ARHCS (Automatic Rainfall Half-Life Cluster System): A Landslides Early Warning System (LEWS) Using Cluster Analysis and Automatic Threshold Definition

A significant portion of Landslide Early Warning Systems (LEWS) relies on the definition of operational thresholds and the monitoring of cumulative rainfall for alert issuance. These thresholds can be obtained in various ways, but most often they are based on previous landslide data. This approach introduces several limitations. For instance, there is a requirement for the location to have been previously monitored in some way to have this type of information recorded. Another significant limitation is the need for information regarding the location and timing of incidents. Despite the current ease of obtaining location information (GPS, drone images, etc.), the timing of the event remains challenging to ascertain for a considerable portion of landslide data. Concerning rainfall monitoring, there are multiple ways to consider it, for instance, examining accumulations over various intervals (1 h, 6 h, 24 h, 72 h), as well as in the calculation of effective rainfall, which represents the precipitation that actually infiltrates the soil. However, in the vast majority of cases, both the thresholds and the rain monitoring approach are defined manually and subjectively, relying on the operators’ experience. This makes the process labor-intensive and time-consuming, hindering the establishment of a truly standardized and rapidly scalable methodology on a large scale. In this work, we propose a Landslides Early Warning System (LEWS) based on the concept of rainfall half-life and the determination of thresholds using Cluster Analysis and data inversion. The system is designed to be applied in extensive monitoring networks, such as the one utilized by Cemaden, Brazil’s National Center for Monitoring and Early Warning of Natural Disasters.

The SNAKE System: CEMADEN’s Landslide Early Warning System (LEWS) Mechanism

In Brazil, the prominent climate-induced disasters are floods and mass movements, with the latter being the most lethal. The spate of major landslide events, especially those in 2011, catalyzed the creation of CEMADEN (National Center for Monitoring and Early Warning of Natural Disasters). This article introduces one of CEMADEN’s pivotal systems for early landslide warnings and traces its developmental timeline. The highlighted SNAKE System epitomizes advancements in digital monitoring, forecasting, and alert mechanisms. By leveraging precipitation data from pluviometers in observed municipalities, the system bolsters early warnings related to potential mass movements, like planar slides and debris flows. Its deployment in CEMADEN’s Situation Room attests to its suitability for overseeing high-risk municipalities, attributed primarily to its robustness and precision.

A Real-time Monitoring and Early Warning System for Landslides in Southwest China

Landslides not only cause property losses,but also kill and injure large numbers of people every year in the mountainous areas. These losses and casualties may be avoided to some extent by early warning systems for landslides. In this paper, a realtime monitoring network and a computer-aided automatic early warning system(EWS) are presented with details of their design and an example of application in the Longjingwan landslide, Kaiyang County, Guizhou Province. Then, according to principle simple method of landslide prediction, the setting of alarm levels and the design of appropriate counter-measures are presented. A four-level early warning system(Zero, Outlook, Attention and Warning) has been adopted, and the velocity threshold was selected as the main warning threshold for the landslide occurrence, but expert judgment is included in the EWS to avoid false alarms. A case study shows the applicability and reliability for landslide risk management, and recommendations are presented for other similar projects.

An Integrated Multi-Level Early Warning Method for Rainfall-Induced Shallow Landslides

Rainfall-induced landslides have occurred frequently in Southwestern China since the Wenchuan earthquake,resulting in massive loss of people’s life and property.Fortunately,landslide early-warning is one of the most important tools for landslide hazard prevention and mitigation.However,the accumulation of historical data of the landslides induced by rainfall is limited in many remote mountain areas and the stability of the slope is easily affected by human engineering activities and environmental changes,leading to difficulties to accurately realize early warning of landslide hazards by statistical methods.The proposed warning method is divided into rainfall warning component and deformation warning component because the deformation induced by rainfall has the characteristic of hysteretic nature.Rainfall,tilted angle and crack width are chosen as monitoring indexes.Rainfall grade level that contains rainfall intensity and duration information is graded according to the variation of the safety factor calculated by 3-D finite difference numerical simulation method,and then is applied using the strength reduction method and unascertained information theory to obtain the deformation grade level of several monitored points.Finally,based on the system reliability theory,we establish a comprehensive landslide warning level method that provides four early warning levels to reflect the safety factor reductions during and post rainfall events.The application of this method at a landslide site yield generally satisfactory results and provide a new method for performing multi-index and multi-level landslide early warnings.

Development of community-based landslide early warning system in the earthquake-affected areas of Nepal Himalaya

In the central Nepal Himalaya,landslides form the major natural hazards annually resulting in many casualties and damage.Structural as well as non-structural measures are in place to minimize the risk of landslide hazard.To reduce the landslide risk,a Landslide Early Warning System(LEWS)as a nonstructural measure has been piloted at Sundrawati village(Kalinchowk rural municipality,Dolakha district)to identify its effectiveness.Intensive discussions with stakeholders,aided by landslide susceptibility map,resulted in a better understanding of surface dynamics and the relationship between rainfall and surface movement.This led to the development of a LEWS comprised of extensometers,soil moisture sensors,rain gauge stations,and solar panels as an energy source that blows siren receiving signals via a micro-controller and interfacing circuit.The data generated through the system is transmitted via a Global System for Mobile Communications(GSM)network to responsible organizations in realtime to circulate the warning to local residents.This LEWS is user-friendly and can be easily operated by a community.The successful pilot early warning system has saved 495 people from 117 households in August 2018.However,landslide monitoring and dissemination of warning information remains a complex process where technical and communications skill should work closely together.

An Early Warning System for Regional Rain-Induced Landslide Hazard

Landslide in alpine regions often causes heavy losses of both human lives and properties, most of the landslides are induced by heavy rainfall. In this paper, we put forward an early warning system of rain-induced landslide. From 2002, we carried on the demonstrative work of landslide monitoring and early warning in Yaan, Sichuan Province, China, and constructed the first county-scale landslide monitoring and early warning region. Yucheng District of Yaan City is located in the west of the Sichuan Basin, right in the intersection of SichuanBasin and the Tibetan Plateau. The slopes are made of Mesozoic sedimentary rock, sandstone inter-bedded with mudstone. Yucheng District has the title “sky funnel” because of the high precipitation, the annual precipitation is about 1750 mm. We carried out detailed landslide survey, and obtained the location, scale, characteristics, influence and triggering factors of the landslides. Then we assessed the regional landslide susceptibility. Based on the evolution law of the landslides, we selected ten factors to study the relationship between the factors and landslide. Using the bi-variate statistics method, we calculated the contribution to landslide from each factor, classified the susceptibility into four categories. We set up the regional rainfall monitoring network with 13 automatic CAWS600R rain gauges. Using the landslide survey data, we studied the rainfall influencing of the regional landslides. The one-day and three-day rainfall controls the occurrence of regional landslide. We also classified the triggering effect of rainfall into four categories. We presented a method to calculate the landslide danger degree using the susceptibility and triggering category. Utilizing the predicted rainfall data and real-time monitored rainfall data, together with the landslide susceptibility map, we developed a WebGIS-based landslide warning system, which greatly strengthened the capability for geohazard control.

Disaster reduction stick equipment: A method for monitoring and early warning of pipeline-landslide hazards

Oil and gas pipelines are of great importance in China,and pipeline security problems pose a serious threat to society and the environment.Pipeline safety has therefore become an integral part of the entire national economy.Landslides are the most harmful type of pipeline accident,and have directed increasing public attention to safety issues.Although some useful results have been obtained in the investigation and prevention of pipeline-landslide hazards,there remains a need for effective monitoring and early warning methods,especially when the complexity of pipeline-landslides is considered.Because oil and gas pipeline-landslides typically occur in the superficial soil layers,monitoring instruments must be easy to install and must cause minimal disturbance to the surrounding soil and pipeline.To address the particular characteristics of pipelinelandslides,we developed a multi-parameter integrated monitoring system called disaster reduction stick equipment.In this paper,we detail this monitoring and early warning system for pipeline-landslide hazards based on an on-site monitoring network and early warning algorithms.The functionality of our system was verified by its successful application to the Chongqing Loujiazhuang pipeline-landslide in China.The results presented here provide guidelines for the monitoring,early warning,and prevention of pipeline geological hazards.

降雨和库水位变动联合作用下滑坡位移加权反分析及监测预警方法

降雨和库水位变动是诱发大型库岸滑坡体持续变形甚至失稳的主要外在因素。针对以上两种因素联合作用下滑坡体的变形特征,本文提出一种基于多源数据的“反演-评估-融合-预警”四步式滑坡监测预警方法。该方法采用趋势项位移进行力学参数反演,有效避免了降雨、库水位变动等外在因素对反演结果的影响。同时,采用熵权值构建了反演目标函数,借此考虑了不同位置监测点对反演结果的影响程度。此外,考虑了滑坡体自身的地质条件和坡体结构对监测预警的影响,综合运用安全系数和综合变形速率构建了具有地质和力学基础的预警模型。采用本文方法对贵州三板溪水电站东岭信滑坡体进行了力学参数反演和监测预警,结果表明:该滑坡体整个监测期间的安全系数在1.25~1.70之间波动,整体状态呈下降趋势,基于安全系数的角度判断其尚处于稳定状态;综合变形速率呈现上、下波动趋势,与安全系数在变化规律上具有高度的相关性,两者可以从整体上表征该滑坡体在降雨和库水位变动联合作用下稳定性的演进过程;当前监测时刻点的综合变形速率是0.0599,根据分级预警模型判断此滑坡体当前处于稳定变形阶段(Ⅰ级预警),与安全系数的判断结果具有一致性。

降雨诱发滑坡的NPR锚索控制机理及临滑预警

为研究降雨滑坡的控制及临滑预警技术,以熊家山滑坡为原型,通过分析边坡变形的时空特征,设计了相似模型试验.以普通预应力锚索(简称PR锚索)和NPR锚索为加固材料,运用高速摄像机、孔压、土压和锚索力传感器,分析了降雨诱发的滑坡规律和锚索控制机理,并结合现场试验探究了滑坡牛顿力变化规律.结果表明:孔隙水压力突降与边坡局部滑塌近似同步;NPR锚索力突降,边坡内部产生局部变形;PR锚索因无法承受大变形被拉断失效,而NPR锚索凭借恒阻大变形特性产生高应力吸能控制功能,促使边坡达到二次平衡状态;NPR锚索在滑坡前表现出明显的“先突增,后突降”的变形-临滑动力学规律;现场试验揭示了滑坡全过程牛顿力演变规律,提前7.5 h发出临滑预警.

基于多源数据与机器学习的降雨型滑坡灾害危险性动态预警

降雨型滑坡灾害是三峡库区发育的主要地质灾害之一,区域滑坡灾害危险性预警对滑坡风险减缓有重要意义。针对区域滑坡灾害危险性预警的时效性等问题,本文以重庆市云阳县为研究区,提出了基于多源数据与机器学习的滑坡灾害危险性动态预警方法(图1)。1滑坡易发性空间预测滑坡易发性预测可准确地预测出潜在滑坡空间分布规律(黄发明等,2020;林高聪等,2023)。

基于普适型监测预警体系中多参数模型研究——以云南省保山市大园子滑坡为例

为全面发展“人防+技防”的地质灾害防治技术体系,构建以普适型监测仪器为主的地质灾害监测预警体系,以云贵高原地区中的云南省保山市大园子滑坡为例,详细介绍了滑坡从方案设计到数据筛选及模型设计等各环节,主要选取降雨、位移变形和裂缝三类参数模型在监测预警体系中的应用,通过采集地质灾害体的关键信息,优化监测预警模型及调整阈值,及时捕捉地质灾害体的变化信息,依靠平台判据模型自主触发预警,完成了普适型监测预警体系建设的关键环节,从而实现了普适型监测预警的效果,极大的提升了地质灾害防治体系的自动化和智能化能力。

崩塌滑坡-堰塞湖-溃决洪水-泥石流灾害链演化特征分析及防治对策研究

某沟谷两岸坡面陡峻,沟谷狭窄,纵坡降较大,在地震、降雨等不利因素影响下,其左岸堆积体上方的崩滑堵沟隐患点可能出现失稳,并发展为崩塌滑坡-堰塞湖-溃决洪水-泥石流灾害链。针对此灾害链不同阶段的演化特征,采用相应的数值模拟模型和数值计算方法进行模拟分析和计算,评价其对沟口桥梁工程的影响,并采取相应的防治对策。经分析计算,崩塌滑坡隐患点距沟底高程落差约1 km,岩体体积约8×10^(6)m^(3),平均厚度约26 m,崩塌滑坡堵河可形成最大水深为14.4 m、面积约为7.19×10^(4)m^(2)、方量约为2.74×10^(5)m^(3)的堰塞湖;堰塞湖溃决形成洪水过程中,桥梁处最大水深为4.43 m(不含原始水位),最大流速为7.54 m/s,峰值流量为807 m^(3)/s;在溃决洪水强烈揭底冲刷和侵蚀的条件下,溃决洪水引发的泥石流在桥梁处的最大水深为7.1 m、最大流速为8 m/s、峰值流量为1685.5 m^(3)/s、最大冲刷深度为16.58 m。为减少该灾害链对桥梁工程的影响,采取河道疏通、岸坡防护和监测预警等防治措施。

降雨诱发的浅表堆积层滑坡成因机理与稳定性预测模型

强降雨引起的滑坡多以浅表堆积层滑移为主,易群发、危害性大,其地下水和土体内部含水率对降雨的水文响应机制复杂,难以精准开展稳定性预测。为研究降雨引起的滑坡水文响应对稳定性的影响,以四川省青川县后山里滑坡为例,开展现场降雨入渗监测、相关性分析和力学分析。通过连续3年的降雨量、土体水分和地下水位等监测,分析降雨入渗-土体体积含水率-地下水位的响应规律,得到降雨量与地下水位的相关关系,并基于无限斜坡稳定性计算公式构建基于降雨量和地下水位埋深的浅层滑坡稳定性预测模型。结果表明:(1)年内地下水呈周期性波动,分为缓慢下降期、快速下降期和快速上升期三个阶段,且降雨量与地下水埋深呈线性负相关,与水位升幅相关性不显著;(2)根据稳定性预测模型确定了该滑坡失稳的临界降雨阈值为81.8mm/d,地下水埋深阈值为0.73m。研究结果可为降雨诱发浅表堆积层滑坡的预警预报提供参考。

四川秦巴山区降雨型滑坡灾害降雨阈值

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。

“阶跃型”滑坡位移监测动态预警系统

“阶跃型”演化规律是滑坡在等速变形阶段受外界因素影响表现出的一种波动特性,具有变形周期长、机制复杂、临灾预警难度大的特点。基于“阶跃型”滑坡变形-时间曲线的特征和规律,通过引入“一个降雨过程”定义滑坡监测序列的多期降雨区间,将预警过程划分为前期+当期降雨和当期降雨两种模式,以诱因-时间-空间作为滑坡预警的重要指标,建立了滑坡综合预警判据模型,据此提出设计麻池村1号滑坡综合动态分级预警系统的方法。并结合麻池村1号滑坡的地质条件、监测数据等信息,深入分析了滑坡的变形演化规律及预警阈值。结果表明:①滑坡变形破坏模式为牵引式,属于典型的降雨诱发的“阶跃型”堆积层滑坡;②前期降雨有效持时为10 d,前期+当期降雨模式下的降雨阈值分别为24、32 mm,当期降雨模式下滑坡的的降雨阈值为37 mm,滑坡的位移速率阈值为12 mm;③以降雨量、位移速率到达阈值时为Ⅲ级黄色预警中心面(线),构建滑坡的综合动态分级预警系统,将滑坡预警从传统的“阈值预警”升级为“过程预警”,使得预警更具针对性。

微震技术在崩塌落石监测预警应用的研究进展

崩塌与落石作为岩质边坡破坏与诱发的典型地质灾害,往往具有突发性、隐蔽性和缺乏明显直接的前兆信息,很难做到准确和及时的应急响应,存在巨大的安全隐患。微震技术依靠远程无接触、低成本、广覆盖和24 h高频率无间断监测能力正逐渐被用于崩塌落石的监测预警中。通过监测和分析崩塌落石发生过程的动态振动信号,及时反馈振动信号的发生位置、时间、能量、以及崩塌落石的体积、质量等物理参数,微震技术最终可以做到对崩塌落石发生前的预测预报、发生时的监测预警和发生后的定位抢险。崩塌落石微震监测预警方法应用的难点主要体现在4个方面:(1)崩塌落石信号的识别和分类,通过对微震监测信号的降噪和特征提取,识别崩塌落石信号,并确定灾害发生和结束时刻;(2)崩塌落石的空间定位,确定崩塌落石发生的空间位置;(3)崩塌落石物理参数的反演,通过监测信号反演计算崩塌落石的速度、体积等物理参数;(4)崩塌落石的预警方法,提出崩塌落石预警模型,预测崩塌落石发生时间及地点。本文将从以上4个方面,结合本文作者最新研究成果,讨论微震技术在崩塌落石监测预警应用的最近研究进展,以及目前存在的缺陷。本文可为微震技术在地灾监测预警应用研究提供参考。

新华滑坡变形演化规律与预警判据

为揭示四川省泸定县新华滑坡在水库蓄水、库水位波动以及季节性降雨综合作用下的变形演化规律,探讨其在多因素耦合作用下滑坡岩土体强度劣化的渐进累积效应。基于对水电工程库区新华滑坡的现场调查及多源长序列监测数据分析,并结合滑坡累计位移-时间(S-t)曲线3阶段演化特征与新华滑坡实测S-t曲线,对该滑坡3个阶跃性变形周期的演化阶段进行了判识与划分,同时采用改进的切线角方法进行了计算与分析,构建了以改进切线角预警阈值为依据的新华滑坡分级预警定量划分标准。研究结果表明,水库蓄水初期水位的短期大幅度上升、水库运行期的水位快速下降及季节性强降雨影响是新华滑坡变形加速的主要诱因。根据分析与探讨,改进切线角预警阈值能在一定程度上为辨别新华滑坡现阶段的稳定状态与潜在风险提供指导,但仍应考虑多重预警指标并综合滑坡体宏观变形破坏迹象进行判定。

基于LSTM_TCN模型的降雨型滑坡时间概率预测及气象预警建模

如果滑坡发生时间信息不完备则会导致滑坡与降雨时序关系错误,以至于降雨阈值模型精度偏低。以重庆市万州区1995-2015年所发生的降雨型滑坡为研究对象,将区内严重缺失历史滑坡时间信息的恒合乡作为验证区,提出了一种基于长短时记忆网络(LSTM)融合时域卷积网络(TCN)的模型方法。该方法通过模拟降雨型滑坡发生时间与降雨量间的非线性关系,重建降雨型滑坡事件在某日发生的时间概率。将重建时间信息后的滑坡事件进行了验证与筛选,应用于累积有效降雨量-降雨历时曲线的合理划分,构建了滑坡气象预警模型。结果表明,本方法所预测滑坡时间概率平均值达到90.33%,高于人工神经网络(ANN)(71.17%)、LSTM(72.75%)和TCN(86.91%)的概率。利用预测概率高于90%的滑坡,将验证区18个时间信息扩充至201个。基于扩充时间信息后的滑坡数据所构建的气象预警模型比仅利用历史滑坡事件具有更合理的预警分级,在严重警告级别上有效预警率提升了42.86%。结果说明该方法可弥补野外调查中灾害数据时间信息不足的问题,为降雨型滑坡气象预警工作提供数据支撑,由此提高气象预警准确率。

基于改进切线角理论的十堰市变质岩滑坡预警阈值研究

十堰市地形地貌复杂,地层岩性多变,变质岩广泛分布,区内滑坡灾害多发。为减轻滑坡灾害带来的严重危害,实现滑坡灾害的准确预警,基于改进切线角理论,结合累积加速度及累积加加速度,选取十堰市4个已发生破坏的变质岩滑坡地表位移监测数据。通过绘制其T-t曲线、累积加速度-时间曲线、累积加加速度-时间曲线,分析得到了十堰市变质岩滑坡蠕滑阶段的综合划分结果。确定临滑预警阈值判据为:T-t曲线切线角大于等于80°、累积加速度大于0,并且增长保持较大正值、累积加加速度大于0,且连续增长或保持高值。研究成果对十堰市变质岩滑坡的监测预警工作具有重要意义,同时也可为其他类型滑坡的监测预警阈值研究提供参考。

基于卷积神经网络的福建省区域滑坡灾害预警模型

福建省滑坡灾害频发,开展区域尺度上的滑坡灾害预警是防灾减灾的重要手段,但由于滑坡成灾机理复杂,传统的区域滑坡预警方法存在精度不足等问题。深度学习是指通过构建神经网络模型进行特征的提取、抽象、表示与学习的技术,是机器学习的一种。卷积神经网络作为一种经典的深度学习算法,具有比传统机器学习更强大的分类能力与表征能力。文章以福建省为研究区,将卷积神经网络引入滑坡灾害预警领域,构建福建省区域滑坡预警模型,过程及结果如下:(1)采用SMOTE优化算法对2010—2018年福建省滑坡灾害样本库进行优化,扩充正样本的个数,将正负样本比例从1∶3.4扩充到1∶2,样本总量达到18040个;(2)构建卷积神经网络模型结构,模型结构包括一个输入层、两个卷积层、两个最大池化层和一个全连接层以及一个输出层;(3)使用卷积神经网络对优化后的样本(2010—2018年样本的80%作为训练集)进行训练,并用贝叶斯优化算法优化模型超参数,得到福建省区域滑坡预警模型;(4)以2010—2018年样本的20%作为测试集对模型进行测试,采用混淆矩阵、ROC曲线进行模型测试,结果显示模型准确度为0.96~0.97,AUC值达到0.977,模型精度与泛化能力良好;(5)以2019年汛期滑坡灾害实况作为正样本,通过时空采样的方法采集负样本,构建2019年区域滑坡样本校验集(样本数603个),对模型进行进一步实况校验,采用混淆矩阵、ROC曲线进行模型校验,结果显示模型准确度为0.75~0.85,AUC值为0.852。虽然仅用了2019年汛期的滑坡实况样本进行校验,但也达到较好的效果。将卷积神经网络算法应用到区域滑坡预警中,为建立区域滑坡预警模型提供了一种新的途径,初步校验表明,模型效果良好,今后将在福建省对模型进行进一步的应用与校验。