Classification Assessment of Debris Flow Susceptibility in Bailong River Basin Based on PU-Bagging Negative Sampling
-
摘要:
为提高白龙江流域泥石流易发分区的可靠性，建立随机森林为基学习器的PU-Bagging负样本采样模型.选取高程、降水量等评价因子，使用逻辑回归、随机森林、支持向量机和XGBoost算法，构建白龙江流域泥石流易发性评价模型.根据混淆矩阵衍生的评价指标、ROC曲线和5种分级方法，对比分析了4种模型的性能，并利用SHAP分析评价因子对模型的贡献程度.结果表明：（1）支持向量机模型结合几何间隔分级方法的灾害识别精度提升了24%.（2）随机森林模型能够识别更多的潜在泥石流样本，而XGBoost模型可减少对非灾害样本的误判.（3）SHAP值对高程变化的敏感性间接反映了高差对泥石流发育的重要性.本研究可以为白龙江流域新型城镇化建设与泥石流防治工程的规划提供数据支撑.
Abstract:To improve the reliability of the debris flow-prone zones in the Bailong River Basin, a PU-Bagging negative sampling model based on the random forest as the base learner is established. Evaluation factors such as elevation and precipitation were selected, and logistic regression, random forest, support vector machine and XGBoost algorithms were used to construct an evaluation model for the susceptibility of debris flows in the Bailong River Basin. Based on the evaluation indicators derived from the confusion matrix, the ROC curve and five classification methods, the performances of the four models were compared and analyzed, and the contribution degree of the evaluation factors to the model was analyzed by using SHAP. The results show follows. (1) The disaster identification accuracy of the support vector machine model combined with the geometric interval classification method has increased by 24%. (2) The random forest model can identify more potential debris flow samples, while the XGBoost model can reduce the misjudgment of non-disaster samples. (3) The sensitivity of SHAP values to elevation changes indirectly reflects the importance of height differences for the development of debris flows. This research can provide data support for the planning of the new urbanization construction and debris flow prevention and control project in the Bailong River Basin.
-
Key words:
- Bailong River Basin /
- negative sampling /
- susceptibility /
- debris flow /
- classification method /
- engineering geology
-
图 1 研究区位置及泥石流分布
Fig. 1. Location of the study area and the distribution of debris flow
图 4 基于随机森林的PU-Bagging负样本采样方法
Fig. 4. PU-Bagging negative sampling method based on random forest
图 8 基于几何间隔划分的白龙江区域易发性评价
a.RF模型；b.LR模型；c.SVM模型；d.XGBoost模型
Fig. 8. Evaluation of Bailong River area based on geometric interval division
图 9 易发性分区统计
a.自然间断点分级法；b.分位数分级法；c.相等间隔分级法；d.几何间隔分级法；e.定义间隔分级法
Fig. 9. Statistics of the prone partition
图 10 不同模型的评价因子SHAP值全局分布
a.RF模型SHAP值分布；b.LR模型SHAP值分布；c.SVM模型SHAP值分布；d.XGBoost模型SHAP值分布
Fig. 10. Global distribution of evaluation factor SHAP values for different models
图 11 甘家沟泥石流易发性概况
a.甘家沟流域物源区易发性分区图；b.甘家沟流域构造断裂图；c.泥石流流域降雨监测布设和拦挡坝图；d.2022年5月至2023年7月甘家沟流量统计结果
Fig. 11. Overview of the occurrence of Ganjiagou debris flow
表 1 泥石流评价因子数据来源
Table 1. Data source of debris flow evaluation factor
评价因子 数据类型 数据来源 高程、坡度、坡向 连续型 地理空间数据云( https://www.gscloud.cn )地层岩性 离散型 国际土壤参考与信息中心( https://www.isric.org/ )不稳定边坡核密度 连续型 全球灾害数据平台( https://www.gddat.cn )降水量 连续型 星图云开放平台( https://open.geovisearth.com )河网密度 连续型 全国地理信息资源目录服务系统( https://www.webmap.cn )地形湿度指数 连续型 地理空间数据云( https://www.gscloud.cn )归一化植被指数 连续型 MODIS植被指数产品(2000—2020年)( https://modis.gsfc.nasa.gov/data/dataprod/mod13.php )土地利用类型 离散型 星图云开放平台( https://open.geovisearth.com )距道路距离 连续型 国家冰川冻土沙漠科学数据中心( http://www.ncdc.ac.cn )
下载: 导出CSV
表 2 负样本采样模型测试集混淆矩阵
Table 2. Confusion matrix of the negative sampling model test set
真实情况 预测结果 泥石流 非泥石流 泥石流 58 6 非泥石流 7 64
下载: 导出CSV
-
Breiman, L., 2001. Random Forests. Machine Learning, 45(1): 5-32. https://doi.org/10.1023/A:1010933404324 Cao, S. A., Guo, Z., Chen, J. L., 2025. Geological Hazard Susceptibility Evaluation Based on Improved Information Model: A Case Study of the G219 National Highway in Zayu County, Xizang. Geological Bulletin of China, 44(1): 185-200(in Chinese with English abstract). Chao, Z. Z., 2023. Study on Erosion Characteristics of Different Cover Slopes in High Debris Flow Area of Bailong River Basin (Dissertation). Lanzhou University, Lanzhou (in Chinese with English abstract). Deng, M. D., Ju, N. P., Wu, T. W., et al., 2024. Evaluation of Susceptibility under Different Landslide Sample Points and Polygonal Expression Modes. Earth Science, 49(5): 1565-1583(in Chinese with English abstract). Du, G. L., Yang, Z. H., Yuan, Y., et al., 2021. Landslide Susceptibility Mapping in the Sichuan-Tibet Traffic Corridor Using Logistic Regression-Information Value Method. Hydrogeology & Engineering Geology, 48(5): 102-111(in Chinese with English abstract). Du, G. L., Zhang, Y. S., Iqbal, J., et al., 2017. Landslide Susceptibility Mapping Using an Integrated Model of Information Value Method and Logistic Regression in the Bailongjiang Watershed, Gansu Province, China. Journal of Mountain Science, 14(2): 249-268. https://doi.org/10.1007/s11629-016-4126-9 Esper Angillieri, M. Y., 2020. Debris Flow Susceptibility Mapping Using Frequency Ratio and Seed Cells, in a Portion of a Mountain International Route, Dry Central Andes of Argentina. CATENA, 189: 104504. https://doi.org/10.1016/j.catena.2020.104504 Gu, T. F., Duan, P., Wang, M. G., et al., 2024. Effects of Non-Landslide Sampling Strategies on Machine Learning Models in Landslide Susceptibility Mapping. Scientific Reports, 14(1): 7201. https://doi.org/10.1038/s41598-024-57964-5 Huang, F. M., Zhang, J., Zhou, C. B., et al., 2020. A Deep Learning Algorithm Using a Fully Connected Sparse Autoencoder Neural Network for Landslide Susceptibility Prediction. Landslides, 17(1): 217-229. https://doi.org/10.1007/s10346-019-01274-9 Huang, Q. L., Chen, W., Fu, X. D., 2018. AHP-RBF Assessment Model of Regional Debris Flow Hazard Supported by Unit Slope. Journal of Zhejiang University (Engineering Science), 52(9): 1667-1675(in Chinese with English abstract). Huang, Y., Zhao, L., 2018. Review on Landslide Susceptibility Mapping Using Support Vector Machines. CATENA, 165: 520-529. https://doi.org/10.1016/j.catena.2018.03.003 Kong, J. X., Zhuang, J. Q., Peng, J. B., et al., 2023. Evaluation of Landslide Susceptibility in Chinese Loess Plateau Based on Ⅳ-RF and Ⅳ-CNN Coupling Models. Earth Science, 48(5): 1711-1729(in Chinese with English abstract). Kumar, D., Thakur, M., Dubey, C. S., et al., 2017. Landslide Susceptibility Mapping & Prediction Using Support Vector Machine for Mandakini River Basin, Garhwal Himalaya, India. Geomorphology, 295: 115-125. https://doi.org/10.1016/j.geomorph.2017.06.013 Li, K., Zhao, J. S., Lin, Y. L., et al., 2022. Assessment of Debris Flow Susceptibility Based on Different Slope Unit Division Methods and BP Neural Network. Bulletin of Surveying and Mapping, (8): 68-74(in Chinese with English abstract). Li, Y. X., Zhang, Y., Su, X. J., et al., 2021. Early Identification and Characteristics of Potential Landslides in the Bailong River Basin Using InSAR Technique. National Remote Sensing Bulletin, 25(2): 677-690(in Chinese with English abstract). doi: 10.11834/jrs.20210094 Li, Z., Chen, N. S., Hou, R. N., et al., 2024. Susceptibility Assessment of Debris Flow Disaster Based on Machine Learning Models in the Loess Area along Yili Valley. The Chinese Journal of Geological Hazard and Control, 35(3): 129-140(in Chinese with English abstract). Lin, X. X., Xiao, G. R., Zhou, H. B., 2023. Landslide Susceptibility Assessment Method Considering Land Use Dynamic Change. Journal of Geo-Information Science, 25(5): 953-966(in Chinese with English abstract). Liu, B., Guo, H. X., Li, J. L., et al., 2024. Application and Interpretability of Ensemble Learning for Landslide Susceptibility Mapping along the Three Gorges Reservoir Area, China. Natural Hazards, 120(5): 4601-4632. https://doi.org/10.1007/s11069-023-06374-3 Liu, C. Z., Wang, J. X., 2024. Research on Classification of Collapse, Landslide and Debris Flow Disaster Chains. Journal of Engineering Geology, 32(5): 1573-1596(in Chinese with English abstract). Liu, G. D., Qin, S. W., Meng, F. Q., et al., 2023. Application of Geographic Information Similarity Based Absence Sampling Method to Debris Flow Susceptibility Mapping. Journal of Engineering Geology, 31(2): 526-537(in Chinese with English abstract). Liu, J., Li, S. L., Chen, T., 2018. Landslide Susceptibility Assesment Based on Optimized Random Forest Model. Geomatics and Information Science of Wuhan University, 43(7): 1085-1091(in Chinese with English abstract). Lv, L., Chen, T., Dou, J., et al., 2022. A Hybrid Ensemble-Based Deep-Learning Framework for Landslide Susceptibility Mapping. International Journal of Applied Earth Observation and Geoinformation, 108: 102713. https://doi.org/10.1016/j.jag.2022.102713 Martinello, C., Cappadonia, C., Conoscenti, C., et al., 2021. Optimal Slope Units Partitioning in Landslide Susceptibility Mapping. Journal of Maps, 17(3): 152-162. https://doi.org/10.1080/17445647.2020.1805807 Merghadi, A., Yunus, A. P., Dou, J., et al., 2020. Machine Learning Methods for Landslide Susceptibility Studies: A Comparative Overview of Algorithm Performance. Earth-Science Reviews, 207: 103225. https://doi.org/10.1016/j.earscirev.2020.103225 Qing, F., Zhao, Y., Chong, Y., et al., 2024. Temporal and Spatial Regularity of Debris Flow Outbreak in Bailong River Basin and Hazard Prediction of River Blocking Disasters. Journal of Lanzhou University (Natural Sciences), 60(4): 488-493(in Chinese with English abstract). Rao, S. S., Leng, X. P., 2024. Debris Flow Susceptibility Evaluation of Liangshan Prefecture Based on the RSIV-RF Model. Bulletin of Geological Science and Technology, 43(1): 275-287(in Chinese with English abstract). Sun, D. L., Gu, Q. Y., Wen, H. J., et al., 2023. Assessment of Landslide Susceptibility along Mountain Highways Based on Different Machine Learning Algorithms and Mapping Units by Hybrid Factors Screening and Sample Optimization. Gondwana Research, 123: 89-106. https://doi.org/10.1016/j.gr.2022.07.013 Sun, D. L., Xu, J. H., Wen, H. J., et al., 2021. Assessment of Landslide Susceptibility Mapping Based on Bayesian Hyperparameter Optimization: A Comparison between Logistic Regression and Random Forest. Engineering Geology, 281: 105972. https://doi.org/10.1016/j.enggeo.2020.105972 Tian, Y., Gao, B., Yin, H., et al., 2024. Handling Imbalanced Samples in Landslide Susceptibility Evaluation. Hydrogeology & Engineering Geology, 51(6): 171-181(in Chinese with English abstract). Bui, D. T., Tuan, T. A., Klempe, H., et al., 2016. Spatial Prediction Models for Shallow Landslide Hazards: A Comparative Assessment of the Efficacy of Support Vector Machines, Artificial Neural Networks, Kernel Logistic Regression, and Logistic Model Tree. Landslides, 13(2): 361-378. https://doi.org/10.1007/s10346-015-0557-6 Wang, Y., Cao, Y., Xu, F. D., et al., 2024. Reservoir Landslide Susceptibility Prediction Considering Non-Landslide Sampling and Ensemble Machine Learning Methods. Earth Science, 49(5): 1619-1635(in Chinese with English abstract). Wu, B., Shi, Z. M., Zheng, H. C., et al., 2024. Impact of Sampling for Landslide Susceptibility Assessment Using Interpretable Machine Learning Models. Bulletin of Engineering Geology and the Environment, 83(11): 461. https://doi.org/10.1007/s10064-024-03980-8 Xiong, K., Adhikari, B. R., Stamatopoulos, C. A., et al., 2020. Comparison of Different Machine Learning Methods for Debris Flow Susceptibility Mapping: A Case Study in the Sichuan Province, China. Remote Sensing, 12(2): 295. https://doi.org/10.3390/rs12020295 Yang, C., Liu, L. L., Huang, F. M., et al., 2023. Machine Learning-Based Landslide Susceptibility Assessment with Optimized Ratio of Landslide to Non-Landslide Samples. Gondwana Research, 123: 198-216. https://doi.org/10.1016/j.gr.2022.05.012 Yu, H. K., Ouyang, J. F., Wang, B. Q., et al., 2024. Susceptibility Assessment of Regional Landslides under Different Sampling Strategies. Safety and Environmental Engineering, 31(5): 122-134, 162(in Chinese with English abstract). Zhou, C., Yin, K. L., Cao, Y., et al., 2020. Landslide Susceptibility Assessment by Applying the Coupling Method of Radial Basis Neural Network and Adaboost: A Case Study from the Three Gorges Reservoir Area. Earth Science, 45(6): 1865-1876(in Chinese with English abstract). Zhou, X. Z., Wen, H. J., Zhang, Y. L., et al., 2021. Landslide Susceptibility Mapping Using Hybrid Random Forest with GeoDetector and RFE for Factor Optimization. Geoscience Frontiers, 12(5): 101211. https://doi.org/10.1016/j.gsf.2021.101211 Zhou, Y. Y., 2023. Risk Assessment and Prediction of Debris Flow Based on the Coupling Mechanism of Eco-Geological Environment in the Bailong River Basin (Dissertation). Lanzhou University, Lanzhou(in Chinese with English abstract). Zhu, H., Zhang, L. M., Xiao, T., et al., 2017. Enhancement of Slope Stability by Vegetation Considering Uncertainties in Root Distribution. Computers and Geotechnics, 85: 84-89. https://doi.org/10.1016/j.compgeo.2016.12.027 曹苏傲, 郭振, 陈佳乐, 2025. 基于改进信息量模型的地质灾害易发性评价: 以西藏察隅县G219国道沿线为例. 地质通报, 44(1): 185-200. 晁增祖, 2023. 白龙江流域泥石流高发区不同覆被坡面侵蚀特征研究(硕士学位论文). 兰州: 兰州大学. 邓明东, 巨能攀, 吴天伟, 等, 2024. 不同滑坡样本点和多边形表达模式下的易发性评价. 地球科学, 49(5): 1565-1583. doi: 10.3799/dqkx.2022.393 杜国梁, 杨志华, 袁颖, 等, 2021. 基于逻辑回归-信息量的川藏交通廊道滑坡易发性评价. 水文地质工程地质, 48(5): 102-111. 黄启乐, 陈伟, 傅旭东, 2018. 斜坡单元支持下区域泥石流危险性AHP-RBF评价模型. 浙江大学学报(工学版), 52(9): 1667-1675. 孔嘉旭, 庄建琦, 彭建兵, 等, 2023. 基于信息量和卷积神经网络的黄土高原滑坡易发性评价. 地球科学, 48(5): 1711-1729. doi: 10.3799/dqkx.2023.006 李坤, 赵俊三, 林伊琳, 等, 2022. 基于不同斜坡单元划分方法和BP神经网络的泥石流易发性评价. 测绘通报, (8): 68-74. 李媛茜, 张毅, 苏晓军, 等, 2021. 白龙江流域潜在滑坡InSAR识别与发育特征研究. 遥感学报, 25(2): 677-690. 李志, 陈宁生, 侯儒宁, 等, 2024. 基于机器学习的伊犁河谷黄土区泥石流易发性评估. 中国地质灾害与防治学报, 35(3): 129-140. 林炫歆, 肖桂荣, 周侯伯, 2023. 顾及土地利用动态变化的滑坡易发性评估方法. 地球信息科学学报, 25(5): 953-966. 刘传正, 王建新, 2024. 崩塌滑坡泥石流灾害链分类研究. 工程地质学报, 32(5): 1573-1596. 刘国栋, 秦胜伍, 孟凡奇, 等, 2023. 基于地理信息相似度的负样本采样策略在泥石流易发性评价中的应用. 工程地质学报, 31(2): 526-537. 刘坚, 李树林, 陈涛, 2018. 基于优化随机森林模型的滑坡易发性评价. 武汉大学学报(信息科学版), 43(7): 1085-1091. 庆丰, 赵岩, 种艳, 等, 2024. 白龙江流域泥石流爆发的时空规律与堵江灾害危险预测. 兰州大学学报(自然科学版), 60(4): 488-493. 饶姗姗, 冷小鹏, 2024. 基于RSIV-RF模型的凉山州泥石流易发性评价. 地质科技通报, 43(1): 275-287. 田尤, 高波, 殷红, 等, 2024. 滑坡易发性评价中样本不均衡问题处理研究. 水文地质工程地质, 51(6): 171-181. 王悦, 曹颖, 许方党, 等, 2024. 考虑非滑坡样本选取和集成机器学习方法的水库滑坡易发性预测. 地球科学, 49(5): 1619-1635. doi: 10.3799/dqkx.2022.407 于海坤, 欧阳九发, 王丙千, 等, 2024. 不同采样策略下的区域滑坡易发性评价. 安全与环境工程, 31(5): 122-134, 162. 周超, 殷坤龙, 曹颖, 等, 2020. 基于集成学习与径向基神经网络耦合模型的三峡库区滑坡易发性评价. 地球科学, 45(6): 1865-1876. doi: 10.3799/dqkx.2020.071 周妍妍, 2023. 基于生态地质环境耦合机制的白龙江流域泥石流灾害风险评价与预测(博士学位论文). 兰州: 兰州大学. -
点击查看大图
计量
- 文章访问数: 83
- HTML全文浏览量: 6
- PDF下载量: 3
- 被引次数: 0
