Semivariogram - Modeling or Adjustment
This page describes the experimental semivariogram adjusting module in two
different ways: Automatic or Visual.
-
The automatic mode uses the Olea et
al. (1996) algorithm, which is based on least square method.
This algorithm gives also the quantitative measure, denominated by Akaike
information ( Akaike, 1974), which
reports which model allows a better adjust.
-
The visual mode
is recommended to specialists which have knowledge and are used to the studied
phenomenon. In this mode, all the parameters are defined by inspection.
-
Besides the adjusting procedures, Automatic or Visual, this
module defines a semivariogram theoretical model to be used by the
Validation and Kriging modules.
The experimental semivariogram adjustment or modeling starts after the
Semivariogram Generation. See the following procedures.
Executing the experimental
semivariogram adjustment:
-
select on the Analysis option, Geostatistics -> Semivariogram
Modeling...;
-
the "Semivariogram Modeling" window is presented;
-
select the Adjusting: Automatic or Visual
type;
-
if the Adjustment
was Automatic, define the Number of Structures (1, 2 or 3) required for the
experimental semivariogram adjustment;
-
if the Adjust was Automatic, define the theoretical models using the selection
buttons Model 1, Model 2 and Model 3 in the window. Select for each model one
of the options Spherical, Exponential, Power, and Gaussian. See about the
theoretical models;
-
press the Apply button and notice two things: that the
Define... button is enabled to establish the model parameters; and the
item(s) in the Adjust Verification list is active.
Adjustment Verification and
Model Parameters definition:
-
press on an item in the Adjustment Verification lists, and notice that the
"Adjusting Model" graphical window will be presented and the "Data Report"
window.
-
The "Adjusting Model" window presents graphically the
theoretical adjusting model (in black) over the experimental semivariogram
(white points). Visually it is possible to say if the adjusting is good or not,
as presented in the figure below.

-
The "Data Report" window presents a set of information, such as: the
theoretical model type selected, the values of Lump Effect, Contribution
and Range which are model compounding parameters. It is also expressed
the Akaike value, which is an adjustment indicator taken; the smaller the value
the better the adjustment. Then, the Lump Effect, Contribution and Range
parameters are always taken related to the smallest Akaike value, as presented
in the Figure below.

-
The next step is to define the semivariogram model based on the
information inside the "Data Report" window. In order to do that, press
the Define... button.
NOTES:
See also:
SPRING - Spatial Analysis
Theoretical Models for the experimental
variogram adjustment
How to Execute? - Semivariogram Generation
How to Execute? - Structural Parameters Definition
How to Execute? - Model Validation
|