Simulations

The Simulations procedure enables you to perform several simulations of numerical variables in the multi-gaussian framework and to compute statistical results from these simulations, such as the average of simulations, confidence intervals, probability of exceeding a threshold, etc. Several simulation algorithms are offered here:

• the Turning Bands Simulations method (TBS),
• the Stochastic Partial Differential Equations method (SPDE),
• the Sequential Gaussian Simulations method (SGS).

All simulations replicate the variogram model given by the input Geostatistical Set. They can be constrained to honor all the active information; in this case, we speak about conditional simulations.

Note: A check is done on multivariate geostatistical sets to protect from numerical instabilities while kriging. This kind of problem can happen when the cross-variogram model is very close to perfect correlation between the variables. In this case, a warning message is printed in the status bar at the bottom of the interface. The same check and an automatic correction is offered in the fitting of the Exploratory Data Analysis.

Conditional simulations are obtained using a two-step procedure:

• Non conditional simulation of the specified variogram model. The number of turning bands is the only parameter involved at this level; this number should be large enough to ensure the quality of the resulting simulations.
• Conditioning to the data with (co-)kriging.