Spatial Correlation and Indicator Kriging

Indicator kriging relies on models that quantify the spatial correlation among data values under the indicator transform. For a given threshold value, the indicator transform function assigns a value of 1 for data less than the threshold and 0 if it is greater. The semivariogram is then calculated using these transformed values.

In indicator kriging, the range of data values may be partitioned by a series of threshold values. For each one of these threshold values, the indicator transform is produced and a geospatial model of correlatoin must be available. The indicator cutoffs are selected from the drop down box at the top of the Cov tab of the Control Panel. (Note: For a new SADA file, OK will be the only option in the drop down box.)

Click on the Edit box next to Type to create new cutoff threshold values or to edit or delete existing values. SADA will open the Indicator Cutoff window.

To create a new value, select New and enter a new cutoff value. To edit a value, select the value in the Indicator Cutoffs window and press Edit. To delete a value, select the value and press Delete. After making changes, press OK. The new set of indicator cutoffs will appear in the drop down box of the Cov tab in ascending order.

For each cutoff threshold, define the major and minor parameters as in the case of ordinary kriging. The major direction is always highlighted in blue while the minor direction is green.