[1]贾 倩a,熊俊楠b*,尚依炜b,等.一种基于堆叠集成学习的山洪危险性评估方法——以汶川地震扰动区为实证[J].山地学报,2025,(2):321-334.[doi:10.16089/j.cnki.1008-2786.000895]
　JIA Qiana,XIONG Junnanb*,SHANG Yiweib,et al.An Innovative Approach to Risk Assessment of Flash Flood Based on Stacked Ensemble Learning: An Empirical Study of the Post-Shock Area of the Wenchuan Earthquake, China[J].Mountain Research,2025,(2):321-334.[doi:10.16089/j.cnki.1008-2786.000895]
点击复制

一种基于堆叠集成学习的山洪危险性评估方法——以汶川地震扰动区为实证()
分享到：

《山地学报》[ISSN:1008-2186/CN:51-1516]

卷:
期数:
2025年第2期
页码:
321-334
栏目:
山地技术
出版日期:
2025-06-25

文章信息/Info

Title:
An Innovative Approach to Risk Assessment of Flash Flood Based on Stacked Ensemble Learning: An Empirical Study of the Post-Shock Area of the Wenchuan Earthquake, China
文章编号:
1008-2786-(2025)2-321-14
作者:
贾 倩1a; 熊俊楠1b*; 尚依炜1b; 肖慧文2; 王启盛3
(1.西南石油大学 a. 地球科学与技术学院; b.土木工程与测绘学院,成都 610500; 2.四川水发勘测设计研究有限公司,成都 610500; 3.四川省第四地质大队,成都 611130)
Author(s):
JIA Qian1a;  XIONG Junnan1b*;  SHANG Yiwei1b;  XIAO Huiwen2;  WANG Qisheng3
(1. a. College of Geosciences and Technology; b. School of Civil Engineering and Geomatics, Southwest Petroleum University, Chengdu 610500, China; 2. Sichuan Water Development Investigation, Design & Research Co., Ltd., Chengdu 610500, China; 3. The 4th Geological Brigade of Sichuan, Chengdu 611130, China)
关键词:
山洪;  集成学习;  汶川地震扰动区
Keywords:
flash flood;  ensemble learning;  the post-shock area of the Wenchuan earthquake
分类号:
P954; X43
DOI:
10.16089/j.cnki.1008-2786.000895
文献标志码:
A
摘要:
山洪危险性评估是灾害链式防控体系的核心环节。利用机器学习算法进行山洪建模及评估山洪风险时,单一模型普遍存在过拟合风险、泛化能力不足等缺陷。集成模型通过模型融合可提升预测性能和模型稳健性,但其建模机理与适用性亟待深入探索与验证。本研究提出一种基于堆叠集成学习的山洪危险性评估方法,以汶川地震重灾区为实证对象,系统构建了包含随机森林(Random Forest, RF)、逻辑回归(Logistic Regression, LR)、梯度提升决策树(Gradient Boosting Decision tree, GBDT)和多层感知机(Multilayer Perceptron, MLP)的单一模型与堆叠集成模型,验证集成模型在山洪危险性评估中的对比优势。主要研究发现:(1)堆叠集成模型的准确率(90.51%)与F1分数(90.37%)显著优于单一模型,召回率与精确率差异仅1.33%,展现出更优的均衡性与综合性能。(2)集成模型的AUC值达96.87%,较RF、LR、GBDT和MLP分别提升1.55%、2.75%、2.13%和1.00%,验证了集成策略对预测精度的提升效果。(3)通过特征重要性分析揭示,距水系距离(0.047)、归一化植被指数(0.029)、坡度(0.028)构成山洪灾害的关键驱动因子,水区高差(0.027)、高程(0.017)及不透水面密度(0.016)次之。(4)在汶川地震扰动区,山洪极高与高危险区集中分布于东部低洼地带与西部高海拔河流沿岸。本研究可为地震灾区山洪风险识别提供创新方法,其构建的集成模型框架可为复杂地形区灾害评估研究提供范式参考。
Abstract:
Risk assessment of flash flood is a prerequisite of flood-related geo-disaster chain prevention and control. Although machine learning algorithms were introduced for flash flood modeling to assess their susceptibility with high efficiency, but a single model generally suffer from the defects of overfitting or insufficient generalization capacity. Some integrated model improve the prediction performance and model robustness through model fusion, but its modeling mechanism and applicability need to be deeply explored and verified.
In this study, it proposed an innovative approach to flash flood risk assessment based on Stacking algorithm, and took an empirical study of the post-shock area of the Wenchuan earthquake in 2008, china. It constructed four single machine learning models(Random Forest, RF; Logistic Regression, LR; Gradient Boosting Decision Tree, GBDT; Multilayer Perceptron, MLP)and a Stacked Ensemble(SE)model which evolved from a fusion of the four machine learning algorithms. Then it systematically validated the comparative advantages of the SE model in the process of flash flood susceptibility assessment.
(1)The stacked ensemble model significantly outperformed single models in terms of accuracy(90.51%)and F1 score(90.37%), with only a 1.33% difference between recall and precision rates, demonstrating superior balance and overall performance.
(2)The AUC(Area Under the Curve)value obtained in the SE model reached 96.87%, an improvement of 1.55%, 2.75%, 2.13%, and 1.00% compared to RF, LR, GBDT, and MLP, respectively, verifying the effectiveness of the ensemble approach in enhancing prediction accuracy.
(3)Feature importance analysis revealed that distance to water systems(0.047), Normalized Difference Vegetation Index(NDVI)(0.029), and slope(0.028)were the key driving factors for flash flood occurrences, followed by water area elevation difference(0.027), elevation(0.017), and impervious surface density(0.016).
(4)In the post-shock area of Wenchuan earthquake, the extremely high and high-risk of flash flood zones were concentrated in the eastern low-lying areas and along high-altitude rivers in the west.
This study provides an innovative approach to flash flood risk identification in earthquake-affected areas, and the ensemble model framework constructed can serve as a paradigm reference for disaster assessment research in complex terrain areas.

参考文献/References:

[1] 肖扬帆. 空间多准则决策与异构空间数据集成的洪水危险性评估[D]. 武汉: 华中科技大学, 2019: 2-10. [XIAO Yangfan. Flood hazard assessment based on spatial multi-criteria decision making and heterogeneous spatial data integration [D]. Wuhan: Huazhong University of Science and Technology, 2019: 2-10]
[2] YOUSSEF A M, PRADHAN B, SEFRY S A. Flash flood susceptibility assessment in Jeddah city(Kingdom of Saudi Arabia)using bivariate and multivariate statistical models [J]. Environmental Earth Sciences, 2016, 75(1): 12. DOI: 10.1007/s12665-015-4830-8
[3] CAO Yifan, JIA Hongliang, XIONG Junnan, et al. Flash flood susceptibility assessment based on geodetector, certainty factor, and logistic regression analyses in Fujian province, China [J]. ISPRS International Journal of Geo-Information, 2020, 9(12): 748. DOI: 10.3390/ijgi9120748
[4] 王雪梅, 郭良, 翟晓燕. 降雨时空不确定性对小流域洪峰模拟的影响评估[J]. 人民黄河, 2024, 46(4): 49-54. [WANG Xuemei, GUO Liang, ZHAI Xiaoyan. Assessing the impact of rainfall spatial-temporal uncertainty on flood peak simulation in small mountainous catchment [J]. Yellow River, 2024, 46(4): 49-54] DOI: 10.3969/j.issn.1000-1379.2024.04.008
[5] 王雨潇, 刘波, 王文鹏, 等. 基于HEC-HMS模型的三峡区间洪水模拟[J]. 长江科学院院报, 2024, 41(6): 76-83. [WANG Yuxiao, LIU Bo, WANG Wenpeng, et al. Simulation of flood in three gorges region based on HEC-HMS model [J]. Journal of Changjiang River Scientific Research Institute, 2024, 41(6): 76-83] DOI: 10.11988/ckyyb.20230047
[6] MOSAVI A, OZTURK P, CHAU K. Flood prediction using machine learning models: Literature review [J]. Water, 2018, 10(11): 1536. DOI: 10.3390/w10111536
[7] YAO Jing, ZHANG Xiaoxiang, LUO Weicong, et al. Applications of stacking/blending ensemble learning approaches for evaluating flash flood susceptibility [J]. International Journal of Applied Earth Observation and Geoinformation, 2022, 112: 102932. DOI: 10.1016/j.jag.2022.102932
[8] BUI D T, TSANGARATOS P, NGO P T T, et al. Flash flood susceptibility modeling using an optimized fuzzy rule based feature selection technique and tree based ensemble methods [J]. Science of the Total Environment, 2019, 668: 1038-1054. DOI: 10.1016/j.scitotenv.2019.02.422
[9] ARABAMERI A, SAHA S, CHEN Wei, et al. Flash flood susceptibility modelling using functional tree and hybrid ensemble techniques [J]. Journal of Hydrology, 2020, 587: 125007. DOI: 10.1016/j.jhydrol.2020.125007
[10] CHEN Jialei, HUANG Guoru, CHEN Wenjie. Towards better flood risk management: Assessing flood risk and investigating the potential mechanism based on machine learning models [J]. Journal of Environmental Management, 2021, 293: 112810. DOI: 10.1016/j.jenvman.2021.112810
[11] 彭立, 刘邵权, 刘淑珍, 等. 汶川地震重灾区10县资源环境承载力研究[J]. 四川大学学报(工程科学版), 2009, 41(3): 294-300. [PENG Li, LIU Shaoquan, LIU Shuzhen, et al. Studies on bearing capacity of resources and environment of 10 counties in area hited by Wenchuan Earthquake [J]. Journal of Sichuan University(Engineering Science Edition), 2009, 41(3): 294-300] DOI: 10.15961/j.jsuese.2009.03.043
[12] 叶帮苹, 冯汉中, 刘志红, 等. 基于Logistic模型的四川山洪流域危险性评价[J]. 成都信息工程大学学报, 2020, 35(5): 573-578. [YE Bangping, FENG Hanzhong, LIU Zhihong, et al. Risk assessment of Sichuan mountain flooding based on logistic model [J]. Journal of Chengdu University of Information Technology, 2020, 35(5): 573-578] DOI: 10.16836/j.cnki.jcuit.2020.05.016
[13] 王钧, 宇岩, 欧国强, 等. 岷江上游汶川地震重灾区山洪灾害危险分区研究[J]. 长江科学院院报, 2017, 34(1): 54-60. [WANG Jun, YU Yan, OU Guoqiang, et al. Flash flood risk zoning of areas hit by Wenchuan Earthquake in the upper reach of Minjiang River [J]. Journal of Yangtze River Scientific Research Institute, 2017, 34(1): 54-60] DOI: 10.11988/ckyyb.20150956
[14] 张圆圆, 毛爽, 张淑伟. 2000—2017年龙门山断裂带生态脆弱性演变研究[J]. 天津农业科学, 2020, 26(2): 22-28. [ZHANG Yuanyuan, MAO Shuang, ZHANG Shuwei. Evolution of ecological vulnerability in Longmenshan fault zone from 2000 to 2017 [J]. Tianjin Agricultural Sciences, 2020, 26(2): 22-28] DOI: 10.3969/j.issn.1006-6500.2020.02.006
[15] 马旭廷. 基于大数据和机器学习的山洪灾害高风险区识别研究[D]. 山东: 山东建筑大学, 2024: 39-42. [MA Xuting. Research on high-risk area identification of flash flood disasters based on big data and machine learning [D]. Shandong: Shandong Jianzhu University, 2024: 39-42] DOI: 10.27273/d.cnki.gsajc.2023.000807
[16] 石辉, 邓念东, 周阳. 随机森林赋权层次分析法的崩塌易发性评价[J]. 科学技术与工程, 2021, 21(25): 10613-10619. [SHI Hui, DENG Niandong, ZHOU Yang. Evaluation of collapse susceptibility based on random forest weighted analytic hierarchy process [J]. Science Technology and Engineering, 2021, 21(25): 10613-10619]
[17] 杜鹏, 陈宁生, 伍康林, 等. 基于随机森林模型的藏东南地区滑坡易发性评价及主控因素分析[J]. 成都理工大学学报(自然科学版), 2024, 51(2): 328-344. [DU Peng, CHEN Ningsheng, WU Kanglin, et al. Evaluation of landslide susceptibility in southeast Tibet based on a random forest model [J]. Journal of Chengdu University of Technology(Science & Technology Edition), 2024, 51(2): 328-344] DOI: 10.3969/j.issn.1671-9727.2024.02.12
[18] 高泽民, 丁明涛, 杨国辉, 等. 川藏铁路孜热—波密段泥石流灾害危险性评价[J]. 工程地质学报, 2021, 29(2): 478-485. [GAO Zemin, DING Mingtao, YANG Guohui, et al. Hazard assessment of debris flow along Zire-Bomi section of Sichuan-Tibet Railway [J]. Journal of Engineering Geology, 2021, 29(2): 478-485] DOI: 10.13544/j.cnki.jeg.2021-0160
[19] 王启盛, 熊俊楠, 程维明, 等. 耦合统计方法、机器学习模型和聚类算法的滑坡易发性评价方法[J]. 地球信息科学学报, 2024, 26(3): 620-637. [WANG Qisheng, XIONG Junnan, CHENG Weiming, et al. Landslide susceptibility mapping methods coupling with statistical methods, machine learning models and clustering algorithms [J]. Journal of Geo- information Science, 2024, 26(3): 620-637] DOI: 10.12082/ dqxxkx.2024.230427
[20] 吴小君, 方秀琴, 任立良, 等. 基于随机森林的山洪灾害风险评估——以江西省为例[J]. 水土保持研究, 2018, 25(3): 142-149. [WU Xiaojun, FANG Xiuqin, REN Liliang, et al. Risk assessment of mountain torrents disaster based on random forest: A case study in Jiangxi province [J]. Research of Soil and Water Conservation, 2018, 25(3): 142-149] DOI: 10.13869/j.cnki.rswc.2018.03.021
[21] COSTACHE R, HONG Haoyuan, PHAM Q B. Comparative assessment of the flash-flood potential within small mountain catchments using bivariate statistics and their novel hybrid integration with machine learning models [J]. Science of the Total Environment, 2020, 711: 134514. DOI: 10.1016/j.scitotenv.2019.134514
[22] HOSSEINI F S, CHOUBIN B, MOSAVI A, et al. Flash-flood hazard assessment using ensembles and Bayesian-based machine learning models: Application of the simulated annealing feature selection method [J]. Science of the Total Environment, 2020, 711: 135161. DOI: 10.1016/j.scitotenv.2019.135161
[23] FRIEDMAN J H. Greedy function approximation: A gradient boosting machine [J]. The Annals of Statistics, 2001, 29(5): 1189-1232. DOI: 10.1214/aos/1013203451
[24] LIANG Weizhang, LUO Suizhi, ZHAO Guoyan, et al. Predicting hard rock pillar stability using GBDT, XGBoost, and LightGBM algorithms [J]. Mathematics, 2020, 8(5): 765. DOI: 10.3390/math8050765
[25] 祝元丽, 冯向阳, 闫庆武, 等. 基于GBDT的望奎县农田土壤有机碳主控因子研究[J]. 中国环境科学, 2024, 44(3): 1407-1417. [ZHU Yuanli, FENG Xiangyang, YAN Qingwu, et al. Spatial distribution and main controlling factors of soil organic carbon under cultivated land based on GBDT model in black soil region of Northeast China [J]. China Environmental Science, 2024, 44(3): 1407-1417] DOI: 10.19674/j.cnki.issn1000-6923.2024.0054
[26] 范天程, 贾云飞, 李云飞, 等. 基于遥感影像与逻辑回归模型的延河流域沟壑分布概率预测[J]. 水土保持研究, 2022, 29(4): 316-321. [FAN Tiancheng, JIA Yunfei, LI Yunfei, et al. Prediction of gully distribution probability in Yanhe basin based on remote sensing image and logistic regression model [J]. Research of Soil and Water Conservation, 2022, 29(4): 316-321] DOI: 10.13869/j.cnki.rswc.20220414.007
[27] QI C R, SU Hao, NIESBNER M, et al. Volumetric and multi-view CNNs for object classification on 3D data [A]. 2016: 5648-5656. DOI: 10.1109/CVPR.2016.609
[28] 曾营, 张迎宾, 张钟远, 等. 基于X-多层感知器耦合模型的滑坡易发性评价——以贵州省松桃自治县为例[J]. 山地学报, 2023, 41(2): 280-294. [ZENG Ying, ZHANG Yingbin, ZHANG Zhongyuan, et al. Landslide susceptibility evaluation based on coupled X-multilayer perceptron model —a case study of Songtao Autonomous county of Guizhou province, China [J]. Mountain Research, 2023, 41(2): 280-294] DOI: 10.16089/j.cnki.1008-2786.000748
[29] 蒲未来, 刘敦龙, 桑学佳, 等. 融合多源异构数据的滑坡变形阶段智能判识方法[J]. 灾害学, 2023, 38(4): 179-186. [PU Weilai, LIU Dunlong, SANG Xuejia, et al. An intelligent identification method of landslide deformation stage based on multi-source heterogeneous data [J]. Journal of Catastrophology, 2023, 38(4): 179-186] DOI: 10.3969/j.issn.1000-811X.2023.04.029
[30] SMYTH P, WOLPERT D. Linearly combining density estimators via stacking [J]. Machine Learning, 1999, 36(1): 59-83. DOI: 10.1023/A:1007511322260
[31] SWETS J A. Measuring the accuracy of diagnostic systems [J]. Science, 1988, 240(4857): 1285-1293. DOI: 10.1126/science.3287615
[32] 中华人民共和国水利部. 中国水旱灾害公报: 2016-2022 [EB/OL]. 北京: 中华人民共和国水利部,(2024- 05-31)[2024- 08- 01]. http://www.mwr.gov.cn/sj/tjgb/zgshzhgb/. [Ministry of Water Resources of the People's Republic of China. China flood and drought disaster bulletin: 2016-2022 [EB/OL]. Beijing: Ministry of Water Resources of the People's Republic of China,(2024- 05-31)[2024- 08- 01]. http://www.mwr.gov.cn/sj/tjgb/zgshzhgb/]
[33] 涂勇, 吴泽斌, 何秉顺. 2011—2019年全国山洪灾害事件特征分析[J]. 中国防汛抗旱, 2020, 30(9/10): 22-25. [TU Yong, WU Zebin, HE Bingshun. Analysis on the characteristics of flash flood disasters in China from 2011 to 2019 [J]. China Flood & Drought Management, 2020, 30(9/10): 22-25] DOI: 10.16867/j.issn.1673-9264.2020239
[34] 时开鑫, 陈跃红, 张晓祥, 等. 基于图聚类神经网络的江西省山洪灾害危险性区划研究[J]. 地理与地理信息科学, 2023, 39(3): 7-15. [SHI Kaixin, CHEN Yuehong, ZHANG Xiaoxiang, et al. Flash flood hazard regionalization based on graph clustering neural network in Jiangxi province, China [J]. Geography and Geo-Information Science, 2023, 39(3): 7-15] DOI: 10.3969/j.issn.1672-0504.2023.03.002
[35] FERNANDEZ D S, LUTZ M A. Urban flood hazard zoning in Tucumán province, Argentina, using GIS and multicriteria decision analysis [J]. Engineering Geology, 2010, 111(1): 90-98. DOI: 10.1016/j.enggeo.2009.12.006
[36] CHAPI K, SINGH V P, Shirzadi A, et al. A novel hybrid artificial intelligence approach for flood susceptibility assessment [J]. Environmental Modelling & Software, 2017, 95: 229-245. DOI: 10.1016/j.envsoft.2017.06.012
[37] TEHRANY M S, PRADHAN B, JEBUR M N. Spatial prediction of flood susceptible areas using rule based decision tree(DT)and a novel ensemble bivariate and multivariate statistical models in GIS [J]. Journal of Hydrology, 2013, 504: 69-79. DOI: 10.1016/j.jhydrol.2013.09.034
[38] MACMILLAN R A, SHARY P A. Chapter 9 Landforms and landform elements in geomorphometry [M]// HENGL T, REUTER H I. Developments in soil science. Amsterdam: Elsevier, 2009, 33: 227-254. DOI: 10.1016/S0166-2481(08)00009-3
[39] WANG Yi, HONG Haoyuan, CHEN Wei, et al. Flood susceptibility mapping in Dingnan county(China)using adaptive neuro-fuzzy inference system with biogeography based optimization and imperialistic competitive algorithm [J]. Journal of Environmental Management, 2019, 247: 712-729. DOI: 10.1016/j.jenvman.2019.06.102
[40] POUDYAL C P, CHANG Chandong, OH H J, et al. Landslide susceptibility maps comparing frequency ratio and artificial neural networks: A case study from the Nepal Himalaya [J]. Environmental Earth Sciences, 2010, 61(5): 1049-1064. DOI: 10.1007/s12665-009-0426-5
[41] CAO Chen, XU Peihua, WANG Yihong, et al. Flash flood hazard susceptibility mapping using frequency ratio and statistical index methods in coalmine subsidence areas [J]. Sustainability, 2016, 8(9): 948. DOI: 10.3390/su8090948
[42] KAISER M, GUNNEMANN S, DISSE M. Regional-scale prediction of pluvial and flash flood susceptible areas using tree-based classifiers [J]. Journal of Hydrology, 2022, 612: 128088. DOI: 10.1016/j.jhydrol.2022.128088
[43] 祁靓雯. 基于GIS技术的紫阳县山洪灾害风险评价[D]. 西安: 长安大学, 2018: 40-41. [QI Liangwen. Mountain flood risk assessment of Ziyang county based on GIS [D]. Xi'an: Chang'an University, 2018: 40-41]
[44] 熊俊楠, 曹依帆, 程维明, 等. 福建省山洪灾害危险性评价[J]. 山地学报, 2019, 37(4): 538-550. [XIONG Junnan, CAO Yifan, CHENG Weiming, et al. Risk assessment of mountain torrent disasters in Fujian province, China [J]. Mountain Research, 2019, 37(4): 538-550] DOI: 10.16089/j.cnki.1008-2786.000446

相似文献/References:

[1]谢湘平,韦方强,谢涛,等.山洪中漂木在拦砂坝前堵塞堆积实验[J].山地学报,2014,(02):249.
　XIE Xiangping,WEI Fangqiang,XIE Tao,et al.Experiment on the Clogging and Deposition of Woody Debris Flowing with Torrents in Front of Debris Dams[J].Mountain Research,2014,(2):249.
[2]谢洪,钟敦伦,王士革,等.1995年康定城区洪灾成因分析[J].山地学报,1997,(02):129.
[3]陶云,唐川,寸灿琼,等.2004-07-05云南德宏州山洪泥石流气象成因分析[J].山地学报,2005,(01):53.
[4]谢湘平,王小军,闫春岭.漂木灾害研究现状及研究展望[J].山地学报,2020,(4):552.[doi:10.16089/j.cnki.1008-2786.000533]
　XIE Xiangping,WANG Xiaojun,YAN Chunling.A Review of the Research on Woody Debris Related Disaster and Its Prospect[J].Mountain Research,2020,(2):552.[doi:10.16089/j.cnki.1008-2786.000533]

备注/Memo

备注/Memo:
收稿日期(Received date): 2024-08-26; 改回日期(Accepted date):2025-01-23
基金项目(Foundation item): 国家重点研发计划专项课题(2023YFC3006701); 四川省科技厅重点研发项目(2024YFHZ0134)。[National Key R&D Program of China(2023YFC3006701); Key R&D Project of Sichuan Science and Technology Department(2024YFHZ0134)]
作者简介(Biography): 贾倩(1998-),女,湖北黄石人,硕士研究生,主要研究方向:机器学习与灾害风险评估。[JIA Qian(1998-), female, born in Huangshi, Hubei Province, M.Sc. candidate, research on machine learning and disaster risk assessment] E-mail: 202221000135@stu.swpu.edu.cn
*通讯作者(Corresponding author): 熊俊楠(1981-),男,博士,教授,主要研究方向:地理信息系统与灾害风险分析。[XIONG Junnan(1981-), male, Ph.D., professor, specialized in GIS and disaster risk analysis]E-mail: neu_xjn@163.com
更新日期/Last Update: 2025-03-30