基于Logistic回归和随机森林的清江流域长阳库岸段堆积层滑坡易发性评价

曾斌; 吕权儒; 寇磊; 艾东; 许汇源; 袁晶晶

doi:10.16031/j.cnki.issn.1003-8035.202205044

基于Logistic回归和随机森林的清江流域长阳库岸段堆积层滑坡易发性评价

doi: 10.16031/j.cnki.issn.1003-8035.202205044

1.
中国地质大学（武汉）环境学院，湖北武汉　430000
2.
湖北省电力勘测设计院有限公司
3.
湖北省地质局第七地质大队，湖北宜昌　443000

计量
- 文章访问数: 329
- HTML全文浏览量: 317
- PDF下载量: 23
- 被引次数: 0
出版历程
- 网络出版日期: 2023-05-16

Susceptibility assessment of colluvium landslides along the Changyang section of Qingjiang River using logistic regression and random forest methods

1.
School of Environmental Studies（China University of Geosciences）, Wuhan, Hubei　430000
2.
PowerChina Hubei electric engineering CO., LTD
3.
The Seventh Geological Brigade of Hubei Geological Bureau

摘要

摘要: 清江流域地质环境背景条件复杂，特别是沿清江库岸地质灾害频发；而过往的地质灾害易发性评价多针对行政区域范围、鲜有针对库岸带的专门评价，并且所采用评价指标体系的针对性以及评价方法的可靠性仍有进一步提升的空间。为构建一套更加符合清江流域库岸带地质灾害发育特征的易发性评价指标体系，获得更加准确、适用性强的易发性区划成果，本文以清江流域渔峡口～资丘段为研究区，以两岸涉水斜坡体为研究对象，以斜坡单元为评价单元，构建坡度、坡向、高差、坡型、NDVI、TWI、斜坡结构类型、工程地质岩组、堆积层厚度、河谷演化类型10个指标组成的易发性评价体系；用基于归一化确定性系数的Logistic回归和随机森林方法构建评价模型并得到不同易发性区划成果。根据评价结果，高易发区主要分布于清江干流渔峡口东～资丘东段左岸顺向斜坡体的中～下部涉水区域，且Logistic回归模型在地形地貌复杂的库岸段适用性要优于随机森林模型。研究表明：所选堆积层厚度及河谷演化类型指标很好地代表了清江库岸段的独特地质背景条件特点；同时在非行政区划范围的特定研究区且当历史滑坡样本数量有限的情况下，Logistic回归模型能够较好的学习灾害发育规律并具备可靠的易发性预测能力。
- 堆积层滑坡 /
- Logistic回归 /
- 随机森林 /
- 易发性评价
Abstract: The geological and environmental background conditions of the Qingjiang River Basin are highly complex, particularly with frequent geological disasters along the Qingjiang reservoir bank. Previous susceptibility assessment for geological disasters was mostly focused on administrative areas and seldom specialized evaluations for the reservoir bank zone. Furthermore, there is still room for improvement in the evaluation index system, as well as in the pertinence and reliability of the evaluation method. To address these shortcomings, a more suitable susceptibility evaluation index system was constructed to obtain accurate and applicable susceptibility zoning results. The Yuxiakou to Ziqiu section of the Qingjiang River Basin was chosen as the research area, with the wading slope body on both sides of the river selected as the research object and the slope unit chosen as the evaluation unit. A susceptibility evaluation system composed of ten indicators, including slope, aspect, elevation range, slope type, NDVI, TWI, slope structure type, engineering geological rock formation, accumulation thickness, and valley evolution, was constructed.The logistic regression and random forest methods were used to construct the evaluation model based on the normalized certain factors, and different susceptibility zoning results were obtained. According to the evaluation results, the high-prone areas were mainly distributed in the middle to lower water wading areas of the left bank from the east of Yuxiakou to the east of Ziqiu, along the main stream of the Qingjiang River. The logistic regression model showed better applicability in the reservoir-bank section with complex topography and landforms. The research revealed that the accumulation thickness and valley evolution indicators were effective in representing the unique geological background conditions of the Qingjiang reservoir bank. The logistic regression model was able to learn the developmental law of disasters and has a reliable susceptibility prediction ability.
- accumulation thickness /
- Logistic regression /
- random forest /
- susceptibility assessment

HTML全文

图 1 研究区地理位置图

Figure 1. Geographical location map of the study area

下载: 全尺寸图片幻灯片

图 2 研究区历史滑坡分布图

Figure 2. Distribution map of historical landslides in the study area

下载: 全尺寸图片幻灯片

图 3 技术路线图

Figure 3. Technology Roadmap

下载: 全尺寸图片幻灯片

图 4 评价指标分类结果图

Figure 4. Classification results of all evaluation index

下载: 全尺寸图片幻灯片

图 5 基于归一化CF值Logistic回归滑坡易发性分区图

Figure 5. Landslide susceptibility zoning map based on normalized CF-valued logistic regression

下载: 全尺寸图片幻灯片

图 6 基于归一化CF值的随机森林滑坡易发性分区图

Figure 6. Landslide susceptibility zoning map based on normalized CF-valued random forest results

下载: 全尺寸图片幻灯片

图 7 两种模型易发性结果ROC曲线

Figure 7. ROC curves of susceptibility results for two models

下载: 全尺寸图片幻灯片

表 1 评价指标来源及处理方式

Table 1. Sources and processing methods of the evaluation indicators

评价指标	来源	数据类型	子类划分方式说明
坡度	5mDEM（栅格）	连续型	（1）NDVI、TWI：用K-means聚类法将指标分为5个子类区间；（2）堆积层厚度：由野外实测堆积层点数据通过经验贝叶斯克里金法插值为面数据，用K-means聚类法分为5个子类区间（3）工程地质岩组分为6个子类区间：Ⅱ-1（坚硬～较坚硬碎屑岩）、Ⅱ-4（较坚硬碎屑岩）、Ⅲ-1（坚硬碳酸盐岩）、Ⅲ-2（坚硬碳酸盐岩夹碎屑岩）、Ⅲ-3（较坚硬碳酸盐岩）、Ⅲ-4（较坚硬碳酸盐岩夹碎屑岩）（4）斜坡结构类型分为4个：顺向坡、斜向坡、横向坡、逆向坡（5）河谷演化类型分为2个：迁移性河谷、持续性河谷（6）坡向分为8个：北-东北-东-东南-南-西南-西-西北（7）坡型分为3个：凸形坡、平直坡、凹形坡
高差	5mDEM（栅格）	连续型
NDVI	Landsat8-OLT（栅格）	连续型
TWI	5mDEM（栅格）	连续型
距断层距离	线数据（矢量）	连续型
堆积层厚度	点数据（矢量）	连续型
工程地质岩组	面数据（矢量）	分类型
斜坡结构类型	面数据（矢量）	分类型
河谷演化类型	面数据（矢量）	分类型
坡向	5mDEM（栅格）	分类型
坡型	5mDEM（栅格）	分类型

下载: 导出CSV

参考文献(21)

[1]	雷深涵,华骐,郭峰,等. 长阳县地质灾害发育特征[J]. 地质灾害与环境保护,2013,24(1):30 − 35. [LEI Shenhan,HUA Qi,GUO Feng,et al. Characteristics of the geological hazards in Changyang[J]. Journal of Geological Hazards and Environment Preservation,2013,24(1):30 − 35. (in Chinese)
[2]	肖尚德,李智民,刘云彪,等. 湖北省清江流域滑坡分布规律与减灾对策研究[J]. 工程地质学报,2012,20(4):514 − 521. [XIAO Shangde,LI Zhimin,LIU Yunbiao,et al. Distribution law and hazard mitigation strategies of landliides in Qingjing River valley,Hubei[J]. Journal of Engineering Geology,2012,20(4):514 − 521. (in Chinese)
[3]	王凯,张少杰,韦方强. 斜坡单元提取方法研究进展和展望[J]. 长江科学院院报,2020,37(6):85 − 93. [WANG Kai,ZHANG Shaojie,WEI Fangqiang. Slope unit extraction methods:advances and prospects[J]. Journal of Yangtze River Scientific Research Institute,2020,37(6):85 − 93. (in Chinese)
[4]	ALVIOLI M,MARCHESINI I,REICHENBACH P,et al. Automatic delineation of geomorphological slope units with r. slopeunits v1.0 and their optimization for landslide susceptibility modeling[J]. Geoscientific Model Development,2016,9(11):3975 − 3991. doi: 10.5194/gmd-9-3975-2016
[5]	SUN Xiaohui,CHEN Jianping,HAN Xudong,et al. Landslide susceptibility mapping along the upper Jinsha River,south-western China:A comparison of hydrological and curvature watershed methods for slope unit classification[J]. Bulletin of Engineering Geology and the Environment,2020,79(9):4657 − 4670. doi: 10.1007/s10064-020-01849-0
[6]	黄立鑫,郝君明,李旺平,等. 基于RBF神经网络−信息量耦合模型的滑坡易发性评价—以甘肃岷县为例[J]. 中国地质灾害与防治学报,2021,32(6):116 − 126. [HUANG Lixin,HAO Junming,LI Wangping,et al. Landslide susceptibility assessment by the coupling method of RBF neural network and information value:A case study in Min Xian,Gansu Province[J]. The Chinese Journal of Geological Hazard and Control,2021,32(6):116 − 126. (in Chinese)
[7]	孙琳,任娜娜,李云安,等. 基于证据权法的公路路基岩溶塌陷危险性评价[J]. 中国地质灾害与防治学报,2019,30(3):94 − 100. [SUN Lin,REN Nana,LI Yun'an,et al. Risk assessment on Karst collapse of the highway subgrade based on weights of evidence method[J]. The Chinese Journal of Geological Hazard and Control,2019,30(3):94 − 100. (in Chinese)
[8]	李光辉,铁永波,白永建,等. 则木河断裂带(普格段)地质灾害发育规律及易发性评价[J]. 中国地质灾害与防治学报,2022,33(3):51. [LI Guanghui,TIE Yongbo,BAI Yongjian,et al. Distribution and susceptibility assessment of geological hazards in Zemuhe fault zone(Puge section)[J]. The Chinese Journal of Geological Hazard and Control,2022,33(3):51. (in Chinese)
[9]	吴赛男,田毅. 我国单体滑坡模拟和区域滑坡易发性评价研究进展[J]. 中国地质灾害与防治学报,2019,30(3):113 − 119. [WU Sainan,TIAN Yi. Review on progress of individual landslide simulation and assessment of reginal landslide susceptibility in China[J]. The Chinese Journal of Geological Hazard and Control,2019,30(3):113 − 119. (in Chinese)
[10]	ZHAO Shuai,ZHAO Zhou. A comparative study of landslide susceptibility mapping using SVM and PSO-SVM models based on grid and slope units[J]. Mathematical Problems in Engineering,2021,2021:1 − 15.
[11]	XIAO Ting,YIN Kunlong,YAO Tianlu,et al. Spatial prediction of landslide susceptibility using GIS-based statistical and machine learning models in Wanzhou County,Three Gorges Reservoir,China[J]. Acta Geochimica,2019,38(5):654 − 669. doi: 10.1007/s11631-019-00341-1
[12]	YU Chenglong,CHEN Jianping. Application of a GIS-based slope unit method for landslide susceptibility mapping in Helong city:comparative assessment of ICM,AHP,and RF model[J]. Symmetry,2020,12(11):1848. doi: 10.3390/sym12111848
[13]	常宏,金维群,王世昌,等. 清江中下游岸坡稳定性对河流地貌过程的响应[J]. 工程地质学报,2011,19(5):756 − 763. [CHANG Hong,JIN Weiqun,WANG Shichang,et al. Slope stability response to river geomorphic process in middle and lower reaches of Qingjiang River[J]. Journal of Engineering Geology,2011,19(5):756 − 763. (in Chinese)
[14]	常宏, 谭建民, 李明. 清江河道变迁与滑坡崩塌[M]. 武汉: 中国地质大学出版社, 2018 CHANG Hong, TAN Jianmin, LI Ming. Changes of Qingjiang River and landslide collapse[M]. Wuhan: China University of Geosciences Press, 2018. （in Chinese with English abstract）
[15]	陈伟. 山区村镇滑坡灾害风险评估研究[D]. 武汉: 武汉大学 CHEN Wei. Study on risk assessment of landslide disaster in mountainous villages and towns[D]. Wuhan: Wuhan University. （in Chinese with English abstract）
[16]	Zizheng,Guo. Landslide susceptibility zonation method based on C5.0 decision tree and K-means cluster algorithms to improve the efficiency of risk management[J]. Geoscience Frontiers,2021,12(6):101249. doi: 10.1016/j.gsf.2021.101249
[17]	王峰. 四川省南江县地质灾害易发性区划研究[D]. 成都: 成都理工大学 WANG Feng. Study on regionalization of geological disaster susceptibility in Nanjiang County, Sichuan Province[D]. Chengdu: Chengdu University of Technology. （in Chinese with English abstract）
[18]	罗路广, 裴向军, 崔圣华, 等. 九寨沟地震滑坡易发性评价因子组合选取研究[J]. 岩石力学与工程学报, 2021, 40（11）: 2306 − 2319. Luo Luguang, Pei Xiangjun, Cui Shenghua, et al. Combined selection of susceptibility assessment factors for Jiuzhaigou earthquake-induced landslides[J]. Chinese Journal of Rock Mechanics and Engineering, 2021, 40（11）: 2306 − 2319. （in Chinese with English abstract）.
[19]	王锐. 基于GIS和Logistic回归模型的降雨型滑坡易发性研究[D]. 杭州: 浙江大学 WANG Rui. Study on the susceptibility of rainfall landslide based on GIS and Logistic regression model[D]. Hangzhou: Zhejiang University. （in Chinese with English abstract）
[20]	吴润泽,胡旭东,梅红波,等. 基于随机森林的滑坡空间易发性评价:以三峡库区湖北段为例[J]. 地球科学,2021(1):321 − 330. [WU Runze,HU Xudong,MEI Hongbo,et al. Spatial susceptibility assessment of landslides based on random forest:a case study from Hubei section in the Three Gorges Reservoir area[J]. Earth Science,2021(1):321 − 330. (in Chinese with English abstract)
[21]	郝国栋. 基于随机森林模型的商南县滑坡易发性评价[D]. 西安: 西安科技大学, HAO Guodong. Evaluation of landslide susceptibility in Shangnan County based on random forest model[D]. Xi'an: Xi'an University of Science and Technology, . （in Chinese with English abstract）