V43D-04
Density Sorting During the Evolution of Continental Crust

Thursday, 17 December 2015: 14:25
310 (Moscone South)
Peter B Kelemen, Columbia University of New York, Palisades, NY, United States, Mark D Behn, Woods Hole Oceanographic Institution, Geology and Geophysics, Woods Hole, MA, United States and Bradley R Hacker, University of California Santa Barbara, Santa Barbara, CA, United States
Abstract:
We consider two settings – in addition to “delamination” of arc lower crust – in which dense, mafic eclogites founder into the convecting mantle while buoyant, felsic lithologies accumulate at the base of evolving continental crust.

Arc processes play a central role in generating continental crust, but it remains uncertain how basaltic arc crust is transformed to andesitic continental crust. Dense, SiO2-poor products of fractionation may founder from the base of arc crust by “delamination”, but lower arc crust after delamination has significantly different trace elements compared to lower continental crust (LCC). In an alternative model, buoyant magmatic rocks generated at arcs are first subducted, mainly via subduction erosion. Upon heating, these buoyant lithologies ascend through the mantle wedge or along a subduction channel, and are “relaminated” at
the base of overlying crust (e.g., Hacker et al EPSL 11, AREPS 15). Average buoyant lavas and plutons
for the Aleutians, Izu-Bonin-Marianas, Kohistan and Talkeetna arcs fall within the range of estimated LCC major and trace elements. Relamination is more efficient in generating continental crust than delamination.

Himalayan cross-sections show Indian crust thrust beneath Tibetan crust, with no intervening mantle. There is a horizontal Moho at ca 80 km depth, extending from thickened Indian crust, across the region where Tibetan crust overlies Indian crust, into thickened Tibetan crust. About half the subducted Indian crust is present, whereas the other half is missing. Data (Vp/Vs; Miocene lavas formed by interaction of continental crust with mantle; xenolith thermometry) indicate 1000°C or more from ca 50 km depth to the Moho since the Miocene. We build on earlier studies (LePichon et al Tectonics 92, T’phys 97; Schulte-Pelkum et al Nature 05; Monsalve et al JGR 08) to advance the hypothesis that rapid growth of garnet occurs at 70-80 km and 1000°C within subducting Indian crust. Dense eclogites founder, while buoyant lithologies accumulate in thickening crust. Mantle return flow and radioactive heating in thick, felsic crust maintains high T. The crustal volume in this region grows at 760 m3/yr/m of strike length. Moho-depth earthquakes may be due to localized deformation and thermal runaway in weak layers and diapir margins.