Classification of ore reserves based on geostatistical and economic parameters

"The proposed classification of ore reserves based on geostatistical and economic parameters is designed to be used as a communication tool to identify for investors, bankers, and financial analysts the degree of risk associated with a specific class of ore reserves. The class designation identifies the degree of accuracy of the estimates for tonnage and grade of annual production volumes of ore. It thus enables decision makers to quantify the risks and uncertainties inherent to the ore reserve in achieving over-all economic goals and annual financial objectives.IntroductionDuring the evaluation, in 1980, of the Creede silver deposit in Colorado, U.S.A., as a potential bulk mining project, the authors were faced with the problem of conveying to senior management the risks and uncertainties that are implied by using common or conventional classification systems to categorize ""ore reserves"". The estimated numbers for tonnages and grade of the projected ore reserves represented the only entry for future revenue from the project. It was, therefore, necessary to convey to management the level of risk associated with those numbers when implementing further stages of development, and to convey what impact significant differences would have on the profitability of the project. The desire to quantify these risks led to a revised definition of ""ore"" and to a re-examination of the classification system of ""ore reserves"". Traditional methods for estimating tonnage and grade of reserves do not include a quantification of the errors of these estimates. The advent of geostatistical methods made this quantification possible. The proper application of geostatistical methods, with adequate geological input and a reliable set of sample data, is essential to its validity.Others have proposed a classification of ore reserves based on geostatistical parameters. Diehl and David (1982) recognized ""that concepts of proven, probable and possible have to be linked to the production rate or time horizon"". Royle (1977) had also addressed this problem. Forsyth (1987) pointed out that bankers are concerned with ""the probability that the variability will fall within a certain degree of accuracy"" especially for grade estimates, and would like to see grades expressed together with the error of the estimate at the 950Jo confidence level."