Abstract

Geodynamics of the early Earth: Venus-like episodic resurfacing?
Peter van Thienen

Because of the strongly different conditions in the mantle of the early Earth regarding temperature and viscosity, present-day geodynamics, dominated by plate tectonics, cannot simply be extrapolated back to the early history of the Earth. Numerical thermochemical convection models including partial melting and a simple mechanism for melt segregation and oceanic crust production were used to investigate an alternative suite of dynamics that may have been in operation in the early Earth. Our modelling results show three processes that may have played an important role in the production and recycling of oceanic crust: (1) Small scale (x * 100 km) convection involving the lower crust and shallow upper mantle with partial melting and thus crustal production in the upwelling limb and delamination of the eclogitic lower crust in the downwelling limb. (2) Large scale resurfacing events in which (nearly) the complete crust sinks into the (eventually lower) mantle, thereby forming a stable reservoir enriched in incompatible elements in the deep mantle. New crust is simultaneously formed at the surface from segregating melt. (3) Intrusion of lower mantle diapirs with a high excess temperature (about 250 K) into the upper mantle, causing massive melting and crustal growth. As crater count studies indicate a global resurfacing event to have taken place on Venus about 500 million years before present, episodic resurfacing has been suggested for Venus. We propose, on the basis of our model results, that the mechanism of large scale resurfacing described above may have been active on both the early Earth and Venus. Important constituents of Archean cratons, formed in the early and hot history of the Earth, are TTG (Tonalite-Trondhjemite-Granodiorite) plutons and greenstone belts. Our numerical models, which were expanded to include partial melting of (meta-) basalt, show that the resurfacing mechanism is capable of producing associations that resemble these granite-greenstone terrains. Partial melting at the base of newly produced crust associated with a large scale resurfacing events may generate felsic melts that are added as intrusives and/or extrusives to the generally mafic crustal succession, adding to what resembles a greenstone belt. Partial melting of metabasalt in the sinking crustal section produces a significant volume of TTG melt that is added to the crust directly above the location of `subduction', presumably in the form of a pluton. The $p,T$-conditions under which partial melting of metabasalt takes place in this scenario are consistent with geochemical trace element data for TTG's, which indicate melting under amphibolite rather than eclogite facies. Other geodynamical settings which have also been investigated, including partial melting in small scale delaminations of the lower crust, at the base of a anomalously thick crust and due to the influx of a lower mantle diapir fail to reproduce this behaviour unequivocally and mostly show melting of metabasalt in the eclogite stability field instead. These results thus provide a mechanism for generating granite-greenstone terrains without requiring the operation of plate tectonics. Although this resurfacing scenario may also have been important in Venus' history, it probably did not produce significant volumes of continental material due to the dryness of this planet.