| Sequential Gaussian simulation (SGS) is a simulation technique used throughout the natural resources industry to construct multiple equiprobable numerical models. The SGS methodology is straightforward, but there are complex interdependent implementation decisions in the transformation from theory to practice.; Some of the implementation issues include: simulation path, search strategy, number of conditioning data, and secondary data affects. The simulation path is random to avoid artifacts. At every unsampled location along this path, the local data is identified by a search and their covariances calculated. This data is then used as input to condition the simulation process. To produce the best possible estimate, the conditioning data should be representative of the surrounding region up to a maximum number of data to avoid excessive CPU time. Secondary data can be considered in this process, but an understanding of how this data will affect the results and problems associated with different techniques will allow the practitioner to make a balanced decision. |
