[1]刘永垚,第宝锋*,詹 宇,等.基于随机森林模型的泥石流易发性评价--以汶川地震重灾区为例[J].山地学报,2018,(05):765-773.[doi:10.16089/j.cnki.1008-2786.000372]
　LIU Yongyao,DI Baofeng*,ZHAN Yu,et al.Debris Flows Susceptibility Assessment in Wenchuan Earthquake Areas Based on Random Forest Algorithm Model[J].Mountain Research,2018,(05):765-773.[doi:10.16089/j.cnki.1008-2786.000372]
点击复制

基于随机森林模型的泥石流易发性评价--以汶川地震重灾区为例()
分享到：

《山地学报》[ISSN:1008-2186/CN:51-1516]

卷:
期数:
2018年05期
页码:
765-773
栏目:
山地灾害
出版日期:
2018-09-30

文章信息/Info

Title:
Debris Flows Susceptibility Assessment in Wenchuan Earthquake Areas Based on Random Forest Algorithm Model
文章编号:
1008-2786-(2018)5-765-09
作者:
刘永垚1; 第宝锋2; 1*; 詹 宇1; Constantine A.Stamatopoulos3
1.四川大学 建筑与环境学院,成都 610065; 2.四川大学 灾后重建与管理学院,成都 610207; 3.Stamatopoulos and Associates Co.and Hellenic Open University, 5 Isavron str, 11471 Athens, Greece
Author(s):
LIU Yongyao1;  DI Baofeng2; 1*;  ZHAN Yu1;  Stamatopoulos C.A.3
1.Collage of Architecture & Environment, Sichuan University, Chengdu 610065, China; 2.Institute for Disaster Management and Reconstruction, Sichuan University, Chengdu 610207, China; 3.Stamatopoulos and Associates Co.and Hellenic Open University, 5
关键词:
泥石流;  易发性;  随机森林算法;  关键因子;  汶川地震
Keywords:
debris flow;  susceptibility;  random forest algorithm;  key variables;  Wenchuan earthquake
分类号:
P694
DOI:
10.16089/j.cnki.1008-2786.000372
文献标志码:
A
摘要:
在区域泥石流易发性研究中,科学确定泥石流易发性主控因子及其贡献率既是关键科学问题,也是区域泥石流预警预报和风险管理的重要基础。本研究选取汶川地震重灾区,引入随机森林算法,以小流域为评价单元,集合多元因子指标体系,建立泥石流易发性评价模型,定量分析了汶川地震重灾区内泥石流关键影响因子及贡献率,并探讨了研究区泥石流易发性的空间分布特征。本文初选了63项评价指标,以模型AUC值变化为基础,筛选出35项指标构成易发性评价指标体系,并用于区域内泥石流易发性主控因子的识别,结果表明:流域高差、流域平均坡度、流域内滑坡面积、平均降雨天数是区域内泥石流易发性主控因子,另外,沟长比降、大于10°积温、年均温、人口密度、村落个数、低覆盖度土地利用方式等在泥石流易发性评价中也发挥着重要作用; 易发性评价结果显示,极高易发区占比达到了22.94%,主要分布于研究区西部,泥石流易发性较高的小流域主要分布在青藏高原向四川盆地过渡的地形急变带,同时也与地震带、断裂带、干旱河谷区域密切相关。模型验证结果表明,平均AUC值达0.84,模型具有很高的稳定性和准确性,说明随机森林算法非常适用于区域泥石流易发性评价研究,机器学习算法结合小流域为单元的方法对区域泥石流易发性评价有效果良好,可为区域尺度灾害易发性及风险评估提供更为有效的方法参考。
Abstract:
As debris flow is one of the major mountain disasters in China, the research on debris flow susceptibility is of great significance for monitoring and risk management.With respect to debris flow susceptibility, identifying the key variables and their relative importance are critical for developing warning systems and managing risk.This paper focused on the most important variables and discussed the spatial characteristics of debris flow susceptibility in the Wenchuan earthquake-hit area by using a machine learning algorithm(i.e., random forest).Watershed was selected as the basic assessment unit, accompanied with a multiple-factor index system, to model debris flow susceptibility.The results of model evaluation showed adequate stability and accuracy, with the 10-fold cross-validation average area-under-curve(AUC)values of 0.84, indicating the random forest model was suitable to evaluate regional debris flow susceptibility.In this paper, on the basis of the change of AUC value, 35 of 63 evaluation indicators were preferred to form a susceptibility evaluation index system, and used to identify the main control factors of debris flow susceptibility in the region.We found that the elevation difference, average slope, landslide, and average rainy days played the most important role in determining the regional debris flow susceptibility.The results of susceptibility assessment indicated that the highest susceptibility watersheds reached 22.94%, was mainly distributed in the western part of the study area.The spatial distributions of the debris flow susceptibility displayed the high-susceptibility watersheds were highly coupled with the locations of the topographical extreme belt, fault zone, seismic belt, and dry valleys.The regional debris flow susceptibility evaluated by the random forest model developed in this study could be applied as a reference for regional-scale disaster susceptibility assessment, and the results also reflected that machine learning algorithm could provide a new method and idea for regional debris flows susceptibility assessment.

参考文献/References:

[1] 崔鹏,庄建琦,陈兴长,等.汶川地震区震后泥石流活动特征与防治对策[J].四川大学学报(工程科学版),2010,42(5):10-19 [CUI Peng, ZHUANG Jianqi, CHEN Xingchang.Charac-teristics and countermeasures of debris flow in Wenchuan area after the earthquake [J].Journal of Sichuan University(Engineering Science Edition), 2010, 42(5): 10-19]
[2] BLAHUT J, HORTON P, STERLACCHINI S, et al.Debris flow hazard modelling on medium scale, Valtellina di Tirano, Italy [J].Natural Hazards and Earth System Science, 2010, 10(11): 2379-2390
[3] CARRARA A, CROSTA G, FRATTINI P.Comparing models of debris-flow susceptibility in the alpine environment [J].Geomorphology, 2008, 94(3/4): 353-378
[4] KAPPES M S, MALET J P, REMAÎTRE A, et al.Assessment of debris-flow susceptibility at medium-scale in the Barcelonnette Basin, France [J].Natural Hazards and Earth System Science, 2011, 11(2): 627-641
[5] 阮沈勇,黄润秋.基于GIS的信息量法模型在地质灾害危险性区划中的应用[J].成都理工大学学报(自然科学版).2001,28(1):89-92 [RUAN Shenyong, HUANG Runqiu.Application of GIS-based information model on assessment of geological hazards risk [J].Journal of Chengdu University of Technology, 2001, 28(1): 89-92]
[6] 徐艳琴,白淑英,徐永明.基于两种方法的攀西泥石流易发性评价对比分析[J].水土保持研究.2018,25(3):285-291 [XU Yanqin, BAI Shuying, XU Yongming.Comparative analysis of debris flow susceptibility assessment based on two methods in Panxi district [J].Research of Soil and Water Conservation, 2018, 25(3): 285-291]
[7] AUTHOR M K A C, GUPTA A S D, GUPTA R P.An artificial neural network approach for landslide hazard zonation in the Bhagirathi(Ganga)Valley, Himalayas [J].International Journal of Remote Sensing, 2004, 25(3): 559-572
[8] 唐川,朱大奎.基于GIS技术的泥石流风险评价研究[J].地理科学, 2002,7(3):300-304 [TANG Chuan, ZHU Dakui.Assessment of debris flow risk of Yunnan Province by using GIS [J].Scientia Geographica Sinica, 2002, 7(3): 300-304]
[9] 王哲,易发成.基于层次分析法的绵阳市地质灾害易发性评价[J].自然灾害学报, 2009,18(1):14-23 [WANG Zhe, YI Facheng.AHP-based evaluation of occurrence easiness of geological disasters in Mianyang City [J].Journal of Natural Disasters, 2009, 18(1): 14-23]
[10] 徐继维,张茂省,范文.地质灾害风险评估综述[J].灾害学, 2015,30(4):130-134 [XU Jiwei, ZHANG Maosheng, FAN Wen.An overview of geological disaster risk assessment [J].Journal of Catastrophology, 2015, 30(4): 130-134]
[11] 邹强,崔鹏,杨伟.G318川藏公路段泥石流危险性评价[J].山地学报, 2013,31(3):342-348 [ZOU Qiang, CUI Peng, YANG Wei.Hazard assessment of debris flows along G318 Sichuan-Tibet highway [J].Journal of Mountain Science, 2013, 31(3): 342-348]
[12] 刘希林.区域泥石流风险评价研究[J].自然灾害学报, 2000,9(1):54-61 [LIU Xilin.Regional risk assessment on debris flow [J].Journal of Natural Disasters, 2000, 9(1): 54-61]
[13] 邹强,唐建喜,李淑松,等.基于水文响应单元的泥石流灾害易发性分区方法[J].山地学报, 2017,35(4):496-505 [ZOU Qiang, TANG Jianxi, LI Shusong, et al.Susceptibility assessment method of debris flows based on hydrological response unit [J].Mountain Research, 2017, 35(4): 496-505]
[14] 熊俊楠,韦方强,刘志奇.四川省小流域泥石流危险性评价[J].地球信息科学学报, 2017,19(12):1604-1612 [XIONG Junnan, WEI Fangqiang, LIU Zhiqi.Hazard assessment of debris flow in Sichuan Province [J].Journal of Geo-Information Science, 2017, 19(12): 1604-1612]
[15] PHAM B T, PRADHAN B, TIEN B D, et al.A comparative study of different machine learning methods for landslide susceptibility assessment: A case study of Uttarakhand area(India)[J].Environmental Modelling & Software, 2016, 84: 240-250
[16] IVERSON L R, PRASAD A M, MATTHEWS S N, et al.Estimating potential habitat for 134 eastern US tree species under six climate scenarios [J].Forest Ecology & Management, 2008, 254(3): 390-406
[17] 刘坚,李树林,陈涛.基于优化随机森林模型的滑坡易发性评价[J].武汉大学学报(信息科学版), 2017, 43(7):1-7 [LIU Jian, LI Shulin, CHEN Tao.Landslide susceptibility mapping by random forest model [J].Geomatics and Information Science of Wuhan University, 2017, 43(7): 1-7]
[18] 李亭,田原,邬伦,等.基于随机森林方法的滑坡灾害危险性区划[J].地理与地理信息科学, 2014,30(6):25-30 [LI Ting, TIAN Yuan, WU Lun.Landslide susceptibility mapping using random forest [J].Geography and Geo-Information Science, 2014, 30(6):25-30]
[19] TRIGILA A, IADANZA C, ESPOSITO C, et al.Comparison of Logistic Regression and Random Forests techniques for shallow landslide susceptibility assessment in Giampilieri(NE Sicily, Italy)[J].Geomorphology, 2015, 249: 119-136
[20] KERN A N, ADDISON P, OOMMEN T, et al.Machine Learning Based Predictive Modeling of Debris Flow Probability Following Wildfire in the Intermountain Western United States [J].Mathematical Geosciences, 2017, 49(6): 717-735
[21] 赖成光,陈晓宏,赵仕威,等.基于随机森林的洪灾风险评价模型及其应用[J].水利学报, 2015,46(1): 58-66 [LAI Chengguang, CHEN Xiaohong, ZHAO Shiwei.A flood risk assessment model based on random forest and its application [J].Journal of Hydraulic Engineering, 2015, 46(1):58-66]
[22] PHAM B T, SHIRZADI A, TIEN Bui D, et al.A hybrid machine learning ensemble approach based on a radial basis function neural network and rotation forest for landslide susceptibility modeling: A case study in the Himalayan area, India [J].International Journal of Sediment Research, 2017
[23] 原立峰.基于SVM的泥石流危险度评价研究[J].地理科学, 2008,28(2):296-300 [YUAN Lifeng.Debris flow hazard hssessment based on SVM [J].Scientia Geographica Sinica, 2008, 28(2):296-300 ]
[24] 曹正凤.随机森林算法优化研究[D].北京:首都经贸大学 2014 [CAO Zhengfeng.Study on optimization of random forests algorithm [D].Beijing: Caital University of Economics and Business, 2014]
[25] 方匡南,吴见彬,朱建平,等.随机森林方法研究综述[J].统计与信息论坛, 2011,26(3):32-38 [FANG Kuangnan, WU Jianbin, ZHU Jianping, et al.A review of technologies on random forests [J].Statistics & Information Forum, 2011, 26(3): 32-38]
[26] 王涛.汶川地震重灾区地质灾害危险性评估研究[D].北京: 中国地质科学院,2010:13-14 [WANG Tao.Study on seismic landslide hazard assessment in Wenchuan earthquake severely afflicted area [D].Beijing: Chinese Academy of Geological Sciences, 2010: 13-14]
[27] 张春山.汶川地震灾区次生灾害隐患排查与工程设计示范[M].北京:中国大地出版社,2009:11-25 [ZHANG Chunshan.Investigation and engineering design demonstration of hidden danger of secondary disasters in Wenchuan earthquake stricken area [M].Beijing: China Land Press, 2009: 11-25]
[28] 许冲.汶川地震滑坡详细编录及其与全球其他地震滑坡事件对比[J].科技导报, 2012,30(25):18-26 [XU Chong.Detailed inventory of landslides triggered by the 2008 Wenchuan earthquake and its comparison with other earthquake events in the world [J].Science & Technology Report, 2012, 30(25): 18-26]
[29] 康志成.中国泥石流研究[M].北京:科学出版社,20004:1-4 [Kang Z, Lee C, Law K, et al.Debris Flow Research in China[M].Beijing: Science Press, 2004: 1-4]
[30] 陈宁生,Javed Iqbal,刘丽红.“8·3”鲁甸地震小震大灾背景下次生泥石流的易发性分析[J].成都理工大学学报(自然科学版), 2016,43(1):102-108 [CHEN Ningsheng, JAVED IQBAL, LIU Lihong.Tendency predicition of debris flow triggered by “8·3” Ludian earthquake of Yunnan, China [J].Journal of Chengdu University of Technology, 2016, 43(1): 102-108]
[31] 张若琳,孟晖,连建发.基于GIS的中国泥石流易发性评价[J].成都理工大学学报(自然科学版),2013,40(4):379-386 [ZHANG Ruolin, MENG Hui, LIAN Jianfa.Susceptibility evaluation of debris-flow in China based on GIS [J].Journal of Chengdu University of Technology, 2013, 40(4): 379-386]

相似文献/References:

[1]蒋志林,朱静,常鸣,等.汶川地震区红椿沟泥石流形成物源量动态演化特征[J].山地学报,2014,(01):81.
　JIANG Zhilin,ZHU Jing,CHANG Ming,et al.Dynamic Evolution Characteristics of Hongchun Gully Source Area of Debris Flow in Wenchuan Earthquake Region[J].Mountain Research,2014,(05):81.
[2]常鸣,唐川,蒋志林,等.强震区都江堰市龙池镇泥石流物源的遥感动态演变[J].山地学报,2014,(01):89.
　CHANG Ming,TANG Chuan,JIANG Zhilin,et al.Dynamic Evolution Process of Sediment Supply for Debris Flow Occurrence in Longchi of Dujiangyan，Wenchuan Earthquake Area[J].Mountain Research,2014,(05):89.
[3]王 钧,欧国强,杨 顺,等.地貌信息熵在地震后泥石流危险性评价中的应用[J].山地学报,2013,(01):83.
　WANG Jun,OU Guoqiang,YANG Shun,et al.Applicability of Geomorphic Information Entropy in the Postearthquake Debris Flow Risk Assessment[J].Mountain Research,2013,(05):83.
[4]王东坡,何思明,葛胜锦,等.“9?07”彝良地震诱发次生山地灾害调查及减灾建议[J].山地学报,2013,(01):101.
　WANG Dongpo,HE Siming,GE Shengjin,et al.Mountain Hazards Induced by the Earthquake of Sep 07，2012 in Yiliang and the Suggestions of Disaster Reduction[J].Mountain Research,2013,(05):101.
[5]喻 武,万 丹,汪书丽,等.藏东南泥石流沉积区植物群落结构和物种多样性特征[J].山地学报,2013,(01):120.
　YU Wu,WAN Dan,WANG Shuli,et al.Community Structure and Species Diversity of Debris Flow Deposition Area in Southeast of Tibet,China[J].Mountain Research,2013,(05):120.
[6]崔鹏,陈晓清,张建强,等.“4·20”芦山7.0级地震次生山地灾害活动特征与趋势[J].山地学报,2013,(03):257.
　CUI Peng,CHEN Xiaoqing,ZHANG Jianqiang,et al.Activities and Tendency of Mountain Hazards Induced by the Ms7.0 Lushan Earthquake，April 20，2013[J].Mountain Research,2013,(05):257.
[7]邹强,崔鹏,杨伟,等.G318川藏公路段泥石流危险性评价[J].山地学报,2013,(03):342.
　ZOU Qiang,CUI Peng,YANG Wei.Hazard Assessment of Debris Flows along G318 Sichuan-Tibet Highway[J].Mountain Research,2013,(05):342.
[8]王根龙,张茂省,于国强,等.舟曲2010年“8·8”特大泥石流灾害致灾因素[J].山地学报,2013,(03):349.
　WANG Genlong,ZHANG Maosheng,YU Guoqiang,et al.Factor Analysis for Catastrophic Debris Flows on August 8，2010 in Zhouqu City of Gansu，China[J].Mountain Research,2013,(05):349.
[9]陈源井,余斌,朱渊,等.地震后泥石流临界雨量变化特征——以汶川地震区小岗剑沟为例[J].山地学报,2013,(03):356.
　CHEN Yuanjing,YU Bin,ZHU Yuan,et al.Characteristics of Critical Rainfall of Debris Flow after Earthquake——A Case Study of the Xiaogangjian Gully[J].Mountain Research,2013,(05):356.
[10]游勇,柳金峰,陈兴长,等.芦山“4·20”地震后宝兴县城打水沟泥石流发育趋势及防治方案[J].山地学报,2013,(04):495.
　YOU Yong,LIU Jinfeng,CHEN Xingzhang.The Potential Tendency and Mitigation Measures of Dashui Gully in Baoxing Coutny after Lushan“4?20”Earthquake of Schuan[J].Mountain Research,2013,(05):495.

备注/Memo

备注/Memo:
收稿日期(Received date):2018-10-24; 改回日期(Accepted date):2018-10-25
基金项目(Foundation item):2017中国和希腊政府间科技合作项目。[2017 Joint China-Greece Intergovernment Science and Technology Cooperation Project]
第一简介(Biography):刘永垚(1993-),男,四川绵阳人,硕士研究生,主要研究方向:水土保持规划与评价。 [LIU Yongyao(1993-), male, born in Mianyang, Sichuan province, M.Sc.candidate, research on soil and water conservation planning] E-mail: liuyongyao93@163.com
*通讯作者(Corresponding author):第宝锋(1977-),男,博士,教授,主要研究方向:山地灾害风险、环境遥感。[DI Baofeng(1977-), male, Ph.D., professor, specialized in mountain disaster risk and remote sensing of environment] E-mail: dibaofeng@scu.edu.cn
更新日期/Last Update: 2018-11-30