四川省南江县地质灾害易发性区划研究

发布时间：2018-10-22 13:01

【摘要】：南江县地质灾害频繁发生,目前编录调查的地质灾害点达到上千处。南江县县域中南部地层为红层,红层属于易滑地层,在外界诱发因素的作用常产生大量地质灾害。近年来,随着全球气候变暖,出现极端强降雨天气的概率变大,例如2011年“9.16”强降雨诱发南江县上千处地质灾害发生。鉴于数量与规模如此巨大的地质灾害,很有必要进行南江县地质灾害易发性区划,为地质灾害防灾减灾工作指明方向。地质灾害易发性区划就是评价各个地区产生地质灾害可能性的大小。本文依托中国地质调查局地质调查工作项目“西南地区重大地质灾害调查与预警区划(12120113010100)”,对研究区进行了详细的野外地质调查,并在此基础进行研究区的地质灾害易发性区划。首先,对南江县区域地质环境条件、地质灾害发育分布规律、形成条件以及诱发因素等进行分析,再根据它们与地质灾害点之间的相关性,选取坡度、剖面曲率、坡高、岩土体类型、斜坡结构类型、地形湿度指数、距水系距离、7~9月份降雨量与距道路距离9个地质灾害灾害易发性评价因子。根据评价指标特征,将其分为基本环境因素与诱发因素两类。选取Logistic回归模型与模糊综合评价模型两种模型进行南江县地质灾害易发性区划。将地质灾害易发性等级分为高易发性、中易发性、低易发性与不易发性四个等级。Logistic回归模型评价指标采用逐步回归的方法进入模型,根据回归分析结果,最终确定坡度、坡高、岩土体类型、距水系距离、7~9月份降雨量与距道路距离6个评价指标作为南江县易发性区划评价指标。采用SPSS软件确定Logistic回归系数,采用地质灾害点密度确定评价因子指标值,运用ArcGIS软件进行叠加分析,最终得到基于Logistic回归模型的南江县地质灾害易发性区划图。模糊综合评价模型采用层次分析法确定评价因子权重,二级模糊综合评判进行地质灾害易发性区划。采用Python语言编程进行评价指标数据的处理,计算评价单元的易发性等级,简化了复杂的数学计算过程,提高了制图效率。采用ROC曲线与Kappa系数评价两种模型区划结果的精度,通过对比分析,基于Logistic回归模型的区划结果精确度更高。因此,选取Logistic回归模型的区划成果图作为南江县地质灾害易发性区划图。评价结果表明:地质灾害高易发区主要分布在南江县中南部红层地区和北部山区低海拔沟谷内;地质灾害中易发区分布在红层地区;地质灾害低易发区主要分布在1000~1500m的斜坡上;地质灾害不易发区主要分布在南江县北部海拔大于1500m的高海拔地区。通过与实际调查结果的对比分析,地质灾害易发性区划结果是合理的。
[Abstract]:Nanjiang County geological disasters occur frequently, the current cataloguing survey of geological hazards to thousands. The south and central strata of Nanjiang County are red beds and the red beds belong to slippery strata. A large number of geological hazards are often caused by the action of external inducing factors. In recent years, with the global warming, the probability of extreme heavy rainfall weather has become greater, for example, "9.16" heavy rainfall induced thousands of geological disasters in Nanjiang County in 2011. In view of the large number and scale of geological disasters, it is necessary to carry out the geological hazard prone regionalization in Nanjiang County, so as to point out the direction of geological disaster prevention and mitigation work. The regionalization of geological hazard vulnerability is to evaluate the possibility of geological hazard in each area. Based on the geological survey project of China Geological Survey Bureau, "investigation and early warning regionalization of major geological hazards in southwest China (12120113010100)," detailed field geological survey has been carried out in this paper. On this basis, the geological hazard susceptibility zoning of the study area is carried out. First of all, the regional geological environment conditions, geological hazard development and distribution, formation conditions and induced factors are analyzed, and then according to the correlation between them and geological hazard points, slope, profile curvature, slope height are selected. The types of rock and soil, the type of slope structure, the index of topography humidity, the distance from the water system, the rainfall in September and the distance from the road are 9 factors to evaluate the vulnerability of geological hazards. According to the characteristics of evaluation index, it can be divided into two categories: basic environmental factors and induced factors. Two models, Logistic regression model and fuzzy comprehensive evaluation model, are selected to regionalize the susceptibility of geological hazards in Nanjiang County. The grade of geological hazard vulnerability is divided into four grades: high susceptibility, moderate vulnerability, low susceptibility and non-susceptibility. The evaluation index of Logistic regression model is entered into the model by stepwise regression method. According to the results of regression analysis, the slope and slope height are finally determined. The types of rock and soil, distance from water system, rainfall in July and September and distance from roads are used as evaluation indexes of susceptibility regionalization in Nanjiang County. The Logistic regression coefficient is determined by SPSS software, the evaluation factor index value is determined by using geological hazard point density, and the superposition analysis is carried out by using ArcGIS software. Finally, the geological hazard susceptibility zoning map of Nanjiang County based on Logistic regression model is obtained. The fuzzy comprehensive evaluation model uses the analytic hierarchy process to determine the weight of the evaluation factors, and the secondary fuzzy comprehensive evaluation is used to regionalize the vulnerability of geological hazards. The Python language is used to process the evaluation index data, and the vulnerability grade of the evaluation unit is calculated, which simplifies the complicated mathematical calculation process and improves the drawing efficiency. The ROC curve and Kappa coefficient are used to evaluate the accuracy of the regionalization results of the two models. By comparison and analysis, the accuracy of the regionalization results based on the Logistic regression model is higher. Therefore, the regionalization result map of Logistic regression model is selected as the map of geological hazard susceptibility in Nanjiang County. The evaluation results show that the high risk areas of geological hazards are mainly distributed in the red beds in the central and southern part of Nanjiang County and in the low elevation gully in the northern mountainous area, the prone areas in the geological hazards are in the red bed areas, and the low risk areas of geological hazards are mainly distributed on the slopes of 1000 ~ 1500m. Geological hazard prone areas are mainly distributed in the high altitude areas of the northern part of Nanjiang County where the altitude is more than 1500m. By comparing with the actual investigation results, the results of geological hazard susceptibility regionalization are reasonable.
【学位授予单位】：成都理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P694

【参考文献】