Resource and reserve evaluation in the presence of imprecise data

"In mineral resource and ore reserve estimation, several sources of information are generally available, such as diamond drillholes, reversed-circulation drillholes, and blastholes, and originate from one or more sampling campaigns and are analyzed by one or more laboratories. In such a situation, the grade measurements may not have the same precision and one should distinguish between hard and imprecise data.This work aims at studying the impact of data imprecision in resource/reserve evaluation through two case studies on simulated deposits: a porphyry copper deposit conditioned to real exploration and production data, and a gold deposit simulated without any conditioning data. Several geostatistical techniques are used to estimate the grades of selective mining units from exploration samples, including kriging, cokriging, lognormal and indicator kriging. Two types of imprecision are examined— measurements with multiplicative errors and measurements defined as intervals. The results are compared to the simulated grades of the selective mining units from a statistical and economic viewpoint.The study concludes that imprecise measurements should never be discarded because they help to improve the estimates, and that simple methods such as kriging are as efficient as more sophisticated ones for resource/reserve assessment. Finally, a few indications on how to diagnose and model the data imprecision are given."