A compositional approach to in-situ evaluation of polymetallic deposits. A case study at Sungun Cu−Mo deposit, NW Iran.

Mineral reserve evaluation is based on compositional data, reporting in units such as percentages, mg/kg or ppm. The nature of such data implies that the values cannot vary independently and geostatistical modeling of their raw form could be misleading. In this study we propose a compositional approach —based on orthonormal isometric logratio (olr) coordinates— to simulate the in-situ value (in US Dollars) at Sungun Cu − Mo deposit in NW Iran. A dataset of 1422 surface samples, analyzed for Ag, As, Au, Cu, Mo, Sb, Pb and Zn is used. This surface covers the current mine pit and is sampled prior to extraction, as a detailed exploration phase. We assumed the sampling surface as an extraction mine bench and separated 34% of the original samples for validation. Comparing with a noncompositional approach, it is prominent that in all extraction categories and for all Cu and Mo cut-off grades, the compositional values are closer to the real ones. Comparison of in-situ value maps revealed that the compositional approach is better able to delineate intense local variability; especially in high grade parts of the deposit. This superiority is also evident in Euclidean and Aitchison distance-distance diagrams of the real and simulated values. It is concluded that, in contrast with known geostatistical methods, the proposed olr-based method delivers more accurate models and, consequently, an improved and more realistic vision of the deposit value is achieved. • A polymetallic mineral resource evaluation deals with compositional data. • Otrhonormal logratio transform is able to summarize all data in arbitrary balances. • Desired balances can be flexibly defined using a Sequential Binary Partition (SBP). • An auxiliary variable recovers the original units after interpolation. • The compositional approach improves the estimation of in-situ value of ore elements. [ABSTRACT FROM AUTHOR]