Recycling of Paleoplacer Gold through Mechanical and Postdepositional Mobilization in the Neoarchean Black Reef Formation, South Africa

Abstract

A B S T R A C TThe source of gold in the ca. 2.66 Ga Black Reef Formation (BRF) has been investigated and constrained through petrographic, mineralogical, geochemical, and high-resolution three-dimensional reflection seismic data combined with drill core and underground geological mapping. The BRF is a strong seismic marker and consists of carbonaceous shale, quartz arenite, and conglomerate. Gold grade in the BRF is primarily controlled by the nature of the host conglomerates. Most of the gold in the BRF conglomerate occurs in native form, and its morphology is highly heterogeneous. Gold was initially introduced through mechanical recycling of underlying Witwatersrand reefs, followed by short-range (millimeter-to centimeter-scale) postdepositional alteration/remobilization associated with the Bushveld Complex and the Vredefort meteorite impact. Although the BRF was subjected to high postdepositional fluid circulation facilitated by high fracture density, the volume of dissolved gold was probably too small to form a large gold deposit, except in areas around the Black Reef/Witwatersrand reefs subcrop positions. Findings from this study demonstrate the importance of both sedimentological controls and impact-related structures in the formation of paleoplacer gold deposits during Neoarchean times.

The Temporal Distribution of the Host Rocks to Gold, the Archean Witwatersrand Basin, South Africa

“…Between the Ventersdorp Supergroup and the overlying Chuniespoort Group of the Transvaal Basin is the ~30 m thick Black Reef Formation of massive to cross-bedded quartz arenites, shales and siltstones, carbon rich shales, and minor conglomerates. The Black Reef Formation is not itself dated, although it is commonly described as 2590 Ma based on the relationship of rocks above and below [49] or 2642 ± 2 Ma [50,51], and we use 2640 Ma to 2590 Ma here. The Black Reef contains detrital zircons in age range from 3515 Ma mostly 3150 Ma to 3050 Ma, and to 2904 Ma as an overgrowth [51].…”

“…The Black Reef Formation is not itself dated, although it is commonly described as 2590 Ma based on the relationship of rocks above and below [49] or 2642 ± 2 Ma [50,51], and we use 2640 Ma to 2590 Ma here. The Black Reef contains detrital zircons in age range from 3515 Ma mostly 3150 Ma to 3050 Ma, and to 2904 Ma as an overgrowth [51]. We are aware of no explanation for the >200 Ma gap between source and deposition, so it remains uncertain whether everything referred to as mineralised Black Reef is correctly the (basal) unit of the Transvaal Supergoup or not [49].…”

Moving towards deep underground mineral resources: Drivers, challenges and potential solutions

“…Sensors can also enable in-situ and rapid estimation of resource grades by using trained models that relate in-mine observational and analytical data (e.g., through remote sensing) to resource potential (e.g., grades and uncertainty) (Daniels, 2015;Nwaila et al, 2019;Samson, 2019;Zhang et al, 2021c). Leveraging this type of technology, in principle, it should be possible to perform more accurate and timely grade control and in-situ assessments to drive rapid feedback between the tactical and operational levels.…”

“…For such data, the ideal approach to modelling is data-driven predictive modelling, such as through machine learning, which is capable of identifying and leveraging complex, multi-dimensional patterns to make inferences. This approach has no counterpart in traditional geostatistics and its application to resource modelling could reveal a range of possibilities, such as rapid point-wise target prediction (e.g., Nwaila et al, 2019;Zhang et al, 2021c). The overall aim is to leverage abundant data and modern algorithms to increase resource model reliability.…”

“…This is a product of geoscientific evolution, as data generation methods changed over the history of resource exploration and extraction. Nonetheless, such issues may be remediated using a combination of modern data science and discipline-specific practices (e.g., Nwaila et al, 2019;Zhang et al, 2021b, c;Mutshafa et al, 2022). The benefits of legacy data cannot be understated.…”