Nearly half of the Earth’s highly incompatible trace elements are housed in the continental crust, which makes up only about 0.3 % of the mass of the mantle. The extreme enrichment of these elements in the continental crust is the integrated result of efficient extraction of these elements from the mantle during crust formation processes. However, if the continental crust gets its trace elements from the mantle, its major element composition says otherwise. The continental crust is too silicic to represent a melt that has been directly extracted from the mantle. Instead, melts from the mantle are basaltic. To resolve this apparent paradox, the basaltic melt must somehow be differentiated into silicic and mafic counterparts, the latter being removed and returned into the mantle, leaving behind the silicic endmember, which goes on to form the continental crust.

This talk will bring together petrology, geochemistry, tectonics, geodynamics, and geophysics in an attempt to fill in the petrogenetic and tectonic details of how continental crust grows and gets its felsic composition. We will investigate the Sierra Nevada and Peninsular Ranges batholiths in California, which are remnants of a Mesozoic continental arc associated with eastward subduction of the Farallon oceanic plate beneath western North America. Roughly speaking, addition of new continental material appears to have occurred in two stages. First, an island arc is accreted onto the margin of cratonal North America. This is followed by retreat of the trench, resulting in subduction of a slab beneath the accreted island arc. Juvenile arc magmatism is built through the accreted terrane. This is then followed by a shallowing of subduction, which causes the arc front to move eastwards into the original cratonal margin. Basaltic magmas, generated in the mantle wedge, intrude through the thicker pre-existing cratonal lithosphere and hence begin differentiation at much greater depths than island arcs. In so doing, they crystallize high MgO garnet pyroxenite lithologies, driving residual liquids to low MgO contents. This residual magma then ascends to lower crustal depths, crystallizing out low MgO garnet pyroxenite lithologies, which drives the liquids toward higher silica. Mixing of this continually evolving magma with partial melts of the surrounding country rock go on to generate even more silicic magmas, such as the granodiorites that make up the bulk of the Sierras and eastern Peninsular Ranges. The garnet pyroxenites are subsequently delaminated, leaving behind a crust that is biased towards felsic compositions.

I will thus propose a model wherein continents grow and get their felsic composition by incremental episodes of island arc accretion followed by continental arc magmatism. We have investigated some Archean garnet pyroxenite xenoliths and tonalite-trondjhemite-granodiorite suites and find that they may have similarities with Phanerozoic continental arcs. The Pb/U ratios of the Sierran garnet pyroxenites are much higher than the “canonical” lower crust values, suggesting that removal/storage of such material in the past could be the much sought-after missing low-‘mu’ reservoir needed to explain why the mantle lies to the right of the geochron on lead isotope diagrams.