基于机器学习的滑坡崩塌地质灾害气象风险预警研究

李阳春; 刘黔云; 李潇; 顾天红; 张楠

doi:10.16031/j.cnki.issn.1003-8035.2021.00-15

基于机器学习的滑坡崩塌地质灾害气象风险预警研究

1.
贵州省地质环境监测院，贵州贵阳　550001

2.
贵州省气象台，贵州贵阳　550001

详细信息

作者简介: 李阳春（1983-），男，湖北武汉人，硕士，高级工程师，研究方向为地质灾害综合防治。E-mail：82066240@qq.com

通讯作者: 张　楠（1987-），男，重庆綦江人，本科，高级工程师，研究方向为地质灾害气象风险预警。E-mail：675400947@qq.com

中图分类号: TP694

收稿日期: 2021-03-29

修回日期: 2021-05-25

刊出日期: 2021-06-25

Exploring early warning and forecasting of meteorological risk of landslide and rockfall induced by meteorological factors by the approach of machine learning

1.
Guizhou Geological Environment Monitoring Institue, Guiyang, Guizhou　550001, China

2.
Meteorological Observatory of Guizhou, Guiyang, Guizhou　550001, China

More Information

Corresponding author: ZHANG Nan, 675400947@qq.com

Received Date: 29 March 2021

Revised Date: 25 May 2021

Publish Date: 25 June 2021

摘要

摘要:
在划分气象风险等级时，传统地质灾害气象风险预警方法忽略了承灾体脆弱性因素，且气象风险预报等级整体偏高，导致高等级风险区空报率较高。基于此，提出基于机器学习的滑坡、崩塌灾害气象风险预警方法。利用信息量法，分析气象因素影响程度。选取坐标点、降雨量、易发生等级，将其作为机器学习人工神经网络的输入节点，判断是否发生崩塌、滑坡灾害；针对地质灾害区域，根据影响程度计算气象引发因子指数，结合滑坡、崩塌灾害潜势度G和承灾体脆弱性M，确定气象风险预警指数R，划分预警级别，完成滑坡、崩塌灾害气象风险预警。实验结果表明，设计方法有效降低了三级预报和四级预警空报率，提升了预警精细化程度。
- 机器学习 /
- 地质灾害；滑坡灾害；崩塌灾害 /
- 气象风险 /
- 预警
Abstract:
In the traditional methods of meteorological risk early warning and forecasting, the vulnerability factors of disaster bearing bodies are ignored when classifying the meteorological risk level, and the meteorological risk prediction level is relatively high, which leads to the high air report rate in high-level risk areas. Based on this, a method of early warning and forecasting of meteorological risk of landslide and collapse geological disasters based on machine learning is proposed. By using the information quantity method, the influence degree of meteorological factors is analyzed, and the coordinate point, rainfall and prone level are selected as input nodes of machine learning artificial neural network to judge whether geological disaster occurs; for the area of ground damage, the meteorological cause sub index is calculated according to the influence degree. Combined with the potential degree of geological disaster and vulnerability of disaster bearing body, the meteorological risk warning index is determined, divide the warning and forecast level, and complete the forecast of geological disaster meteorological risk. The experimental results show that the proposed method can effectively reduce the three-level forecast air report rate and the fourth level air alarm rate, and improve the precision of the early warning forecast.
- machine learning /
- geological hazard /
- landslide; rockfall; meteorological risk /
- early warning and forecastin

HTML全文

图 1 贵州省地质灾害易发区分布示意图

Figure 1.

下载: 全尺寸图片幻灯片

图 2 滑坡、崩塌灾害机器学习神经网络结构

Figure 2.

下载: 全尺寸图片幻灯片

图 3 贵州省降水量变化

Figure 3.

下载: 全尺寸图片幻灯片

图 4 崩塌预警结果

Figure 4.

下载: 全尺寸图片幻灯片

图 5 滑坡预警结果

Figure 5.

下载: 全尺寸图片幻灯片

表 1 滑坡、崩塌灾害高易发区气象风险预警级别

Table 1. Early warning level of meteorological risk in high areas prone to geological disasters

累积降水 /mm	预报小雨 0.01～10	预报中雨 10～25	预报大雨 25～50	预报暴雨 50～100	预报大暴雨 ≥100
≤30		蓝色	黄色	橙色	红色
30～50	蓝色	黄色	橙色	红色	红色
50～100	黄色	橙色	红色	红色	红色
≥100	橙色	红色	红色	红色	红色

下载: 导出CSV

表 2 滑坡、崩塌灾害中易发区气象风险预警级别

Table 2. Warning level of meteorological risk in areas prone to geological disasters

累积降水 /mm	预报小雨 0.01～10	预报中雨 10～25	预报大雨 25～50	预报暴雨 50～100	预报大暴雨 ≥100
≤30			蓝色	黄色	橙色
30～50		蓝色	黄色	橙色	红色
50～100	蓝色	黄色	橙色	红色	红色
≥100	黄色	橙色	红色	红色	红色

下载: 导出CSV

表 3 滑坡、崩塌灾害低易发区气象风险预警级别

Table 3. Early warning level of meteorological risk in low areas prone to geological disasters

累积降水 /mm	预报小雨 0.01～10	预报中雨 10～25	预报大雨 25～50	预报暴雨 50～100	预报大暴雨 ≥100
≤30				蓝色	黄色
30～50			蓝色	黄色	橙色
50～100		蓝色	黄色	橙色	红色
≥100	蓝色	黄色	橙色	红色	红色
注：其中预报降水为24 h预报降雨量，累积降水为最近五天累计降雨量。

下载: 导出CSV

表 4 贵州省当日临界雨量和5日临界雨量

Table 4. Critical rainfall and mm rainfall of 5 th Day of Guizhou Province

	灾害易发区域	一级	二级	三级	四级
当日临界雨量 /m	不易发区	92	55	37	28
	低易发区	110	67	45	34
	中易发区	132	79	53	40
	高易发区	254	151	101	76
5日临界雨量 /m	不易发区	223	133	89	67
	低易发区	243	157	103	77
	中易发区	262	157	105	79
	高易发区	304	181	121	91

下载: 导出CSV

表 5 贵州省典型地质灾害统计数据

Table 5. Statistical data of typical geological disasters in Guizhou Province

灾害点类型	灾害点数量/个	分布市镇数量/个	占灾害点总数比例/%
滑坡	1032	58	85.7%
崩塌	111	19	9.2%
泥石流	29	12	2.4%
地面塌陷	25	8	2.1%
地裂缝	7	3	0.5%

下载: 导出CSV

表 6 崩塌预警空报率

Table 6. Empty reporting rate of collapse early warning and forecast

	设计方法	常规方法1	常规方法2
一级预报/%	0	0	0
二级预报/%	0	0	0
三级预报/%	8.27	14.92	17.92
四级预警/%	7.26	13.29	19.26

下载: 导出CSV

表 7 滑坡预警空报率

Table 7. Empty reporting rate of landslide early warning and forecast

	设计方法	常规方法1	常规方法2
一级预报/%	0	0	0
二级预报/%	0	0	1.21
三级预报/%	9.92	14.96	16.92
四级预警/%	6.12	14.63	17.29

下载: 导出CSV

参考文献(17)

[1]	申欣凯, 吕义清, 张静, 等. 基于ArcGIS的山西省暴雨灾害风险评估[J]. 国土资源科技管理,2020,37(1):61 − 73. [SHEN Xinkai, LYU Yiqing, ZHANG Jing, et al. Risk assessment of rainstorm disaster in Shanxi Province based on ArcGIS[J]. Scientific and Technological Management of Land and Resources,2020,37(1):61 − 73. (in Chinese with English abstract)
[2]	罗贵东, 于竹娟, 张驹. 基于概率分布的达州市地质灾害发生频次与降水量级的关系[J]. 成都信息工程大学学报,2020,35(6):697 − 703. [LUO Guidong, YU Zujuan, ZHANG Ju. The relationship between the frequency of geological disasters and the precipitation in Dazhou based on probability distribution[J]. Journal of Chengdu University of Information Technology,2020,35(6):697 − 703. (in Chinese with English abstract)
[3]	李宇梅, 杨寅, 狄靖月, 等. 全国地质灾害气象风险精细化网格预报方法及其应用[J]. 气象,2020,46(10):1310 − 1319. [LI Yumei, YANG Yin, DI Jingyue, et al. Meteorological risk assessment method of geological disaster in China and its mesh refinement application[J]. Meteorological Monthly,2020,46(10):1310 − 1319. (in Chinese with English abstract) doi: 10.7519/j.issn.1000-0526.2020.10.005
[4]	兰腾达. 宁德市地质灾害气象风险预警系统应用与展望[J]. 福建地质,2020,39(2):155 − 160. [LAN Tengda. Application and prospect of early warning of meteorological risk system for geological hazards, Ningde City[J]. Geology of Fujian,2020,39(2):155 − 160. (in Chinese with English abstract) doi: 10.3969/j.issn.1001-3970.2020.02.008
[5]	薛俊卓, 詹辉, 张国良, 等. 长输油气管道河沟段水毁灾害气象风险评价研究[J]. 安全与环境工程,2020,27(2):168 − 174. [XUE Junzhuo, ZHAN Hui, ZHANG Guoliang, et al. Meteorological risk assessment of river channel washout disaster of long-distance oil and gas pipelines[J]. Safety and Environmental Engineering,2020,27(2):168 − 174. (in Chinese with English abstract)
[6]	黎力, 朱永亮, 陈亮, 等. 重庆市地质灾害气象风险预警问题研究[J]. 安徽农业科学,2019,47(24):234 − 236. [LI Li, ZHU Yongliang, CHEN Liang, et al. Study on geological hazard meteorological risk early warning in Chongqing[J]. Journal of Anhui Agricultural Sciences,2019,47(24):234 − 236. (in Chinese with English abstract) doi: 10.3969/j.issn.0517-6611.2019.24.069
[7]	罗鸿东, 李瑞冬, 张勃, 等. 基于信息量法的地质灾害气象风险预警模型:以甘肃省陇南地区为例[J]. 地学前缘,2019,26(6):289 − 297. [LUO Hongdong, LI Ruidong, ZHANG Bo, et al. An early warning model system for predicting meteorological risk associated with geological disasters in the Longnan area, Gansu Province based on the information value method[J]. Earth Science Frontiers,2019,26(6):289 − 297. (in Chinese with English abstract)
[8]	杨竹云, 李华宏, 胡娟, 等. 昭通市地质灾害特征及气象风险预警模型研究[J]. 云南大学学报(自然科学版),2019,41(4):753 − 764. [YANG Zuyun, LI Huahong, HU Juan, et al. Research on features of geological hazard and meteorological risk early-warning model in Zhaotong[J]. Journal of Yunnan University(Natural Sciences Edition),2019,41(4):753 − 764. (in Chinese with English abstract)
[9]	方然可, 刘艳辉, 苏永超, 等. 基于逻辑回归的四川青川县区域滑坡灾害预警模型[J]. 水文地质工程地质,2021,48(1):181 − 187. [FANG Ranke, LIU Yanhui, SU Yongchao, et al. A early warning model of regional landslide in Qingchuan County, Sichuan Province based on logistic regression[J]. Hydrogeology & Engineering Geology,2021,48(1):181 − 187. (in Chinese with English abstract)
[10]	杨寅, 包红军, 彭涛. 台风“鲇鱼”强降水引发的地质灾害气象风险预警检验与分析[J]. 暴雨灾害,2019,38(3):221 − 228. [YANG Yin, BAO Hongjun, PENG Tao. Verification and analysis of meteorological early warning of geological hazards during precipitation of Typhoon "MEGI"[J]. Torrential Rain and Disasters,2019,38(3):221 − 228. (in Chinese with English abstract)
[11]	包红军, 曹勇, 林建, 等. 山洪灾害气象预警业务技术进展[J]. 中国防汛抗旱,2020,30(1):40 − 47. [BAO Hongjun, CAO Yong, LIN Jian, et al. Operational technology advances in meteorological early warning for flash flood disasters[J]. China Flood & Drought Management,2020,30(1):40 − 47. (in Chinese with English abstract)
[12]	方楠, 黄清瀚, 丁雨鑫, 等. 浙江省近3 a气象类灾害预警信息特征[J]. 浙江农业科学,2020,61(6):1246 − 1250. [FANG Nan, HUANG Qinghan, DING Yuxin, et al. The characteristics of meteorological disaster early warning information in Zhejiang Province in recent 3 years[J]. Journal of Zhejiang Agricultural Sciences,2020,61(6):1246 − 1250. (in Chinese with English abstract)
[13]	高树静, 董廷坤, 王程龙. 基于机器学习的车道线检测系统仿真与优化[J]. 计算机仿真,2020,37(2):140 − 143. [GAO Shujing, DONG Tingkun, WANG Chenglong. Simulation and optimization of lane detection system based on machine learning[J]. Computer Simulation,2020,37(2):140 − 143. (in Chinese with English abstract) doi: 10.3969/j.issn.1006-9348.2020.02.029
[14]	李爽爽. 精细化暴雨监测预报及风险预警系统的开发与应用[J]. 浙江气象,2020,41(3):29 − 35. [LI Shuangshuang. Development and application of refined rainstorm monitoring and forecasting and risk early warning system[J]. Journal of Zhejiang Meteorology,2020,41(3):29 − 35. (in Chinese with English abstract)
[15]	刘云, 康卉君. 江西崩塌滑坡泥石流灾害空间时间分布特征分析[J]. 中国地质灾害与防治学报,2020,31(4):107 − 112. [LIU Yun, KANG Huijun. Spatial-temporal distribution of landslide, rockfall and debris flow hazards in Jiangxi Province[J]. The Chinese Journal of Geological Hazard and Control,2020,31(4):107 − 112. (in Chinese with English abstract)
[16]	程素珍, 路璐, 翟淑花, 等. 2004—2018年北京市突发地质灾害时空分布特点和监测预警状况[J]. 中国地质灾害与防治学报,2020,31(6):38 − 46. [CHENG Suzhen, LU Lu, ZHAI Shuhua, et al. Temporal-spatial distribution and monitoring and early warning of sudden geological disasters in Beijing during the period of 2004 to 2018[J]. The Chinese Journal of Geological Hazard and Control,2020,31(6):38 − 46. (in Chinese with English abstract)
[17]	李滨, 殷跃平, 高杨, 等. 西南岩溶山区大型崩滑灾害研究的关键问题[J]. 水文地质工程地质,2020,27(4):5 − 13. [LI Bin, YIN Yueping, GAO Yang, et al. Critical issues in rock avalanches in the karst mountain areas of southwest China[J]. Hydrogeology & Engineering Geology,2020,27(4):5 − 13. (in Chinese with English abstract)