基于GIS的兰州地区滑坡灾害易发性评价

发布时间：2018-04-19 00:36

本文选题：滑坡 + 孕灾环境　；参考：《兰州理工大学》2017年硕士论文

【摘要】：兰州地区作为西北重要的工业与商贸中心城市,由于其地质环境复杂、人类工程活动明显,导致近年来该地区地质灾害频繁发生,严重威胁当地人们的生命和财产安全。针对这一情况,本文选取兰州地区及周边作为研究对象,基于ArcGIS平台,利用兰州地区1:5万DEM数据与遥感影像提取坡度、坡向等致灾因子,结合致灾因子图层与历史灾害数据进行了研究区地质灾害空间分布特征分析;综合利用敏感性指数(SI)、确定性系数(CF)和相关系数方法进行致灾因子在区内地质灾害研究中的适宜性分析;形成一套普遍适用的因子建立体系;根据筛选出的致灾因子,结合研究区的地质、地貌数据,利用概率指数模型(PI)、信息量模型(IM)和逻辑回归模型(LR)对兰州及周边地区的地质灾害进行了危险性评价研究并进行了精度分析;通过兰州地区2000年、2008年、2013年遥感影像利用ENVI平台分析了十三年来兰州地区植被覆盖度变化,最终得出结论:(1)对兰州地区进行的灾害空间分布特征分析表明,在研究区,地质灾害通常易发生在坡度25°~40°、高程1700m~2600m、15m~45m地形起伏度、切割深度5m~30m的东南、南、西南凸形阳坡;(2)本研究通过致灾因子在研究区的适应性筛选,建立了一套对研究区普遍适用的因子建立体系,并根据此因子建立体系,筛选出坡度、坡向、高程、坡度变率、地形起伏度和地表切割度作为研究区地质灾害易发性评价的最终致灾因子;(3)利用三种灾害评价模型进行的灾害易发性评价与精度分析发现,三种模型下地质灾害的敏感性强弱分布趋势基本一致,其中,信息量模型所对应的ROC曲线下面积较大,逻辑回归(Logistic)模型次之,概率指数模型略低。模型精度评价表明,信息量模型精度略高于Logistic回归模型,概率指数模型评价精度最低,即对研究区而言,采用信息量模型进行灾害易发性评价的结果最为准确。(4)灾害易发性评价结果表明,研究区内,位于地质灾害极高易发区的县(区)为:城关区、西固区、永登县;位于高易发区的县(区)有:七里河区、安宁区;位于易发区的县(区)为皋兰县;相对较稳定(不易发区)的县(区)为榆中县。(5)分析兰州13年间植被覆盖度变化情况发现,兰州市南北两山植被总体呈增长的趋势,其中的劣等植被覆盖面积变化最大;2013年兰州南北两山植被已经进入很好的生长期,两山绿化和生态恢复产生了成效。由于人们的绿化意识不断提高,政府的大力投资和对南北两山进行的生态保护、恢复等工作,对两山的绿化、植被恢复方面起了重要作用。
[Abstract]:Lanzhou is an important industrial and trade center city in the northwest of China. Because of its complex geological environment and obvious human engineering activities, geological disasters occur frequently in this area in recent years, which seriously threaten the safety of local people's lives and property.In view of this situation, this paper selects Lanzhou area and its periphery as the research object, based on the ArcGIS platform, uses the Lanzhou area 1: 50,000 DEM data and the remote sensing image to extract the slope, the slope direction and so on disaster causing factors, etc.The spatial distribution characteristics of geological hazards in the study area are analyzed based on the disaster-causing factor layer and historical disaster data.The suitability of disaster factors in the study of geological hazards in the region is analyzed by using the sensitivity index (SI), the deterministic coefficient (CFC) and the correlation coefficient method, and a set of universally applicable factors is established.Based on the geological and geomorphological data of the study area, the risk assessment and accuracy analysis of geological hazards in Lanzhou and its surrounding areas are carried out by using probabilistic exponential model (Pi), information quantity model (IMM) and logical regression model (LR).Based on the remote sensing images of Lanzhou in 2000, 2008 and 2013, the changes of vegetation coverage in Lanzhou during the past 13 years were analyzed by using ENVI platform. Finally, a conclusion was drawn that the spatial distribution of disasters in Lanzhou area was analyzed by the analysis of the spatial distribution characteristics of disasters in Lanzhou area.Geological hazards usually occur at 25 掳~ 40 掳slope, 1 700 m ~ 2 600 m ~ 2 600 m ~ (15) m ~ (45 m) topographic undulation, cut depth of 5m~30m in southeast, south and southwest convex sunny slope.) in this study, the adaptive selection of disaster factors in the study area is carried out.In this paper, a system of establishing factors which is generally applicable to the study area is established, and according to this system, the slope, direction, elevation, gradient change rate are screened out.Topographic fluctuation and surface cutting are the ultimate hazard factors for the assessment of geological hazard vulnerability in the study area. The results of the assessment and accuracy analysis of the three kinds of disaster assessment models are as follows.The distribution trend of the sensitivity of geological hazards under the three models is basically the same. Among them, the area under the ROC curve corresponding to the information content model is larger, the logical regression logistic model is the second, and the probability index model is slightly lower.The accuracy evaluation of the model shows that the accuracy of the information quantity model is slightly higher than that of the Logistic regression model, and the probabilistic exponential model has the lowest accuracy, that is, for the study area,The results of disaster vulnerability evaluation based on information quantity model show that the counties (districts) located in extremely high geological hazard prone areas are Chengguan District, Xigu District and Yongdeng County. The results show that the evaluation results of disaster vulnerability are as follows: (1) in the study area, the counties (districts) are: Chengguan District, Xigu District, Yongdeng County;The counties (districts) located in the high susceptible areas are: Qili River District, Anning District; Gaolan County, which is located in the susceptible area; and Yuzhong County, Yizhong County, which is relatively stable (not susceptible area).The vegetation of the north and south mountains of Lanzhou showed an increasing trend in general, in which the inferior vegetation cover area changed the most; in 2013, the vegetation of the north and south mountains of Lanzhou had entered a very good growing period, and the greening and ecological restoration of the two mountains had produced results.Because people's consciousness of greening has been improved, the government's great investment, ecological protection and restoration of the north and south mountains have played an important role in the greening of the two mountains and the restoration of vegetation.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P642.22

【参考文献】