Influence of Incomplete Landslide Data on Susceptibility Modeling and Suggestions for Improvement

doi:10.11871/jfdc.issn.2096-742X.2025.01.014

Abstract

Abstract:

[Objective] China is one of the countries with the most frequent landslide disasters in the world. It is necessary to establish a reliable landslide susceptibility model suitable for the whole country to determine the areas with high landslide hazards, formulate appropriate disaster prevention and reduction strategies, and reduce the loss of people's lives and property. [Methods] Given the difficulty in obtaining completely unbiased landslide data in such a large area of China, this study selected 10 influencing factors such as slope, aspect, profile curvature, plan curvature, road density, river density, soil moisture, lithology, land use, and geological environment division as the driving data and designed Model Scheme 1 (Based on LightGBM and ignoring the effects of incomplete landslide data), Model Scheme 2 (Based on LightGBM and excluding factors associated with landslide incompleteness) and Model Scheme 3 (Based on TBMM and including the variables describing landslide incompleteness, i.e. land use and geological environment division, as random effect terms) to assess landslide susceptibility separately to explore the impact of incomplete landslide data on the modelling of landslide susceptibility in China, the impact of incomplete landslide data on the modelling of landslide susceptibility in China, and the measure to counteract the effect of such bias. [Results] The results show that, in the context of large regional susceptibility modeling, (1) although the model schemes that simply ignore or exclude the factors associated with existing landslide data deficiencies have higher statistical performance, they will lead to geomorphically incoherent landslide susceptibility prediction results; (2) the mixed effects model can effectively reduce the bias impact caused by incomplete landslide data. [Conclusions] This study provides a new idea for landslide susceptibility mapping under the background of incomplete landslide data and contributes to assessing China’s overall mass movement susceptibility situation and assisting policymakers in master planning for risk mitigation.

Key words: landslide, susceptibility map, inventory bias, LightGBM, TBMM

HE Yifei, ZHANG Yaonan. Influence of Incomplete Landslide Data on Susceptibility Modeling and Suggestions for Improvement[J]. Frontiers of Data and Computing, 2025, 7(1): 186-202, https://cstr.cn/32002.14.jfdc.CN10-1649/TP.2025.01.014.

Figures/Tables 15

Fig.1

Table 1

Fig.2

Fig.3

Table 2

Fig.4

Table 3

Table 4

Fig.5

Fig.6

Table 5

Fig.7

Fig.8

Fig.9

Fig.10

References 65

[1]	LIN Q, WANG Y. Spatial and temporal analysis of a fatal landslide inventory in China from 1950 to 2016[J]. Landslides, 2018, 15(12): 2357-2372.
[2]	陶伟, 孙岳. 基于GIS滑坡地质灾害易发性评价方法综述[J]. 世界有色金属, 2020, 561(21): 157-159.
[3]	POURGHASEMI H R, TEIMOORI YANSARI Z, PANAGOS P, et al. Analysis and evaluation of landslide susceptibility: a review on articles published during 2005—2016 (periods of 2005—2012 and 2013—2016)[J]. Arabian Journal of Geosciences, 2018, 11(9): 1-12.
[4]	WANG D, HAO M, CHEN S, et al. Assessment of landslide susceptibility and risk factors in China[J]. Natural hazards, 2021, 108(3): 3045-3059.
[5]	GAPRINDASHVILI G, VAN WESTEN C J. Generation of a national landslide hazard and risk map for the country of Georgia[J]. Natural hazards, 2016, 80(1): 69-101.
[6]	SAROGLOU C. GIS-based rockfall susceptibility zoning in Greece[J]. Geosciences, 2019, 9(4): 163.
[7]	TITTI G, BORGATTI L, ZOU Q, et al. Landslide susceptibility in the Belt and Road Countries: continental step of a multi-scale approach[J]. Environmental Earth Sciences, 2021, 80(18): 1-18.
[8]	BROECKX J, VANMAERCKE M, DUCHATEAU R, et al. A data-based landslide susceptibility map of Africa[J]. Earth-Science Reviews, 2018, 185: 102-121.
[9]	GüNTHER A, REICHENBACH P, MALET J P, et al. Tier-based approaches for landslide susceptibility assessment in Europe[J]. Landslides, 2013, 10(5): 529-546.
[10]	LIN L, LIN Q, WANG Y. Landslide susceptibility mapping on a global scale using the method of logistic regression[J]. Natural Hazards and Earth System Sciences, 2017, 17(8): 1411-1424.
[11]	HONG Y, ADLER R, HUFFMAN G. Use of satellite remote sensing data in the mapping of global landslide susceptibility[J]. Natural hazards, 2007, 43(2): 245-256.
[12]	COROMINAS J, VAN WESTEN C, FRATTINI P, et al. Recommendations for the quantitative analysis of landslide risk[J]. Bulletin of engineering geology and the environment, 2014, 73(2): 209-263.
[13]	周宇, 常鸣, 孙文静, 等. 基于改进证据权重法的北海道地震同震滑坡易发性评价[J]. 地理与地理信息科学, 2022, 38(1): 138-144.
[14]	李郎平, 兰恒星, 郭长宝, 等. 基于改进频率比法的川藏铁路沿线及邻区地质灾害易发性分区评价[J]. 现代地质, 2017, 31(5): 911-929.
[15]	罗金. 基于各类机器学习方法的滑坡易发性评价及软件系统开发[D]. 西安: 长安大学, 2021.
[16]	侯鹏鹂. 岷江上游区域泥石流发育规律与易发性评价[D]. 绵阳: 西南科技大学, 2022.
[17]	GOETZ J, BRENNING A, PETSCHKO H, et al. Evaluating machine learning and statistical prediction techniques for landslide susceptibility modeling[J]. Computers & geosciences, 2015, 81: 1-11.
[18]	陈芯宇, 师芸, 温永啸, 等. 基于确定性系数与支持向量机的滑坡易发性评价[J]. 科学技术与工程, 2023, 23(2): 518-527.
[19]	艾骁. 基于机器学习的地震滑坡易发性评估模型构建[D]. 哈尔滨: 中国地震局工程力学研究所, 2021.
[20]	万洋, 郭捷, 马凤山, 等. 基于最大熵模型的中尼交通廊道滑坡易发性分析[J]. 中国地质灾害与防治学报, 2022, 33(2): 88-95.
[21]	邵健. 基于机器学习的滑坡敏感性评价研究[D]. 贵阳: 贵州大学, 2022.
[22]	张越, 宋炜炜. 基于BP神经网络和决策树的昆明市东川区滑坡空间易发性评价[J]. 国土与自然资源研究, 2023, 203(2): 67-70.
[23]	段宇英, 汤军, 刘远刚, 等. 基于随机森林的山西省柳林县黄土滑坡空间敏感性评价[J]. 地理科学, 2022, 42(2): 343-351. doi: 10.13249/j.cnki.sgs.2022.02.016
[24]	HARP E L, KEEFER D K, SATO H P, et al. Landslide inventories: the essential part of seismic landslide hazard analyses[J]. Engineering Geology, 2011, 122(1-2): 9-21.
[25]	STEGER S, MAIR V, KOFLER C, et al. Correlation does not imply geomorphic causation in data-driven landslide susceptibility modelling-Benefits of exploring landslide data collection effects[J]. Science of the total environment, 2021, 776: 145935.
[26]	KIRSCHBAUM D, STANLEY T, ZHOU Y. Spatial and temporal analysis of a global landslide catalog[J]. Geomorphology, 2015, 249: 4-15.
[27]	STEGER S, SCHMALTZ E, GLADE T. The (f) utility to account for pre-failure topography in data-driven landslide susceptibility modelling[J]. Geomorphology, 2020, 354: 107041.
[28]	STEGER S, BRENNING A, BELL R, et al. Exploring discrepancies between quantitative validation results and the geomorphic plausibility of statistical landslide susceptibility maps[J]. Geomorphology, 2016, 262: 8-23.
[29]	LIMA P, STEGER S, GLADE T. Counteracting flawed landslide data in statistically based landslide susceptibility modelling for very large areas: a national-scale assessment for Austria[J]. Landslides, 2021, 18(11): 3531-3546.
[30]	STEGER S, MAIR V, KOFLER C, et al. The necessity to consider the landslide data origin in statistically-based spatial predictive modelling-A landslide intervention index for South Tyrol (Italy)[C]. EGU General Assembly Conference Abstracts, 2020: 3440.
[31]	STEGER S, BRENNING A, BELL R, et al. The influence of systematically incomplete shallow landslide inventories on statistical susceptibility models and suggestions for improvements[J]. Landslides, 2017, 14: 1767-1781.
[32]	DINGEMANSE N J, DOCHTERMANN N A. Quantifying individual variation in behaviour: mixed-effect modelling approaches[J]. Journal of Animal Ecology, 2013, 82(1): 39-54. doi: 10.1111/1365-2656.12013 pmid: 23171297
[33]	SIGRIST F. Gaussian process boosting[J]. arXiv preprint arXiv:200402653, 2020.
[34]	杨迁, 王雁林, 马园园. 2001—2019年中国地质灾害分布规律及引发因素分析[J]. 地质灾害与环境保护, 2020, 31(4): 43-48.
[35]	REICHENBACH P, ROSSI M, MALAMUD B, et al. A review of statistically-based landslide susceptibility models[J]. Earth-Science Reviews, 2018: S001282 5217305652.
[36]	PELLICANI R, VAN WESTEN C J, SPILOTRO G. Assessing landslide exposure in areas with limited landslide information[J]. Landslides, 2014, 11(3): 463-480.
[37]	LIN Q, LIMA P, STEGER S, et al. National-scale data-driven rainfall induced landslide susceptibility mapping for China by accounting for incomplete landslide data[J]. Geoscience Frontiers, 2021, 12(6): 101248.
[38]	YANG Y, YANG J, XU C, et al. Local-scale landslide susceptibility mapping using the B-GeoSVC model[J]. Landslides, 2019, 16(7): 1301-1312. doi: 10.1007/s10346-019-01174-y
[39]	LUO W, LIU C C. Innovative landslide susceptibility mapping supported by geomorphon and geographical detector methods[J]. Landslides, 2018, 15(3): 465-474.
[40]	汪潮. 滇西地区花岗岩风化层滑坡的形成条件及防治对策研究[D]. 昆明: 昆明理工大学, 2016.
[41]	POURGHASEMI H R, ROSSI M. Landslide susceptibility modeling in a landslide prone area in Mazandarn Province, north of Iran: a comparison between GLM, GAM, MARS, and M-AHP methods[J]. Theoretical and Applied Climatology, 2017, 130(1): 609-633.
[42]	周玉龙. 融入时序InSAR形变的毛尔盖库区滑坡易发性评价研究[D]. 成都: 成都理工大学, 2020.
[43]	FARR T G, KOBRICK M. Shuttle Radar Topography Mission produces a wealth of data[J]. Eos, Transactions American Geophysical Union, 2000, 81(48): 583-585.
[44]	周鑫. 河南省渑池县地质灾害发育特征及形成条件研究[J]. 华北自然资源, 2020, 96(3): 107-109.
[45]	白志刚, 刘启蒙, 刘瑜. 基于熵指数与随机森林模型的滑坡易发性评价[J]. 人民长江, 2022, 53(10): 95-102.
[46]	蒋万钰. 基于贝叶斯概率的降雨数据稀缺山区滑坡预警模型研究[D]. 兰州: 兰州大学, 2022.
[47]	TRABUCCO A, ZOMER R. Global High-Resolution Soil-Water Balance[J]. Figshare Fileset, 2019, 10: m9.
[48]	HARTMANN J, MOOSDORF N. The new global lithological map database GLiM: A representation of rock properties at the Earth surface[J]. Geochemistry, Geophysics, Geosystems, 2012, 13(12): 1-37.
[49]	DU J, GLADE T, WOLDAI T, et al. Landslide susceptibility assessment based on an incomplete landslide inventory in the Jilong Valley, Tibet, Chinese Himalayas[J]. Engineering Geology, 2020, 270: 105572.
[50]	SUN D, SHI S, WEN H, et al. A hybrid optimization method of factor screening predicated on GeoDetector and Random Forest for Landslide Susceptibility Mapping[J]. Geomorphology, 2021, 379: 107623.
[51]	KE G, MENG Q, FINLEY T, et al. Lightgbm: A highly efficient gradient boosting decision tree[J]. Advances in neural information processing systems, 2017: 3149-3157.
[52]	赵泽园, 罗菲. 基于LightGBM模型的区域滑坡危险性评价研究[J]. 内蒙古煤炭经济, 2020, 298(5): 48-49.
[53]	SZKLO M, NIETO F J. Epidemiology: beyond the basics[M]. American Journal of Epidemiology, 2001, 153(8): 821-822.
[54]	SIGRIST F. Latent Gaussian model boosting[J]. arXiv preprint arXiv:210508966, 2021.
[55]	BRENNING A. Spatial cross-validation and bootstrap for the assessment of prediction rules in remote sensing: The R package sperrorest[C]. proceedings of the 2012 IEEE international geoscience and remote sensing symposium, 2012: 5362-5375.
[56]	BUITINCK L, LOUPPE G, BLONDEL M, et al. API design for machine learning software: experiences from the scikit-learn project[J]. arXiv preprint arXiv:13090238, 2013.
[57]	王伟, 袁雯宇, 邹丽芳, 等. 基于滑坡敏感性评价的库区水动力型滑坡区域综合预警研究[J]. 岩石力学与工程学报, 2022, 41(3): 479-491.
[58]	MANDAL K, SAHA S, MANDAL S. Applying deep learning and benchmark machine learning algorithms for landslide susceptibility modelling in Rorachu river basin of Sikkim Himalaya, India[J]. Geoscience Frontiers, 2021, 12(5): 101203.
[59]	KORUP O, STOLLE A. Landslide prediction from machine learning[J]. Geology today, 2014, 30(1): 26-33.
[60]	CHEN W, CHAI H, SUN X, et al. A GIS-based comparative study of frequency ratio, statistical index and weights-of-evidence models in landslide susceptibility mapping[J]. Arabian Journal of Geosciences, 2016, 9(3): 1-16.
[61]	PETSCHKO H, BELL R, GLADE T. Effectiveness of visually analyzing LiDAR DTM derivatives for earth and debris slide inventory mapping for statistical susceptibility modeling[J]. Landslides, 2016, 13(5): 857-872.
[62]	PETSCHKO H, BRENNING A, BELL R, et al. Assessing the quality of landslide susceptibility maps-case study Lower Austria[J]. Natural Hazards and Earth System Sciences, 2014, 14(1): 95-118.
[63]	STEGER S, BRENNING A, BELL R, et al. The propagation of inventory-based positional errors into statistical landslide susceptibility models[J]. Natural Hazards and Earth System Sciences, 2016, 16(12): 2729-2745.
[64]	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.
[65]	LIU L, LI S, LI X, et al. An integrated approach for landslide susceptibility mapping by considering spatial correlation and fractal distribution of clustered landslide data[J]. Landslides, 2019, 16(4): 715-728.

数据名称	原始数据类型	数据来源
滑坡数据	点	中国地质调查局
坡度	栅格（90 m）	提取自DEM（https://srtm.csi.cgiar.org）
坡向	栅格（90 m）	提取自DEM
剖面曲率	栅格（90 m）	提取自DEM
平面曲率	栅格（90 m）	提取自DEM
道路密度	线	https://www.tianditu.gov.cn
河流密度	线	https://www.tianditu.gov.cn
土壤湿度	栅格（1 km）	https://cgiarcsi.community
岩性	面	https://www.uni-hamburg.de
土地利用	栅格（1 km）	https://www.resdc.cn
地质环境分区	面	https://geocloud.cgs.gov.cn

变量	模型方案1 LightGBM	模型方案2 LightGBM	模型方案3 TBMM
8个基本预测因子	训练和预测	训练和预测	训练和预测
土地利用	训练和预测	不使用	只用于训练
地质环境分区	训练和预测	不使用	只用于训练

参数	方案1	方案2	方案3
learning_rate	0.05	0.05	0.8
max_depth	30	10	5
min_data_in_leaf	80	80	30
num_boost_round	400	400	300

预测因子	容差	是方差膨胀系数
坡度	0.4179	2.3926
坡向	0.3188	3.1363
剖面曲率	0.8342	1.1988
平面曲率	0.8672	1.1532
道路密度	0.8087	1.2351
河流密度	0.7535	1.3272
土壤湿度	0.2291	4.3645
岩性	0.4291	2.3306

模型方案	CV AUROC中位数（Q1-Q3）	SCV AUROC中位数（Q1-Q3）	差值（CV-SCV）
模型1	0.918 (0.918-0.920)	0.868 (0.737-0.886)	0.050
模型2	0.885 (0.884-0.887)	0.793 (0.665-0.863)	0.092
模型3	0.841 (0.837-0.844)	0.823 (0.678-0.867)	0.018