Block Estimation At Various Stages Of Deposit Development

Elementary estimation methods in which blocks are assigned values according to a geometrical area of influence concept are investigated for a tabular deposit. This quantitative assessment of the estimation technique is made using Matheron's geostatistics which allows the variance of the error between actual block values and estimated values to be found once the semi-variogram is known. For the deposit in question it is shown that the elementary technique makes poor use of the available information. There is a 20% relative error associated with it, whereas by using the optimal technique of kriging this can be reduced to 9%. An infill drilling program, which might have been suggested on area of influence results, would reduce the estimation variance only marginally over that of the current campaign; the relative error being lowered to 6%. Thus the geostatistical analysis shows this would be an unnecessary expense. The accuracy of an earlier drilling stage is also studied and it is shown that with the information available at the earlier stage block values could have been known with a relative error of 13%. Drilling could have stopped then if this were reckoned a tolerable error level. These results are all obtained for blocks whose size is dictated by the mining engineer and is set at the current grid size. In the area of influence approach smaller blocks are considered as the grid becomes finer and a study of the accuracy of area of influence estimates and kriged estimates for differing block sizes is presented. Between drilling stages more information from which the semi- variogram is calculated becomes available. The sensitivity of the estimation to changes in the semi-variogram parameters between earlier and current drilling programs is investigated.