Resource classification – a case study from the Joffre-hosted iron ore of BHP Billiton's Mount Whaleback operations

Abstract

It is recognised by the current JORC (1999) code that resource classification involves the interaction of numerous qualitative and quantitative criteria such as data quality, geological continuity and grade continuity. No prescribed criteria and rules will work for all situations or even between different ore types within the same deposit. Also, these criteria and rules are sometimes applied without a clear understanding of their appropriateness, accuracy or correct implementation. For these reasons, case studies are useful to evaluate and compare criteria commonly used to assist in resource classification. In this paper, blasthole data from a selected area of Joffre Member hosted ore of the Brockman Iron Formation at the Mount Whaleback orebody are used as the basis of a case study for the above-mentioned purposes.Iron grade, in percent, is interpolated from a blasthole dataset into a block model using ordinary kriging. Samples are then removed from this blasthole dataset to produce a random sample grid, a semi-random sample grid and a regular sample grid. Block iron grades are obtained using nearest neighbour,inverse distance squared and ordinary kriging estimation methods and sequential gaussian simulation, using these three blasthole data subsets as inputs. This provides three groups of estimates and one group of simulations with the ordinary kriging estimate based on the complete blasthole dataset being considered to represent the true estimate of iron grades. Also generated during this process are measurements of the expected error for each block grade. These measures of error range from the simple, such as drill spacing, through to more advanced methodologies such as kriging efficiency and simulation. The grade and error determined from estimation and simulation are then compared to the true grade and true error using graphs and statistics. In order of increasing accuracy, the block grade determination methods were nearest neighbour, inverse distance weighting, ordinary kriging and finally the average of multiple simulations. Simulation in some instances doubled the accuracy of individual block grades when compared to nearest neighbour and inverse distance weighting estimates. It was found that many methodologies for determining the error of individual block grades performed equally well with only methods such as average drillhole spacing and classification by search ellipse pass number performing poorly. An approach on how to convert the above-mentioned error determinations of individual blocks into a meaningful JORC classification is also discussed.Although advanced non-linear resource estimates are applicable, most iron ore mines are still using relatively straightforward methods. The use of blasthole data and some simple linear estimation methods and simple linear-based error estimates makes this study repeatable for most iron ore sites and their resource geologists. This style of investigation is recommended as a useful approach for the competent person to apply to their deposit and thus better select, implement and understand the criteria used for resource classification and provide more consistency and confidence in the resource classification process.