New developments in borehole valuations for new gold mines and undeveloped sections of existing mines

Using a very large number of chip-sample values from a mined-out area of the Hartebeestfontein mine and accepting these as equivalent to borehole values (also areas of 1 km by 1 km as equivalent to 'mines'), the authors apply a series of borehole-valuation procedures to these 'mines', Large undeveloped sections of existing mines are also regarded as the equivalent of such 'mines'. These procedures start with the orthodox arithmetic mean and then introduce additional information in the form of the Iog normal frequency distribution model for Sichel's 't' estimator. Knowledge of the average log-variance within mines leads to the tH estimator. The mean grade of the 'family' of mines to which a new mine belongs is then introduced via regression or elementary kriging. Next, further information on the macro spatial structure of the grade is used, together with the grades of adjacent existing mines in ordinary and simple macro-kriging procedures; finally, via the distribution model, the grade is seen as a function of the logs of the grades and their variances. These are introduced as spatial variables, and both are estimated by macro-Iog normal ordinary and simple kriging procedures with grades and variances from adjacent mines'. Comparisons are also made between macro-kriging on lognormal transformed values versus untransformed values. Conclusions are drawn on the relative importance of the various items of additional information used, on the procedures followed, and on the significance of the advantages gained