This work deals with the geostatistical simulation of copper grade in Sungun deposit (northwestern Iran). The deposit is composed of two main rock types: Porphyry ore, which is the dominant rock type, and waste dykes that sporadically cut the porphyry. Because the dykes have a small thickness and because the exploration drill hole data is scarce, it is impractical to separate the two rock types for the geostatistical modeling and conditional simulation of copper grade. Consequently, the realizations lack realism as they do not reproduce the presence of low-grade dykes cutting the porphyry ore. To solve this problem, it is proposed to introduce secondary rock type data and to use them as conditioning information for simulating the copper grade. Specifically, based on the geological knowledge of the deposit, a network of secondary data locations is defined, for which the rock type (porphyry or dyke) is assumed to be known. Using a multivariate Gaussian model, the distribution of the copper grade at these locations is determined, conditionally to the copper grade drill hole data and to the rock type information for both the drill holes and secondary data. Finally, the copper grade can be simulated at the secondary data locations first, then at the remaining locations in the deposit. The resulting copper grade realizations show a substantial improvement, insofar as they reproduce waste dykes cutting the porphyry ore in agreement with the rock type model. Furthermore, the use of rock type data for conditioning the grade realizations reduces the uncertainty in the unknown copper grade.
展开▼
