Model Type

To specify a semi variogram model type and its parameters, right click in the Model parameters input box to enter azimuth and plunge values for the main, 2nd and 3rd directional variograms.

To open the Semi Variogram model dialog, right-click in the main, 2nd and 3rd direction input boxes.

The model type you select will determine which of the following inputs are enabled. In some cases the names and the number of inputs will change.

All response boxes have spin controls associated with them and these provide a quick way of adjusting the values. The controls are set up to increment the increment value (change gear) over 10 consecutive increments by a factor of 10. You can therefore use these to quickly make both fine and coarse adjustments.

Nugget

Enter the Nugget value. The Nugget value (effect) is the variance at distance zero. This is always less than the sill. The nugget effect arises because the regionalised variable is erratic over a very short distance that the semi variogram goes from zero to the nugget effect in a distance less than the sampling interval.

Slope / Cycle distance

Enter the distance between peaks and troughs in Hole Effect or Paddington Mix variogram models.

Power / Decay

Enter a damping factor for Hole Effect or Paddington Mix variogram models. This is an inverse function - the smaller the factor the faster the damping.

Final Sill

For Final sill, enter the semi variogram (Y axis) value of the final or highest sill in multiple component variogram models.

Variogram directions

Define the variogram direction by entering azimuth and plunge values. These values are used internally to define the model's Anisotropy.

Number of components

Note that some semi variogram model types only allow one component to be defined. Where multiple components are allowed, the number of components can be set (or updated) and the associated range, partial sill, and model type can be entered for each component.

Range

Enter the range of the components of the semi variogram model. The range is the distance where there ceases to be an identifiable relationship between sample pairs.

Partial Sill

Enter the Y coordinate (semi variogram) value of the sill for each component of the model. This is constant for a dataset.

Model type

The type of model should have a general shape which resembles the experimental semi variogram. See: Types of semi variogram model

Forms

Once you have refined the model's characteristics, save them as a form set. This form set will then be accessible in Semi Variograms and the 3D Block Kriging functions.

Test

(Optionally) Click the Test button to test the “goodness of fit”. A Noel Cressie statistic, a form of least squares prediction, is calculated for each direction and appears in the numeric display.

It is calculated as the sum of the squared difference between each experimental point and the model value at the same distance. The lower the value, the better the fit. Note that the actual value is heavily dependent on the data.

See Also