The Correlation And Its Variance Principle Among Multiple Terrain Factors Derived From DEM In The Loess Plateau Of Northern Shaanxi

Geography h as been turned into an important developing area since 1980's which the researches about geographical process and mechanism have been the key topics. The analysis methods and the morphologic description of the earth's surface have come though the different periods from simple to complex, from qualitative to quantitative and from simulated express to digital express. The analysis of terrain factors is the foundation of the digital terrain analysis based on DEM. Terrain factors are the basic geographical elements which directly influence the transfer and the transition of the surface material and indicate the form and the development of basic surface environmental pattern. Different terrain factors express the undulation and variance of the true surface in different aspects. The relationships among them can be proved existed and play a key role in revealing the mechanism and development of the terrain and geomorphologic situation in great extent. The reallocation of terrain material and energy is indicated and the form and the development of terrain are reflected by the relationships and their variance principles.The 1:10,000 map scale DEMs of high resolution and high precision are selected as the basic data. The Statistical Analysis and Gray Correlation Analysis methods are also adopted. On the basis of the former results, via the contrast and analysis of multiple physiognomy regions in the Loess Plateau of Northern Shaanxi, the elemental terrain factors which can solely describe the relief are firstly put forward. The relationships between them and other terrain factors and their varianceprinciples are discussed then. At last, the mechanism is analyzed. First of all, the definition of terrain factors' correlation is defined and the classification system is put forward. In the next place, the theories and methods abstracting terrain information from spatial data are systematically analyzed and summarized in the way of the definition, expression and abstracting arithmetic of terrain factors, then the most suitable terrain factors and their realization arithmetic are selected and the suitable model are established. Subsequently, the stabilizing conditions to keep the correlation in its order are analyzed in the aspects of the model, the sampling manner, the size of the swatch and the count of the swatches. Based on this, the two random basic terrain factors from the selected are analyzed for their relationship which is test later so as to ascertain the basic terrain factors. Ultimately, the correlation between these factors and the gully density is analyzed and the correlation model is established. Furthermore, the correlation is contrasted in different physiognomy regions and the mechanism and its variance principle are analyzed combined with physiognomy theory. It is shown that the correlation between terrain factors can reflect the physiognomy and the terrain developing mode well. This result is hoping to be helpful in Loess Plateau physiognomy research.