A hybrid data-driven approach for rainfall-induced landslide susceptibility mapping: Physically-based probabilistic model with convolutional neural network

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 pr...

Full description

Saved in:
Bibliographic Details
Main Authors: Hong-Zhi Cui, Bin Tong, Tao Wang, Jie Dou, Jian Ji
Format: Article
Language:English
Published: Elsevier 2025-08-01
Series:Journal of Rock Mechanics and Geotechnical Engineering
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S167477552400355X
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849335631986032640
author Hong-Zhi Cui
Bin Tong
Tao Wang
Jie Dou
Jian Ji
author_facet Hong-Zhi Cui
Bin Tong
Tao Wang
Jie Dou
Jian Ji
author_sort Hong-Zhi Cui
collection DOAJ
description 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.
format Article
id doaj-art-d5e9bd9ad1c94a93be3c80b4f1f99a12
institution Kabale University
issn 1674-7755
language English
publishDate 2025-08-01
publisher Elsevier
record_format Article
series Journal of Rock Mechanics and Geotechnical Engineering
spelling doaj-art-d5e9bd9ad1c94a93be3c80b4f1f99a122025-08-20T03:45:11ZengElsevierJournal of Rock Mechanics and Geotechnical Engineering1674-77552025-08-011784933495110.1016/j.jrmge.2024.08.005A hybrid data-driven approach for rainfall-induced landslide susceptibility mapping: Physically-based probabilistic model with convolutional neural networkHong-Zhi Cui0Bin Tong1Tao Wang2Jie Dou3Jian Ji4Key Laboratory of Ministry of Education for Geomechanics and Embankment Engineering, Hohai University, Nanjing, 210024, China; Division of Geotechnical Engineering and Geosciences, Department of Civil and Environmental Engineering, Universitat Politecnica de Catalunya, Barcelona, 08034, SpainKey Laboratory of Ministry of Education for Geomechanics and Embankment Engineering, Hohai University, Nanjing, 210024, ChinaKey Laboratory of Ministry of Education for Geomechanics and Embankment Engineering, Hohai University, Nanjing, 210024, ChinaBadong National Observation and Research Station of Geohazards, China University of Geosciences, Wuhan, 430074, ChinaKey Laboratory of Ministry of Education for Geomechanics and Embankment Engineering, Hohai University, Nanjing, 210024, China; Corresponding author.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.http://www.sciencedirect.com/science/article/pii/S167477552400355XRainfall landslidesLandslide susceptibility mappingHybrid modelPhysically-based modelConvolution neural network (CNN)Probability of failure (POF)
spellingShingle Hong-Zhi Cui
Bin Tong
Tao Wang
Jie Dou
Jian Ji
A hybrid data-driven approach for rainfall-induced landslide susceptibility mapping: Physically-based probabilistic model with convolutional neural network
Journal of Rock Mechanics and Geotechnical Engineering
Rainfall landslides
Landslide susceptibility mapping
Hybrid model
Physically-based model
Convolution neural network (CNN)
Probability of failure (POF)
title A hybrid data-driven approach for rainfall-induced landslide susceptibility mapping: Physically-based probabilistic model with convolutional neural network
title_full A hybrid data-driven approach for rainfall-induced landslide susceptibility mapping: Physically-based probabilistic model with convolutional neural network
title_fullStr A hybrid data-driven approach for rainfall-induced landslide susceptibility mapping: Physically-based probabilistic model with convolutional neural network
title_full_unstemmed A hybrid data-driven approach for rainfall-induced landslide susceptibility mapping: Physically-based probabilistic model with convolutional neural network
title_short A hybrid data-driven approach for rainfall-induced landslide susceptibility mapping: Physically-based probabilistic model with convolutional neural network
title_sort hybrid data driven approach for rainfall induced landslide susceptibility mapping physically based probabilistic model with convolutional neural network
topic Rainfall landslides
Landslide susceptibility mapping
Hybrid model
Physically-based model
Convolution neural network (CNN)
Probability of failure (POF)
url http://www.sciencedirect.com/science/article/pii/S167477552400355X
work_keys_str_mv AT hongzhicui ahybriddatadrivenapproachforrainfallinducedlandslidesusceptibilitymappingphysicallybasedprobabilisticmodelwithconvolutionalneuralnetwork
AT bintong ahybriddatadrivenapproachforrainfallinducedlandslidesusceptibilitymappingphysicallybasedprobabilisticmodelwithconvolutionalneuralnetwork
AT taowang ahybriddatadrivenapproachforrainfallinducedlandslidesusceptibilitymappingphysicallybasedprobabilisticmodelwithconvolutionalneuralnetwork
AT jiedou ahybriddatadrivenapproachforrainfallinducedlandslidesusceptibilitymappingphysicallybasedprobabilisticmodelwithconvolutionalneuralnetwork
AT jianji ahybriddatadrivenapproachforrainfallinducedlandslidesusceptibilitymappingphysicallybasedprobabilisticmodelwithconvolutionalneuralnetwork
AT hongzhicui hybriddatadrivenapproachforrainfallinducedlandslidesusceptibilitymappingphysicallybasedprobabilisticmodelwithconvolutionalneuralnetwork
AT bintong hybriddatadrivenapproachforrainfallinducedlandslidesusceptibilitymappingphysicallybasedprobabilisticmodelwithconvolutionalneuralnetwork
AT taowang hybriddatadrivenapproachforrainfallinducedlandslidesusceptibilitymappingphysicallybasedprobabilisticmodelwithconvolutionalneuralnetwork
AT jiedou hybriddatadrivenapproachforrainfallinducedlandslidesusceptibilitymappingphysicallybasedprobabilisticmodelwithconvolutionalneuralnetwork
AT jianji hybriddatadrivenapproachforrainfallinducedlandslidesusceptibilitymappingphysicallybasedprobabilisticmodelwithconvolutionalneuralnetwork