Hidden Markov model for spatial analysis of three-dimensional mineralization distribution: Insights into magma flow and mineral exploration targets in the Jinchuan Ni-Cu-(PGE) sulfide deposit, China.

The spatial distribution of mineralization includes significant clues about the magma/fluid flow process for ore formation. However, given the complex nature of the ore-forming process and its weak signature in the mineralization data, revealing the magma/fluid flow process in their mineralization footprint presents a considerable challenge. In this paper, we present a hidden Markov model (HMM) for the analysis of three-dimensional (3D) spatial distribution of mineralization. The HMM assumes the magma flow is a Markov process that is "hidden" in the "observable" mineralization features. Under this setting, the HMM quantifies the states of magma flow and links them as a whole in a probabilistic manner. This allows for the comprehensive combination of mineralization information, facilitating holistic inference of magma flow hidden in the mineralization signal. Based on the maximum likelihood principle, the magma trajectories and the associated injection points are inferred from the HMM. The proposed model is applied to spatial analysis of the mineralization in the world-class Jinchuan Ni-Cu-(PGE) magmatic sulfide deposit in NW China. Based on the 3D mineralization data interpolated from drill assays, the HMM infers magma trajectories and the associated injection points at Jinchuan, which indicate the Jinchuan deposit was formed by a multiple magma conduit system that comprises four independent injection points of magma ingress. According to the identified magma trajectories and injection points, four sites beneath the injection points and along the extension of magma trajectories are considered promising for future mineral exploration. With the insights gained from this case study, the proposed HMM can serve as an effective tool in the exploration information system to unveil magma/fluid flow history and guide exploration efforts toward concealed, deep-seated mineralization. • A Hidden Markov model for spatial analysis of three-dimensional mineralization distribution. • A spatial analysis approach to gain insights into magma flow in their mineralization footprint. • Magma trajectories are inferred by maximizing their posterior probability. • A promising tool for converting data to insight in exploration information system. • Jinchuan Ni–Cu-(PGE) deposit was possibly formed by a multiple magma conduit subsystem. [ABSTRACT FROM AUTHOR]