Growth of the continental crust: a planetary-mantle perspective

Abstract

This paper addresses the net growth through time of the total continental crust (CC), not to be confused with sporadic growth of individual continental masses (e.g., by accretion of island arcs), a process that does not necessarily require net growth. Geochemists tend to emphasize differences between CC and oceanic crust (OC), but from a whole-planet perspective CC and OC are remarkably similar in composition and density. The CC/OC distinction is canonically defined in terms of elevation. However, the Earth's bimodal elevation pattern, clearly reflects a genetic distinction. The OC participates directly in mantle convection, whereas the CC is the remaining, relatively stable portion of the crust. The Earth's advective cooling mechanism requires massive subduction of lithosphere, which due to the inherent buoyancy of crust tends to require a low crust/total lithosphere thickness ratio. Given relatively constant lithosphere thickness (except near thermally buoyant mid-ocean ridges), the Earth responds by heaping most of its crust into a few thick continents, allowing subduction to freely operate over a large area of relatively thin crust. This crustal thickness bimodality engenders the bimodal elevation pattern. Prevailing wisdom tends to hold that the rate of CC growth has been slow and ongoing; typical models show CC volume roughly proportional to 1/ h, where h is the total heat generation by the Earth's K, Th and U. However, such models are generally vague as to the mechanism that supposedly holds back growth, even after 4.5 Ga. The strong T-dependence of mantle rheology probably ensured that the mantle took only a few million years to cool within a few hundred degrees of its modern T. Gradual compositional evolution of the CC from a plausible initial composition (e.g., basaltic, anorthositic) might have slightly lowered the mean CC density, but this effect is of second-order importance in comparison to the general increase in crust/total lithosphere thickness ratio (for any given crustal thickness) engendered by temporal cooling of the mantle. Intense cratering, occurring coevally with early intense crust/mantle recycling, probably reconstituted virtually all of the CC extant prior to ~ 3.9 Ga, but CC was probably not recycled into the mantle any faster than it was regenerated. Lunar samples and meteorites provide strong circumstantial evidence that in primordial times the outer few hundred kilometers of all terrestrial planets were largely molten. Outcomes of primordial magmatism must have been strongly dependent on the pressure regimes within the planetary interiors. In general, the far greater dT/ dP of silicate melting curves vs. adiabats probably tended to foster a layered magmasphere, with a partially molten or paracumulate layer, constrained by dT/ dP relationships to a P range of roughly 3 GPa, surmounted by fully molten “magma ocean” layer, also limited in thickness by efficient convective cooling. In the Earth, P-stabilized garnet and Al-rich pyroxene were key factors limiting ultimate yield of crustal feldspar. This effect has been examined with quantitative models of high-P fractional crystallization (including, in some models, periodic replenishment of the magma ocean from below). Results indicate that a plausibly deep (~ 300–400 km) magma ocean would produce a crust comparable in both composition and thickness to that of the modern Earth. Mildly worrisome discrepancies involve CaO (for which high-P pyroxene compositions are poorly constrained) and SiO 2. The crust probably evolved to its present SiO 2-rich composition through differentiation catalyzed by H 2O, which presumably was not as abundant in the magma ocean as in the crust soon afterward. Nominal model results for mean crustal thickness are typically ~ 1.2–1.3 × the modern global mean, but for this parameter the model is designed to give conservatively high results, so the discrepancy is probably not significant. After the chaotic first few million years of solar system history, I suggest that planetary size and composition, for more than time, govern crustal volume.