Classification of reserve in Sungun mine based on Archimedean copulas estimates

Incorrect estimation of grade values is a major problem in the Sungun copper mine of Iran that the kriging estimates and blast hole data show inconsistencies due to biased estimates of kriging in this case. Therefore, re-evaluation of mineable reserve of the Sungun mine was considered through spatial copula because of its advantages over kriging. Spatial dependence of variable under study was analyzed through the calculation of empirical copulas and variograms. A Jackknife test was implemented by dividing the observed values into data and validation sets. Validation sets were estimated through Archimedean copulas and ordinary kriging methods, and the estimated values were compared to the real ones. Copula showed better accuracy and precision than ordinary kriging due to better reproduction of mean values and having lower mean square errors. Moreover, cumulative distribution functions of the copula estimates were closer to those of observations. Ordinary kriging showed biased estimates, especially for highly skewed data with extreme values. The copula method showed higher local precision due to having smaller estimation errors. Based on the estimated values and their confidence intervals, the blocks were classified as A, B, C1, C2, and C3 categories. The copula method showed its superiority over ordinary kriging by giving higher number of blocks categorized as A and B. Moreover, the average Cu value of the mineable blocks, which is 0.62% in the blast hole data, was respectively calculated through copula and ordinary kriging to be 0.69% and 0.88%.