Study On The Landslide Spatial Distribution Of Lixian County Under The Support Of GIS And R Language

As a common sudden geological disaster,landslides will cause immeasurable loss to people's lives and property,so it's necessary to study landslide hazards.The occurrence of landslide is not only related to external factors,but also more importantly related to topographic factors,geological factors and other internal factors.There are certain spatial relationships among the internal factors of landslides in a region.The traditional statistics do not involve the study of sample data and sample locations.Therefore,the spatial relationship of these data cannot be obtained.And the birth of spatial statistics can properly solve the problems of sample data and its location.This paper adopts a research tools combing GIS and R language,and starts from the spatial statistical analysis method,discusses the spatial dependence of landslide properties in the study area,and the distribution characteristics of landslides.This paper interprets the remote sensing data of Li County in July 2017,and combined with the landslide data provided by the Land and Resources Bureau of the county in January of the same year and the field investigation data at the same time,181 landslide samples of Lixian County were obtained through comparison and screening.From the DEM and geological maps of the study area,geological and topographical attribute data related to landslides were obtained:relative elevation,slope,aspect and stratigraphic data.Under the condition of spatial statistical analysis,the terrain factor which is easy to quantify is selected as the analysis index,and non-numerical factors are tentatively analyzed.The following results are obtained:(1)Through the global spatial autocorrelation analysis,it is found that the distribution of landslide relative elevation,slope,and aspect values in space presents a positive state of aggregation,indicating that the occurrence of landslides is affected by the interaction of various local factors.However,the positive autocorrelation of relative elevation and slope can be seen from the value of Moran's I,and the positive autocorrelation of the aspect is weaker,indicating that the aspect topographical factors in the study area have little influence on the distribution of landslides(2)Through autocorrelation analysis(clustering/outlier type analysis),in terms of relative elevation,it appears that the west-height and east-low aggregation phenomenon,but there is no obvious transition period from east to west,indicating that the landslides in the central part of the study area are relatively random in elevation,and the impact of elevation on landslides is not significant.In terms of slope,most of the landslides in the study area occurred in places with low slopes,and a few landslides occurred in places with steep slopes.Combined with the strata map,landslides with steep slopes may be related to the softer lithology of the local texture.According to the data of the two sets of aspect,the landslide tends to occur on shady aspect or semi-sunny aspect,but the distribution is not very obvious.Combined with the global spatial autocorrelation coefficient,it can be shown that the influence of the aspect on the landslide in the study area is weaker than the other two topographic factors.(3)Through the spatial statistical analysis and results of aspect,it can be found that the method of assigning values to non-numerical data is feasible in spatial statistical analysis.