3.1 Geospatial Interpolation
The geo-statistical techniques are different from the classical
statistical techniques. Geo-statistical methods are more reliable here
because these methods assume that there is an auto correlation among the
data sets. When precipitation occurs, it is obvious that the area closer
to them will also be getting ample amount of rainfall, thus it indicates
that there is always some relation for the precipitation data in the
region surrounded by an area experiencing precipitation.
ArcGIS uses geo statistical extension to incorporate geospatial
interpolation. There is several interpolation techniques that this
extension provides all of which are based on the same principle of
Tobler law. It states that all things are inter-related while things
that are nearer are more similar among themselves as compared to things
which are distant from each other. Therefore, geo-statistical
interpolation is based on notion of spatial autocorrelation. When the
interpolation is done using this technique first it determines the
spatial autocorrelation among the spatial data. The correlation
determines
- Resemblance of data (objects) in the search neighborhood.
- Degree of spatial correlation of data with itself in space.
- How the variables are dependent and their degree.
All the interpolation methods give different result. No two methods will
give same result.
- Deterministic Interpolation: Creates a surface on the basis of
observed data sets in the search radius. It uses mathematical
equations to determine the smoothness of the observed interpolation
surface grid.
- Geo-statistical Interpolation:
Creates a surface on the basis of
observed data sets in the search radius. It uses mathematical
equations to determine the smoothness of the observed interpolation
surface grid.