Authors:
Matthew J. Samson, Nevada Gold Mines, Matthew.Samson@nevadagoldmines.com
Diogo Silva, Resource Modeling Solutions, dsilva@geologicai.com
Lisa Davenport, Nevada Gold Mines, Lisa.Davenport@nevadagoldmines.com
George Benthen, Nevada Gold Mines, George.Benthen@nevadagoldmines.com
Keywords: Cokriging, Copper Speciation, Resource Estimation, Supergene Enrichment, Ore Routing, Metallurgical Recovery, Multivariate Estimation
The economic viability of porphyry copper deposits is highly sensitive to weathered zones of primary hypogene sulfide assemblages where bulk leach processing methods are favorable. The success of feasibility studies hinges on the accurate discrimination of metallurgical copper ore-types that are controlled by mineralogy. An analytical proxy for quantifying the mineralogy of altered and weathered zones is acid-soluble copper and cyanide-soluble copper, enabling accurate geometallurgical domaining and estimation to inform economic and processing sensitivities. Though a common gap in legacy assets is a lack of geometallurgical information beyond grade assays as these methods weren’t developed until the late 1990s to early 2000s. It may also reflect a singular focus of grade during exploration drilling often neglecting geometallurgical and other datasets.
This paper presents an interdisciplinary geostatistical framework utilizing Standardized Ordinary Cokriging with an intrinsic model of coregionalization to bridge the gap between sparse geometallurgical data with large copper assay datasets to inform robust resource estimation.
The methodology leverages the mineralogical zonation of the supergene weathering profile to inform robust estimation domains that fulfill first-order stationarity. Characterization of the cross-correlation between copper with acid- and cyanide-soluble copper, a multivariate framework is defined utilizing data-dense copper data to inform spatial data gaps of the latter. By implementing a intrinsic Model of coregionalization, the heterotopic nature of drill hole data is effectively considered, ensuring the spatial continuity of primary economic variables informs the estimation of under sampled metallurgical proxies.
The collaborative impact of this approach is demonstrated through a significant reduction in estimation error. Validation results via leave-one-out cross-validation show a reduction in root mean squared error for copper soluble assays. This increased precision directly impacts the bottom line by preventing Type 1 and Type 2 Error.
The transition from univariate to multivariate estimation ensures the estimation is no longer biased by grade data alone, rather it is informed by a balance of advanced geostatistical tools constrained to geologically rational domains, representing a technically rigorous outcome for operational optimization across a multidisciplinary value-chain. By honoring the complex geological and metallurgical interdependencies of the deposit, this framework provides the foundational data required for informed decision-making in mineral processing and long-term economic planning.
A consistent gap in mature assets is a singular focus on collecting grade assays of economic metals of interest, often neglecting critical geometallurgical data to inform economic and processing sensitivities.