Landslide Susceptibility mapping using random forest and extreme gradient boosting: A case study of Fengjie, Chongqing

Geological Journal, Journal Year: 2023, Volume and Issue: 58(6), P. 2372 - 2387

Published: Feb. 7, 2023

Landslide susceptibility analysis can provide theoretical support for landslide risk management. However, some analyses are not sufficiently interpretable. Moreover, the accuracy of many research methods needs to be improved. Therefore, this study supplement these deficiencies. This aims evaluation effects random forest (RF) and extreme gradient boosting (XGBoost) classifier models on susceptibility, compare their applicability in Fengjie County, Chongqing, a typical landslide‐prone area southwest China. Firstly, 1624 landslides information from 1980 2020 were obtained through field investigation, geospatial database 16 conditional factors had been constructed. Secondly, non‐landslide points selected form complete data set RF XGBoost established. Finally, under ROC curve (AUC) value, accuracy, F ‐score used two models. The results show that even though both classifiers have highly accurate model performs better. In comparison, has higher AUC value 0.866, its approximately 2% than XGBoost. land use, elevation, lithology County contribute occurrence landslides. is due human engineering activities (such as reclamation, housing construction) resulting low slope stability widely distributed sandstone, siltstone, mudstone layers owing permeability planes weakness.

Language: Английский

---

Ensemble learning framework for landslide susceptibility mapping: Different basic classifier and ensemble strategy

Geoscience Frontiers, Journal Year: 2023, Volume and Issue: 14(6), P. 101645 - 101645

Published: June 7, 2023

The application of ensemble learning models has been continuously improved in recent landslide susceptibility research, but most studies have no unified framework. Moreover, few papers discussed the applicability model mapping at township level. This study aims defining a robust framework that can become benchmark method for future research dealing with comparison different models. For this purpose, present work focuses on three basic classifiers: decision tree (DT), support vector machine (SVM), and multi-layer perceptron neural network (MLPNN) two homogeneous such as random forest (RF) extreme gradient boosting (XGBoost). hierarchical construction deep relied leading technologies (i.e., homogeneous/heterogeneous bagging, boosting, stacking strategy) to provide more accurate effective spatial probability occurrence. selected area is Dazhou town, located Jurassic red-strata Three Gorges Reservoir Area China, which strategic economic currently characterized by widespread risk. Based long-term field investigation, inventory counting thirty-three slow-moving polygons was drawn. results show do not necessarily perform better; instance, Bagging based DT-SVM-MLPNN-XGBoost performed worse than single XGBoost model. Amongst eleven tested models, Stacking RF-XGBoost model, ensemble, showed highest capability predicting landslide-affected areas. Besides, factor behaviors DT, SVM, MLPNN, RF reflected characteristics landslides reservoir area, wherein unfavorable lithological conditions intense human engineering activities water level fluctuation, residential construction, farmland development) are proven be key triggers. presented approach could used occurrence prediction similar regions other fields.

Language: Английский

---

Uncertainties of landslide susceptibility prediction considering different landslide types

Journal of Rock Mechanics and Geotechnical Engineering, Journal Year: 2023, Volume and Issue: 15(11), P. 2954 - 2972

Published: March 20, 2023

Most literature related to landslide susceptibility prediction only considers a single type of landslide, such as colluvial rock fall or debris flow, rather than different types, which greatly affects performance. To construct efficient considering Huichang County in China is taken example. Firstly, 105 falls, 350 landslides and 11 environmental factors are identified. Then four machine learning models, namely logistic regression, multi-layer perception, support vector C5.0 decision tree applied for modeling landslide. Thirdly, three (LSP) models types based on with excellent performance constructed generate final susceptibility: (i) united method, combines all directly; (ii) probability statistical couples analyses indices under formula; (iii) maximum comparison selects the index through comparing predicted landslides. Finally, uncertainties assessed by accuracy, mean value standard deviation. It concluded that LSP results coupled basically conform spatial occurrence patterns County. The method has best performance, followed method. More cases needed verify this result in-depth. superior taking into account.

Language: Английский

---

Slope stability prediction based on a long short-term memory neural network: comparisons with convolutional neural networks, support vector machines and random forest models

International Journal of Coal Science & Technology, Journal Year: 2023, Volume and Issue: 10(1)

Published: April 10, 2023

Abstract The numerical simulation and slope stability prediction are the focus of disaster research. Recently, machine learning models commonly used in prediction. However, these have some problems, such as poor nonlinear performance, local optimum incomplete factors feature extraction. These issues can affect accuracy Therefore, a deep algorithm called Long short-term memory (LSTM) has been innovatively proposed to predict stability. Taking Ganzhou City China study area, landslide inventory their characteristics geotechnical parameters, height angle analyzed. Based on characteristics, typical soil slopes constructed using Geo-Studio software. Five control affecting stability, including height, angle, internal friction cohesion volumetric weight, selected form different construct model input variables. Then, limit equilibrium method is calculate coefficients under factors. Each coefficient its corresponding sample. As result, total 2160 training samples 450 testing constructed. sample sets imported into LSTM for modelling compared with support vector (SVM), random forest (RF) convolutional neural network (CNN). results show that overcomes problem difficulty extracting global features. Furthermore, better performance SVM, RF CNN models.

Language: Английский

---

Landslide susceptibility mapping in Three Gorges Reservoir area based on GIS and boosting decision tree model

Stochastic Environmental Research and Risk Assessment, Journal Year: 2023, Volume and Issue: 37(6), P. 2283 - 2303

Published: Feb. 8, 2023

Language: Английский

---

A novel swarm intelligence: cuckoo optimization algorithm (COA) and SailFish optimizer (SFO) in landslide susceptibility assessment

Stochastic Environmental Research and Risk Assessment, Journal Year: 2023, Volume and Issue: 37(5), P. 1717 - 1743

Published: Jan. 29, 2023

Language: Английский

---

Unmanned Aerial Vehicles (UAVs) in Landslide Investigation and Monitoring: A Review

Drones, Journal Year: 2024, Volume and Issue: 8(1), P. 30 - 30

Published: Jan. 22, 2024

Over the past decade, Unmanned Aerial Vehicles (UAVs) have emerged as essential tools for landslide studies, particularly in on-site investigations. This paper reviews UAV applications with a focus on static geological characteristics, monitoring temporal and spatial dynamics, responses post-events. We discuss functions limitations of various types UAVs sensors (RGB cameras, multi-spectral thermal IR SAR, LiDAR), outlining their roles data processing methods applications. review focuses UAVs’ geology surveys, emphasizing mapping, modeling characterization. For change monitoring, it provides an overview evolution through UAV-based shedding light dynamic processes. Moreover, this underscores crucial role emergent response scenarios, detailing strategies automated detection using machine learning algorithms. The discussion challenges opportunities highlights need ongoing technology advancements, addressing regulatory hurdles, hover time limitations, 3D reconstruction accuracy potential integration technologies like swarms.

Language: Английский

---

Physics-informed optimization for a data-driven approach in landslide susceptibility evaluation

Journal of Rock Mechanics and Geotechnical Engineering, Journal Year: 2024, Volume and Issue: 16(8), P. 3192 - 3205

Published: March 16, 2024

Landslide susceptibility mapping is an integral part of geological hazard analysis. Recently, the emphasis many studies has been on data-driven models, notably those derived from machine learning, owing to their aptitude for tackling complex non-linear problems. However, prevailing models often disregard qualitative research, leading limited interpretability and mistakes in extracting negative samples, i.e. inaccurate non-landslide samples. In this study, Scoops 3D (a three-dimensional slope stability analysis tool) was utilized conduct a assessment Yunyang section Three Gorges Reservoir area. The depth bedrock predicted utilizing Convolutional Neural Network (CNN), incorporating local boreholes building insights prior research. Random Forest (RF) algorithm subsequently used execute landslide proposed methodology demonstrated notable increase 29.25% evaluation metric, area under receiver operating characteristic curve (ROC-AUC), outperforming benchmark model. Furthermore, map generated by model superior interpretability. This result not only validates effectiveness amalgamating mathematical mechanistic such analyses, but it also carries substantial academic practical implications.

Language: Английский

---

Enhancing landslide susceptibility mapping incorporating landslide typology via stacking ensemble machine learning in Three Gorges Reservoir, China

Geoscience Frontiers, Journal Year: 2024, Volume and Issue: 15(4), P. 101802 - 101802

Published: Jan. 29, 2024

Different types of landslides exhibit distinct relationships with environmental conditioning factors. Therefore, in regions where multiple coexist, it is required to separate landslide for susceptibility mapping (LSM). In this paper, a landslide-prone area located Chongqing Province within the middle and upper reaches Three Gorges Reservoir (TGRA), China, was selected as study area. 733 were classified into three types: reservoir-affected landslides, non-reservoir-affected rockfalls. Four inventory datasets 15 conditional factors trained by Machine Learning models (logistic regression, random forest, support vector machine), Deep (DL) model. After comparing using receiver operating characteristics (ROC), indexes acquired best performing These then used input generate final map based on Stacking method. The results revealed that DL model showed performance LSM without considering types, achieving an under curve (AUC) 0.854 testing 0.922 training. Moreover, when we separated LSM, AUC improved 0.026 0.044 Thus, paper demonstrates different can significantly improve quality maps. maps turn, be valuable tools evaluating mitigating hazards.

Language: Английский

---

Regional early warning model for rainfall induced landslide based on slope unit in Chongqing, China

Engineering Geology, Journal Year: 2024, Volume and Issue: 333, P. 107464 - 107464

Published: Feb. 29, 2024

Language: Английский

---