基于GIS技术的区域滑坡灾害危险性预测

基于GIS技术,利用信息量模型开展区域滑坡灾害危险性预测研究,编制滑坡灾害易发分区图,为滑坡灾害的风险预测及实时预警预报提供基础资料。以浙江省永嘉县为例,利用MAPGIS二次开发得到的信息量专业模块,结合永嘉县历史滑坡灾害和2004年以来新发生的灾害点,分别评价了研究区的历史滑坡灾害危险性和现状滑坡灾害危险性;提出用历史滑坡灾害危险性图件结合新发生的灾害点来验证评价模型;将历史灾害点和新灾害点结合生成滑坡灾害危险性预测图件的预测过程;研究成果经在永嘉县的实际验证分析,2004年后3次台风期间（2004年的“云娜” 台风,2005年的“海棠”和“麦莎”台风）发生的有准确地点的滑坡灾害点全部位于滑坡灾害易发区内,表明采用的模型具有较好的实用性和可靠性;采用历史统计和快速聚类相结合的方法进行危险性等级的划分,克服了前人研究工作中人为划分易发区的缺陷,更科学、客观。     

基于GIS的滑坡灾害危险性区划研究——以重庆市万州区为例

为了弥补滑坡灾害危险性区划研究中影响因子和等级划分的不确定性,结合前人研究成果,依据斜坡几何形态、岩性、地质构造、河流侵蚀、土地利用类型、人类工程活动、降水条件等影响因子与研究区实际已发生的滑坡灾害数之间的关系,编制重庆市万州区滑坡灾害危险性评价标准,并基于GIS技术和信息量模型法,计算滑坡评价因子的信息量,就万州区滑坡危险性进行区划,最后基于乡镇行政区对该区滑坡危险性区划进行细化。结果表明：建设用地、坡高为90～200 m的地形、1 024～1 060 mm的年降雨量以及侏罗系中统上沙溪庙组岩层等因素对万州区滑坡发生影响较大;根据滑坡灾害危险性评价标准,万州区滑坡灾害被划分为高、中、低、极低等4个危险区;应用信息量模型法得到的万州区滑坡危险性区划与实际情况比较吻合;高危险区和中危险区面积分别为564.4 km2和848.6 km2,分别占万州区总面积的16.3%和24.5%,主要分布于长江干流及支流两岸的居民相对集中区以及公路干线地段;高危险和中危险乡镇主要分布在万州区经济较为发达的长江干流两岸,尤其是左岸的黄柏乡、太龙镇、天城镇、李河镇等以及万州主城区。     

基于逻辑回归模型的汶川地震滑坡危险性评价与检验

以2008年5月12日汶川地震区为研究区,基于高分辨率航片与卫星影像开展地震滑坡目视解译,制作了汶川地震滑坡编录图.选择坡度、坡向、高程、与水系距离、与公路距离、与映秀-北川断裂距离、地震烈度、岩性共8个影响因子开展地震滑坡危险性评价工作.滑坡样本采用前期48007处滑坡编录点数据,不滑样本为在基于证据权重模型的滑坡危险性评价结果的低危险区与极低危险区随机选择的48000个点.基于这8个影响因子与逻辑回归模型,建立了汶川地震滑坡危险性索引图.采用这48007个滑坡样本点与汶川地震滑坡最新编录的增加滑坡,分别进行模型的成功率与预测率检验.结果表明,模型成功率为81.739％,预测率达到86.278％.     

不同日降雨工况下万州区滑坡灾害危险性分析

以三峡库区万州区为例,选择具有代表性的地质环境指标,分析各指标等级,利用逻辑回归、支持向量机和决策树3种数理统计模型,计算全区滑坡灾害易发性程度,分析3种日降雨工况下滑坡的发生概率,得到各日降雨工况下万州区滑坡灾害危险性分布图。确定了支持向量机模型为万州区滑坡灾害易发性分析的最优模型;万州区滑坡灾害高易发区和高危险区主要表现出沿河道水系呈带状分布、沿高程垂直分布、在城镇区集中分布的特点;特定工况下,万州区滑坡灾害危险性随着日降雨量增大而增大。     

逻辑回归和信息量模型在山区滑坡易发性评价中适宜性的分析

以万山区为例,在区域滑坡孕灾条件的基础上,筛选工程地质岩组、斜坡结构、平均坡度、地貌、距构造距离及距河流距离共6个易发条件因子,选取逻辑回归模型和信息量模型对山区滑坡进行易发性评价。结果显示逻辑回归模型中中高易发区面积占比分别为1578%和1970%,82%的地质灾害点落在该区域内;信息量模型中中高易发区面积占比为1241%、2519%,包含了区域88%的滑坡灾害点。最后通过实际发生的灾害点在各易发区的分布情况进行检验,逻辑回归模型中灾害点落在高易发区的比例远小于信息量模型,且高易发等级中灾害点实际发生的比值较小,说明针对山区区域滑坡地质灾害易发性评价结果预测上,信息量模型的评价结果更为客观准确。     

基于信息量与Logistic模型的矿区滑坡灾害易发性评价

针对矿区长期煤矿开采引起的滑坡灾害频发问题,快速高效地模拟和评价矿致滑坡灾害易发性是实现采矿地区科学防灾减灾的关键。基于此,本文应用信息量与Logistic回归模型结合多源高分辨率光学遥感数据等,选取相对高差、坡度、坡向、距断层距离、NDVI、距采空区距离6个滑坡影响因子来评价采煤矿区滑坡灾害易发性。结果表明:(1)信息量与Logistic回归模型耦合的综合预测准确率为96%,信息量模型滑坡预测准确率为95%,实验结果表明耦合模型的预测精度优于单一信息量评价模型,评价模型的合理性和预测精度皆符合检验要求;(2)研究结果也表明了采用信息量+Logistic回归模型耦合能较为客观准确、快速高效地评价地下采矿引起的滑坡灾害易发范围,评价结果可为类似地区高效快速划定滑坡灾害易发区间提供技术支撑。     

基于GIS的陕西省旬阳地区滑坡灾害危险性区划

西部地区是我国地质灾害的重灾区。随着西部大开发战略的实施，该地区即将开展大规模的基础建设、能源开发等。区域内的经济发展与地质灾害的矛盾将不可避免地暴露出来。为解决这一问题，论文选取中国滑坡重灾区的江汉流域开展灾害危险性区划应用研究。研究区选在旬阳地区的县城近郊，通过MAPGIS软件平台及其二次开发的滑坡灾害分析系统，采用规则网格单元划分方法，运用信息量模型对该区斜坡稳定性进行了．空间定量预测，并依信息量法的结果编制了该区的危险性预测分区图。为政府部门进行土地规划决策、避免在地质灾害易发区进行大规模土地开发和工程建设提供了科学依据。     

基于ArcGIS模型的滑坡地质灾害敏感性评价——以金沙江上游为例

针对一些灾害评价模型过于复杂,影响因子选择过多的情况,在前人研究的基础上,选取坡度、工程地层岩性、断裂构造、地形起伏度4个因素作为滑坡地质灾害的影响因子。以金沙江上游白玉至巴塘段为研究区域,利用GIS技术,基于信息量模型,分析不同因子对滑坡地质灾害的敏感性,并利用信息量值绘制出滑坡灾害敏感图。结果表明,滑坡灾害点分布与敏感性分级具有显著相关性。评价方法可对类似金沙江上游这样的山地区域性滑坡灾害敏感性评价提供思路和参考。     

基于GIS技术的巴东新城区滑坡灾害危险性区划

基于ArcGIS8软件平台开发了三峡库区巴东县新城区滑坡灾害信息系统,通过全面分析巴东新城区滑坡灾害的地形条件,工程地质岩组、构造与斜坡结构类型、人类工程活动、水的作用等影响因素,建立了相应的滑坡灾害危险性评价指标体系。采用基于GIS技术的信息量模型和敏感性评价方法,实现了巴东县新城区滑坡灾害危险性区划,其中,高危险区面积3.30km^2,占7.196;中危险区面积5.77km^2,占12.4％;低危险区面积16.40km^2,占35.3％;基本安全面积21.05km^2,占45.2％,可以作为巴东新城区城镇建设规划和减灾防灾的参考依据。     

基于Random Forest 和AHP的贵德县北部山区滑坡危险性评价

滑坡危险性评价是滑坡灾害防治和管理的重要依据。文章基于层次分析法和随机森林模型,结合距离函数法,探索性地提出了一种新的组合赋权法（RF-AHP）。采用RF-AHP对青海省贵德县北部山区滑坡进行了危险性评价,对比探讨了AHP、RF和RF-AHP三种模型评价结果与实际滑坡灾害的吻合性,结果表明:（1）RF-AHP在高危险区和极高危险区面积占比38.38%的情况下,包括了60.13%的滑坡灾害,结果准确性相比AHP和RF两种模型有较大提升;（2）随着危险性等级的逐步提高,RF-AHP区划结果中相应分区的灾害实际发生的比率也随之增高,并对三种方法出现结果差异的客观原因进行了分析讨论,证明RF-AHP适用于滑坡危险性评价工作。     

浙江省永嘉县区域滑坡灾害人口易损性评价和伤亡风险预测

区域滑坡灾害人口易损性及人口伤亡风险预测研究是区域滑坡灾害预警预报工作的一个重要环节,该研究对提高预警预报工作的针对性和有效性具有关键作用.在对浙江省永嘉县有关资料进行分析的基础上,从研究区人口年龄结构、居民对滑坡灾害风险的防范意识、政府对滑坡灾害的重视程度及滑坡灾害预警预报体系的完善程度4个方面评价了研究区人口易损性,并给出了计算人口易损性的公式,据此得到了永嘉县人口易损性分布图.根据永嘉县的实际情况,提出了耕地人口密度的概念.综合人口易损性分布图、人口密度分布图和滑坡灾害易发性预测图得到了研究区受威胁人口伤亡风险预测图,为当地政府职能部门实施滑坡灾害风险的控制和管理提供决策依据.     

Landslide hazard zonation using analytical hierarchy process along National Highway-3 in mid Himalayas of Himachal Pradesh,India

Landslide zonation studies emphasize on preparation of landslide hazard zonation maps considering major instability factors contributing to occurrence of landslides. This paper deals with geographic information system-based landslide hazard zonation in mid Himalayas of Himachal Pradesh from Mandi to Kullu by considering nine relevant instability factors to develop the hazard zonation map. Analytical hierarchy process was applied to assign relative weightages over all ranges of instability factors of the slopes in study area. To generate landslide hazard zonation map, layers in geographic information system were created corresponding to each instability factor. An inventory of existing major landslides in the study area was prepared and combined with the landslide hazard zonation map for validation purpose. The validation of the model was made using area under curve technique and reveals good agreement between the produced hazard map and previous landslide inventory with prediction accuracy of 79.08%. The landslide hazard zonation map was classified by natural break classifier into very low hazard, low hazard, moderate hazard, high hazard and very high landslide hazard classes in geographic information system depending upon the frequency of occurrence of landslides in each class. The resultant hazard zonation map shows that 14.30% of the area lies in very high hazard zone followed by 15.97% in high hazard zone. The proposed model provides the best-fit classification using hierarchical approach for the causative factors of landslides having complex structure. The developed hazard zonation map is useful for landslide preparedness, land-use planning, and social-economic and sustainable development of the region.     

Landslide hazard zoning along Himalayan Kaghan Valley of Pakistan��by integration of GPS, GIS, and remote sensing technology

This paper presents a methodology for developing a landslide hazard zonation map by integration of global positioning system (GPS), geographic information system (GIS), and remote sensing (RS) for Western Himalayan Kaghan Valley of Pakistan. The landslides in the study area have been located and mapped by using GPS. Eleven causative factors such as landuse, elevation, geology, rainfall intensity, slope inclination, soil, slope aspect, distances from main road, distances from secondary roads, and distances from main river and those from trunk streams were analyzed for occurrence of landslides. These factors were used with a modified form of pixel-based information value model to obtain landslide hazard zones. The matrix analysis was performed in remote sensing to produce a landslide hazard zonation map. The causative factors with the highest effect of landslide occurrence were landuse, rainfall intensity, distances from main road, distances from secondary roads, and distances from main river and those from trunk streams. In conclusion, we found that landslide occurrence was only in moderate, high, or very high hazard zones, and no landslides were in low or very low hazard zones showing 100% accuracy of our results. The landslide hazard zonation map showed that the current main road of the valley was in the zones of high or very high hazard. Two new safe road routes were suggested by using the GIS technology.     

Influence of Shannon’s entropy on landslide-causing parameters for vulnerability study and zonation—a case study in Sikkim, India

Landslide is a common hazard in the hilly regions, which causes heavy losses to life and properties every year. Since 1980, various researches and analyses have been carried out in the geographic information systems (GIS) environment to identify factors responsible for causing landslides. The important conditioning factors identified by the researchers are slope, geological, geomorphologic structures, and land use coupled with triggering factors like rainfall and a few of the anthropogenic activities. Almost all landslides vulnerability studies carried out so far used parameters of landslide events of the past as essential inputs and advanced methods like information value, regression analysis, fuzzy logic, etc. The present research is an attempt to investigate the landslide vulnerabilities in different slope areas with simple and realistic method of assignments of weights to the parameters based on experts?? opinion and generic logic, without using the parameters of past landslide events as inputs. The identified factors were assigned appropriate weights based on experts?? opinion and these weights were further balanced with respect to the Shannon??s entropy of their occurrences within the study area. The study area was finally classified into three zones namely least vulnerable zone, moderately vulnerable zone, and most vulnerable zone. When compared with the actual landslide history of the past, it was found that Shannon??s entropy applied zonation model matched to real landslide events with higher value of landslide density as compared to the model developed without Shannon??s entropy.     

基于WEB的浙江省降雨型滑坡预警预报系统

滑坡发生的最主要诱发因素是降雨.基于浙江省淳安、磐安、庆元和永嘉4个县的历史滑坡资料, 在研究降雨量、降雨强度和降雨过程与滑坡灾害的空间分布、时间上的对应关系, 建立起滑坡灾害时空分布与降雨过程的统计关系, 确定区域性滑坡的临界降雨量和降雨强度阀值的基础上, 开发出了降雨型滑坡预警预报系统, 其中采用临界降雨量模型和有效降雨量模型来进行预警预报.系统留给浙江省气象台一个传送数据的接口, 气象台将每天按时上传降雨数据, 数据保存在后台数据库MicrosoftSQLServer中.建成的系统能够自动获取数据库中的数据生成时间与降雨量实时曲线, 当降雨量达到降雨强度阀值时, 触发MAPGIS图件, 在Internet上发布区域预警预报信息, 并提供预警措施.      

Landslide hazard mapping along national highway-154A in Himachal Pradesh,India using information value and frequency ratio

For the socio-economic development of a country, the highway network plays a pivotal role. It has therefore become an imperative to have landslide hazard assessment along these roads to provide safety. The current study presents landslide hazard zonation maps, based on the information value method and frequency ratio method using GIS on 1:50,000 scale by generating the information about the landslide influencing factors. The study was carried out in the year 2017 on a part of Ravi river catchment along one of the landslide-prone Chamba to Bharmour road corridor of NH-154A in Himachal Pradesh, India. A number of landslide triggering geo-environmental factors like “slope, aspect, relative relief, soil, curvature, Land Use and Land Cover (LULC), lithology, drainage density, and lineament density” were selected for landslide hazard mapping based on landslide inventory. The landslide inventory has been developed using satellite imagery, Google earth and by doing exhaustive field surveys. A digital elevation model was used to generate slope gradient, slope aspect, curvature, and relative relief map of the study area. The other information, i.e., soil maps, geological maps, and toposheets, have been collected from various departments. The landslide hazard zonation map was categorized namely “very high hazard, high hazard, medium hazard, low hazard, and very low hazard.” The results from these two methods have been validated using area under curve (AUC) method. It has been found that hazard zonation map prepared using frequency ratio model had a prediction rate of 75.37% while map prepared using information value method had prediction rate of 78.87%. Hence, on the basis of prediction rate, the landslide hazard zonation map, obtained using information value method, was experienced to be more suitable for the study area.     

基于逻辑回归模型和确定性系数的崩滑流危险性区划

崩滑流是崩塌、滑坡和泥石流地质灾害的总称。本文根据逻辑回归模型和贵州省崩滑流地质灾害发生的确定性系数CF,统计贵州省内崩滑流发生概率与其影响因子之间的函数关系; 并利用GIS技术编制贵州省崩滑流地质灾害危险性区划图。首先根据影响因子子集中已发崩滑流灾害面积和影响因子子集面积来计算崩滑流地质灾害发生的确定性系数CF; 其次将灾害是否发生作为因变量,影响因子子集发生崩滑流地质灾害的确定性系数CF作为自变量,应用逻辑回归模型统计分析它们之间的函数关系; 然后利用GIS技术计算研究区内各独立属性单元发生崩滑流地质灾害的概率p,按p值10等分标准将研究区划分为10个危险性等级区,并绘制贵州省崩滑流地质灾害危险性区划图; 最后用已发崩滑流地质灾害的分布数据来检验危险性区划的效果。研究结果表明:本文根据逻辑回归模型和崩滑流地质灾害发生的确定性系数CF,将贵州省分为Ⅰ～Ⅹ的10个崩滑流地质灾害危险性等级区与实际情况基本符合,能够良好地反映贵州省境内发生崩滑流地质灾害的难易程度。     

证据权法在区域滑坡危险性评价中的应用以贵州省为例

以GIS为技术平台,采用证据权法对研究区进行了滑坡地质灾害危险性分析。综合分析历史滑坡数据及其环境因素和触发因素,数据源主要有地形图、DEM、地质图,选取地层岩性、构造、高程、坡度、坡向、地形起伏度、道路、水系作为危险性评价因子。首先应用ArcGIS软件对数据源进行处理,提取各个评价因子图层,并对每个图层进行分级、缓冲区分析等处理,建立若干证据层。然后将历史灾害点与评价因子进行空间关联分析,计算每个评价因子等级的权重,最后计算出评价单元的危险性指数,并将危险性分为极高危险区、高危险区、中等危险区、低危险区。采用成功率曲线法对证据权法评价精度进行验证,结果表明本次评价的精度为71%。利用历史滑坡数据对评价结果进行验证,结果显示评价结果与实际情况较为吻合,说明证据权可以客观定量地评价各影响因子对滑坡的影响程度,该方法应用于区域地质灾害危险性评价比较有效。     

降雨型滑坡时间概率的逻辑回归拟合及连续概率滑坡危险性建模

提高降雨型滑坡危险性预警精度和空间辨识度具有重要意义.以江西宁都县1980—2001年156个降雨型滑坡为例,首先基于传统的EE-D（early effective rainfall-rainfall duration）阈值法计算不同降雨诱发滑坡的时间概率级别;然后以各级别临界降雨阈值曲线对应的时间概率为因变量,并以对应的前期有效降雨量（early effective rainfall,EE）和降雨历时（D）为自变量,采用逻辑回归拟合出上述因变量与自变量之间的非线性关系,得到降雨诱发滑坡的连续概率值;之后对比C5.0决策树和多层感知器的滑坡易发性预测性能;最后利用降雨诱发滑坡的连续概率值与易发性图相耦合以实现连续概率滑坡危险性预警.结果显示:（1）宁都降雨型滑坡连续概率值的逻辑回归方程为1/P=1+e4.062+0.747 4×D-0.079 44×EE,其拟合优度为0.983;（2）2002—2003年的20处用于连续概率阈值测试的降雨型滑坡大都落在连续概率值大于0.7的区域,只有4处落在小于0.7的区域;（3）C5.0决策树预测滑坡易发性的精度显著高于多层感知器;（4）近5年的4次降雨型滑坡的连续概率危险性值都在0.8以上,且高和极高预警区的面积较传统滑坡危险性分区更小.可见连续概率滑坡危险性预警法相较于传统危险性分区法具有更高的预警精度和空间辨识度,且通过叠加滑坡易发性图及其临界降雨阈值可开展实时滑坡危险性预警制图.