Integrating soft data into geostatistical modeling of geometallurgical variables: Implications for modeling the copper oxide ratio in copper porphyry deposits

Abstract

Geostatistical modeling involves continuous and categorical variables that are relevant in various earth science applications. While many studies have focused on modeling continuous variables such as ore and mineral grades with geological domains, the choice of modeling technique depends on the abruptness of the variation across these domains. In cases in which continuous variables display soft fluctuations across domains, a joint simulation approach is advocated. This approach leverages the spatial and local cross-correlations of continuous variables using categorical formations. However, patchy and unstructured categorical simulation outcomes may not accurately represent long-range geological settings. In the context of geometallurgical modeling, continuous variables represent geometallurgical parameters, while categorical variables signify geometallurgical domains, such as mineralization zones. The generation of patchy and unstructured geometallurgical domains remains a challenge when soft boundaries are present. Another challenge in geometallurgical parameter modeling relates to their non-additive nature, rendering block-support simulations through traditional block-kriging or block-simulation techniques ineffective. This study introduces an enhanced geostatistical modeling and co-simulation method aimed at co-simulating copper (Cu), copper oxide (CuO), and the copper oxide ratio (COR), which is strongly correlated with mineralization zones in copper porphyry deposits. The proposed method incorporates soft information from mineralization zones in the form of interpretive geological models and a change of variable technique to address the non-additive nature of the COR, which typically hinders block-support simulations, distinguishing it from CuO and Cu, which are additive variables. The methodology aims to provide a reliable framework for the spatial modeling of geometallurgical parameters, optimizing mineral processing plants and mine planning.