COMPOSITIONAL KRIGING APPLIED TO THE RESERVE ESTIMATION OF A GRANITE DEPOSIT

Making an accurate estimate of quality distribution in a granite deposit is essential, both from a financial point of view, to determine the profitability of the site, and from an environmental perspective, to focus operations on the most profitable areas thereby reducing the extent of land affected by such work. Granite is extracted in blocks whose profitability and value depend on the final size of the slabs, which is an important factor in defining quality. This article uses a variant of disjunctive kriging in order to determine the quality of granite in one of the largest reserves in the world-the Porriño deposit located in northwest Spain. This method, unlike classical disjunctive kriging, considers random variables that are not necessarily binary. The advantage of using this technique compared to the classical statistical cokriging technique is that all the qualities are considered as variables with the same importance and that the sum of quality percentages in a block is one hundred percent. The validity of the method was tested in a cross-validation process. The results compared favourably with those obtained using ordinary cokriging and fuzzy kriging.