Article
Engineering, Environmental
Wenjuan Li, Zhice Fang, Yi Wang
Summary: This paper introduces a hybrid framework integrating stacking ensemble with convolutional neural network (CNN) and recurrent neural network (RNN) for landslide spatial prediction in the Three Gorges Reservoir area, China. Experimental results demonstrate that the proposed framework achieves the best predictive capability in terms of AUC compared to CNN, RNN, and logistic regression, which is significant for landslide disaster management and assessment.
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT
(2022)
Article
Engineering, Environmental
Fasheng Miao, Fancheng Zhao, Yiping Wu, Linwei Li, Akos Torok
Summary: In this study, a boosting-C5.0 decision tree model is used to prepare regional landslide susceptibility mapping (LSM) in the Three Gorges Reservoir area. The results show that landslide susceptibility is divided into four levels: low, moderate, high, and very high, with the boosting-C5.0 model performing the best. This study demonstrates the feasibility of machine learning in landslide susceptibility assessment and provides a basis for risk management and control of geological disasters.
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT
(2023)
Article
Engineering, Environmental
Jian Hu, Kaibin Xu, Genglong Wang, Youcun Liu, Muhammad Asim Khan, Yimin Mao, Maosheng Zhang
Summary: The study proposes a novel method for constructing landslide susceptibility maps using the OPTICS algorithm and Hausdorff distance. This approach effectively divides mapping units into multiple subclasses and categorizes them into five susceptibility levels based on landslide density values. Applying this method can significantly enhance the assessment of landslide susceptibility.
BULLETIN OF ENGINEERING GEOLOGY AND THE ENVIRONMENT
(2021)
Article
Environmental Sciences
Zhice Fang, Yi Wang, Gonghao Duan, Ling Peng
Summary: This study introduces a new ensemble framework to predict landslide susceptibility by combining decision trees with the rotation forest technique. By selecting training and validation sets based on historical landslide locations, screening landslide conditioning factors, producing training subsets, and integrating all DTs classification results using RF ensemble technique, the framework effectively improves the spatial prediction of landslides. The experimental results show that the proposed ensemble methods outperform traditional DTs and other popular ensemble methods in terms of predictive values.
Article
Geochemistry & Geophysics
Cheng Chen, Lei Fan
Summary: This study proposes an interpretable DL model called Deep-Attention-LSF, which assigns significance scores to contributing factors at local levels for attributing landslide susceptibility. The Deep-Attention-LSF model outperforms other models in predicting landslide occurrence and provides reasonable explanations for landslide attributions.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Computer Science, Interdisciplinary Applications
Huijuan Zhang, Yingxu Song, Shiluo Xu, Yueshun He, Zhiwen Li, Xianyu Yu, Ye Liang, Weicheng Wu, Yue Wang
Summary: This study investigates the application of a class-weighted algorithm with LR, LightGBM, and RF models in landslide susceptibility evaluation. Results show that the weighted models outperform the unweighted ones, with the WRF model exhibiting the best performance. The insights from this research will be useful for improving landslide susceptibility mapping and prevention strategies in the Wanzhou section.
COMPUTERS & GEOSCIENCES
(2022)
Article
Geosciences, Multidisciplinary
Jinchang Shi
Summary: This study used a geographic information system (GIS) platform and the maximum entropy (MaxEnt) model to develop a landslide susceptibility mapping for the China National Highway 109 New Line Expressway. The results showed high reliability of the mapping, with a ROC value of 82.1% and %LRclass of 2.25. Furthermore, the study compared the landslide probability levels of two roads and found that the new line had lower risk compared to the existing line. This research provides valuable insights for decision-making in similar linear engineering projects.
FRONTIERS IN EARTH SCIENCE
(2022)
Article
Engineering, Geological
Chao Zhou, Ying Cao, Xie Hu, Kunlong Yin, Yue Wang, Filippo Catani
Summary: This study proposes a new method to obtain dynamic landslide hazard maps by utilizing ground deformation measured by SAR imagery. By combining spatial probability of landslide occurrence and temporal probability under different rainfall conditions, a preliminary hazard map is initialized. The final hazard map is determined by considering deformation velocities. The proposed method reduces false-negative and false-positive errors in landslide hazard mapping and provides higher accuracy.
Article
Environmental Sciences
Zhu Liang, Changming Wang, Zhijie Duan, Hailiang Liu, Xiaoyang Liu, Kaleem Ullah Jan Khan
Summary: This study employed a hybrid model that utilized the advantages of both supervised and unsupervised learning, and through a two-stage modeling process, constructed a robust landslide prediction model with improved performance.
Article
Engineering, Environmental
Jingjing Long, Yong Liu, Changdong Li, Zhiyong Fu, Haikuan Zhang
Summary: This study identified Jurassic facility-sliding strata as a fundamental factor affecting rainfall-reservoir induced landslides in western Hubei Province, China Three Gorges Reservoir area. A novel hybrid model based on the two steps self-organizing mapping-random forest (two steps SOM-RF) algorithm was proposed to address the problem of identifying true landslides and non-landslides. The results showed that considering true landslides and non-landslides is effective in producing a more accurate landslide susceptibility map with superior prediction skill and higher reliability.
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT
(2021)
Article
Engineering, Environmental
Areeba Qazi, Kanwarpreet Singh, Dinesh Kumar Vishwakarma, Hazem Ghassan Abdo
Summary: The aim of this study is to determine and evaluate the landslide susceptibility zonation of Kinnaur district in HP, providing preventive and remedial measures. Three statistical methods were used to map the susceptibility of landslide hazard. The landslide inventory was created using data acquired from the Geological Survey of India. Various geo-environmental factors were selected to create landslide susceptibility maps, which were categorized into five categories based on mapping data. The study emphasizes the importance of identifying landslide areas and assisting local authorities in implementing precautionary measures and improving contingency plans.
BULLETIN OF ENGINEERING GEOLOGY AND THE ENVIRONMENT
(2023)
Article
Green & Sustainable Science & Technology
Prafull Singh, Ankit Sharma, Ujjwal Sur, Praveen Kumar Rai
Summary: This study conducted a comparative assessment between statistical information value (SIV) and index of entropy (IOE) to determine their effectiveness in landslide susceptibility mapping in the Bhanupali-Beri region. The index of entropy model showed better results in predicting landslide susceptibility, with landuse-landcover, topographic wetness index (TWI), lineament density, geomorphology, and slope identified as the most significant factors contributing to landslide occurrence.
ENVIRONMENT DEVELOPMENT AND SUSTAINABILITY
(2021)
Article
Remote Sensing
Liang Lv, Tao Chen, Jie Dou, Antonio Plaza
Summary: Landslide susceptibility mapping is crucial for landslide prevention. This paper proposes a hybrid framework using ensemble learning methods and deep learning models. The proposed model shows improved accuracy and stability compared to single deep learning models, and achieves high performance in testing. Elevation factor is found to be the most influential in landslide susceptibility mapping.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2022)
Article
Green & Sustainable Science & Technology
Aihua Wei, Kaining Yu, Fenggang Dai, Fuji Gu, Wanxi Zhang, Yu Liu
Summary: This study compares popular ensemble machine learning-based models and applies them to landslide susceptibility mapping. The results show that several ensemble models can appropriately predict landslide susceptibility maps, with the XGBoost model performing the best.
Article
Environmental Sciences
Ali Polat
Summary: This study utilized the LSAT toolbox to produce landslide susceptibility maps for the Akincilar region, achieving prediction rates ranging from 70% to 73% using five different methods. LSAT proved to be effective in reducing time-consuming processes associated with constructing LSM, allowing for quick and automated data preparation, visualization of modeling results, and accuracy assessment of LSM.
ENVIRONMENTAL EARTH SCIENCES
(2021)
Article
Engineering, Geological
Kun Fang, Minghao Miao, Huiming Tang, Shixun Jia, Ao Dong, Pengju An, Bocheng Zhang
Summary: This study analyzes the failure mechanism of a physical slope model under excavation through multi-field monitoring. The results show that the relative displacement and triangular shear plane reflect the deformation behavior of the slope, and the arch ring expands and deforms during excavation. The failure time of the slope can be effectively predicted using the inverse velocity method. Multi-field monitoring can reveal the behavior of the slope model from different perspectives and provide new insights into the failure mechanism of the slope.
Article
Thermodynamics
Xinxin Li, Chengyu Li, Wenping Gong, Yanjie Zhang, Junchao Wang
Summary: In this study, a discrete fracture network modeling procedure based on finite element method is proposed to simulate the uncertainty of heat extraction performance in a complex fracture network. Probabilistic analysis of heat production performance using the Monte Carlo simulation method indicates that the performance can be more effectively assessed and the uncertainties involved cannot be ignored.
Article
Engineering, Civil
Jie Tan, Changdong Li, Jia-Qing Zhou, Huiming Tang
Summary: This study systematically compares the fluid flow in filled fractures with open fractures and porous media based on pore-scale simulation results. The filling medium has a dual role in altering the flow regime, and a new semi-empirical model is proposed to estimate the fluid flow through filled fractures.
JOURNAL OF HYDROLOGY
(2023)
Article
Engineering, Geological
Kun Fang, Ao Dong, Huiming Tang, Pengju An, Bocheng Zhang, Minghao Miao, Bingdong Ding, Xiaolong Hu
Summary: Physical modelling is an effective method for studying landslides under laboratory conditions. This study evaluates the performance of a multismartphone measurement system for landslide model tests. The system shows high trueness and precision at millimetre and submillimetre levels. It can be used to monitor landslides and accurately measure deformation.
Review
Geosciences, Multidisciplinary
Kun Fang, Huiming Tang, Changdong Li, Xuexue Su, Pengju An, Sixuan Sun
Summary: This article provides an overview of the application of centrifuge modelling in landslide science. It discusses the experimental principles, various triggering factors for landslide models, and methods for mitigating landslides in centrifuge. The behaviors of centrifuge models, including deformation and failure mechanisms, are also discussed. Based on this review, a best-practice methodology for preparing a centrifuge landslide test is proposed, along with suggestions for further research efforts.
GEOSCIENCE FRONTIERS
(2023)
Article
Engineering, Geological
Fumeng Zhao, Wenping Gong, Huiming Tang, Shiva P. Pudasaini, Tianhe Ren, Zhan Cheng
Summary: Land subsidence caused by over-exploitation of groundwater resources poses a significant hazard in many large cities globally. Assessing infrastructure risks under the threat of land subsidence is crucial for urban planning and design. This study proposes an integrated approach combining land subsidence, ground fissures, and elements at risk. Time-series Interferometric Synthetic Aperture Radar (InSAR) is used to study ground surface deformation, and differential settlement is assessed using an angular distortion index. Land use classification analysis is conducted to identify potentially affected elements using optical and radar images with an object-based approach. Finally, a risk matrix integrating differential settlement, ground fissures, and land use classification results is employed to assess land subsidence risk. The effectiveness of the proposed method is demonstrated through a risk assessment of land subsidence in Xi'an, China, and the advantages of synergetic land use classification over pixel-based classification are illustrated.
ENGINEERING GEOLOGY
(2023)
Article
Geosciences, Multidisciplinary
Kun Fang, Huiming Tang, Jichen Zhu, Zijin Fu, Pengju An, Bocheng Zhang, Chunyan Tang
Summary: A simplified geomechanical model was proposed to investigate the deformation behaviors of necking-type slopes. Three physical models of necking-type slopes were built according to the model, and preliminary calculations related to the arching effect were conducted. The evolution stages of necking-type slopes were presented based on the formation and disappearance of the arching effect. The proposed model and physical models provide guidance for the establishment of geomechanical and physical models of landslides.
JOURNAL OF EARTH SCIENCE
(2023)
Article
Engineering, Civil
Yunfeng Ge, Jenny Liu, Xiong Zhang, Huiming Tang, Xiaolong Xia
Summary: An innovative approach was developed to automatically measure cracks on concrete using 3D point clouds collected by a terrestrial laser scanner. The approach integrates various techniques to characterize the cracks and accurately determine their dimensions. The developed method was validated using two cases of surface cracks, and it showed good agreement with manual measurements and projection method.
JOURNAL OF INFRASTRUCTURE SYSTEMS
(2023)
Article
Engineering, Geological
Yanjie Zhang, Bilal M. Ayyub, Wenping Gong, Huiming Tang
Summary: Landslides frequently disrupt roadway networks in mountainous regions worldwide. This paper establishes a risk assessment framework based on risk theory and complex network theory to combine landslide susceptibility mapping along roadways and impact assessment on the roadway network. The proposed framework was illustrated through a case study of the roadway network in Fengjie County, China, and identified critical road segments highly susceptible to landslides. The results can support adaptive strategies for landslide mitigation, preparedness, and emergency response services, as well as improve roadway plans to reduce exposure and associated consequences.
Article
Engineering, Geological
Yunfeng Ge, Qian Chen, Huiming Tang, Bei Cao, Wakeel Hussain
Summary: This paper proposes a method for rapid and quantitative estimation of geological strength index (GSI) using 3D point clouds generated through non-contact measurement methods. The method includes acquiring point clouds using a terrestrial laser scanner, identifying discontinuities through artificial neural networks (ANN) and density-based spatial clustering of applications with noise (DBSCAN), extracting geometric information for the detected discontinuities, and estimating GSI according to the detection and characterization of discontinuities. The application results show that 3D point clouds can provide an objective and efficient way to obtain GSI values of rock mass and can be used as a potential alternative to the traditional GSI estimation method.
ROCK MECHANICS AND ROCK ENGINEERING
(2023)
Article
Environmental Sciences
Fumeng Zhao, Wenping Gong, Tianhe Ren, Jun Chen, Huiming Tang, Tianzheng Li
Summary: This study proposes a method that integrates InSAR and the random forest method for improved permafrost stability mapping on the Tibetan Plateau. The ground deformation rate is first studied using InSAR, and the relationship between environmental factors and permafrost stability is mapped using the random forest method. The trained random forest model is then applied to map permafrost stability in the whole study area.
Article
Computer Science, Interdisciplinary Applications
Lei Xing, Wenping Gong, Bing Li, Chao Zhao, Huiming Tang, Lei Wang
Summary: Earthquake-induced rock slope failure is a destructive geohazard that poses serious threats to human lives and properties. The potential hazard of a rock slope under earthquake loading can be assessed by considering the characteristics of failure and runout, which are affected by inherent fractures. Traditional deterministic analysis of rock slope hazard may lead to uncertainties due to differences between arbitrarily chosen discrete fracture networks (DFNs) and true DFNs. To address this issue, this paper proposes a probabilistic analysis that uses multiple realizations of systematically generated DFNs, obtained through Monte Carlo simulation, as inputs for discrete element method (DEM) modeling of slope behavior, thus explicitly considering and evaluating the uncertainty induced by DFN modeling.
COMPUTERS AND GEOTECHNICS
(2023)
Article
Engineering, Geological
Jia-Qing Zhou, Fu-Shuo Gan, Changdong Li, Huiming Tang
Summary: Through massive direct numerical simulations, we have found a seemingly irrational dependency of inertial permeability on the volumetric flux range. This dependency is closely related to the evolution of microscale eddies and macroscale flow regimes. Based on this finding, we propose the concept of global inertial permeability and establish two parametrized criterion models for calculating it. These findings are of great significance for accurately evaluating the hydraulic conductivity of rock fractures and have applications in various geophysical and industrial fields.
ENGINEERING GEOLOGY
(2023)
Article
Engineering, Geological
Yang Ye, Changdong Li, Yawu Zeng, Huiming Tang
Summary: In this study, a three-dimensional discrete element model considering random microcracks was used to investigate the variability of mechanical parameters of rock. The results showed that the presence of microcracks significantly affected the coefficient of variations and average values of the mechanical parameters. The parametric study indicated that the effect of random microcracks on the coefficient of variations was significant, while the effect of grain size variation was negligible.
INTERNATIONAL JOURNAL OF ROCK MECHANICS AND MINING SCIENCES
(2023)