地理科学进展 ›› 2006, Vol. 25 ›› Issue (4): 1-12.doi: 10.11820/dlkxjz.2006.04.001

• 地理灾害预警研究 •    下一篇

基于专家知识的滑坡危险性模糊评估方法

朱阿兴1,2, 裴韬1, 乔建平3, 陈永波3, 周成虎1, 蔡强国1   

  1. 科院地理科学与资源研究所, 北京100101|2 美国威斯康星大学地理系, 麦迪逊53706|3 中科院成都山地研究所, 成都610041
  • 收稿日期:2006-06-01 修回日期:2006-07-01 出版日期:2006-07-25 发布日期:2006-07-25
  • 作者简介:朱阿兴(1962-), 男, 博士, 研究员, 研究方向为:地理信息系统.E-mail:axing@lreis.ac.cn
  • 基金资助:

    国家重点基础研究发展计划项目( 2006CB701305) ; 中科院"百人计划"项目; 国家自然科学基金 ( 40225004)

A Landslide Susceptibility Mapping Approach Using Exper t Knowledge and Fuzzy Logic Under GIS

ZHU Axing1,2, PEI Tao1, QIAO JianPing3, CHEN Yongbo3, ZHOU Chenghu1, CAI Qiangguo1   

  1. 1. Institute of Geographic Sciences and Natural Resources Research, CAS, Beijing 100101, China|
    2. Department of Geography, University of Wisconsin Madison, 550N, Park Street, Madison 53706, USA|
    3. Institute of Mountain Hazards and Environment, CAS, Chengdu 610041, China
  • Received:2006-06-01 Revised:2006-07-01 Online:2006-07-25 Published:2006-07-25

摘要:

滑坡发生的影响因素众多, 其危险性与各因素之间的关系多呈非线性关系, 同时各因素之 间也存在或强或弱的相关性, 而目前的危险性评价方法难以体现这些要求。本文提出了一种借助 滑坡专家知识并利用模糊推理理论进行滑坡危险性评价的方法。该方法通过建立了①坡度与岩 层倾角之差和坡向与岩层倾向之差、②坡度和岩性、③临空面和岩性、④坡形和岩性等四种环境 因子组合, 以此将不同环境因子之间的相关性融入各组合模型中, 并将四种组合所得的模糊危险 度进行叠加用于滑坡危险度的模糊评价。环境组合模型中的参数利用专家经验给出。将该方法应 用于三峡库区云阳- 巫山段, 得到了滑坡危险性的分级分布图。从滑坡危险性分布图上可清楚发 现, 本方法所计算出的危险性值在滑坡发生的地区明显高于未发生滑坡的地区, 该结果可以用于 城镇建设和重要基础规划设施的参考。

关键词: 滑坡, 滑坡危险性, 模糊推理, 三峡, 专家知识

Abstract:

Current methods cannot meet the needs of evaluating landslide susceptibility because they cannot incorporate the nonlinear relations between landslide and the inducing factors as well as the different forms of relationships. A new approach integrating the expert knowledge and fuzzy logic is proposed in this paper to evaluate the landslide susceptibility. This method first encodes the knowledge from domain experts into four combinations of environmental factors: 1) the difference between slope gradient and the strata gradient as well as the difference between slope orientation and strata orientation, 2) slope gradient and lithology, 3) slope height and lithology, and 4) slope shape and lithology, and thus incorporates the relativity between factors into these combination models. Information on variables involved in these combinations is characterized using a set of GIS techniques. The susceptibility of a given location hence can be evaluated by evaluating each combination and overlaying the values from these four combinations. Some parameters in these models are determined by statistical analysis, and others are provided by landslide experts. The approach is applied to a section of the Three Gorges area from Yunyang to Wushan, and a susceptibility map is produced. The result shows that the computed susceptibility values are much higher over areas with landslides than areas without landslides. We conclude that our approach is capable of capturing the landslide susceptibility.

Key words: fuzzy induction, landslide, slope susceptibility, the knowledge from expert, three gorges

中图分类号: 

  • P931