Multi-branch network assessment and dynamic change analysis of wide-area landslide susceptibility

doi:10.11947/j.AGCS.2025.20240014

Abstract

Abstract:

In response to the issue where convolutional neural networks (CNN) in landslide susceptibility assessments overly focus on specific factors due to data channel stacking, this study proposes a multi-branch data fusion model for landslide susceptibility assessment. The model integrates multi-source remote sensing data features through a multi-branch structure and an adaptive weighting mechanism, and then it leverages deep CNNs to fully extract the semantic information of evaluation factors for accurate landslide susceptibility assessment. The experiment selected the southeastern Qinghai-Tibet Plateau as a typical study area, and comparative analyses with the random forest, shallow CNN, and ResNet101 models showed that the proposed multi-branch network model offers superior performance in wide-area landslide susceptibility assessment. The model outperforms existing models in accuracy, precision, recall, F₁ score, area under the curve (AUC), and frequency ratio accuracy, with respective values of 0.88, 0.89, 0.92, 0.90, 0.92, and 0.97. Based on these results, this study further investigates the intrinsic relationship between fluctuations in environmental factors, such as vegetation and rainfall, and changes in the landslide susceptibility index by combining the landslide susceptibility assessment results from five consecutive years. The study also explored temporal and spatial variations in the landslide susceptibility index using the coefficient of variation. The results indicate that in the past five years, the Minjiang-Daduhe River Basin, Yalong River Basin, and Yarlung Tsangpo River Basin have experienced a trend of first increasing and then decreasing landslide risk, influenced by the normalized difference vegetation index (NDVI) and local rainfall variations. In contrast, the Jinsha River Basin and Nujiang-Lancang River Basin have shown smaller fluctuations in vegetation and rainfall, with landslide susceptibility levels generally remaining at medium to high-risk levels. The proposed model in this study can be used as references for landslide risk assessments in similar regions.

Key words: landslide susceptibility assessments, multi-branch data fusion, convolutional neural networks, multi-source remote sensing, dynamic changes

CLC Number:

P237

Jichao LÜ, Rui ZHANG, Xu HE, Ruikai HONG, Age SHAMA, Guoxiang LIU. Multi-branch network assessment and dynamic change analysis of wide-area landslide susceptibility[J]. Acta Geodaetica et Cartographica Sinica, 2025, 54(1): 104-122.

Figures/Tables 20

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

Fig. 6

Fig. 7

Tab. 1

Fig. 8

Tab. 2

Fig. 9

Fig. 10

Tab. 3

Tab. 4

Fig. 11

Tab. 5

Fig. 12

Fig. 13

Fig. 14

Fig. 15

References 53

[1]	CHEN Cheng, FAN Lei. An attribution deep learning interpretation model for landslide susceptibility mapping in the Three Gorges Reservoir area[J]. IEEE Transactions on Geoscience and Remote Sensing, 2023, 61: 3323668.
[2]	ZHANG Rui, LÜ Jichao, YANG Yunjie, et al. Analysis of the impact of terrain factors and data fusion methods on uncertainty in intelligent landslide detection[J]. Landslides, 2024, 21(8): 1849-1864.
[3]	LÜ Jichao, ZHANG Rui, SHAMA A, et al. Exploring the spatial patterns of landslide susceptibility assessment using interpretable Shapley method: mechanisms of landslide formation in the Sichuan-Tibet region[J]. Journal of Environmental Management, 2024, 366: 121921.
[4]	CHEN Li, MA Peifeng, YU Chang, et al. Landslide susceptibility assessment in multiple urban slope settings with a landslide inventory augmented by InSAR techniques[J]. Engineering Geology, 2023, 327: 107342.
[5]	LIU Qiang, TANG Aiping, HUANG Delong. Exploring the uncertainty of landslide susceptibility assessment caused by the number of non-landslides[J]. CATENA, 2023, 227: 107109.
[6]	HUANG Wubiao, DING Mingtao, LI Zhenhong, et al. Landslide susceptibility mapping and dynamic response along the Sichuan-Tibet transportation corridor using deep learning algorithms[J]. CATENA, 2023, 222: 106866.
[7]	FANG Zhice, WANG Yi, PENG Ling, et al. Integration of convolutional neural network and conventional machine learning classifiers for landslide susceptibility mapping[J]. Computers & Geosciences, 2020, 139: 104470.
[8]	VAN BEEK L P H, VAN ASCH T W J. Regional assessment of the effects of land-use change on landslide hazard by means of physically based modelling[J]. Natural Hazards, 2004, 31(1): 289-304.
[9]	姬建, 崔红志, 佟斌, 等. 基于物理过程不确定性的降雨诱发浅层滑坡易发性快速区划：GIS-FORM技术开发与应用[J]. 岩石力学与工程学报, 2024, 43: 838-850.
	JI Jian, CUI Hongzhi, TONG Bin, et al. Fast zoning of rainfall-induced shallow landslide susceptibility based on physical process uncertainty: development and application of GIS-FORM[J]. Chinese Journal of Rock Mechanics and Engineering, 2024, 43: 838-850.
[10]	KAYASTHA P, DHITAL M R, DE SMEDT F. Application of the analytical hierarchy process (AHP) for landslide susceptibility mapping: a case study from the Tinau watershed, west Nepal[J]. Computers & Geosciences, 2013, 52: 398-408.
[11]	HONG Haoyuan, ILIA I, TSANGARATOS P, et al. A hybrid fuzzy weight of evidence method in landslide susceptibility analysis on the Wuyuan area, China[J]. Geomorphology, 2017, 290: 1-16.
[12]	LEE S, PRADHAN B. Landslide hazard mapping at Selangor, Malaysia using frequency ratio and logistic regression models[J]. Landslides, 2007, 4(1): 33-41.
[13]	黄发明, 欧阳慰平, 蒋水华, 等. 考虑机器学习建模中训练/测试集时空划分原则的滑坡易发性预测建模[J]. 地球科学, 2024, 49(5): 1607-1618.
	HUANG Faming, OUYANG Weiping, JIANG Shuihua, et al. Landslide susceptibility prediction considering spatio-temporal division principle of training/testing datasets in machine learning models[J]. Earth Science, 2024, 49(5): 1607-1618.
[14]	李正, 冷亮, 孙永鑫, 等. 基于信息量-机器学习耦合模型的水电梯级开发流域滑坡易发性评价[J]. 测绘通报, 2024(): 237-241.
	LI Zheng, LENG Liang, SUN Yongxin, et al. Landslide susceptibility assessment in the river cascade development basin based on the IV-LM coupling model[J]. Bulletin of Surveying and Mapping, 2024(): 237-241.
[15]	YOUSSEF A M, POURGHASEMI H R. Landslide susceptibility mapping using machine learning algorithms and comparison of their performance at Abha basin, Asir region, Saudi Arabia[J]. Geoscience Frontiers, 2021, 12(2): 639-655.
[16]	AYALEW L, YAMAGISHI H. The application of GIS-based logistic regression for landslide susceptibility mapping in the Kakuda-Yahiko mountains, central Japan[J]. Geomorphology, 2005, 65(1/2): 15-31.
[17]	SUN Deliang, WEN Haijia, WANG Danzhou, et al. A random forest model of landslide susceptibility mapping based on hyperparameter optimization using Bayes algorithm[J]. Geomorphology, 2020, 362: 107201.
[18]	SUN Deliang, XU Jiahui, WEN Haijia, et al. Assessment of landslide susceptibility mapping based on Bayesian hyperparameter optimization: a comparison between logistic regression and random forest[J]. Engineering Geology, 2021, 281: 105972.
[19]	BADOLA S, MISHRA V N, PARKASH S. Landslide susceptibility mapping using XGBoost machine learning method[C]//Proceedings of 2023 International Conference on Machine Intelligence for GeoAnalytics and Remote Sensing. Hyderabad: IEEE, 2023: 1-4.
[20]	KAVZOGLU T, TEKE A. Advanced hyperparameter optimization for improved spatial prediction of shallow landslides using extreme gradient boosting (XGBoost)[J]. Bulletin of Engineering Geology and the Environment, 2022, 81(5): 201.
[21]	MARJANOVIĆ M, KOVAČEVIĆ M, BAJAT B, et al. Landslide susceptibility assessment using SVM machine learning algorithm[J]. Engineering Geology, 2011, 123(3): 225-234.
[22]	HUANG Yu, ZHAO Lu. Review on landslide susceptibility mapping using support vector machines[J]. CATENA, 2018, 165: 520-529.
[23]	LEE S, RYU J H, WON J S, et al. Determination and application of the weights for landslide susceptibility mapping using an artificial neural network[J]. Engineering Geology, 2004, 71(3/4): 289-302.
[24]	DI NAPOLI M, CAROTENUTO F, CEVASCO A, et al. Machine learning ensemble modelling as a tool to improve landslide susceptibility mapping reliability[J]. Landslides, 2020, 17(8): 1897-1914.
[25]	HUANG Faming, ZHANG Jing, ZHOU Chuangbing, et al. A deep learning algorithm using a fully connected sparse autoencoder neural network for landslide susceptibility prediction[J]. Landslides, 2020, 17(1): 217-229.
[26]	YEON Y K, HAN J G, RYU K H. Landslide susceptibility mapping in Injae, Korea, using a decision tree[J]. Engineering Geology, 2010, 116(3/4): 274-283.
[27]	PHAM B T, NGUYEN-THOI T, QI Chongchong, et al. Coupling RBF neural network with ensemble learning techniques for landslide susceptibility mapping[J]. CATENA, 2020, 195: 104805.
[28]	YI Yaning, ZHANG Zhijie, ZHANG Wanchang, et al. Landslide susceptibility mapping using multiscale sampling strategy and convolutional neural network: a case study in Jiuzhaigou region[J]. CATENA, 2020, 195: 104851.
[29]	YANG Xin, LIU Rui, YANG Mei, et al. Incorporating landslide spatial information and correlated features among conditioning factors for landslide susceptibility mapping[J]. Remote Sensing, 2021, 13(11): 2166.
[30]	HAKIM W L, REZAIE F, NUR A S, et al. Convolutional neural network (CNN) with metaheuristic optimization algorithms for landslide susceptibility mapping in Icheon, South Korea[J]. Journal of Environmental Management, 2022, 305: 114367.
[31]	WANG Yi, FANG Zhice, HONG Haoyuan. Comparison of convolutional neural networks for landslide susceptibility mapping in Yanshan county, China[J]. Science of the Total Environment, 2019, 666: 975-993.
[32]	谭林, 张璐璐, 魏鑫, 等. 基于U-Net语义分割网络的区域滑坡易发性评价方法和跨地区泛化能力研究[J/OL]. 土木工程学报: 1-14[2023-12-12]. https://doi.org/10.15951/j.tmgcxb.23110923.
	TAN Lin, ZHANG Lulu, WEI Xin, et al. Study on regional landslide susceptibility assessment method based on U-Net semantic segmentation network and its cross-generalization ability[J/OL]. China Civil Engineering Journal: 1-14[2023-12-12]. https://doi.org/10.15951/j.tmgcxb.23110923.
[33]	CHEN Yangyang, MING Dongping, LING Xiao, et al. Landslide susceptibility mapping using feature fusion-based CPCNN-ML in Lantau island, Hong Kong[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2021, 14: 3625-3639.
[34]	ASLAM B, ZAFAR A, KHALIL U. Comparative analysis of multiple conventional neural networks for landslide susceptibility mapping[J]. Natural Hazards, 2023, 115(1): 673-707.
[35]	GE Yunfeng, LIU Geng, TANG Huiming, et al. Comparative analysis of five convolutional neural networks for landslide susceptibility assessment[J]. Bulletin of Engineering Geology and the Environment, 2023, 82(10): 377.
[36]	金必晶, 曾韬睿, 桂蕾, 等. 考虑未来土地利用动态情景的滑坡易发性制图[J]. 地球信息科学学报, 2024, 26(6): 1486-1499.
	JIN Bijing, ZENG Taorui, GUI Lei, et al. Mapping the landslide susceptibility considering future land use dynamics scenario[J]. Journal of Geo-Information Science, 2024, 26(6): 1486-1499.
[37]	GAO Binghai, HE Yi, CHEN Xueye, et al. A deep neural network framework for landslide susceptibility mapping by considering time-series rainfall[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024, 17: 5946-5969.
[38]	林炫歆, 肖桂荣, 周侯伯. 顾及土地利用动态变化的滑坡易发性评估方法[J]. 地球信息科学学报, 2023, 25(5): 953-966.
	LIN Xuanxin, XIAO Guirong, ZHOU Houbo. Landslide susceptibility assessment method considering land use dynamic change[J]. Journal of Geo-information Science, 2023, 25(5): 953-966.
[39]	HE Yi, ZHAO Zhanao, ZHU Qing, et al. An integrated neural network method for landslide susceptibility assessment based on time-series InSAR deformation dynamic features[J]. International Journal of Digital Earth, 2024, 17(1): 2295408.
[40]	高秉海, 何毅, 张立峰, 等. 顾及InSAR形变的CNN滑坡易发性动态评估：以刘家峡水库区域为例[J]. 岩石力学与工程学报, 2023, 42(2): 450-465.
	GAO Binghai, HE Yi, ZHANG Lifeng, et al. Dynamic evaluation of landslide susceptibility by CNN considering InSAR deformation: a case study of Liujiaxia reservoir[J]. Chinese Journal of Rock Mechanics and Engineering, 2023, 42(2): 450-465.
[41]	王启盛, 熊俊楠, 程维明, 等. 耦合统计方法、机器学习模型和聚类算法的滑坡易发性评价方法[J]. 地球信息科学学报, 2024, 26(3): 620-637.
	WANG Qisheng, XIONG Junnan, CHENG Weiming, et al. Landslide susceptibility mapping methods coupling with statistical methods, machine learning models and clustering algorithms[J]. Journal of Geo-information Science, 2024, 26(3): 620-637.
[42]	赵占骜, 王继周, 毛曦, 等. 多维CNN耦合的滑坡易发性评价方法[J]. 武汉大学学报(信息科学版), 2024, 49(8): 1466-1481.
	ZHAO Zhan'ao, WANG Jizhou, MAO Xi, et al. A multi-dimensional CNN coupled landslide susceptibility assessment method[J]. Geomatics and Information Science of Wuhan University, 2024, 49(8): 1466-1481.
[43]	ZHAO Zeyang, CHEN Tao, DOU Jie, et al. Landslide susceptibility mapping considering landslide local-global features based on CNN and Transformer[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024, 17: 7475-7489.
[44]	ALQADHI S, MALLICK J, ALKAHTANI M. Integrated deep learning with explainable artificial intelligence for enhanced landslide management[J]. Natural Hazards, 2024, 120(2): 1343-1365.
[45]	YANG Qiyuan, WANG Xianmin, YIN Jing, et al. A novel CGBoost deep learning algorithm for coseismic landslide susceptibility prediction[J]. Geoscience Frontiers, 2024, 15(2): 101770.
[46]	郑德凤, 高敏, 闫成林, 等. 基于卷积神经网络的滑坡易发性评价：以辽南仙人洞国家级自然保护区为例[J]. 地球科学, 2024, 49(5): 1654-1664.
	ZHENG Defeng, GAO Min, YAN Chenglin, et al. Susceptibility assessment of landslides based on convolutional neural network model: a case study from Xianrendong national nature reserve in southern Liaoning province[J]. Earth Science, 2024, 49(5): 1654-1664.
[47]	FANG Bo, CHEN Gang, PAN Li, et al. GAN-based Siamese framework for landslide inventory mapping using bi-temporal optical remote sensing images[J]. IEEE Geoscience and Remote Sensing Letters, 2021, 18(3): 391-395.
[48]	HE Kaiming, ZHANG Xiangyu, REN Shaoqing, et al. Spatial pyramid pooling in deep convolutional networks for visual recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(9): 1904-1916.
[49]	THI NGO P T, PANAHI M, KHOSRAVI K, et al. Evaluation of deep learning algorithms for national scale landslide susceptibility mapping of Iran[J]. Geoscience Frontiers, 2021, 12(2): 505-519.
[50]	GLOROT X, BENGIO Y. Understanding the difficulty of training deep feedforward neural networks[EB/OL]. [2023-12-12]. http://proceedings.mlr.press/v9/glorot10a/glorot10a.pdf.
[51]	LOSHCHILOV I, HUTTER F. SGDR: stochastic gradient descent with warm restarts[EB/OL]. [2023-12-12]. https://doi.org/10.48550/arXiv.1608.03983.
[52]	崔鹏, 邹强. 川藏交通廊道山地灾害演化规律与工程风险[M]. 北京: 科学出版社, 2021.
	CUI Peng, ZOU Qiang. Evolution law and engineering risk of mountain disasters in Sichuan-Tibet traffic corridor[M]. Beijing: Science Press, 2021.
[53]	甘孜州水利局. 甘孜州2019年水资源公报[R/OL]. [2023-12-12]. https://slj.gzz.gov.cn/tzgs/article/533037.
	Ganzi State Water Resources Bureau. Ganzi state water resources bulletin 2019[R/OL]. [2023-12-12]. https://slj.gzz.gov.cn/tzgs/article/533037.

数据	来源	数据	来源
DEM	https://gdex.cr.usgs.	地表粗糙度	30 m SRTM DEM
地层岩性	http://dcc.ngac.org.cn	地形起伏度	30 m SRTM DEM
断层	https://geocloud.cgs	地形湿度指数	30 m SRTM DEM
地面峰值加速度	http://gmo.gfz-potsdam.de/	曲率	30 m SRTM DEM
河流	https://www.openstreetmap.org/	平面曲率	30 m SRTM DEM
道路	https://www.openstreetmap.org/	剖面曲率	30 m SRTM DEM
降雨	https://gpm.nasa.gov	坡向	30 m SRTM DEM
NDVI	GEE/Landsat-8	坡度	30 m SRTM DEM

模型	准确度	精确度	召回率	F₁值
随机森林	0.77	0.78	0.78	0.78
浅层CNN	0.70	0.66	0.87	0.75
ResNet101	0.74	0.72	0.77	0.74
多分支网络模型	0.88	0.89	0.92	0.90

模型	易发性等级	分级栅格数	百分比/（%）	滑坡点数	百分比/（%）	频率比
	极低易发区	103 224 760	33.93	29	1.70	0.05
	低易发区	89 398 692	29.38	135	7.93	0.27
随机森林	中等易发区	56 627 969	18.61	276	16.21	0.87
	高易发区	41 894 540	13.77	650	38.19	2.77
	极高易发区	13 109 914	4.31	612	35.96	8.34
	极低易发区	23 860 212	7.84	14	0.82	0.10
	低易发区	150 562 987	49.48	320	18.80	0.38
浅层CNN	中等易发区	58 941 341	19.37	312	18.33	0.95
	高易发区	66 053 031	21.70	915	53.76	2.48
	极高易发区	4 838 304	4.30	141	8.28	5.21
	极低易发区	77 173 862	25.36	112	6.58	0.26
	低易发区	69 152 928	22.73	158	9.28	0.41
ResNet101	中等易发区	65 258 032	21.45	347	20.39	0.96
	高易发区	59 334 304	19.50	598	35.14	1.80
	极高易发区	33 336 749	10.96	487	28.61	2.61
	极低易发区	151 964 854	49.94	26	1.52	0.03
	低易发区	47 612 221	15.64	42	2.46	0.16
多分支网络模型	中等易发区	58 740 480	19.30	126	7.40	0.38
	高易发区	33 216 672	10.91	386	22.67	2.08
	极高易发区	12 721 648	4.18	1122	65.92	15.77

模型	准确度	精确度	召回率	F₁值
去除自适应定权机制	0.84	0.81	0.90	0.85
去除伪孪生网络结构	0.82	0.79	0.88	0.83

模型	易发性等级	分级栅格数	百分比/（%）	滑坡点数	百分比/（%）	频率比
去除自适应定权机制	极低易发区	78 971 539	25.96	29	1.70	0.07
	低易发区	61 998 992	20.38	70	4.11	0.20
	中等易发区	50 522 720	16.61	143	8.40	0.51
	高易发区	46 883 968	15.41	372	21.86	1.42
	极高易发区	65 878 656	21.65	1088	63.92	2.95
去除伪孪生网络结构	极低易发区	89 695 094	29.48	135	7.93	0.27
	低易发区	70 354 349	23.12	200	11.75	0.51
	中等易发区	62 070 096	20.40	318	18.68	0.92
	高易发区	43 871 088	14.42	419	24.62	1.71
	极高易发区	38 265 248	12.58	630	37.02	2.94