安徽省歙县地质灾害易发性区划与气象预警研究

发布时间：2018-02-04 15:53

本文关键词： Logistic 回归模型评价指标易发性区划气象预警模型　出处：《成都理工大学》2015年硕士论文　论文类型：学位论文

【摘要】：安徽省黄山市歙县地处皖南山区,区域地质环境脆弱,加之不合理的人类工程活动进一步恶化地质环境,在降雨作用下导致灾害频发,造成地质灾害表现出“点多面广、突发性强、灾害损失大”等特点,因此地质灾害防治与预警工作显得尤为重要。但是现有地质灾害预警预报过分强调气象预警,没有实现降雨与地质灾害形成的地质条件的有机结合,预警预报精度不高。本文以研究区502个历史灾害点及其部分灾害点对应的降雨数据为基础,深入研究了黄山市歙县地质灾害易发性区划方法、降雨诱发地质灾害的时空分布与降雨的关系,完成了研究区地质灾害易发性区划,确定了降雨临界值表达式,创建了区域地质灾害气象预警方法,实现了对地质灾害区域气象预警。本文取得的主要研究成果如下:(1)歙县易发地质灾害类型主要为滑坡、崩塌和不稳定斜坡,主要呈现出主要分布在偏东、东北的城镇,沿控制性山脉、断层具有分带性,沿道路线性分布明显,低海拔地区具有不均匀性,与人类工程活动强烈程度和降雨量大小呈明显正相关关系。(2)选取影响该区域地质灾害发生的10个因子(高程、坡高、坡度、坡体形态、地层岩性、断层分布、河流切割、植被覆盖率、道路密度、居民点分布)建立了地质灾害易发性评价指标体系。根据研究区历史地质灾害点在各评价指标内的点密度和点比例突变点作为该指标的二次分级点,并对各级指标进行量化归一。(3)采用Logistic回归模型建立了研究区地质灾害易发性区划评价模型,以研究区历史灾害点和随机非灾害点作为统计样本,建立歙县地质灾害易发性(Pa)的概率计算公式,并绘制了研究区地质灾害易发性区划图。区划结果与研究区地质灾害发育实情相符合,效果较好。(4)确定了降雨单因素诱发地质灾害的概率(Pb)计算公式和降雨量临界表达式。以前期降雨量与当日日降雨量作为评级指标,建立了降雨单因素诱发地质灾害Logistic回归模型,由此确定了前期累计降雨量的范围、降雨单因素诱发地质灾害的概率(Pb)计算公式和降雨量临界表达式。(5)建立了地质灾害的气象预警预报模型,将研究区地质灾害易发性概率值Pa作为空间因子,将降雨单因素诱发地质灾害概率值Pb作为时间和空间的辅助因子,通过Pa与Pb进行乘叠,得到地质灾害发生的实时空间预报概率值Pc。根据Pc值将降雨诱发地质灾害时空预警预报等级分为5级,不同等级对应不同程度的防御措施。
[Abstract]:Shexian County, Huangshan City, Anhui Province is located in southern Anhui mountain area, the regional geological environment is fragile, coupled with unreasonable human engineering activities to further deteriorate the geological environment, resulting in frequent disasters under the action of rainfall. Geological hazards are characterized by "wide points, strong sudden occurrence and large disaster losses". Therefore, geological disaster prevention and early warning work is particularly important, but the existing geological disaster early warning and forecast too much emphasis on meteorological early warning, did not achieve the organic combination of rainfall and geological hazards formed by the geological conditions. The prediction accuracy is not high. Based on the rainfall data of 502 historical disaster sites and some disaster sites in the study area, the method of geological hazard susceptibility zoning in Shexian County, Huangshan City, is studied in this paper. The relationship between spatial and temporal distribution of geological hazards induced by rainfall and rainfall, the regionalization of geological hazard susceptibility in the study area is completed, the expression of rainfall critical value is determined, and the method of regional geological hazard meteorological early warning is established. The main research results obtained in this paper are as follows: 1) the main types of geological hazards in Shexian County are landslide, collapse and unstable slope, mainly distributed in the east. The northeast town, along the control mountain range, the fault has the zonation, along the road the linear distribution is obvious, the low elevation area has the inhomogeneity. There are 10 factors (elevation, slope height, slope degree, slope shape, stratigraphic lithology) that affect the occurrence of geological hazards in this area, which have a significant positive correlation with the intensity of human engineering activities and rainfall. Fault distribution, river cutting, vegetation coverage, road density. The evaluation index system of geological hazard vulnerability is established. According to the point density and the point proportion mutation point of the historical geological hazard site in each evaluation index in the study area, the secondary grading point of the index is established. The evaluation model of geological hazard susceptibility regionalization in the study area is established by using Logistic regression model. The probability calculation formula of geological hazard susceptibility in Shexian County is established by taking historical disaster points and random non-disaster points as statistical samples in the study area. The zoning map of the susceptibility of geological hazards in the study area is drawn. The results of the regionalization are in accordance with the development of geological hazards in the study area. The calculation formula of probability and critical expression of rainfall induced geological hazard by single factor rainfall are determined. The index of rating is pre-precipitation and daily rainfall on the same day. The Logistic regression model of geological hazards induced by single factor rainfall was established, and the range of accumulated rainfall was determined. The calculation formula of probability Pb) and the critical expression of rainfall induced geological hazard by single factor rainfall. 5) the meteorological early warning and forecasting model of geological hazard is established. The probabilistic value Pa of geological hazard in the study area is taken as spatial factor, and the probability value of geological hazard induced by rainfall is regarded as the auxiliary factor of time and space, and the probability value of geological hazard induced by rainfall is multiplied by Pa and Pb. The probability value of real time spatial prediction of geological hazards is obtained. According to the values of PC, the spatial and temporal early warning and prediction levels of rainfall induced geological hazards are divided into 5 levels, and different grades correspond to different degree of defense measures.
【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】