集中降雨影响下辽南仙人洞国家级自然保护区滑坡灾害多因子风险评价  

Multi-factor risk assessment of landslide disasters under concentrated rainfall in Xianrendong national nature reserve in southern Liaoning Province

作　　者：郑德凤[1,2] 潘美伊 高敏[1] 闫成林 李媛媛[1] 年廷凯[3] ZHENG Defeng;PAN Meiyi;GAO Min;YAN Chenglin;LI Yuanyuan;NIAN Tingkai(School of Geography,Liaoning Normal University,Dalian Liaoning 116029,China;Dalian Key Laboratory of Agro-Meteorological Disaster Risk Prevention and Control,Liaoning Normal University,Dalian Liaoning 116029,China;School of Construction Engineering,Dalian University of Technology,Dalian Liaoning 116024,China)

机构地区：[1]辽宁师范大学地理科学学院,辽宁大连116029 [2]辽宁师范大学大连市农业气象灾害风险防控重点实验室,辽宁大连116029 [3]大连理工大学建设工程学院,辽宁大连116024

出　　处：《地质科技通报》2025年第2期48-58,共11页Bulletin of Geological Science and Technology

基　　金：国家自然科学基金项目(42077272);辽宁师范大学高端科研成果培育资助计划项目(23GDL007)。

摘　　要：位于辽宁省南部的仙人洞国家级自然保护区及毗邻区域,近年来在集中降雨期间多次发生滑坡等地质灾害;然而,针对集中降雨影响下该区域滑坡灾害易发性与风险评价还很少探究,因此亟待开展深入研究,这对有效减轻保护区地质灾害危害、提升地质灾害应急和风险防控能力具有重要意义。首先利用SMOTE-Tomek综合采样法、耦合XGBoost模型,对地形地貌、地质岩性、水文气象和人类工程活动因子等12个指标进行深入分析,以获取滑坡灾害易发性评价结果;其次,着重考虑短时间集中降雨和连续性降雨影响,利用研究区周围4个气象站的日降雨数据,计算2018-2023年平均暴雨强度和3 d累计降雨量,作为危险性评价指标,实现对滑坡灾害危险性的量化评估。在此基础上,考虑承灾体的易损性特点和区域防灾减灾能力,构建适用于研究区的多因子滑坡灾害风险评价模型,进而得到滑坡灾害风险评价分区图。结果表明:较高及以上等级风险区约占研究区总面积的10%,主要分布在仙人洞镇东北部、步云山乡北部、长岭镇北部和荷花山镇西南部。未来应重点关注滑坡地质灾害风险较高的仙人洞镇和步云山乡,加强监测和预警工作,为整个研究区的地质灾害防控提供决策支持。[Objective]Xianrendong National Nature Reserve in southern Liaoning Province and its adjacent areas have experienced numerous geological disasters,such as landslides,during periods of concentrated rainfall in recent years.However,the vulnerability and risk assessment of landslide disasters in this region under the influence of concentrated rainfall have been rarely explored.Therefore,it is urgent to conduct in-depth research,which is of great significance for effectively reducing the harm of geological disasters in the reserve and enhancing the capacity for emergency response and risk prevention and control.[Method]Firstly,using the SMOTE-Tomek comprehensive sampling method coupled with an XGBoost model,12 indices including topography,geology and lithology,hydrometeorology,and human engineering activities are deeply analyzed to obtain the evaluation results of landslide susceptibility.Secondly,focusing on the impact of short-term concentrated rainfall and continuous rainfall,daily rainfall data from four weather stations surrounding the study area are used to calculate annual average rainstorm intensity from 2018 to 2023 and three-day cumulative rainfall as risk assessment indicators,enabling the quantitative assessment of landslide hazards.On this basis,considering the vulnerability characteristics of the disaster-bearing body and regional disaster prevention and mitigation capabilities,a multi-factor landslide disaster risk assessment model suitable for the study area is constructed,resulting in a landslide disaster risk zoning map.[Result]The results show that high-risk and above areas account for approximately 10%of the total study area,primarily distributed in the northeast of Xianrendong Town,the north of Buyunshan Town,the north of Changling Town,and the southwest of Hehuashan Town.[Conclusion]In the future,efforts should focus on high-risk areas such as Xianrendong Town,strengthening monitoring and early warning systems to provide decision support for the prevention and control of geological disasters acr

关 键 词：滑坡灾害 易发性评价 风险评价 SMOTE-Tomek算法 XGBoost模型 防灾减灾能力 自然保护区 集中降雨 

分 类 号：P642.22[天文地球—工程地质学]