GSA Annual Meeting in Phoenix, Arizona, USA - 2019

Paper No. 163-1
Presentation Time: 8:15 AM


LATYPOV, Rais and KRUGER, Willem, School of Geosciences, Wits University, Private Bag 3, Johannesburg, 2050, South Africa

At the heart of magma differentiation on Earth – irrespective of whether this occurs in shallow magma chambers, magma oceans/impact melt sheets or planetary cores – is a physical separation of chemically distinct crystals and liquids. Exactly how this separation takes place to cause chemical differentiation of magmas is a central question of modern volcanology and igneous petrology. Critical information necessary to address this issue is locked in the internal structure of solidification fronts – the partially crystalline zones of magma that occur along the margins of all magmatic systems and are thermally contained between the solidus and liquidus isotherms. Here we present the first ever records of fossilized solidification fronts from the Bushveld Complex – the largest basaltic magma chamber in the Earth’s crust. These were discovered by two-dimensional geochemical mapping of massive magnetitites on field outcrops. The mapping revealed that nucleation and crystallization start directly at the chamber floor as concentrically-zoned nodes which, with continued lateral growth, coalesce into a single, seemingly homogeneous, layer. The steep chemical gradients across the in situ growth nodes imply that magma differentiation in the chamber was remarkably effective and likely induced by a convective instability of a liquid boundary layer at the crystal-liquid interface. This process results in complete removal of a chemical boundary layer from growing crystals to form solid adcumulates directly at the chamber floor. Our direct observations and numerical modelling indicate no compaction in solidification fronts of the Bushveld Complex and imply that perfect fractional crystallization in basaltic chambers can be achieved by convection induced by an instability of a chemical boundary layer around in situ growing crystals. We speculate therefore that this type of compositional convection may represent a very effective mechanism of liquid evolution in magmatic systems. Our data also indicate that a mushy zone in the Bushveld Complex was either non-existent (in magnetitite) or, at best, only a few metres thick (in silicate rocks) which strongly argues for a classical ‘magma chamber’ paradigm that has been underpinning models of the Earth’s magmatism for over a century.