Discriminating between pyroxenite and peridotite sources for continental flood basalts (CFB) in southern Africa using olivine chemistry

Abstract

Continental Flood Basalts (CFB) result from voluminous outpourings of magma that often precede continental break-up. Notwithstanding the petrogenetic importance of CFBs, the nature of the mantle source for such magmas is contentious, particularly with regard to picrites with Ni-rich olivine phenocrysts. Previous studies have suggested that Ni-rich olivines associated with plume volcanism in regions of thickened (>90 km) lithosphere are related to either source mineralogy differences (peridotite versus pyroxenite) or change in olivine-melt partitioning due to pressure increase. In order to evaluate these two hypotheses, we present trace element data for olivines from the Karoo CFB Tuli and Mwenezi picrites and the Etendeka CFB Horingbaai/LTZ-L type picrites, all of which erupted in regions of thickened (>90 km) lithosphere in southern Africa. Karoo picrite olivines are Ni-rich, Ca- and Mn-poor, and have low (<1.4) 100*Mn/Fe. These compositions are consistent with a pyroxenitic source. Etendeka Horingbaai/LTZ-L picrite olivines do not show Ni-enrichment, but are characterized by high Al and Cr, and high (>1.4) 100*Mn/Fe, which is more consistent with high temperature melting of a dominantly peridotitic source. We also show that the Karoo and Etendeka olivines are characterized by distinct Mn/Zn ratios of <13 and >15, respectively.In addition, bulk rock geochemical data compilations and previously reported olivine δ18O for Karoo and Etendeka CFBs are discussed in order to further constrain source components based on previously described pyroxenite melt geochemical indices such as MgO–CaO systematics, FeO/MnO, Zn/Fe, and FC3MS (FeO/CaO–3*MgO/SiO2). These geochemical indices suggest a pyroxenite-dominated source for Karoo CFBs as well as for Etendeka ferropicrites whereas a peridotite-dominated source is indicated for Etendeka Horingbaai/LTZ-L type picrites analyzed in this study. Based on our data, Ni-enrichment of olivine in plume-related magmas in regions of thickened lithosphere in southern Africa is not ubiquitous. We therefore suggest that mineralogical variation of the source is a more likely major control of olivine chemistry and parent melt variations for Karoo and Etendeka CFBs. We also show that olivine Mn–Zn correlations are a useful discriminator for source variation and recommend the use of olivine Mn/Zn<13 for a pyroxenite-dominated source relative to olivine Mn/Zn>15 for a peridotite-dominated source.