基于证据权与逻辑回归耦合的新疆伊犁河谷地区滑坡易发性评价

    EVALUATION OF LANDSLIDE SUSCEPTIBILITY IN ILI VALLEY, XINJIANG BASED ON THE COUPLING OF WOE MODEL AND LOGISTIC REGRESSION

    • 摘要: 伊犁地处中亚内陆腹地,分布着许多高山峻岭,自然地理条件和地质环境复杂,地质灾害(尤其是滑坡)频发,严重影响着该地区正常的生产生活,有必要开展滑坡地质灾害易发性评价。本文在前期工作成果的基础上,结合该地区影响滑坡发生条件,选取高程、坡度、温度、降雨、地形湿度指数、工程地质岩组、距道路距离、距水系距离以及距断层距离等9个因子作为评价因子,采用证据权和逻辑回归(Logistic)耦合模型,通过GIS与统计学相结合的技术方法开展伊犁地区滑坡的易发性评价。研究结果表明:选取的9个因子均对滑坡的发生具有较强的相关性,在证据权模型的基础上与逻辑回归模型耦合具有较好的评价效果。经过评价得出,地质灾害极高易发区与高易发区面积占全区总面积25.74%,主要分布于中低山丘陵地区。依据逻辑回归计算结果,该地区滑坡的发生与河流、人类工程活动以及坡度关系最为密切。经ROC曲线检验,该耦合模型评价精度高达89.7%(AUC=0.897),通过现场检验,本文模型得出的易发性评价结果更为精确合理且符合实际情况,可以为该地区的地质灾害预防和治理工作提供重要参考。

       

      Abstract: Ili is located in the inland hinterland of Central Asia. It has many high mountains and mountains,complex natural geographical conditions and geological environment,frequent geological disasters(especially landslides),which seriously affect the normal production and life of the region. It is necessary to carry out landslide geological disaster susceptibility assessment. Based on the results of the previous work,combined with the conditions affecting landslides in the area,nine factors were selected as evaluation factors,including elevation,slope,temperature,rainfall,topographic moisture index,engineering geological rock group,distance from road,distance from water system and distance from fault,and the vulnerability evaluation of landslides in Ili area was carried out by combining GIS and statistics by using the coupling model of evidence weight and logistic. The results show that the nine factors selected have a strong correlation on the occurrence of landslide,and have a good evaluation effect on the basis of the evidence weight model and the logistic regression model. After evaluation,it was concluded that the area of extremely high risk of geological disasters and high prone areas accounted for 25.74% of the total area of the whole area,mainly distributed in medium and low hilly areas. According to logistic regression calculations,the occurrence of landslides in this area is most closely related to rivers,human engineering activities and slope. After the ROC curve test,the evaluation accuracy of the coupling model is as high as 89.7%(AUC=0.897),and the vulnerability evaluation results obtained by the proposed model are more accurate and reasonable and in line with the actual situation,which can provide an important reference for the prevention and treatment of geological disasters in this region.

       

    /

    返回文章
    返回