The following example will use the GeoEAS variogram analysis programs Prevar and Vario.  These run from the DOS window. 

An alternative to GeoEAS Vario is a true Window's program for variogram analysis which uses datasets in GeoEAS file format. It is called Variowin.   For more information on variowin, 

see http://www-sst.unil.ch/research/variowin/index.html

and http://curie.ei.jrc.it/software/variowin.htm

Variowin has many capabilities that GeoEAS Vario does not have, but GeoEAS Vario is adequate for the purposes of this analysis.  An explanation of GeoEAS Vario follows.

The purpose of variogram analysis is to develop a mathematical model of the spatial continuity of a variable.  To do understand the spatial continuity, we need to look at pairs of data that are separated by different distances.  The separation distance between points is called the lag distance.  The first step in a variogram analysis in GeoEAS is to create a file that consists of all possible pairs of data points, and for each pair the distance and direction between the points.  This file is called the pair comparison file. It has the DOS extension .pcf.

To create the pcf file, Open a DOS Window and type "Prevar" at the DOS prompt in the GeoEAS directory.

The following screen appears:

1.  Be sure the File Prefix is correct with a "\" a the end.

2.  Enter the Data File name,  press enter and choose a name for the Pair Comparison File (or better accept the default name).  For variogram analysis we have selected a file that contains all of our data on S strain incidence in Arizona.  This includes data collected prior to 1998.

3. The 283 records produces 39903 possible pairs of locations.  Because GeoEAS limits the number of possible pairs to 16384,  we use the limit menu option to select only those pairs that are less than 50000 m (50 km) apart.

4.  Press Execute and the .pcf file is created.

Return to the DOS prompt and type vario:

The following menu screen appears:

Verify that the Prefix (1), File name (2), and Variable (3) are as desired, and press  Options/Execute (4).

The Options Menu Window appears:

Adjust the Lag Spacing (1) or change the Directions option (2) to look for directional effects, and Press Execute (3) to see the Variogram Results window.  It is best to try out several different lag spacings by iterating between the Options window and Results window.  The 1500 m increment shown above was selected by iteration.

The results menu window appears.  One can choose from several types of estimators of the spatial continuity.  The default is the variogram estimator, which we most frequently use.  The next step is to plot and model the variogram.  We select "Model" from the menu and press enter.

We select the Model options (1) and enter the parameters for a variogram model.  Next we Plot (2) the model for easy comparison with the sample variogram results (3).
The final variogram model is selected by iteration with the model parameters (1) selected to give an approximate fit to the sample variogram.  The variogram model can also be based on past experience or other knowledge about the spatial continuity of the variable.

 
In conclusion, we select a nested variogram model with a nugget of 200, a spherical function with range 2500 m and sill 400 and a spherical function with range 25000 m and sill 200.  This choice is consistent with previous experience for Aspergillus flavus S strain incidence in Arizona.  With the model selected, we are ready for ordinary kriging to create a surface map of incidence.

Return to previous menu.