【作者】 付建飞；

【导师】 王恩德；

【作者基本信息】 东北大学 ， 矿产普查与勘探， 2007， 博士

【摘要】 自然灾害尤其是地质灾害是人类可持续发展中面临的一个巨大的挑战,中国地质灾害种类多、分布广、危害大。特别是在汛期,受气象因素的影响,崩塌、滑坡、泥石流等突发性地质灾害频发,占地质灾害总数的80%以上。据1995年至2002年以来的统计资料显示,全国每年仅因突发性地质灾害造成一千多人死亡,经济损失高达几十亿元。巨大的灾害损失,不仅直接造成了人民生命财产的损失,而且冲击了社会发展的各项事业。辽宁省位于我国东北地区南部,处在中朝准地台和天山—兴蒙地槽褶皱系两个一级大地构造单元内,地质环境复杂,自然变异强烈,地质灾害发生频繁多,为全国地质灾害重灾区。巨大而频繁的灾害,严重地威胁着人民的生命和财产安全。因此,进行辽宁省滑坡和泥石流地质灾害的潜在性研究具有重要的理论和实际意义。本文以GIS为平台,利用地统计方法、Yamamoto检测法、权重证据法和Random forest等分析了辽宁省地质灾害的潜在性分布。研究取得的主要成果如下：(1)利用地统计方法、Yamamoto检测法、气候倾向率和趋势系数方法研究辽宁省过去40年的气候的时空变化,并利用加拿大气候模拟和分析中心(CCCma, Canadian Centre for Climate Modelling and Analysis)推出的第二代全球气候耦合模型CGCM2数据研究未来100年辽宁省气候的时空变化,研究表明辽宁省的气候有暖干化趋势,暖化期主要体现在春冬两季,干旱期体现在夏冬两季。突变检测表明,辽宁省气温有突变,其中夏季突变最明显,而降水基本无突变。空间分析表明,辽宁省的年均温由西南向东北方向、年降水由东向西方向都有逐渐递减的趋势。同样气温和降水的趋势系数也存在空间变异性,并且四季变化不一致。未来降水增加的可能性较大。未来百年的气温变化趋势比较明显,一直呈增加的趋势。(2)以GIS为平台,通过空间叠加取样和Buffer方法等研究辽宁省地层岩性、地质构造、岩浆岩、植被土壤等因子和地质灾害的空间关系,并探寻引起地质灾害的地质条件主导因素。研究表明辽宁省地质条件与地质灾害的空间关系密切,凤城凸起的辽东裂谷的多旋回演化和裂谷内的断裂为泥石流提供了大量的松散碎屑物质；受构造控制的北镇凸起断块,区内北东及北北东向的极为发育断裂和两侧有一宽度较大的韧性剪切带为地质灾害的提供了良好的发育条件。同样,植被和土壤也与辽宁省滑坡泥石流灾害有着较密切的空间关系。(3)以GIS为平台,利用权重证据法对辽宁省地质灾害的敏感性评价表明：岩浆岩、地质岩性和植被类型对地质灾害的影响最大,权重证据法在辽宁省地质灾害的敏感性评价中效果较好。权重指数图显示辽东山区为辽宁省地质灾害的高敏感区,其次辽西山区和辽宁省东北部为地质灾害的次敏感区,中部平原为地质灾害的低敏感区,另外沈阳市区附近有较高的敏感区。(4)在敏感性评价的基础上,选用地震、降水和气温为诱发因子,利用权重证据法研究辽宁省目前的地质灾害危险性评价。同时,以CGCM2提供的未来的气候条件为基础,进行未来50年和100年的地质灾害危险性评价,并进行三个时段危险性分区图的对比分析。研究表明辽东山区是辽宁省地质灾害危险性最大的区域,而同样,在沈阳市区附近存在着一个较小的危险性较高的区域,其它的较高的危险性区域是辽西和辽宁省东北部。未来地质灾害发生概率较大的区域将逐渐增大。到2050辽西地质灾害发生概率较高区域将减少,而从辽东中部山区向外围扩展地质灾害发生概率较高区域将增多,同时,在辽宁省东北部危险性高值区域也将增加。而到2100年,辽宁省危险性高值增加区域主要在辽宁省东北部。(5)在辽宁省危险性评价的基础上,选择危险性较高的凤城市为研究对象,收集RS数据、高程数据、日降水数据(从1971到2000年),地质数据和土壤数据,利用Random forest方法进行了凤城市地质灾害降水预警预测,研究表明：Random forest方法较好地预测了凤城市的降水预警图,凤城市1和2日累计降水对预警预测具有指示意义。相对来说,西北部降水诱发地质灾害的阈值较小,东南部降水诱发地质灾害的阈值较大。更多还原

【Abstract】 Natural disaster, especially the geologic hazard, is a great challenge in the processes of sustainable development of human being. The geologic hazards in China are is in a great variety, wide distribution and with great damage. In the flood season, driven by the climatic factors, avalanche, landslip, debris flow etc. are occurred frequently. These kinds of geologic hazard occupying above 80% of entire geologic hazard. According to the records of statistic data from 1995 to 2002, the geologic hazard would result in the death of more than 1 thousand people and losing of decades of billion Yuan per year, which would also have bad influence on all enterprise.Liaoning province is located in the south of Northeastern China, and is in the tectonic units between China-Korea paraplatform and Tianshan-Xingmeng geosyncline. There are complex geologic environment, high heterogeneity and various geologic hazard with wide distribution and big impaction. Since 1998 to 2002, there are 62 geologic hazards, which occurred in 13 cities. These had threaten the safety of people’s life and property. Thus, it is very important to study the potential distribution of geologic hazard in Liaoning Province. This dissertation analysed the potential distribution of geologic hazard using geostatistic, Yamamoto detection, Weight of Evidence and Random Forest on the GIS platform, and drew these conclusions:(1) Geostatistic, Yamamoto detection, climate tendency rate and tendency coefficient were used to explore the spatial and temporal changes of the climate in last 40 years. The CGCM2 model, brought forward by Canadian Centre for Climate Modelling and Analysis was used to supply the future climate scinarios. The research showed that there are warming and drying tendency of Liaoning province in last 40 years. The warming tendency is obvious in spring and winter and drying tendency is obvious in summer and winter. Yamamoto detection showed that there are break in the temperature, especially in summer. There are no break in the precipitation. The spatial analysis showed that, the mean annual temperature have the tendency of degressive from southwest to northeast, and the annual precipitation have the dendency of degressive from east to west. There are spatial variability of the tendency coefficient of temperature and precipitation. In future 100 years, the precipitation and temperature would both increase.(2) Spatial overlap and Butter analysis were used to study the relationships of stratum lithology, geological structure, magma rock, vegetation and soil and geologic hazard, aiming to explore the driven factors of geologic hazard. Research showed that there are intimate spatial relationship between geologic structure and geologic hazards. Liaodong Rift and the fracture structures in the rift supplied a lot of relax and chipping material for debreis flow. Beizheng heave, controlled by geologic structure, also supplied the condition for geologic hazard. At the same time, vegetation and soil factors also have relation to the landslide and debris flow of Liaoning Province.(3) Weight of Evidence was used to study the sensitivity of geologic hazard of Liaoning Province. Resutls showed that, magma rock, stratum lithology and vegetation are the most three important factors controlling the geologic hazard. Weight index map showed that Liaodong Mountains were the highly sensitive area of Liaoning Province, Liaoxi Mountains were next, while the middle plain was the lowly sensitive area. In addition, there are a small highly sensitive area near shenyang city.(4) On the basis of sensitive assessment analysis, we use earthquake, precipitation and temperature as the main factos to assess the fatalness of the geologic hazard by weight of evidence. At the same time, based on the future climate scenarios supplied by CGCM2, we assess the fatalness of the geologic hazard in future 50 and 100 years, then we compared the fatalness of the geologic hazard in three periods. The results showed that Liaodong Mountain is the area with highest fatalness and there are also a small area near shenyang city with higher fatalness. Other area with higher fatalness were Liaoxi Mountains and the northeast part of Liaoning Province. The high fatalness area in future would increase, the area with high fatalness in Liaoxi Mountain would decrease, but the area with high fatalness in Liaodong mountian and northeastern part would increase in 2050 year. The area with high fatalness in northeastern part of Liaoning Province would increase in 2100 year.(5) On the basis of analysis the fatalness assessment, fengchen city, with higher fatalness was chosen as the study are in small scale. The TM image, elevation, day precipitation data (from 1971 to 2000), geologic data and soil data of this area were collected to predict the relationship between precipitation and occurrence rate of geologic hazard by Random Forest technique. The results showed that 1 day and 2 day cummulative precipitation have a directive meaning for geologic hazard. In total area, the cutoff value in northwest part is small, while it is big in southeast part of the area.更多还原