‘Andesite Model’ of Continental Crust Formation Revisited

Abstract:

Fifty years ago, Ross Taylor observed that the chemical composition of average continental crust and orogenic andesites were similar for many major and trace elements. However, almost immediately it was recognized that this generalization does not apply to many oceanic island arcs because they are too mafic and too depleted in the most incompatible elements, including K, Th, and U. Although intra-crustal differentiation and delamination might produce sufficiently evolved middle and lower crust, it is an inefficient way to increase incompatible element concentrations in intermediate magma. There are alternative processes. First, because the vast majority of arc rocks are volcaniclastic sediments rather than igneous rocks, mechanical mixing is a dominant process during arc evolution. Although its effects are similar to those of magma mixing and assimilation, it has fewer physical constraints and is more likely. The finer-grained the sediment, the more mixing is likely. Drilling in the Izu arc shows that the majority of the sediment is andesitic mud even though the majority of lavas are basalt and rhyolite. Volcaniclastic sediments may be a better match to continental crust than the lavas that were the original basis of the andesite model. Second, some reararc lavas and sediments are more like continental crust than those from volcanic fronts, and reararc crust is at least as thick. Third, aeolian dust in arc muds may be a larger mass fraction of recycled continental crust in oceanic arcs than the subducted pelagic sediment component in arc magmas. The net effect of these features adds nuance to the andesite model but is not enough to overcome the deficit of incompatible elements in Cenozoic oceanic arcs. Accreting them still makes the continents less continental in composition.