Ninth International Geostatistics Congress, Oslo, Norway
June 11 – 15, 2012

No Paper Available


Mining 1

Abstract No.:



Non-MultiGaussian Multivariate Simulations with Guaranteed Reproduction of Inter-Variable Correlations


AJ Cornah, QG (AU)
J E Vann, Quantitative Group (AU)


Stochastic modelling of interdependent continuous spatial attributes is routinely carried out in the minerals industry through multiGaussian conditional simulation algorithms. However, transformed conditioning data frequently violate multiGaussian assumptions in practice, resulting in poor inter-variable correlation reproduction in resulting simulations. In addition the maximum entropy property that is imposed on the resulting simulations is not universally appropriate. A new point Direct Sequential Cosimulation algorithm is proposed where pairwise simulated values are drawn directly from the discrete multivariate conditional distribution under an assumption of intrinsic correlation; local Ordinary Kriging weights are used to inform the draw probability. It allows the generation of multivariate simulations with two key advantages over multiGaussian methods: (1) inter-variable correlations are assured because the pairwise inter-variable dependencies within the untransformed conditioning data are imbedded directly into each realisation; and (2) the resultant stochastic models are not constrained by the maximum entropy properties of multiGaussian geostatistical simulation tools.




    Copyright 2012 International Geostatistics Congress

Produced by X-CD Technologies Inc.