Research On Terrain Differentiation Of The Loess Plateau In Northern Shaanxi Province Based On The Geostatistics

Loess Plateau is located in the middle North region of mainland China and is the typical region of the distribution of loess in China. The regulars of regional differentiation are obvious. It exhibits very important theoretical significance and application prospects to study on the terrain’s spatial differentiation regulars of the Loess Plateau. Fragment and revolution of the Earth surface can be revealed, the spatial pattern of soil erosion in the Loess Plateau can be further verified, ecological rehabilitation of Loess Plateau and regional comprehensive government can be guided.Compared with the traditional statistical methods, geostatistics has the advantage in spatial statistical analysis. Geostatistics can reflect spatial distribution, variation and correlation of regionalized variables effectively. It also can explain the effects of natural and anthropogenic processe on spatial variation of variables. And this method provides a new way to study characteristics of terrain spatial variation in Loess Plateau of Northern Shaanxi.This study selected 47 experimental plots in loess tableland, loess Liang, loess hilly and other typical landform type areas in Loess Plateau of Northern Shaanxi. Based on the theory of geostatistics, range, sill, nugget of variation function are used to study spatial variation of three topography parameters:elevation, slope, aspect, which are derived from SRTM DEM. The main contents are as follows:(1) through the exploratory analysis of the data, it is found that most of the elevation data of the sample area are negative skewness and platykurtic. Elevation and slope of sample areas accord with normal distribution. In the aspect of area. All kinds of terrain factors appear in the sample areas have global trend, and the global trend of elevation and slope can be fit by two degree polynomial. The global trend of slope be fit by a degree polynomial. Elevation has a strong spatial autocorrelation, but the spatial autocorrelation of slope and aspect is not obvious. (2) According to the statistical indicators as root mean square prediction error and average standard error, the research established variation function’s optimal theoretical model of terrain factors in different plots, and got the characteristic function parameters, such as range, sill, nugget, etc.(3) The variation function characteristic parameter analysis shows that the range of values in the higher terrain in southern and northern, central region is relatively low, indicating North Shaanxi Loess Plateau in the southern and Northern all terrain factors from a large range, the central region correlation range is small. From the south to the north elevation and slope base value decreases, the total spatial heterogeneity the degree of decrease, slope to the base value change is relatively flat, the degree of spatial heterogeneity is not high. The nugget effect said regional variable spatial correlation degree, the central region has moderate spatial correlation, the southern and Northern has a strong spatial correlation. The terrain factors of area calculation results show that the fractal dimension, the terrain small North and south, the surface complexity is weak; the central region terrain surface with higher complexity.(4) Analysis of the terrain factor anisotropy, anisotropic analysis of different terrain factors. Scale analysis of the variation of elevation, landform elevation has two scales.