We use ordinary kriging to provide a surface map of Aspergillus flavus S strain incidence in Arizona.   Ordinary kriging is a regression technique for estimation and is designed to give a best linear unbiased estimate of a variable at an unsampled location.  Usually our purpose in using ordinary kriging is less ambitious.  We frequently use it to get a regional overview of a variable, because it smoothes out the local variability by calculating a moving spatial average.  The average, in the case of kriging, is  weighted according to our understanding of the spatial continuity as reflected in the variogram model. (See our explanation of selecting a variogram model using the GeoEAS program vario.)   We use the range of the variogram model for the radius of  the search neighborhood for the kriging. The search neighborhood specifies which data points are included in the variogram-weighted moving spatial average.  This means the data points used in the the estimation are correlated with the unsampled location to be estimated, based on our model of the spatial continuity (the variogram model).  On future pages, we plan compare the the surface map generated by ordinary kriging with maps generated by indicator kriging and by inverse distance weighting.

To start the kriging program, open the DOS window, move to the directory with GeoEAS, and type "krige".

Verify that the File Prefix is correct and chose "Option/Execute" from the opening menu (below):

The Krige Options menu is shown below: 

Be sure the data file name is correct (1) and choose the appropriate grid parameters (2).  The grid parameters (2) are chosen with a number of considerations in mind.  In GeoEAS, there is a maximum of 100 rows and 100 columns.  Therefore, you want to choose the grid spacing so that you can span the region with 100 columns.  Also, usually we want to link the output of kriging to an ArcView polygon shapefile (see later), so we choose the orgin, spacing, and number for the grid to match the ArcView polygon shapefile.  Consider the range of the variogram model  in choosing the search parameters (3), highlight the Variable/Models choice (4), and press enter to go to the Variable/Models page. 

At the variogram models page, choose "new variable" and select percent S for the variable and choose "edit" to enter the variogram model.  Enter the model selected from the variogram program.  Then choose Quit to return to the Krige Options window. 

At the Krige options window, highlight "Execute" (*) and press enter.  The results of the kriging calculations are stored in the output file (here s9899web.grd).  Choose quit to exit.  The s9899web.grd file can be viewed with the Postplot program (type "postplot" at the dos prompt in the geoeas subdirectory) or it can be processed so that it can be viewed in ArcView.

Return to main menu for details.