Study On Spatial Interpolation Model Of Soil Thickness In Typical Karst Mountainous Region

China is one of the countries which suffer the most serious harm of karst rocky desertification in the world. Yunnan Province is one of the eight provinces which suffer these seriously. The carst area in Yunnan Province is 11.09km2, accounted 28.93% of the total land area. The karst rocky desertification causes many ecological environment problems, including serious soil erosion, vegetation coverage reduced, Sustained reduction of cultivated land area. It makes a serious threat to the ecological environment security in karst mountain areas of Southwest of our country. The thickness of the soil directly restricts the growth condition of vegetation in karst mountain. It reflects the soil erosion of the karst mountain intuitively. Soil quality can be improved and adjusted by artificial way, but the soil thickness cannot be recovered in a short time. The research of spatial distribution characteristics of karst mountain soil thickness is an important basis of the comprehensive management project of rocky desertification.This paper takes the Babao Town in Wenshan, a typical karst mountainous region of Yunnan Province as the study area. Based on the analysis of the study area land use status, we determine the research object is the soil thickness of three land types (farmland, forest land, bare soil) in the study area. Based on the distribution and class characteristics of land types,3 different type land of each class are selected as research object. In the specific research, soil thickness data can be obtained through field investigation. Histogram and normal QQ figure can be use to make a normal distribution test of all kinds of the soil thickness data. On the basis of normal distribution test, we make a heterogeneity analysis of soil thickness of each sample, including descriptive statistics analysis, spatial variability and trend analysis.The paper research content and the main achievements include:1) Use the descriptive statistics analysis to make a statistics and distributed analysis of soil thickness data of each sample.2) Use the spatial variability analysis to make spatial structure analysis of soil thickness data, in the condition of data is classified as normal distribution under the K-S nonparametric test.3) Use the trend analysis to mainly revealed the general regularity of spatial distribution of soil thickness in the study area, reflect the main characteristics of the change in the space. removed the global trend in the modeling of spatial interpolation.Based on the above analysis, the focus and innovation of the research work is mainly reflected in this:Using Deterministic interpolation including Inverse distance weighted interpolation, local polynomial interpolation, global polynomial interpolation, the radial basis function interpolation, and the Statistical interpolation method including Ordinary Kriging interpolation, Simple Kriging interpolation, Universal Kriging interpolation and Co-Kriging interpolation interpolation to do spatial interpolation modeling and model simulation analysis for sample area’s soil thickness data. And make an accuracy test and comparison analysis of the result of the spatial interpolation of soil thickness by Crosscheck and the Validation Data set test. Through these studies to find a suitable, optimal spatial interpolation method for the three typical class in study area. Universal Kriging interpolation is the optimal method for farmland soil thickness simulation, Universal kriging interpolation is the optimal method of forest land soil thickness simulation. Radial basis function (planar spline function) is the interpolation method for simulate the optimal bare soil thickness.