基于不同统计模型的肯尼亚滑坡危险性评价

周苏华; 付宇航; 邢静康; 彭爱泉; 蒋明奕

doi:10.16031/j.cnki.issn.1003-8035.202206006

基于不同统计模型的肯尼亚滑坡危险性评价

doi: 10.16031/j.cnki.issn.1003-8035.202206006

1.
湖南大学土木工程学院，湖南长沙　410006
2.
湖南大学建筑安全与节能教育部重点实验室，湖南长沙　410006
3.
贵州交通职业技术学院机械电子工程系，贵州贵阳　550081

基金项目: 国家自然科学基金青年基金项目（51708199） ;贵州省科技支撑计划项目（2020-4Y047） ;贵州省交通运输厅科技项目（2017-143-054）;福建省地质灾害重点实验室自主课题（KLGHZ202104）；创新平台与人才计划-湖湘高层次人才聚集工程-创新团队（2019RS1030）

详细信息

作者简介:
周苏华（1987-），男，江苏盐城人，副教授，博士，从事岩土工程风险评价研究。E-mail：zhousuhua@hnu.edu.cn

中图分类号: P642.22;
计量
- 文章访问数: 743
- HTML全文浏览量: 372
- PDF下载量: 22
- 被引次数: 0
出版历程
- 收稿日期: 2022-06-07
- 修回日期: 2022-09-30
- 网络出版日期: 2023-05-16

Assessment of landslide Hazard Risk in Kenya based on Different Statistical Models

1.
College of Civil Engineering, Hunan University, Changsha, Hunan　410006, China
2.
Key Laboratory of Building Safety and Energy Efficiency of the Ministry of Education, Changsha Hunan　410006, China
3.
Department of mechanical and electrical engineering, Guizhou Communications Polytechnic, Guiyang Guizhou　550081, China

摘要

摘要: 肯尼亚是我国“一带一路”倡议在东非重要支点。受高原裂谷地形和显著的雨旱季节影响，肯尼亚地质灾害频发。本文以肯尼亚的历史滑坡数据为样本，选取高度、坡度、坡向、地貌、平面曲率、土壤类型、年平均降雨量、水流强度指数、地形湿度指数及土地利用类型作为评价指标，分别基于信息量模型（IV）、逻辑回归模型（LR）和极限学习机模型（ELM）对肯尼亚滑坡灾害进行危险性区划，其中ELM分别考虑了SIG、SIN和HARDLIM函数作为激活函数进行讨论。主要结论如下：（1）肯尼亚滑坡灾害高危险性及以上等级区域集中分布在西南部的高原和高原-裂谷过渡地带；（2）采用ROC曲线对模型精度进行评价，各模型的AUC值分别为IV（0.977）、LR（0.965）、ELM-SIG（0.859）、ELM-SIN（0.900）、ELM-HARDLIM（0.941），评价结果有效；（3）综合PR曲线结果判定，LR模型的召回率和精确率都处于较高的水平，优于其他模型；（4）肯尼亚内罗毕省（Nairobi）、中部省（Central）、尼扬扎省（Nyanza）和西部省（Western）四个省份高危险性区域占比较大。
- 肯尼亚 /
- 危险性 /
- 信息量 /
- 逻辑回归 /
- 机器学习
Abstract: Kenya is an important fulcrum of China's Belt and Road Initiative in East Africa. However, due to its plateau rift terrain and aboriginal rain and drought season, geological disasters occur frequently in Kenya. The study used historical landslide data in Kenya as samples and selected several evaluation indexes, including elevation, slope, aspect, landform, plane curvature, soil type, annual average rainfall, stream power index, terrain witness index, and land use type. The landslide risk in Kenya was evaluated based on the information value model (IV), logistic regression model (LR), and extreme learning machine model (ELM), with the ELM model considering SIG, SIN, and HARDLIM functions as activation functions for discussion. The main findings are as follows: (1) The high-risk and above-grade areas of landslide disasters in Kenya are mainly concentrated in the plateau and plateau-rift transition zone in the southwest. (2) The ROC curve was used to evaluate the accuracy of the models, and the AUC values of the IV(0.977), LR(0.965), ELM-SIG(0.859), ELM-SIN(0.900), and ELM-HARDLIM(0.941) models illustrate their validity. (3) Considering the PR curve results comprehensively, the recall rate and precision rate of the LR model are at a high level, marking it better than other models. (4) Nairobi, Central, Nyanza and Western provinces in Kenya account for a significant proportion of the high-risk and above-grade areas of landslide disasters.
- Kenya /
- risk /
- information value /
- logistic regression /
- machine learning

HTML全文

图 1 肯尼亚概况

Figure 1. Distribution map of geological hazards in Kenya

下载: 全尺寸图片幻灯片

图 2 评价因素

Figure 2. Classification chart of each evaluation factor in Kenya

下载: 全尺寸图片幻灯片

图 3 神经网络结构示意图

Figure 3. Schematic diagram of neural network structure

下载: 全尺寸图片幻灯片

图 4 评价流程

Figure 4. Evaluation process chart

下载: 全尺寸图片幻灯片

图 5 极限学习机模型评价结果

Figure 5. Risk mapping evaluation results of the ELM model

下载: 全尺寸图片幻灯片

图 6 不同模型ROC曲线和PR曲线

Figure 6. Comparison of the ROC curves and PR curves for different models

下载: 全尺寸图片幻灯片

图 7 肯尼亚省份危险性区划分布

Figure 7. Distribution of province hazard zone in Kenya

下载: 全尺寸图片幻灯片

表 1 信息量模型系数

Table 1. Summary table for coefficients of the IV model

因素	因子分级	信息量	因素	因子分级	信息量
高程（m）	0−50	−1.341	坡度（°）	0−5	−2.212
	50−200	0.000		5−15	0.315
	200−500	−1.941		15−25	1.552
	500−1000	−2.813		25−35	3.671
	1000−2000	0.452		35−45	4.889
	>2000	2.316		>45	5.356
坡向	平	0.000	地貌	洼地	1.547
	北	0.282		山麓	0.340
	东北	0.322		高原	1.389
	东北	0.183		平原	−2.277
	东南	−0.032		谷底	0.000
	南	−0.215		悬崖	1.062
	西南	−0.707		丘陵	0.913
	西	−0.089		山谷	2.353
	西北	0.052		山脊	1.476
土壤类型	黏土	−0.124		水体	0.000
	壤土	0.115	年平均降雨量（mm）	<400	−1.762
	砂土	−1.948		400−800	−0.729
	高含量黏土	0.450		800−1200	0.850
TWI	7−12	1.666		1200−1600	2.240
	12−14	−1.423		1600−2000	0.503
	14−16	−2.041		2000-2400	0.607
	16−20	−1.785		>2400	0.000
	20−32	−2.319	SPI	2-5	−2.889
土地利用类型	农业用地	0.800		5−7	−1.562
	荒地	−1.365		7−9	0.972
	灌木丛	−1.570		9−12	1.107
	林地	2.043		12−23	0.250
	草地	0.000	平面曲率	凸	−0.089
	沼泽	0.000		平	−1.026
	城镇	2.148		凹	0.187

下载: 导出CSV

表 2 逻辑回归模型系数

Table 2. Summary table for coefficients of the LR model

因素	系数	因素	系数
高程	1.683	土壤类型	−0.048
坡度	0.754	TWI	−1.125
坡向	−0.097	SPI	1.481
地貌	0.229	年平均降雨量	1.466
平面曲率	0.047	土地利用类型	0.026

下载: 导出CSV

表 3 不同模型灾害分布统计结果

Table 3. Statistical results of disasters distribution for different models

危险性分区评估模型		极低危险性	低危险性	中危险性	高危险性	极高危险性
LR模型	面积占比	72.70%	11.90%	5.50%	4.40%	5.40%
	数量占比	2.57%	0.93%	1.17%	10.51%	84.81%
	灾害比重	0.035	0.078	0.213	2.389	15.705
IV模型	面积占比	35.30%	29.80%	16.90%	11.00%	6.90%
	数量占比	0.93%	0.93%	2.80%	17.52%	77.80%
	灾害比重	0.026	0.031	0.1657	1.593	11.275
ELM模型（SIG函数）	面积占比	1.30%	76.30%	19.40%	1.10%	1.90%
	数量占比	1.87%	8.41%	88.08%	1.17%	0.47%
	灾害比重	1.438	0.110	4.540	1.064	0.247
ELM模型（SIN函数）	面积占比	1.10%	38.30%	43.60%	12.80%	4.20%
	数量占比	0.70%	2.80%	9.11%	33.88%	53.50%
	灾害比重	0.636	0.073	0.208	2.647	12.738
ELM模型（HARDLIM函数）	面积占比	5.20%	34.30%	34.50%	18.00%	8.00%
	数量占比	0.93%	0.93%	2.80%	17.52%	77.80%
	灾害比重	0.178	0.027	0.081	0.973	9.725

下载: 导出CSV

参考文献(26)

[1]	BATALA L K,YU Wangxing,KHAN A,et al. Natural disasters' influence on industrial growth,foreign direct investment,and export performance in the South Asian region of Belt and road initiative[J]. Natural Hazards,2021,108(2):1853 − 1876. doi: 10.1007/s11069-021-04759-w
[2]	BAI Yuanli,WIERZBICKI T. Application of extended Mohr-Coulomb criterion to ductile fracture[J]. International Journal of Fracture,2010,161(1):1 − 20. doi: 10.1007/s10704-009-9422-8
[3]	ROGERS J D,CHUNG J. Applying Terzaghi’s method of slope characterization to the recognition of Holocene land slippage[J]. Geomorphology,2016,265:24 − 44. doi: 10.1016/j.geomorph.2016.04.020
[4]	AHMED A,UGAI K,YANG Qing qing. Assessment of 3D slope stability analysis methods based on 3D simplified janbu and hovland methods[J]. International Journal of Geomechanics,2012,12(2):81 − 89. doi: 10.1061/(ASCE)GM.1943-5622.0000117
[5]	WANG Chun ming,LIU Chun yuan,WU Mai,et al. Research on soil-like slope instability based on FEM strength reduction[J]. Applied Mechanics and Materials,2013,438/439:1244 − 1248. doi: 10.4028/www.scientific.net/AMM.438-439.1244
[6]	吴信才,白玉琪,郭玲玲. 地理信息系统(GIS)发展现状及展望[J]. 计算机工程与应用,2000,36(4):8 − 9. [WU Xincai,BAI Yuqi,GUO Lingling. Development and prospect of geographic information system[J]. Computer Engineering and Applications,2000,36(4):8 − 9. (in Chinese)
[7]	SON J,SUH J,PARK H D. GIS-based landslide susceptibility assessment in Seoul,South Korea,applying the radius of influence to frequency ratio analysis[J]. Environmental Earth Sciences,2016,75(4):310. doi: 10.1007/s12665-015-5149-1
[8]	屠水云,张钟远,付弘流,等. 基于CF与CF-LR模型的地质灾害易发性评价[J]. 中国地质灾害与防治学报,2022,33(2):96 − 104. [TU Shuiyun,ZHANG Zhongyuan,FU Hongliu,et al. Geological hazard susceptibility evaluation based on CF and CF-LR model[J]. The Chinese Journal of Geological Hazard and Control,2022,33(2):96 − 104. (in Chinese)
[9]	刘福臻,王灵,肖东升. 机器学习模型在滑坡易发性评价中的应用[J]. 中国地质灾害与防治学报,2021,32(6):98 − 106. [LIU Fuzhen,WANG Ling,XIAO Dongsheng. Application of machine learning model in landslide susceptibility evaluation[J]. The Chinese Journal of Geological Hazard and Control,2021,32(6):98 − 106. (in Chinese)
[10]	LIU Rui,LI Luyao,PIRASTEH S,et al. The performance quality of LR,SVM,and RF for earthquake-induced landslides susceptibility mapping incorporating remote sensing imagery[J]. Arabian Journal of Geosciences,2021,14(4):1 − 15.
[11]	WANG Xi,WANG Shuangyin,QI Jiashuo. Open-channel landslide hazard assessment based on AHP and fuzzy comprehensive evaluation[J]. Water Supply,2020,20(8):3687 − 3696. doi: 10.2166/ws.2020.176
[12]	吴易雯,李莹杰,张列宇,等. 基于主客观赋权模糊综合评价法的湖泊水生态系统健康评价[J]. 湖泊科学,2017,29(5):1091 − 1102. [WU Yiwen,LI Yingjie,ZHANG Lieyu,et al. Assessment of lakes ecosystem health based on objective and subjective weighting combined with fuzzy comprehensive evaluation[J]. Journal of Lake Sciences,2017,29(5):1091 − 1102. (in Chinese with English abstract) doi: 10.18307/2017.0507
[13]	张俊,殷坤龙,王佳佳,等. 三峡库区万州区滑坡灾害易发性评价研究[J]. 岩石力学与工程学报,2016,35(2):284 − 296. [ZHANG Jun,YIN Kunlong,WANG Jiajia,et al. Evaluation of landslide susceptibility for Wanzhou district of Three Gorges Reservoir[J]. Chinese Journal of Rock Mechanics and Engineering,2016,35(2):284 − 296. (in Chinese)
[14]	CHEN Wei,LI Wenping,CHAI Huichan,et al. GIS-based landslide susceptibility mapping using analytical hierarchy process (AHP) and certainty factor (CF) models for the Baozhong region of Baoji City,China[J]. Environmental Earth Sciences,2016,75(1):63. doi: 10.1007/s12665-015-4795-7
[15]	刘璐瑶,高惠瑛,李照. 基于CF与Logistic回归模型耦合的永嘉县滑坡易发性评价[J]. 中国海洋大学学报(自然科学版),2021,51(10):121 − 129. [LIU Luyao,GAO Huiying,LI Zhao. Landslide susceptibility assessment based on coupling of CF model and logistic regression model in Yongjia County[J]. Periodical of Ocean University of China,2021,51(10):121 − 129. (in Chinese)
[16]	国家市场监督管理总局, 国家标准化管理委员会. 地质灾害危险性评估规范: GB/T 40112—2021[S]. 北京: 中国标准出版社, 2021 Standardization Administration of the People's Republic of China. Specifications for risk assessment of geological hazard: GB/T 40112—2021[S]. Beijing: Standards Press of China, 2021. （in Chinese）
[17]	杨先全,周苏华,邢静康,等. 肯尼亚滑坡灾害分布特征及敏感性区划[J]. 中国地质灾害与防治学报,2019,30(5):65 − 74. [YANG Xianquan,ZHOU Suhua,XING Jingkang,et al. Distribution patterns and susceptibility mapping of landslides in Kenya[J]. The Chinese Journal of Geological Hazard and Control,2019,30(5):65 − 74. (in Chinese)
[18]	朱丛瑞. 肯尼亚建国后的环境问题研究[D]. 昆明: 云南师范大学 ZHU Congrui. Research on environmental problems after the founding of Kenya[D]. Kunming: Yunnan Normal University. （in Chinese with English abstract）
[19]	肯尼亚信息、通信和技术部.滑坡统计资料[M/OL]. [2015-5-18]. https://www.ict.go.ke/wp-content/uploadsKenya. Ministry of Information, Communication and Technology of Kenya. Landslide Statistics[M/OL]. [2015-5-18]. https://www.ict.go.ke/wp-content/uploadsKenya
[20]	全球毁灭性滑坡数据库.肯尼亚地区滑坡数据[M/OL]. [2019-6-18]. https://blogs.agu.org/landslideblog/. Global Database of Fatal Landslides.Kenya Historical Landslide Data[M/OL]. [2019-6-18]. https://blogs.agu.org/landslideblog/.
[21]	BROECKX J,VANMAERCKE M,DUCHATEAU R,et al. A data-based landslide susceptibility map of Africa[J]. Earth-Science Reviews,2018,185:102 − 121. doi: 10.1016/j.earscirev.2018.05.002
[22]	NASA 地球公开数据. ASTERGDEMV2卫星全球高程公开共享数据[M/OL]. [2013-11-30].https://visibleearth.nasa.gov/. NASA Earth Open Data. ASTERGDEMV2 satellite global elevation data[M/OL]. [2013-11-30]. https://visibleearth.nasa.gov/.
[23]	开放非洲数据库. 肯尼亚地区地表径流、地貌、年降雨量等共享数据[M/OL] 4-1-20]. www. Openafrica. org. [Open Africa Database. Shared data on surface runoff, landforms, and annual rainfall in Kenya [M/OL]. [2014-1-20]. www. Openafrica. org
[24]	世界资源研究所公开数据. 肯尼亚地区土地利用类型共享数据[M/OL] 4-1-20]. www. wri. org. [World Resources Institute Open Data. Shared data on land use types in Kenya [M/OL]. [2014-1-20]. www. wri. org
[25]	CHENG Yongzhen,HUANG Xiaoming. Effect of mineral additives on the behavior of an expansive soil for use in highway subgrade soils[J]. Applied Sciences,2018,9(1):30. doi: 10.3390/app9010030
[26]	SWETS J A. Measuring the accuracy of diagnostic systems[J]. Science,1988,240(4857):1285 − 1293. doi: 10.1126/science.3287615