Definition and Characteristics of Geostatistics
Regionalized Variable
Based on the regionalized variable theory, geostatistics explores the natural phenomenon with a certain structure and randomness scattering in space. When a variable presents its spatial distribution, it is called “regionalization”; moreover, this variable often responds to the features of a certain spatial phenomenon. This phenomenon described by a regionalized variable is called “Regionalized Phenomenon”.
Features of Regionalized Variable Theory
We cannot use traditional statistical methods to analyze spatial variations, because the spatial variations are usually irregular. In this case, since Z(x) has two important features: randomness and structure (the spatial correlation). Regionalized variables with the two features have absolute advantages in exploring the spatial structure and spatial succession of a natural phenomenon. “Randomness” means a regionalized variable is a random function with local, random and anomalous characters. “Structure” means a regionalized variable has normal or average structures. This means that there is an autocorrelation between variables at point x and x=h (h= spatial distance) and the values of Z(x) and Z(x+h). This autocorrelation is decided by the distance (h) between the two points and the features of the variable.
The structure and randomness of a regionalized variable always have mathematical or statistical significance. Moreover, when we process a concrete variable, it also has spatial limitations, continuity and anisotropy.
➢ Spatial Limitations
When a regionalized variable is applied in a specific spatial range, this range is called “Geometric Region of Regionalized Variable”. Its attribute performance in this range is the most obvious; that out of the range is obscure or zero. |
➢Continuity
Different regionalized variables have different kinds of continuity. The continuity is described through the variance between the adjacent points. Under some special conditions, the continuity does not exist. For example, the nitrogen availability in forest soil may have huge differences between two very close points. This differences show us the discontinuity. This phenomenon is also called “Nugget Effect”. |
➢ Anisotropy
If regionalized variables are all equal in all directions, this is called “Isotropy”. On the other hand, if regionalized variables are different in all directions, this is called “Anisotropy”. The purpose of distinguishing between isotropy and anisotropy from data is to consider the level of autocorrelation of regionalized variables within a certain range. Beyond this range, the autocorrelation will decrease progressively or disappear. |
©2017 Supergeo Technologies Inc. All rights reserved.