四川秦巴山区降雨型滑坡灾害降雨阈值  

Precipitation Threshold for Rainfall-Type Landslides in the Qinba Mountains Area,Sichuan Province,China

作　　者：王智昊 杨赛霓[1,2] 姚可桢 佟彬[5] 唐得胜 WANG Zhihao;YANG Saini;YAO Kezhen;TONG Bin;TANG Desheng(Joint International Research Laboratory of CatastropheSimulation and Systemic Risk Governance,Zhuhai 519087,Guangdong,China;School of National Safety and Emergency Management,Beijing Normal University,Beijing 100875,China;Faculty of Geographical Science,Beijing Normal University,Beijing 100875,China;Laboratory of Environmental Change and Natural Disaster,Ministry of Education,Beijing Normal University,Beijing 100875,China;China Institute of Geoenvironment Monitoring,Beijing 100081,China;Sichuan Institute of Land Space Ecological Restoration and Geological Disaster Prevention,Chengdu 610081,Sichuan,China)

机构地区：[1]北京师范大学教育部巨灾模拟与系统性风险应对国际合作联合实验室,广东珠海519087 [2]北京师范大学国家安全与应急管理学院,北京100875 [3]北京师范大学地理科学学部,北京100875 [4]北京师范大学环境演变与自然灾害教育部重点实验室,北京100875 [5]中国地质环境监测院,北京100081 [6]四川省国土空间生态修复与地质灾害防治研究院,成都610081

出　　处：《山地学报》2024年第2期238-248,共11页Mountain Research

基　　金：第二次青藏高原综合科学考察研究(2019QZKK0906)。

摘　　要：四川省秦巴山区因其复杂地形和频发的极端降雨,成为中国滑坡灾害高发区,降雨是其主要自然致灾因素之一。尽管现有区域滑坡预警系统提供了基本的监测,但在准确性方面仍有提升空间。本研究通过分析2000—2020年间1850起滑坡灾害及同期逐时降雨数据,识别影响滑坡发生的关键降雨特征,采用混淆矩阵和技能得分等评价指标,拟合筛选确定最佳阈值曲线。研究表明:(1)拟合确定了诱发滑坡灾害的最佳降雨阈值曲线,并将其应用于2021年滑坡灾害预报,预测正确率达到87.72%。(2)通过滑坡隐患点编目数据进行滑坡危险性区划,中高及高危险性区域(4级和5级)的滑坡灾害点占比51.46%,中至高危险性区域(3级至5级)的灾害点占比91.23%,显示出阈值曲线在空间预测上的有效性。最佳降雨阈值曲线的应用显著提高了滑坡预警系统的准确性和可靠性。研究成果有助于优化现行滑坡预警系统,提高灾害管理效率,可为滑坡灾害预防和减灾工作提供科学支持。The Qinba Mountains area in Sichuan province,China,has been known for high incidences of landslides partly due to complex local terrain and frequent microclimate extremes,but precipitation is certainly one of the main natural factors causing landslides.Although the existing regional landslide warning system provides regular monitoring services,there is still room for improvement in its accuracy.In this study,1850 landslide events and simultaneous hourly precipitation data from 2000—2020 were collected and regressively analyzed to identify the key rainfall patterns governing landslide occurrences;an optimal threshold curve of precipitation was determined using combined indicators such as confusion matrix and skill scores.(1)The optimal precipitation threshold curve for inducing landslides in the Qinba Mountains area was determined by fitting,and validated by the case of landslide forecast in 2021,with a prediction accuracy of 87.72%.(2)By applying the optimal precipitation threshold curve to landslide hazard mapping through catalog data of field survey,it was found that 51.46%of landslides were in high-risk zones(classes 4 and 5)and 91.23%were in medium to high-risk zones(classes 3 to 5),justifying the effectiveness of the threshold curve for spatial prediction.This study provides important scientific references for the optimization of the current landslide early warning system,improving the efficiency of disaster management,and supporting landslide disaster prevention and mitigation.

关 键 词：降雨型滑坡 降雨阈值 气象预警 秦巴山区 

分 类 号：X43[环境科学与工程—灾害防治]