Please wait a minute...
Journal of ZheJiang University (Engineering Science)  2021, Vol. 55 Issue (3): 472-482    DOI: 10.3785/j.issn.1008-973X.2021.03.007
    
Landslides hazard warning based on decision tree and effective rainfall intensity
Fa-ming HUANG(),Zhong-shan CAO,Chi YAO(),Qing-hui JIANG,Jia-wu CHEN
1. School of Civil Engineering and Architecture, Nanchang University, Nanchang 330031
2. School of Civil Engineering, Wuhan University, Wuhan, 430000
Download: HTML     PDF(1526KB) HTML
Export: BibTeX | EndNote (RIS)      

Abstract  

Information value (IV), back-propagation neural network (BPNN) and C5.0 decision tree models were used to implement landslide susceptibility prediction (LSP) for comparisons by taking the Xunwu County of Jiangxi Province as a case, EI-D model was proposed to calculate the critical rainfall thresholds of these landslides based on the concept of early effective rainfall. The results of EI-D were compared with the results of conventional I-D model for uncertainty analysis. LSP results were coupled with the EI-D model to realize the landslides hazard warning with the warning accuracy further verified. Results show that the C5.0 decision tree has higher LSP accuracy than BPNN, followed by IV model. The temporal probability prediction accuracy of EI-D model is superior to the I-D model. The present model based on LSP and EI-D can effectively achieve real-time rainfall-induced landslides early warning.



Key wordsrainfall-induced landslide      landslide hazard warning      landslide susceptibility prediction (LSP)      effective rainfall intensity      C5.0 decision tree     
Received: 18 January 2020      Published: 25 April 2021
CLC:  P 694  
Fund:  国家自然科学基金资助项目(41807285,41762020,51879127,51769014);中国博士后基金资助项目(2019M652287);江西省自然科学基金资助项目(20192BAB216034,20192ACB2102,20192ACB20020);江西省博士后基金资助项目(2019KY08)
Corresponding Authors: Chi YAO     E-mail: faminghuang@ncu.edu.cn;chi.yao@ncu.edu.cn
Cite this article:

Fa-ming HUANG,Zhong-shan CAO,Chi YAO,Qing-hui JIANG,Jia-wu CHEN. Landslides hazard warning based on decision tree and effective rainfall intensity. Journal of ZheJiang University (Engineering Science), 2021, 55(3): 472-482.

URL:

http://www.zjujournals.com/eng/10.3785/j.issn.1008-973X.2021.03.007     OR     http://www.zjujournals.com/eng/Y2021/V55/I3/472


基于决策树和有效降雨强度的滑坡危险性预警

以江西省寻乌县为例,采用信息量(IV)、反向传播神经网络(BPNN)和C5.0决策树模型进行滑坡易发性预测(LSP),比较不同模型的预测性能;基于有效降雨量的概念提出有效降雨强度-历时(EI-D)模型,计算滑坡临界降雨阈值并将其与传统的降雨强度-历时(I-D)阈值做对比;将LSP结果与EI-D模型耦合,实现滑坡灾害预警并进一步验证了预警精度. 结果表明:C5.0决策树的LSP精度高于BPNN和IV,EI-D阈值的预测效果优于I-D模型,且基于滑坡易发性和EI-D阈值的模型能有效实现降雨型滑坡的实时预报.


关键词: 降雨型滑坡,  滑坡危险性预警,  滑坡易发性预测(LSP),  有效降雨强度,  C5.0决策树 
T1(叠加S1~S5 T2(叠加S1~S5 T3(叠加S1~S5 T4(叠加S1~S5 T5(叠加S1~S5
不预警区(1级) 不预警区(1级) 不预警区(1级) 不预警区(1级) 不预警区(2级)
不预警区(1级) 不预警区(1级) 不预警区(2级) 2级预警区 3级预警区
不预警区(1级) 不预警区(2级) 3级预警区 3级预警区 4级预警区
不预警区(1级) 不预警区(2级) 3级预警区 4级预警区 5级预警区
不预警区(2级) 3级预警区 4级预警区 5级预警区 5级预警区
Tab.1 Hazards warning levels based on susceptibility levels and time probability of rainfall-induced landslides
预警级别 当日滑坡发生情况 相应措施
Ⅰ级或不预警 可能性极小,无危害 无需采取措施
Ⅱ级或不预警 可能性较小,基本无危害 对重要滑坡隐患点定时监测
Ⅲ级预警(注意) 可能性中等,规模和危害中等 注意监测滑坡,采取防御措施,提醒灾区人员关注灾害动态
Ⅳ级预警(预警) 可能性较大,规模或危害较大 加强对灾害点的监测,对滑坡危险区应开展预防应急措施
Ⅴ级预警(警报) 可能性很大,规模或危害严重 全天候监测滑坡,建立防御措施、救灾体系和紧急疏散通道等
Tab.2 Significance of warning level and corresponding measures
Fig.1 Xunwu county location and landslide location
Fig.2 Landslide-related environmental factors(aspect and profile curvature are not present)
环境因子 属性区间 因子类型 全区格栅 滑坡格栅 频率比
数量 占比/% 数量 占比/%
坡度/(°) 0~4.8 连续型 399 507 15.57 125 3.86 0.25
坡度/(°) 4.8~8.6 连续型 526 595 20.52 657 20.31 0.99
坡度/(°) 8.6~12.4 连续型 528 932 20.62 932 28.81 1.40
坡度/(°) 12.4~16.1 连续型 434 885 16.95 721 22.29 1.31
坡度/(°) 16.1~20.1 连续型 322 276 12.56 475 14.68 1.17
坡度/(°) 20.1~24.5 连续型 203 715 7.94 228 7.05 0.89
坡度/(°) 24.5~30.2 连续型 111 214 4.34 81 2.50 0.58
坡度/(°) 30.2~53.4 连续型 38 661 1.51 16 0.49 0.33
距河流距离/m 900~3 000 离散型 884 820 34.49 483 14.93 0.43
距河流距离/m 600~900 离散型 477 428 18.61 362 11.19 0.60
距河流距离/m 300~600 离散型 563 869 21.98 538 16.63 0.76
距河流距离/m 0~300 离散型 639 627 24.93 1 852 57.25 2.30
岩土类型 变质岩类 离散型 592 129 23.08 1 050 32.46 1.41
岩土类型 碳酸岩类 离散型 1 727 958 67.35 2 036 62.94 0.93
岩土类型 碎屑岩类 离散型 245 695 9.58 149 4.61 0.48
NDVI 0~0.33 连续型 10 451 0.41 6 0.19 0.46
NDVI 0.33~0.45 连续型 48 261 1.88 33 1.02 0.54
NDVI 0.45~0.55 连续型 112 791 4.40 201 6.21 1.41
NDVI 0.55~0.63 连续型 287 167 11.19 582 17.99 1.61
NDVI 0.63~0.67 连续型 526 616 20.53 805 24.88 1.21
NDVI 0.67~0.73 连续型 661 306 25.77 876 27.08 1.05
NDVI 0.73~0.80 连续型 568 890 22.17 554 17.13 0.77
NDVI 0.80~1.00 连续型 350 303 13.65 178 5.50 0.40
Tab.3 Frequency ratios of environmental factors
Fig.3 AUC values of all three models
Fig.4 Landslide susceptibility map produced by models
Fig.5 Monthly mean rainfall and corresponding landslide occurrence frequency from 1970 to 2010
$\alpha $ R0 $R_1^\prime $ $R_2^\prime $ $R_3^\prime $ $R_4^\prime $ $R_5^\prime $ $R_6^\prime $ $R_7^\prime $ $R_8^\prime $ $R_9^\prime $
注:**表示在0.01显著性水平下,显著相关.
0.9 0.517** 0.561** 0.604** 0.612** 0.613** 0.607** 0.585** 0.570** 0.560** 0.558**
0.8 0.517** 0.567** 0.612** 0.621** 0.623** 0.622** 0.611** 0.604** 0.602** 0.601**
0.7 0.517** 0.564** 0.612** 0.627** 0.623** 0.620** 0.614** 0.607** 0.605** 0.605**
0.6 0.517** 0.561** 0.607** 0.622** 0.613** 0.608** 0.609** 0.609** 0.609** 0.609**
0.5 0.517** 0.557** 0.598** 0.611** 0.601** 0.602** 0.601** 0.601** 0.601** 0.601**
Tab.4 Correlation analysis of rainfall and landslide in different days and different coefficients
Fig.6 Classification diagram of two threshold models
Fig.7 Accuracy of two threshold models
降雨时间 Rc /mm Re /mm
第一日 57.27 57.27
第二日 92.38 107.42
第三日 118.75 161.23
第四日 140.27 223.97
Tab.5 Macroscopic threshold of landslide warning
滑坡发生日期 R7 /mm R6 /mm R5 /mm R4 /mm R3 /mm R2 /mm R1 /mm R0 /mm Rc /mm EI /(mm·d?1 D /d
1991?06?21 4.0 0.0 0.0 0.0 23.6 69.9 42.8 45.8 118.1 29.5 4
2000?06?20 0.0 0.0 0.0 0.0 19.4 67.5 61.0 18.5 100.9 25.2 4
2000?07?07 1.9 0.0 0.2 0.0 0.0 28.6 46.1 46.9 93.2 31.1 3
2000?08?25 0.3 0.0 0.0 0.0 0.0 10.5 23.2 143.0 164.4 54.8 3
2001?06?05 0.0 18.5 0.0 0.0 0.0 36.0 74.0 41.6 111.0 37.0 3
2002?07?02 0.0 0.0 0.0 0.0 0.0 21.9 41.0 42.0 81.4 27.1 3
2003?05?17 23.5 0.0 0.0 4.9 14.0 9.1 4.2 106.1 119.5 23.9 5
2004?04?24 0.5 0.0 0.0 0.0 28.3 40.6 79.8 5.6 91.1 22.8 4
2004?06?12 0.3 0.0 10.4 43.2 49.5 23.4 31.2 31.1 93.5 15.6 6
2004?07?08 10.1 1.0 0.0 0.0 0.0 0.0 0.0 77.4 77.4 77.4 1
Tab.6 Early rainfall of each landslide occurrence date
Fig.8 Probability position of landslide in sample point
Fig.9 Landslide hazard warning maps of Xunwu County with rainfall threshold of T1 to T5
[1]   黄发明, 殷坤龙, 张桂荣, 等 多变量PSO-SVM模型预测滑坡地下水位[J]. 浙江大学学报: 工学版, 2015, 49 (6): 1193- 1200
HUANG Fa-ming, YIN Kun-long, ZHANG Gui-rong, et al Prediction of groundwater level in landslide using multivariable PSO-SVM model[J]. Journal of Zhejiang University: Engineering Science, 2015, 49 (6): 1193- 1200
[2]   HUANG F, YIN K, HUANG J, et al Landslide susceptibility mapping based on self-organizing-map network and extreme learning machine[J]. Engineering Geology, 2017, 223: 11- 22
doi: 10.1016/j.enggeo.2017.04.013
[3]   黄发明, 殷坤龙, 张桂荣, 等 基于相空间重构和小波分析-粒子群向量机的滑坡地下水位预测[J]. 地球科学: 中国地质大学学报, 2015, 40 (7): 1254- 1265
HUANG Fa-ming, YIN Kun-long, ZHANG Gui-rong, et al Landslide groundwater level time series prediction based on phase space reconstruction and wavelet analysis-support vector machine optimized by PSO algorithm[J]. Earth Science: Journal of China University of Geosciences, 2015, 40 (7): 1254- 1265
doi: 10.3799/dqkx.2015.105
[4]   刘永莉, 孙红月, 于洋, 等 基于BOTDR监测技术抗滑桩上滑坡推力确定[J]. 浙江大学学报: 工学版, 2012, 46 (5): 798- 803
LIU Yong-li, SUN Hong-yue, YU Yang, et al Determining landslide thrust on anti-sliding pile by BOTDR monitoring technology[J]. Journal of Zhejiang University: Engineering Science, 2012, 46 (5): 798- 803
[5]   GUO Z, YIN K, GUI L, et al Regional rainfall warning system for landslides with creep deformation in Three Gorges using a statistical black box model[J]. Scientific Reports, 2019, 9 (1): 8962
doi: 10.1038/s41598-019-45403-9
[6]   LI D, HUANG F, YAN L, et al Landslide susceptibility prediction using particle-swarm-optimized multilayer perceptron: comparisons with multilayer-perceptron-only, BP neural network, and information value models[J]. Applied Science, 2019, 9 (18): 3664
doi: 10.3390/app9183664
[7]   HUANG F, CAO Z, GUO J, et al Comparisons of heuristic, general statistical and machine learning models for landslide susceptibility prediction and mapping[J]. Catena, 2020, 191: 104580
doi: 10.1016/j.catena.2020.104580
[8]   WANG?L?J,?SAWADA?K,?MORIGUCHI?S.?Landslide-susceptibility analysis ?using ?light ?detection ?and ?ranging-derived digital elevation?models?and?logistic?regression?models:?a?case?study?inMizunami?City,?Japan [EB/OL]. [2020-01-10]. http://socolar.com/Article/Index?=100028583553&jid=100000013536.
[9]   ZHU L, HUANG L, FAN L, et al Landslide susceptibility prediction modeling based on remote sensing and a novel deep learning algorithm of a cascade-parallel recurrent neural network[J]. Sensors (Basel), 2020, 20 (6): 1576
doi: 10.3390/s20061576
[10]   CHANG Z, DU Z, ZHANG F, et al Landslide susceptibility prediction based on remote sensing images and GIS: comparisons of supervised and unsupervised machine learning models[J]. Remote Sensing, 2020, 12 (3): 502
doi: 10.3390/rs12030502
[11]   周超, 殷坤龙, 黄发明 混沌序列WA-ELM耦合模型在滑坡位移预测中的应用[J]. 岩土力学, 2015, 36 (9): 2674- 2680
ZHOU Chao, YIN Kun-long, HUANG Fa-ming Application of the chaotic sequence WA-ELM coupling model in landslide displacement prediction[J]. Rock and Soil Mechanics, 2015, 36 (9): 2674- 2680
[12]   MA J, TANG H, LIU X, et al Establishment of a deformation forecasting model for a step-like landslide based on decision tree C5.0 and two-step cluster algorithms: a case study in the Three Gorges Reservoir area, China[J]. Landslides, 2017, 14 (3): 1275- 1281
doi: 10.1007/s10346-017-0804-0
[13]   SEGONI S, PICIULLO L, GARIANO S L A review of the recent literature on rainfall thresholds for landslide occurrence[J]. Landslides, 2018, 15 (8): 1483- 1501
doi: 10.1007/s10346-018-0966-4
[14]   GUZZETTI F, REICHENBACH P, CARDINALI M, et al Probabilistic landslide hazard assessment at the basin scale[J]. Geomorphology, 2005, 72 (1-4): 272- 299
doi: 10.1016/j.geomorph.2005.06.002
[15]   桂蕾. 三峡库区万州区滑坡发育规律及风险研究[J]. 北京: 中国地质大学, 2014.
GUI Lei. Research on landslide development regularities and risk in Wan Zhou District, Three Gorges Reservoir [D]. Beijing: China University of Geoscience, 2014.
[16]   ZHAN L T, LI H, CHEN Y M, et al Parametric analyses of intensity-duration curve for predicting rainfall-induced landslides in residual soil slope in Southeastern coastal areas of China[J]. Rock and Soil Mechanics, 2012, 33 (3): 872- 880
[17]   陈洪凯, 魏来, 谭玲 降雨型滑坡经验性降雨阈值研究综述[J]. 重庆交通大学学报: 自然科学版, 2012, 31 (5): 990- 996
CHEN Hong-kai, WEI Lai, TAN Ling Review of research on empirical rainfall threshold of rainfall-induced landslide[J]. Journal of Chongqing Jiaotong University: Natural Sciences, 2012, 31 (5): 990- 996
[18]   MATHEW J, BABU D G, KUNDU S, et al Integrating intensity–duration-based rainfall threshold and antecedent rainfall-based probability estimate towards generating early warning for rainfall-induced landslides in parts of the Garhwal Himalaya, India[J]. Landslides, 2013, 11 (4): 575- 588
[19]   鲍其云, 麻土华, 李长江, 等 浙江62个丘陵山区县引发滑坡的降雨强度——历时阈值[J]. 科技通报, 2016, 32 (5): 48- 55
BAO Qi-yun, MA Tu-hua, LI Chang-jiang, et al Rainfall intensity-duration thresholds for the initiation of landslides in 62 hilly and mountainous counties of Zhejiang Province[J]. Bulletin of Science and Technology, 2016, 32 (5): 48- 55
[20]   ROSSI M, LUCIANI S,VALIGI D, et al Statistical approaches for the definition of landslide rainfall thresholds and their uncertainty using rain gauge and satellite data[J]. Geomorphology, 2017, 285: 16- 27
doi: 10.1016/j.geomorph.2017.02.001
[21]   吴益平, 张秋霞, 唐辉明, 等 基于有效降雨强度的滑坡灾害危险性预警[J]. 地球科学: 中国地质大学学报, 2014, 39 (7): 889- 895
WU Yi-ping, ZHANG Qiu-xia, TANG Hui-ming, et al Landslide hazard warning based on effective rainfall intensity[J]. Earth Science: Journal of China University of Geosciences, 2014, 39 (7): 889- 895
doi: 10.3799/dqkx.2014.083
[22]   HUANG Y, ZHAO L Review on landslide susceptibility mapping using support vector machines[J]. Catena, 2018, 165: 520- 529
doi: 10.1016/j.catena.2018.03.003
[23]   LIU W, LUO X, HUANG F, et al Uncertainty of the soil–water characteristic curve and its effects on slope seepage and stability analysis under conditions of rainfall using the Markov chain Monte Carlo method[J]. Water, 2017, 9 (10): 758
[24]   刘长殿, 孙红月, 康剑伟, 等 土体的充气阻渗试验[J]. 浙江大学学报: 工学版, 2014, 48 (2): 236- 241
LIU Chang-dian, SUN Hong-yue, KANG Jian-wei, et al Experimental investigation of seepage barrier effect by air-inflation in soil[J]. Journal of Zhejiang University: Engineering Science, 2014, 48 (2): 236- 241
[25]   YANG H, WEI F, MA Z, et al Rainfall threshold for landslide activity in Dazhou, southwest China[J]. Landslides, 2020, 17 (1): 61- 77
[26]   王芳, 殷坤龙, 桂蕾, 等 不同日降雨工况下万州区滑坡灾害危险性分析[J]. 地质科技情报, 2018, 37 (1): 190- 195
WANG Fang, YIN Kun-long, GUI Lei, et al Landslide hazard analysis under different daily rainfall conditions in Wanzhou District[J]. Geological Science and Technology Information, 2018, 37 (1): 190- 195
[1] HUANG Fa-ming, YIN Kun-long, ZHANG Gui-rong, TANG Zhi-zheng, ZHANG Jun.
Prediction of groundwater level in landslide using multivariable PSO-SVM model
[J]. Journal of ZheJiang University (Engineering Science), 2015, 49(6): 1193-1200.