MINERAL RESOURCE ESTIMATION AND MINE PLANNING IN A GOLD-COPPER DEPOSIT: A MULTIVARIATE GEOSTATISTICAL PRACTICE

Abstract

In mining practices, the life of every mine begins with resource estimation and predicting the possible income from operations. Thus, the first step identifies the potential implementation of mine into actual operation. Due to these reasons, geostatistics and resource estimation are crucial part in mining engineering. Moreover, geostatistics offers an appropriate method for accurately determining several properties of a deposit that are critical for mine design. However, because of the complexity and high cost of sampling variables in mines, some challenges could be faced in determining the grades of present minerals. For instance, the estimation of a variable in a univariate deposit is much simpler than in the case of a multivariate deposit. In this regard, geostatistics has been evolving throughout the years to propose some solutions to this issue. For instance, the computation of cokriging instead of kriging is a good tool in determining the grades with lower error as it takes into consideration the correlation between two or several variables. In this study, the case of a real-life Australian multivariate deposit of gold and copper was taken into discussion. After computing ordinary kriging and receiving a block model, it was found out that there is a presence of a high correlation between Cu and Au. Thus, after realization of an idea of ordinary cokriging with keeping all the steps and details, the mine planning of both methods was done and compared. Therefore, it was concluded that in resource estimation, ordinary cokriging provided us with much better results with lower variance, leading to the superior outcomes in mine planning part including economic aspects. The results of this study showed that by using cokriging instead of kriging, it is possible to increase the NVP (net present value) of a mine project by approximately $120.000.000. However, after proceeding the modelling, some drawbacks of deterministic methods were established. For instance, smoothing effect and negative values in kriging and cokriging, as well as not taking into consideration the time value in mine planning step. Thus, the future research should be focused on the side of stochastic methods.