Principal component analysis of trace element data from the Upper Zone of the Bushveld Complex : constraints on possible tectonic provenance

Abstract

The Kaapvaal Craton in South Africa is characterised by ultramafic to mafic complexes, and several mafic intrusions are associated closely with different sections of the craton. The focus of this research is the Transvaal basin in the Kaapvaal Craton that hosts the Bushveld Complex. This complex is the most extensive mafic layered intrusion on Earth, with the magmas ranging from Fe-rich to Si-rich. The extensive body of research on this area is a good foundation for the current study, which focuses on the Upper Zone and upper Main Zone (UUMZ) of the complex. The Bushveld Complex is endowed with rich mineral deposits, lending it robust economic significance, allowing substantial contributions to the global economy through various avenues relating to mineral processing and distribution. Objectively, this research project seeks to provide tectonic constraints on the origin of the magma emplaced in the UUMZ through geochemical analyses of trace elements and rare earth element (REE) data from Bierkraal boreholes (western limb) and Magnetite Layer 21 (eastern limb), respectively. In this research approach, the geochemistry of basaltic rocks is considered a predictive indicator in determining the likely tectonic setting in which the parental magma originated. A global database of basaltic rock data was employed for this project, namely the GEOROC database, which was subdivided according to tectonic settings (convergent plate margin, oceanic island, submarine ridge, and oceanic plateau) for comparing the geochemical signatures against that of the Bushveld Complex. The overall data were analysed following a multidisciplinary approach, i.e. combining geostatistics and programming. Principal component analysis (PCA) was conducted of the Bierkraal trace element data (sampled in the western limb of the complex) by employing various software applications to generate a probability distribution function of the variables, i.e. the trace elements in this context. Initially, PCA identifies the elements that impart the most significant variation to the dataset. These elements function as indicators and identifiers of specific geological and chemical processes. Overall, the geochemical data from the Bierkraal area and those from Layer 21 are crucial for narrowing down the potential tectonic settings exhibiting similar signatures. The settings and locations most synonymous to those of the Bushveld appear to be those of the Kaula and Nihau islands of Hawaii. In addition, Hawaii data exhibit more depletions in elements such as Nb and Y relative to those of the UUMZ of the Bushveld Complex. The same data show overlaps with data from the Fiji and Tonga islands. These oceanic island-island arc settings are influenced primarily by mantle hotspots/plumes. On the other hand, the REE apatite data show a more generalised overlap with most identified tectonic settings within arc-related settings. The geochemical signatures showing the most similarity are those from the lower 300 m segment of the Upper Zone, characterised by multiple magnetite layers, and those from the East Scotia Ridge spreading centre, located approximately 1 288 km to the east of the tip of South America. This ridge is a back-arc tectonic environment associated with diverging plate motions and subsequent rifting, allowing molten rock to push upward and build new crust. Both geochemical signatures from the Bierkraal trace element data and REE apatite data can be related to an enriched magmatic source (i.e. E-MORB). Conventionally, an enriched geochemical signature is associated with magma derived from the lower parts of the mantle. However, taking the geological location and age of the Bushveld complex into consideration, the E-MORB geochemical signature identified in the Upper Zone could be created by the melting of Archaean crust during the upwelling of depleted (N-MORB) magma.