Continental crust formation: Numerical modelling of chemical evolution and geological implications

Abstract

Oceanic plateaus develop by decompression melting of mantle plumes and have contributed to the growth of the continental crust throughout Earth's evolution. Occasional large-scale partial melting events of parts of the asthenosphere during the Archean produced large domains of precursor crustal material. The fractionation of arc-related crust during the Proterozoic and Phanerozoic contributed to the growth of continental crust. However, it remains unclear whether the continents or their precursors formed during episodic events or whether the gaps in zircon age records are a function of varying preservation potential. This study demonstrates that the formation of the continental crust was intrinsically tied to the thermoconvective evolution of the Earth's mantle. Our numerical solutions for the full set of physical balance equations of convection in a spherical shell mantle, combined with simplified equations of chemical continent–mantle differentiation, demonstrate that the actual rate of continental growth is not uniform through time. The kinetic energy of solid-state mantle creep (Ekin) slowly decreases with superposed episodic but not periodic maxima. In addition, laterally averaged surface heat flow (qob) behaves similarly but shows peaks that lag by 15–30Ma compared with the Ekin peaks. Peak values of continental growth are delayed by 75–100Ma relative to the qob maxima. The calculated present-day qob and total continental mass values agree well with observed values. Each episode of continental growth is separated from the next by an interval of quiescence that is not the result of variations in mantle creep velocity but instead reflects the fact that the peridotite solidus is not only a function of pressure but also of local water abundance. A period of differentiation results in a reduction in regional water concentrations, thereby increasing the temperature of the peridotite solidus and the regional viscosity of the mantle. By plausibly varying the parameters in our model, we were able to reproduce the intervals of the observed frequency peaks of zircon age determinations without essentially changing any of the other results. The results yield a calculated integrated continental growth curve that resembles the curves of GLAM, Begg et al. (2009), Belousova et al. (2010), and Dhuime et al. (2012), although our curve is less smooth and contains distinct variations that are not evident in these other curves.