Mapping Of Landslide Susceptibility Based On GIS In Yongjing County Of Gansu Province

The types of landslide disasters are diverse,the disaster-causing mechanism is complex and the result of landslide is extremely destructive,so that active prevention and mitigation for disasters has become a central and difficult issue in international research.Landslide susceptibility research is the basis of landslide risk assessment work and has received much attention in academia.Relevant scholars have considered many internal and external factors for susceptibility research,but there is short of quantitative analysis of which factors will make the best results.As the research progresses,more factors will be proposed to trigger the landslide.It is imperative to find the best combination of factors according to local geological conditions.This paper takes Yongjing County as an example to introduce the evaluation factor selection process and its application effect,including:(1)To compare the advantages and disadvantages of the slope unit and the terrain unit is presented based on mapping principle,mapping step and mapping expression.Besides,using the landslide success rate curve for illustrating the dominant position of the terrain unit.As a result,It is considered that the terrain unit generally performs better than slope unit.(2)A method for selecting the best combination of factors using the stepwise discriminant analysis and factor analysis is proposed.The stepwise discriminant analysis method is used to construct the core group consisting of four factors: type of slope,surface roughness,distance to river and distance to road.The factor analysis method is used to construct five major factors,and its characterized by the degree of surface erosion,the degree of surface fragmentation,the degree of surface variability,the degree of topography,and the distribution of lithology.Combined with the core 4,the factor analysis method proposes five factors: type of slope,surface roughness,lithology,distance to river and distance to road.Reasonably explain the relationship between the combination of factors selected by the two methods and geological environmental conditions.(3)Four factors combination and their comparison groups 4a~4d experiments and five factors combination and their comparison groups 5a~5d were designed respectively.Then constructing artificial neural network model to calculate the weight of each group of factors,and complete the two groups of experiments respectively using terrain unit.The landslide susceptibility mapping of the two sets of experiments were completed,and the results were verified by the landslide success rate curve.The results showed that the screening results of the factors group were better than the comparison groups,and the mapping effect of the five factors group was slightly better than the four factors group,which verified the rationality of the statistical methods used.So,these methods can be applied to the selection of evaluation factors in landslide susceptibility mapping.