Landslide inventory and susceptibility assessment using multiple statistical approaches along the Karakoram highway,northern Pakistan

China-Pakistan Economic Corridor(CPEC)is a framework of regional connectivity,which will not only benefit China and Pakistan but will have positive impact on Iran,Afghanistan,India,Central Asian Republic,and the region.The surrounding area in CPEC is prone to frequent disruption by geological hazards mainly landslides in northern Pakistan.Comprehensive landslide inventory and susceptibility assessment are rarely available to utilize for landslide mitigation strategies.This study aims to utilize the high-resolution satellite images to develop a comprehensive landslide inventory and subsequently develop landslide susceptibility maps using multiple techniques.The very high-resolution(VHR)satellite images are utilized to develop a landslide inventory using the visual image classification techniques,historic records and field observations.A total of 1632 landslides are mapped in the area.Four statistical models i.e.,frequency ratio,artificial neural network,weights of evidence and logistic regression were used for landslide susceptibility modeling by comparing the landslide inventory with the topographic parameters,geological features,drainage and road network.The developed landslides susceptibility maps were verified using the area under curve(AUC)method.The prediction power of the model was assessed by the prediction rate curve.The success rate curves show 93%,92.8%,92.7%and 87.4%accuracy of susceptibility maps for frequency ratio,artificial neural network,weights of evidence and logistic regression,respectively.The developed landslide inventory and susceptibility maps can be used for land use planning and landslide mitigation strategies.     

优化循环神经网络在滑坡位移预测中的应用

以中国典型黄土滑坡域甘肃黑方台党川6#滑坡体为例,基于滑坡体北斗和位移计时序监测数据,首先利用深度学习框架Tensorflow分别构建3种循环神经网络滑坡位移预测模型：简单循环神经网络(simple recurrent neural network,SimpleRNN)、长短期记忆网络(long short-term memory,LSTM)和门控循环单元(gated recurrent unit,GRU),并进一步针对循环神经网络在参数设置时多采用经验手动调参或采用网格搜索法,易造成人为主观影响较大和计算效率低下的突出问题,引入遗传算法（genetic algorithm,GA）优化循环神经网络参数的自动最佳化选取,分别构建3种基于遗传算法改进的循环神经网络滑坡位移高精度预测模型：GA-SimpleRNN、GA-LSTM、GA-GRU。研究结果表明,改进参数自动寻优后的3种循环神经网络预测模型具有更优的预测性能,特别是GA-GRU模型预测精度最高,更适用于滑坡体长时序位移的高精度预测。     

基于SSA-Adam-BP神经网络模型的堰塞坝稳定性预测

现有的堰塞坝稳定性预测模型多为线性模型, 无法充分考虑堰塞坝稳定性与其形态特征和水域条件之间的复杂非线性关系。鉴于此, 结合反向传播神经网络模型和樽海鞘优化算法, 提出了一种新型的堰塞坝稳定性预测模型SSA-Adam-BP。该模型通过网格搜索法选取确定模型结构的最佳超参数组合, 进而利用交叉验证和绘制ROC曲线的方式分别对采用不同优化算法的模型进行评估。使用开源数据库中的全球153例堰塞坝数据对模型的实际应用进行了说明及验证。与传统线性模型的对比表明神经网络模型预测准确率较高, 具有较低的误报率。将SSA与Adam优化算法结合提高了BP模型的全局搜索能力, 其平均交叉验证准确率达到了91.73%, 能够使用较少的参数实现对堰塞坝稳定性快速准确的预测。SSA-Adam-BP模型对近年来典型工程的稳定性能够准确预测, 具有一定的实用性和系统平台推广应用价值。      

水库滑坡成因机制研究进展与展望

滑坡是水库库区主要地质灾害类型之一,开展水库滑坡成因机制研究具有重要理论意义和工程应用价值.利用WebofScience(WoS)数据库和VOSviewer文献计量工具对1999-2018年已发表的969篇以水库滑坡为主题的相关论文进行研究趋势分析.文献计量分析表明三峡库区滑坡稳定性和变形研究是未来水库滑坡成因机制研究主要趋势.从库水对滑坡的宏观力学作用方式、库水作用下岩土体渗流应力耦合机理和库水对岩土体劣化作用过程等方面,对国内外水库滑坡成因机制研究的主要成果与进展进行了综述.综合现有的研究成果指出水库滑坡在精细化地质建模、岩土体多场耦合特征参数获取和岸坡长期演化评价等方面尚存在不足.基于上述问题,提出水库滑坡成因机制研究应以多场信息监测为重要手段,立足多学科交叉,采用大数据融合与挖掘和人工智能技术等解决水库滑坡长期演化趋势难题.考虑水库滑坡所处地质环境的复杂性,建议未来应在水库滑坡立体精细地质建模、多场关联监测、地质结构多场多尺度演变过程、基于监测数据大数据分析的滑坡预警阈值确定和原位试验综合平台构建等方面进一步深入研究.     

基于物理模型试验的多层滑带堆积层滑坡—抗滑桩受力特征

研究抗滑桩的受力特征是进行抗滑桩设计工作的关键。我国三峡库区部分堆积层滑坡发育多层滑带,而目前抗滑桩的设计方法仅针对单层滑坡,因此,对多层滑带堆积层滑坡—抗滑桩受力特征的研究具有重要意义。基于三峡库区堆积层滑坡工程地质特征,开展了多层滑带堆积层滑坡物理试验模型,在滑坡的后缘施加推力来模拟滑坡演化过程,同时监测滑坡—抗滑桩体系的多场信息。试验结果表明,在多层滑带堆积层滑坡的演化过程中,桩身受力表现出了很好的规律性。根据坡表位移的变化趋势,将滑坡演化分为4个阶段,在此基础上对桩身受力进行分析。滑坡推力分布图式中出现了4个极大值,土体抗力分布图式中出现唯一最大值,该试验结果为抗滑桩的设计提供了一定的理论支持。      

Visual analysis of the evolution and focus in landslide research field

This paper analysed the evolution of landslide research and research foci in different countries. The data comprise 3105 landslide SCI articles published between January 1977 and June 2015 from the Web of Science. The data are extracted under interaction constraints of the journal title, category, and keywords. The complex network method is used for the analysis. First, from the perspective of topics and methods, the evolution is systematically assessed by generating a co-citation network of the articles and a semantic cluster analysis. Second, from the perspective of topics and landsliderelated disasters, the focus in different countries is discussed by generating co-occurrence networks. These networks are the co-occurrence of the countries and keywords, and the co-occurrence of countries and landslide-related disaster phrases. The main conclusions are as follows:(1) landslide susceptibility analysis and methods of machine learning are popular research topics and methods, respectively. The topics change through time, and the article output is influenced by increasing landslide-related disasters, increasing economic losses and casualties, a desire for a more complete and accurate landslide inventory, and the use of effective methods, such as geographical information Science(GIS) and machine learning.(2) The research focus in each country is related with the country-specific disasters or economic costs caused by landslides to some degree. In addition to Italy and the USA, China is the country most commonly affected by landslides, and it should develop its own landslide database and complete in-depth studies of disaster mitigation.     

基于GEE和U-net模型的同震滑坡识别方法

地震滑坡解译是震后重建的重要基础工作,主要通过室内人工遥感解译和室外野外调查确定。地震滑坡相比其他地物来说更为复杂,很难通过简单指数识别。室内遥感解译通过滑坡后壁、侧壁和堆积等纹理特征进行识别,大面积同震滑坡解译工作往往耗费大量人力和物力,且耗时长,难以满足灾害应急需求。本研究利用U-net神经网络模型,结合Google Earth Engine（GEE）云平台和人工智能学习平台Tensorflow,以地震局解译的汶川滑坡作为样本数据,以震后30 m分辨率的Landsat影像、高程、坡度以及NDVI数据作为模型输入参数,自动识别并获取了汶川地震后的同震滑坡数据,同时比较了不同参数组合情况下U-net神经网络模型的分割识别精度。研究表明：① U-net模型可以用于以Landsat影像为基础数据的同震滑坡快速自动识别;② 随着高程、坡度以及NDVI等输入参数增加,模型分割精度在逐渐提高,但假阳性结果也会出现增多,震后滑坡影像+高程+坡度+NDVI的输入参数组合精度最高;③ 在细节上,模型在多参数组合的情况下,大型滑坡能够很好被识别,一些较小型的滑坡受制于影像分辨率的影响,分割精度较差。为了更好识别小型滑坡,后续研究可能需提高影像的分辨率。此外,GEE云平台大大提高了训练样本获取的效率,为科研人员快速进行基于神经网络与遥感数据的地物识别研究提供了条件。     

重大滑坡预测预报研究进展与展望

我国是世界上滑坡灾害最严重的国家之一, 重大滑坡灾害严重威胁人民生命财产安全和国家重大战略实施。滑坡精准预测预报是防灾减灾的前提, 也是亟待突破的世界性科学难题。以重大滑坡预测预报为目标, 聚焦滑坡演化过程与物理力学机制核心科学问题, 凝炼了滑坡启滑关联机制、滑坡启滑物理力学机制、滑坡过程预测预报理论3个关键科学问题, 提出了如下研究思路: 以系统论、控制论和信息论为指导, 依托大型野外试验场, 采用现场原型试验与多场关联监测、大型物理模型试验、多场耦合模拟等技术手段, 以滑坡孕育过程为基础, 提出了重大滑坡的启滑分类; 揭示锁固解锁型、静态液化型和动水驱动型滑坡启滑物理力学机制, 建立相应的启滑判据; 构建重大滑坡数值预报模式与实时预报平台, 创立基于物理力学过程的滑坡预测预报理论。通过实施, 可奠定上述3类滑坡预测预报的地质、力学与物理基础, 引领重大滑坡预测预报研究, 保障国家重大战略的顺利实施, 契合国家防灾减灾重大需求。      

Dynamic prediction of landslide displacement using singular spectrum analysis and stack long short-term memory network

An accurate landslide displacement prediction is an important part of landslide warning system. Aiming at the dynamic characteristics of landslide evolution and the shortcomings of traditional static prediction models, this paper proposes a dynamic prediction model of landslide displacement based on singular spectrum analysis(SSA) and stack long short-term memory(SLSTM) network. The SSA is used to decompose the landslide accumulated displacement time series data into trend term and periodic term displacement subsequences. A cubic polynomial function is used to predict the trend term displacement subsequence, and the SLSTM neural network is used to predict the periodic term displacement subsequence. At the same time, the Bayesian optimization algorithm is used to determine that the SLSTM network input sequence length is 12 and the number of hidden layer nodes is 18. The SLSTM network is updated by adding predicted values to the training set to achieve dynamic displacement prediction. Finally, the accumulated landslide displacement is obtained by superimposing the predicted value of each displacement subsequence. The proposed model was verified on the Xintan landslide in Hubei Province, China. The results show that when predicting the displacement of the periodic term, the SLSTM network has higher prediction accuracy than the support vector machine(SVM) and auto regressive integrated moving average(ARIMA). The mean relative error(MRE) is reduced by 4.099% and 3.548% respectively, while the root mean square error(RMSE) is reduced by 5.830 mm and 3.854 mm respectively. It is concluded that the SLSTM network model can better simulate the dynamic characteristics of landslides.     

Displacement characteristic of landslides reinforced with flexible piles: field and physical model test

A field monitoring system was established in an active river bank landslide in the Three Gorges area, China, and a consecutive monitoring for about 5 years were conducted to understand the displacement characteristics of flexible piles and the surrounding soil. It was found that piles deformed elastically under reservoir operation, and the soil in front of piles was gradually separated from piles. The movement of the pile heads exceeded that of the soil between and behind piles. This phenomenon was further studied by a large-scale physical model test to gain insights into the pile-soil interaction. The displacement relationship between pile heads and the surrounding soil is in good agreement with the field data. The physical model test shows that the deformation process of pile-reinforced landslides can be divided into two stages: firstly, when the piles head movement exceeds soil movement, the soil arching is mainly affected by the deflection of the piles, the arches between and behind piles bent upwards;but when the soil movement exceeds piles head movement, the arches near the upslope and downslope bent downwards and upwards, respectively. Furthermore, the different deformation of two adjacent piles and the pile stiffness influenced the arch’s shape and formation;the flexible piles exhibit great coordinated deformation with the landslide, and caused the soil arch on the downslope.     

Rapid susceptibility mapping of co-seismic landslides triggered by the 2013 Lushan Earthquake using the regression model developed for the 2008 Wenchuan Earthquake

The primary objective of landslide susceptibility mapping is the prediction of potential landslides in landslide-prone areas.The predictive power of a landslide susceptibility mapping model could be tested in an adjacent area of similar geoenvironmental conditions to find out the reliability.Both the 2008 Wenchuan Earthquake and the 2013 Lushan Earthquake occurred in the Longmen Mountain seismic zone,with similar topographical and geological conditions.The two earthquakes are both featured by thrust fault and similar seismic mechanism.This paper adopted the susceptibility mapping model of co-seismic landslides triggered by Wenchuan earthquake to predict the spatial distribution of landslides induced by Lushan earthquake.Six influencing parameters were taken into consideration: distance from the seismic fault,slope gradient,lithology,distance from drainage,elevation and Peak Ground Acceleration(PGA).The preliminary results suggested that the zones with high susceptibility of coseismic landslides were mainly distributed in the mountainous areas of Lushan,Baoxing and Tianquan counties.The co-seismic landslide susceptibility map was completed in two days after the quake and sent to the field investigators to provide guidance for rescue and relief work.The predictive power of the susceptibility map was validated by ROC curve analysis method using 2037 co-seismic landslides in the epicenter area.The AUC value of 0.710 indicated that the susceptibility model derived from Wenchuan Earthquake landslides showed good accuracy in predicting the landslides triggered by Lushan earthquake.     

金沙江结合带巴塘段滑坡群InSAR探测识别与形变特征

金沙江结合带结构破碎,软弱岩层发育,流域性特大高位地质灾害频繁发生.针对该区域开展大范围滑坡调查与监测研究,对减灾防灾具有重要意义.以金沙江结合带巴塘段为试验区,采用堆叠InSAR技术分别利用升轨、降轨Sentinel-1 A卫星数据对该区域滑坡隐患开展了调查研究.在此基础上,以中心绒乡滑坡群为重点研究区,利用多维小基...     

Evaluating effectiveness of frequency ratio,fuzzy logic and logistic regression models in assessing landslide susceptibility: a case from Rudraprayag district,India

Rudraprayag in Garhwal Himalayan division is one of the most vulnerable districts to landslides in India. Heavy rainfall, steep slope and developmental activities are important factors for the occurrence of landslides in the district. Therefore, specific assessment of landslide susceptibility and its accuracy at regional level is essential for disaster management and proper land use planning. The article evaluates effectiveness of frequency ratio, fuzzy logic and logistic regression models for assessing landslide susceptibility in Rudraprayag district of Uttarakhand state, India. A landslide inventory map was prepared and verified by field data. Fourteen landslide parameters and generated inventory map were utilized to prepare landslide susceptibility maps through frequency ratio, fuzzy logic and logistic regression models. Landslide susceptibility maps generated through these models were classified into very high, high, medium, low and very low categories using natural breaks classification. Receiver operating characteristics (ROC) curve, spatially agreed area approach and seed cell area index (SCAI) method were used to validate the landslide models. Validation results revealed that fuzzy logic model was found to be more effective in assessing landslide susceptibility in the study area. The landslide susceptibility map generated through fuzzy logic model can be best utilized for landslide disaster management and effective land use planning.     

霍西煤矿区地表滑坡灾害敏感性数值建模与等级划分

本文以山西省霍西煤矿区为研究区,利用遥感和GIS方法对滑坡灾害的敏感性进行了数值建模与定量评价。利用交叉检验方法构建了径向基核函数支持向量机滑坡敏感性评价模型,并基于拟合精度对模型进行了定量评价;对各评价因子在模型中的重要性进行对比分析;基于空间分辨率为30m的评价因子,通过径向基核函数支持向量机模型获得了霍西煤矿区滑坡敏感性指数值,并利用分位数法将霍西煤矿区的滑坡敏感性分为极高、高、中和低4个等级。结果表明：拟合精度建模阶段和验证阶段分别为87.22%和70.12%;与滑坡敏感性关系最密切的5个评价因子依次是岩性、距道路距离、坡向、高程和土地利用类型;极高和高敏感区域分布了93.49%的滑坡点,面积占总面积的50.99%,是比较合理的分级方案。本研究不仅可以为研究区人工边坡调查和煤矿资源合理开采提供借鉴,对相似矿区的相关工作也具有参考价值。     

Co-seismic landslide inventory and susceptibility mapping in the 2008 Wenchuan earthquake disaster area, China

The Ms 8.0 May 12,2008 Wenchuan earthquake triggered tens of thousands of landslides.The widespread landslides have caused serious casualties and property losses,and posed a great threat to post-earthquake reconstruction.A spatial database,inventoried 43,842 landslides with a total area of 632 km 2,was developed by interpretation of multi-resolution remote sensing images.The landslides can be classified into three categories:swallow,disrupted slides and falls;deep-seated slides and falls,and rock avalanches.The correlation between landslides distribution and the influencing parameters including distance from co-seismic fault,lithology,slope gradient,elevation,peak ground acceleration(PGA) and distance from drainage were analyzed.The distance from co-seismic fault was the most significant parameter followed by slope gradient and PGA was the least significant one.A logistic regression model combined with bivariate statistical analysis(BSA) was adopted for landslide susceptibility mapping.The study area was classified into five categories of landslide susceptibility:very low,low,medium,high and very high.92.0% of the study area belongs to low and very low categories with corresponding 9.0% of the total inventoried landslides.Medium susceptible zones make up 4.2% of the area with 17.7% of the total landslides.The rest of the area was classified into high and very high categories,which makes up 3.9% of the area with corresponding 73.3% of the total landslides.Although the susceptibility map can reveal the likelihood of future landslides and debris flows,and it is helpful for the rebuilding process and future zoning issues.     

Centrifuge model test of an irrigation-induced loess landslide in the Heifangtai loess platform,Northwest China

The Heifangtai platform in Northwest China is famous for irrigation-induced loess landslides. This study conducted a centrifuge model test with reference to an irrigation-induced loess landslide that occurred in Heifangtai in 2011. The loess slope model was constructed by whittling a cubic loess block obtaining from the landslide site. The irrigation water was simulated by applying continuous infiltration from back of the slope. The deformation, earth pressure, and pore pressure were investigated during test by a series of transducers. For this particular study, the results showed that the failure processes were characterized by retrogressive landslides and cracks. The time dependent reductions of cohesion and internal friction angle at basal layer with increasing pore-water pressure were responsible for these failures. The foot part of slope is very important for slope instability and hazard prevention in the study area, where concentration of earth pressure and generation of high pore-water pressures would form before failures. The measurements of earth pressure and pore-water pressure might be effective for early warning in the study area.     

Landslide distribution and sliding mode control along the Anninghe fault zone at the eastern edge of the Tibetan Plateau

Tibetan Plateau is known as the roof of the world. Due to the continuous uplift of the Tibetan Plateau, many active fault zones are present. These active fault zones such as the Anninghe fault zone have a significant influence on the formation of special geomorphology and the distribution of geological hazards at the eastern edge of the Tibetan Plateau. The Anninghe fault zone is a key part of the Y-shaped fault pattern in the Sichuan-Yunnan block of China. In this paper, high-resolution topographic data, multitemporal remote sensing images, numerical calculations, seismic records, and comprehensive field investigations were employed to study the landslide distribution along the active part of the Anninghe. The influence of active faults on the lithology, rock mass structures and slope stress fields were also studied. The results show that the faults within the Anninghe fault zone have damaged the structure and integrity of the slope rock mass, reduced the mechanical strength of the rock mass and controlled the slope failure modes. The faults have also controlled the stress field, the distribution of the plastic strain zone and the maximum shear strain zone of the slope, thus have promoted the formation and evolution of landslides. We find that the studied landslides are linearly distributed along the Anninghe fault zone, and more than 80% of these landslides are within 2-3 km of the fault rupture zone. Moreover, the Anninghe fault zone provides abundant substance for landslides or debris flows. This paper presents four types of sliding mode control of the Anninghe fault zone, e.g., constituting the whole landslide body, controlling the lateral boundary of the landslide, controlling the crown of the landslide, and constituting the toe of the landslide. The results presented merit close attention as a valuable reference source for local infrastructure planning and engineering projects.     

Numerical simulation of a high-speed landslide in Chenjiaba,Beichuan, China

High-speed landslide is a catastrophic geological disaster in the mountainous area of southwest China. To predict the movement process of landslide reactivation in Chenjiaba town, Beichuan county, Sichuan province, China, we simulated the movement process of two landslide failures in Chenjiaba via rapid mass movement simulation and unmanned aerial vehicle images(UAV), and obtained the movement characteristic parameters of the landslides. According to a back analysis, the most remarkable fitting rheological parameters were friction coefficient(μ=0.18) and turbulence(). The parameter of landslide pressure was applied as the zoning index of landslide hazard to obtain the influence zone and hazard zoning map of the Chenjiaba landslide. Results show that the Duba River was blocked quickly with a landslide accumulation at the maximum height of 44.14 mwhen the Chenjiaba deposits lost stability. The hazard zoning map indicated that the landslide hazard degree is positively correlated with the slope.This landslide assessment is a quantitative hazard assessment method based on a landslide movement process and is suitable for high-speed landslide. Such method can provide a scientific basis for urban construction and planning in the landslide hazard area to avoid hazards effectively.     

GIS-based evaluation on the fault motion-induced coseismic landslides

Earthquake-induced potential landslides are commonly estimated using landslide susceptibility maps. Nevertheless, the fault location is not identified and the ground motion caused by it is unavailable in the map. Thus, potential coseismic landslides for a specific fault motion-induced earthquake could not be predicted using the map. It is meaningful to incorporate the fault location and ground motion characteristics into the landslide predication model. A new method for a specific fault motion-induced coseismic landslide prediction model using GIS (Geographic Information System) is proposed herein. Location of mountain ridges, slope gradients over 45 o , PVGA (Peak Vertical Ground Accelerations) exceeded 0.15 g, and PHGA (Peak Horizontal Ground Accelerations) exceeded 0.25 g of slope units were representing locations that initiated landslides during the 1999 Chi-Chi earthquake in Taiwan. These coseismic landslide characteristics were used to identify areas where landslides occurred during Meishan fault motion-induced strong ground motions in Chiayi County in Taiwan. The strong ground motion (over 8 Gal in the database, 1 Gal = 0.01 m/s 2 , and 1 g = 981 Gal) characteristics were evaluated by the fault length, site distance to the fault, and topography, and their attenuation relations are presented in GIS. The results of the analysis show that coseismic landslide areas could be identified promptly using GIS. The earthquake intensity and focus depth have visible effects on ground motion. The shallower the focus depth, the larger the magnitude increase of the landslides. The GIS-based landslide predication method is valuable combining the geomorphic characteristics and ground motion attenuation relationships for a potential region landslide hazard assessment and in disaster mitigation planning.     

Landslide initiation and runout susceptibility modeling in the context of hill cutting and rapid urbanization: a combined approach of weights of evidence and spatial multi-criteria

Rainfall induced landslides are a common threat to the communities living on dangerous hill-slopes in Chittagong Metropolitan Area, Bangladesh. Extreme population pressure, indiscriminate hill cutting, increased precipitation events due to global warming and associated unplanned urbanization in the hills are exaggerating landslide events. The aim of this article is to prepare a scientifically accurate landslide susceptibility map by combining landslide initiation and runout maps. Land cover, slope, soil permeability, surface geology, precipitation, aspect, and distance to hill cut, road cut, drainage and stream network factor maps were selected by conditional independence test. The locations of 56 landslides were collected by field surveying. A weight of evidence (WoE) method was applied to calculate the positive (presence of landslides) and negative (absence of landslides) factor weights. A combination of analytical hierarchical process (AHP) and fuzzy membership standardization (weighs from 0 to 1) was applied for performing a spatial multi-criteria evaluation. Expert opinion guided the decision rule for AHP. The Flow-R tool that allows modeling landslide runout from the initiation sources was applied. The flow direction was calculated using the modified Holmgren’s algorithm. The AHP landslide initiation and runout susceptibility maps were used to prepare a combined landslide susceptibility map. The relative operating characteristic curve was used for model validation purpose. The accuracy of WoE, AHP, and combined susceptibility map was calculated 96%, 97%, and 98%, respectively.