基于GIS和专家知识的滇西南地区滑坡敏感性模糊逻辑推理方法

(1.云南师范大学 地理学部,云南 昆明650500; 2.云南省地震局,云南 昆明 650224; 3.中国地震局地质研究所,北京 100029)

滑坡敏感性; 专家知识; 模糊逻辑; 约束算术平均法; 滇西南地区

A Fuzzy Logic Modeling of Landslide Susceptibility Mapping in Southwest Yunnan Province Based on GIS and Expert Knowledge
BAI Xianfu1,2,DAI Yuqian1,YE Liaoyuan1,NIE Gaozhong3,XU Shuo2,LUO Weidong2,YU Jiang2,XU Xin2

(1.Faculty of Geography Science,Yunnan Normal University,Kunming 650050,Yunnan,China)(2.Yunnan Earthquake Agency,Kunming 650224,Yunnan,China)(3.Institute of Geology,China Earthquake Administration,Beijing 100029,China)

landslide susceptibility; expert knowledge; Fuzzy Logic modeling; Constrained Arithmetic Mean method; southwest Yunnan Province

DOI: 10.20015/j.cnki.ISSN1000-0666.2022.0013

备注

为了充分识别和有效减轻滑坡灾害风险,对滇西南南涧(约470 km2)和凤庆—昌宁(约2 300 km2)两个研究区开展了基于GIS和专家知识的滑坡敏感性模糊逻辑评价研究。通过检查模型计算得到的历史滑坡点敏感性值与整个研究区域的滑坡敏感性平均值是否不同来评价本方法的性能,用Z值检查来测试差异的统计显著性。计算结果显示,南涧地区的Z值为4.1,相应的P值小于0.001,表明通过模型计算得到的滑坡敏感性值是该区域滑坡事件发生的良好指标; 凤庆—昌宁地区的Z值为8.93,相应的P值小于0.001。在此基础上,采用自然断点法对滑坡敏感性值进行分类,根据分类结果将滑坡敏感性水平划分成5个等级:极低(0.0~0.001)、较低(0.001~0.051)、中等(0.051~0.394)、较高(0.394~0.557)和极高(0.557~1.0)。敏感性极低和较低的地区没有发现历史滑坡记录; 敏感性极高地区的历史滑坡密度约是敏感性较高地区的4倍,约为敏感性中等地区的10倍。凤庆—昌宁地区的研究结果表明,从区域专家群中提取的滑坡敏感性与环境因子关系的知识可以外延到滇西南其它地区。
In order to fully identify and effectively reduce the risk of landslide hazards,we presented an approach to landslide susceptibility mapping in southwest Yunnan Province based on the expert knowledge.The approach was tested in two study areas in southwest Yunnan:Nanjian(about 470 km2)and Fengqing-Changning(about 2 300 km2)[KG-*3].The performance of our approach was evaluated by examining if the mean value of the computed susceptibility values at landslide sites was statistically different from that of the entire study area.A Z-score test was used to examine the statistical significance of the difference.The computed Z-score for the Nanjian area was 4.1 and the corresponding P-value was less than 0.001.This suggested that the computed landslide susceptibility values were good indicators of landslide occurrences.In the Fengqing-Changning study area,the computed Z-score was 8.93 and the corresponding P-value was less than 0.001.In addition,we divided the susceptibility value into five levels by the Natural Break Point method.The result sets like:very low(0.0-0.001),low(0.001-0.051),moderate(0.051-0.394),high(0.394-0.557)and very high(0.557-1.0).No landslides were found in areas of very low and low susceptibility.Landslide density was about four times higher in areas of very high susceptibility than that in the high susceptibility areas,and about 10 times as high as that in the moderate susceptibility areas.Results from these case studies suggest that our approach based on expert knowledge is workable in mapping landslide susceptibility and that its performance is maintained when it is applied to a new area.