In the Parameters panel for Gauss functions, define Standard Deviation and Scale.

Standard Deviation controls the range for Side Length. Note, that the standard deviation here directly refers to the corresponding Gauss function, while the standard deviation in the Options tab defines the value range in the generated Gaussian Random Field.

Scale controls the range for the Covariance, being a factor multiplied to the Gauss function.

In the following example, observe the impact of scale and standard deviation on the graphs. The generator input visualization was turned off by right-clicking in the plot and deselecting Generator Input

To learn more about how to edit plots in GeoDict refer to the Result Viewer handbook.

For Linear functions define the intercept of the linear function with the Y-axis (Y0) and the X-axis (X0) and multiply this function with the factor given for Scale.

While loading the autocorrelation values in the Options tab is usually the easiest possibility, autocorrelations can also be loaded from GeoDict result files generated with Compute Metrics, available from the Project Status by right-clicking on the volume field to analyze.

In the Compute Metrics dialog select Minimum, Maximum, Mean Value, Standard Deviation and Variance and Autocorrelation and define the parameters as needed.

For example, if the autocorrelation in the current field is different for the three directions, compute the metrics for each direction. Therefore, for Direction select X-, Y- or Z-Direction instead of Isotropic.

Clicking Run computes the standard deviation, mean value and autocorrelation function for the chosen volume field.

In the Gaussian Random Field Generator enter the Standard Deviation and Mean manually in the Options tab as given in the Compute Metrics result file.

Then, move to each of the X, Y and Z tabs and load the computed autocorrelations.

Therefore, click Load Correlation and browse for the Compute Metrics result.

If an autocorrelation was loaded either from the current volume field or from a metrics result, in the plot a graph for the Measured Covariance is added and the buttons Guess Parameters and Improve Parameters become available. Additionally, below the plot the Error between the measured covariance and the current generator input is displayed. It is computed as the sum of squared deviations for all data points in the curve.

The number and type of functions needed to fit the measured covariance must be defined. After adding the needed number of Gauss or Linear functions, clicking Guess Parameters tries different parameters to fit the measured covariance. Automatically, the first best guess of the algorithm is entered for the parameters.

It can be necessary to click Improve Parameters several times, if the Generator Input does not fit the Measured Covariance with the desired accuracy.