events,28.3%)are also recognized,and earthquakes may be the most common trigger.Due to the limited data,these landslide type classifications and landslide triggers are perhaps immature,and further systematic analysis is needed.

Freely accessible inventory and spatial distribution of large-scale landslides in Xianyang City,Shaanxi Province,China

In this study, we used high-resolution optical satellite images on the Google Earth platform to map large-scale landslides in Xianyang City, Shaanxi Province, China. After mapping, a comprehensive and detailed large-scale landslide inventory that contains 2 924 large-scale landslides was obtained. We analyzed the spatial distribu-tion of landslides with seven influencing factors, including elevation, slope angle, aspect, curvature, lithology, distance to a river, and distance to the fault. Landslide Number, Landslide Area, Landslide Number Density(LND), and Landslide Area Percentage(LAP) were selected as indexes for the spatial distribution analysis. The results show that the number and area of landslides in the elevation range of 1 000–1 200 m is the highest. The highest number of landslides was observed in the slope angle of 25°–30°. North-facing slopes are prone to sliding. The area and number of landslides are the largest when the slope curvature ranges from-1.28 to 0. The LND and LAP reach their maxima when the slope curvature is less than-2.56. Areas covered by the Tertiary stratum with weakened fine-grained sandstone and siltstone show the highest LND and LAP values. Regarding distance to a river, the LAP peaks in the range of 300–600 m, whereas the LND peaks in an area larger than 2100 m. The values of LND and LNP rise as the distance from the faults increases, except for the locations 30 km away from active faults. This phenomenon is because active faults in this area pass through the plain areas, while landslides mostly occur in mountainous areas. The cataloging of landslide development in Xianyang City provides a significant scientific foundation for future research on landslides. In addition, the spatial distribution results are useful for landslide hazard prevention decisions and provide valuable references in this area.

Improvement of large-scale-region landslide susceptibility mapping accuracy by transfer learning

Machine-learning methodologies have increasingly been embraced in landslide susceptibility assessment.However,the considerable time and financial burdens of landslide inventories often result in persistent data scarcity,which frequently impedes the generation of accurate and informative landslide susceptibility maps.Addressing this challenge,this study compiled a nationwide dataset and developed a transfer learning-based model to evaluate landslide susceptibility in the Chongqing region specifically.Notably,the proposed model,calibrated with the warmup-cosine annealing(WCA)learning rate strategy,demonstrated remarkable predictive capabilities,particularly in scenarios marked by data limitations and when training data were normalized using parameters from the source region.This is evidenced by the area under the receiver operating characteristic curve(AUC)values,which exhibited significant improvements of 51.00%,24.40%and 2.15%,respectively,compared to a deep learning model,in contexts where only 1%,5%and 10%of data from the target region were used for retraining.Simultaneously,there were reductions in loss of 16.12%,27.61%and 15.44%,respectively,in these instances.

Landslide Susceptibility Mapping Using RBFN-Based Ensemble Machine Learning Models

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.

Large-deformation finite-element modeling of seismic landslide runout: 3D probabilistic analysis with cross-correlated random field

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.

Dynamic intelligent prediction approach for landslide displacement based on biological growth models and CNN-LSTM

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.

Landslide model tests with a miniature 2D principal stress sensor

Understanding the stress distribution derived from monitoring the principal stress(PS)in slopes is of great importance.In this study,a miniature sensor for quantifying the two-dimensional(2D)PS in landslide model tests is proposed.The fundamental principle and design of the sensor are demonstrated.The sensor comprises three earth pressure gages and one gyroscope,with the utilization of three-dimensional(3D)printing technology.The difficulties of installation location during model preparation and sensor rotation during testing can be effectively overcome using this sensor.Two different arrangements of the sensors are tested in verification tests.Additionally,the application of the sensor in an excavated-induced slope model is tested.The results demonstrate that the sensor exhibits commendable performance and achieves a desirable level of accuracy,with a principal stress angle error of±5°in the verification tests.The stress transformation of the slope model,generated by excavation,is demonstrated in the application test by monitoring the two miniature principal stress(MPS)sensors.The sensor has a significant potential for measuring primary stress in landslide model tests and other geotechnical model experiments.

Influence of material characteristics on the failure mode and process of landslide dam

Landslide dams,as frequent natural hazards,pose significant risks to human lives,property,and ecological environments.The grading characteristics and density of dam materials play a crucial role in determining the stability of landslide dams and the potential for dam breaches.To explore the failure mechanisms and evolutionary processes of landslide dams with varying soil properties,this study conducted a series of flume experiments,considering different grain compositions and material densities.The results demonstrated that grading characteristics significantly influence landslide dam stability,affecting failure patterns,breach processes,and final breach morphologies.Fine-graded materials exhibited a sequence of surface erosion,head-cut erosion,and subsequent surface erosion during the breach process,while well-graded materials typically experienced head-cut erosion followed by surface erosion.In coarse-graded dams,the high permeability of coarse particles allowed the dam to remain stable,as inflows matched outflows.The dam breach model experiments also showed that increasing material density effectively delayed the breach and reduced peak breach flow discharge.Furthermore,higher fine particle content led to a reduction in the residual dam height and the base slope of the final breach profile,although the relationship between peak breach discharge and the content of fine and coarse particles was nonlinear.To better understand breach morphology evolution under different soil characteristics and hydraulic conditions,three key points were identified—erosion point,control point,and scouring point.This study,by examining the evolution of failure patterns,breach processes,and breach flow discharges under various grading and density conditions,offers valuable insights into the mechanisms behind landslide dam failures.

Spatial distribution of landslides in response to the geomorphometric constraints in Darma Valley,Kumaun Himalaya

The Kumaun Himalaya is well-known as a geologically and tectonically complex region that amplifies mass wasting processes,particularly landslides.This study attempts to investigate the interplay between landslide distribution and the lithotectonic regime of Darma Valley,Kumaun Himalaya.A landslide inventory comprising 295 landslides in the area has been prepared and several morphotectonic proxies such as valley floor width to height ratio(Vf),stream length gradient index(SL),and hypsometric integral(HI)have been used to infer tectonic regime.Morphometric analysis,including basic,linear,aerial,and relief aspects,of 59 fourth-order sub-basins,has been carried out to estimate erosion potential in the study area.The result demonstrates that 46.77%of the landslides lie in very high,20.32%in high,21.29%in medium,and 11.61%in low erosion potential zones respectively.In order to determine the key parameters controlling erosion potential,two multivariate statistical methods namely Principal Component Analysis(PCA)and Agglomerative Hierarchical Clustering(AHC)were utilized.PCA reveals that the Higher Himalayan Zone(HHZ)has the highest erosion potential due to the presence of elongated sub-basins characterized by steep slopes and high relief.The clusters created through AHC exhibit positive PCA values,indicating a robust correlation between PCA and AHC.Furthermore,the landslide density map shows two major landslide hotspots.One of these hotspots lies in the vicinity of highly active Munsiyari Thrust(MT),while the other is in the Pandukeshwar formation within the MT's hanging wall,characterized by a high exhumation rate.High SL and low Vf values along these hotspots further corroborate that the occurrence of landslides in the study area is influenced by tectonic activity.This study,by identifying erosionprone areas and elucidating the implications of tectonic activity on landslide distribution,empowers policymakers and government agencies to develop strategies for hazard assessment and effective landslide risk mitigation,consequently safeguarding lives and communities.

Using fracture mechanics method to analyze the failure mechanism and equilibrium equation of interfacial loess-mudstone landslides

Loess-mudstone landslides are common in the Loess Plateau.Investigations into the mechanical theory of loess-mudstone landslides have become a challenging undertaking due to the distinctive interfacial properties of loess-mudstone and the unique water sensitivity characteristics of mudstone.Hence,it is imperative to develop innovative mechanical models and mathematical equations specifically tailored to loess-mudstone landslides.In this study,we analyze the fracture mechanism of the loess-mudstone sliding zone using plastic fracture mechanics and develop a unique fracture yield model.To calculate the energy release rate during the expansion of the loess-mudstone interface tip region,the shear fracture energy G is applied,which reflects both the yield failure criterion and the fracture failure criterion.To better understand the instability mechanism of loess-mudstone landslides,equilibrium equations based on G are established for tractive,compressive,and tensile loess-mudstone landslides.Based on the equilibrium equation,the critical length Lc of the sliding zone can be used for the safety evaluation of loess-mudstone landslides.In this way,this study proposes a new method for determining the failure mechanism and equilibrium equation of loessmudstone landslides,which resolves their starting mechanism,mechanical equilibrium equations,and safety evaluation indicators,thus justifying the scientific significance and practical value of this research.

Model tests and numerical analysis of emergency treatment of cohesionless soil landslide with quick-setting polyurethane

Shotcrete is one of the common solutions for shallow sliding.It works by forming a protective layer with high strength and cementing the loose soil particles on the slope surface to prevent shallow sliding.However,the solidification time of conventional cement paste is long when shotcrete is used to treat cohesionless soil landslide.The idea of reinforcing slope with polyurethane solidified soil(i.e.,mixture of polyurethane and sand)was proposed.Model tests and finite element analysis were carried out to study the effectiveness of the proposed new method on the emergency treatment of cohesionless soil landslide.Surcharge loading on the crest of the slope was applied step by step until landslide was triggered so as to test and compare the stability and bearing capacity of slope models with different conditions.The simulated slope displacements were relatively close to the measured results,and the simulated slope deformation characteristics were in good agreement with the observed phenomena,which verifies the accuracy of the numerical method.Under the condition of surcharge loading on the crest of the slope,the unreinforced slope slid when the surcharge loading exceeded 30 k Pa,which presented a failure mode of local instability and collapse at the shallow layer of slope top.The reinforced slope remained stable even when the surcharge loading reached 48 k Pa.The displacement of the reinforced slope was reduced by more than 95%.Overall,this study verifies the effectiveness of polyurethane in the emergency treatment of cohesionless soil landslide and should have broad application prospects in the field of geological disasters concerning the safety of people's live.

Forecasting step-like landslide displacement through diverse monitoring frequencies

The precision of landslide displacement prediction is crucial for effective landslide prevention and mitigation strategies.However,the role of surface monitoring frequency in influencing prediction accuracy has been largely neglected.This study examined the effect of varying monitoring frequencies on the accuracy of displacement predictions by using the Baijiabao landslide in the Three Gorges Reservoir Area(TGRA)as a case study.We collected surface automatic monitoring data at different intervals,ranging from daily to monthly.The Ensemble Empirical Mode Decomposition(EEMD)algorithm was utilized to dissect the accumulated displacements into periodic and trend components at each monitoring frequency.Polynomial fitting was applied to forecast the trend component while the periodic component was predicted with two state-of-the-art neural network models:Long Short-Term Memory(LSTM)and Gated Recurrent Unit(GRU).The predictions from these models were integrated to derive cumulative displacement forecasts,enabling a comparative analysis of prediction accuracy across different monitoring frequencies.The results demonstrate that the proposed models achieve high accuracy in landslide displacement forecasting,with optimal performance observed at moderate monitoring intervals.Intriguingly,the daily mean average error(MAE)decreases sharply with increasing monitoring frequency,reaching a plateau.These findings were corroborated by a parallel analysis of the Bazimen landslide,suggesting that moderate monitoring intervals of approximately 7 to 15 days are most conducive to achieving enhanced prediction accuracy compared to both daily and monthly intervals.

Large-scale laboratory investigation of pillar-support interaction

Underground mine pillars provide natural stability to the mine area,allowing safe operations for workers and machinery.Extensive prior research has been conducted to understand pillar failure mechanics and design safe pillar layouts.However,limited studies(mostly based on empirical field observation and small-scale laboratory tests)have considered pillar-support interactions under monotonic loading conditions for the design of pillar-support systems.This study used a series of large-scale laboratory compression tests on porous limestone blocks to analyze rock and support behavior at a sufficiently large scale(specimens with edge length of 0.5 m)for incorporation of actual support elements,with consideration of different w/h ratios.Both unsupported and supported(grouted rebar rockbolt and wire mesh)tests were conducted,and the surface deformations of the specimens were monitored using three-dimensional(3D)digital image correlation(DIC).Rockbolts instrumented with distributed fiber optic strain sensors were used to study rockbolt strain distribution,load mobilization,and localized deformation at different w/h ratios.Both axial and bending strains were observed in the rockbolts,which became more prominent in the post-peak region of the stress-strain curve.

Large-scale Rock Landslide Slip Surface Location Analysis and Strength Parameters Evaluation

Large-scale rock landslides have huge impacts on various large-scale rock engineering and project operations. They are also important aspects evaluating geological disasters. In the initial evaluations on the stability of large-scale rock landslides, in most cases, it is difficult to conduct evaluation or to have accurate evaluations because most of large-scale rock landslides are huge in size, high in slopes, and located in the canyon of mountains, which makes the exploration very difficult and thus hard to get credible data on slip surface form, location, depth and strength. This paper describes the Badi landslide happened along the Lancang River, and systematically introduces methods to analyze and verify large-scale slip surface form using terrain conditions surrounding the large-scale landslide, shape of the slide walls, and development patterns of streams and gully. This paper also introduces ways to obtain strength parameters of slip surface with the soil in the slide zone by using the principles of stress state, principles of gravity compaction, structure regeneration and strength regeneration. It is confirmed that analyzed results to the slip surface are basically consistent with the exploration results. The methods introduced here have been successfully applied to evaluate the stability of Badi large-scale rock landslide and have been applied in engineering practices.

Formation and reactivation mechanisms of large-scale ancient landslides in the Longwu River basin in the northeast Tibetan Plateau, China

The northeastern Tibetan Plateau exhibits steep topography and strong internal or external dynamic geological effect and is frequently subjected to strong earthquakes and heavy rainfall. The geological evolution has resulted in a wide distribution of ancient landslides, which has become a hotspot for studying ancient landslide formation and reactivation. In recent decades, several ancient landslides on both banks of the Longwu River, Qinghai Province, China were reactivated, causing serious economic losses and casualties. This study conducted remote sensing interpretation and ground surveys on these ancient landslides. Totally 59 ancient landslides were identified, and the formation mechanism, evolution process, and resurrection mechanism of the Longwu Xishan No.2 ancient landslide were analyzed by means of a detailed field geological survey, drilling, and series of experimental tests such as the particle size distribution test, the Xray diffraction test and the mechanical properties test. The results show that the formation of these ancient landslides is closely associated with the uplift of the Tibetan Plateau and the erosion of the Longwu River. Firstly, the intermittent uplift of the Tibetan Plateau lead to the diversion and downcutting of the Longwu River basin, which forms the alternate slope topography with steep and slow slopes, thereby providing favourable topography and slope structure conditions for the formation of landslides. Secondly, 34.5% clay-mineral content in the Neoproterozoic mudstone with 32.7% particle size less than 0.005 mm, and the corrosion and softening effects of the Neogene mudstone with high clay mineral content under the erosion of water provides favourable material conditions for the formation of landslides. Thirdly, rainfall and human activities are the primary triggering factors for the revival of this ancient landslide group. It is revealed that the evolution process of the ancient landslides on both banks of the Longwu River can be divided into five stages namely tectonic rapid uplift slope formation, river erosion creep-sliding deformation, slope instability critical status, landslide failure-movement-accumulation, and slope reactivation under rainfall erosion and engineering excavation.

Stability and deformation of Xiaozhuang landslide: A large-scale creeping landslide in Gansu, China

Information about the long-term spatiotemporal evolution of landslides can improve our understanding of the landslide development process and can help prevent landslide disasters.This paper describes the Xiaozhuang landslide triggered by a historical earthquake and rainfall in Tianshui,Northwest China.The landslide is dominated by rotational-sliding movement.Several new failures and many fissures formed in the landslide area as a result of the 2013 Ms 6.6 Minxian earthquake and rainfall.Accordingly,field investigations,borehole drilling,geotechnical laboratory tests,and numerical calculations were conducted to study the mechanism of the landslide and to forecast its stability.A triaxial creep test of the slip soil indicates that the axial deformation of the mudstone increases with increasing water content.Numerical simulations suggest that failure is prone to occur within the deep part of the landslide under earthquake conditions.If the input seismic acceleration exceeds 0.2 g,the landslide will become unstable.Furthermore,the horizontal peak ground acceleration near the surface of the landslide is greater than that at depth.During a strong earthquake,the unstable regions are primarily located in the middle of the landslide and at its crest.When the rainfall intensity rate is 200 mm/d,the factor of safety is 1.319 and a dangerous zone appears in the lower and middle parts of the landslide.

Uncertainties in landslide susceptibility prediction:Influence rule of different levels of errors in landslide spatial position

The accuracy of landslide susceptibility prediction(LSP)mainly depends on the precision of the landslide spatial position.However,the spatial position error of landslide survey is inevitable,resulting in considerable uncertainties in LSP modeling.To overcome this drawback,this study explores the influence of positional errors of landslide spatial position on LSP uncertainties,and then innovatively proposes a semi-supervised machine learning model to reduce the landslide spatial position error.This paper collected 16 environmental factors and 337 landslides with accurate spatial positions taking Shangyou County of China as an example.The 30e110 m error-based multilayer perceptron(MLP)and random forest(RF)models for LSP are established by randomly offsetting the original landslide by 30,50,70,90 and 110 m.The LSP uncertainties are analyzed by the LSP accuracy and distribution characteristics.Finally,a semi-supervised model is proposed to relieve the LSP uncertainties.Results show that:(1)The LSP accuracies of error-based RF/MLP models decrease with the increase of landslide position errors,and are lower than those of original data-based models;(2)70 m error-based models can still reflect the overall distribution characteristics of landslide susceptibility indices,thus original landslides with certain position errors are acceptable for LSP;(3)Semi-supervised machine learning model can efficiently reduce the landslide position errors and thus improve the LSP accuracies.

How do the landslide and non-landslide sampling strategies impact landslide susceptibility assessment? d A catchment-scale case study from China

The aim of this study is to investigate the impacts of the sampling strategy of landslide and non-landslide on the performance of landslide susceptibility assessment(LSA).The study area is the Feiyun catchment in Wenzhou City,Southeast China.Two types of landslides samples,combined with seven non-landslide sampling strategies,resulted in a total of 14 scenarios.The corresponding landslide susceptibility map(LSM)for each scenario was generated using the random forest model.The receiver operating characteristic(ROC)curve and statistical indicators were calculated and used to assess the impact of the dataset sampling strategy.The results showed that higher accuracies were achieved when using the landslide core as positive samples,combined with non-landslide sampling from the very low zone or buffer zone.The results reveal the influence of landslide and non-landslide sampling strategies on the accuracy of LSA,which provides a reference for subsequent researchers aiming to obtain a more reasonable LSM.