Eclogites lower mantle

Similar results have been reported by Mattern et al. (2002), using more recent equations of state for lower-mantle minerals and incorporating the solubility of alumina in silicate perovskite. They also used a three-layered slab model (midocean ridge basalt (MORE) over harzburgite over pyrolite), but with a MORE composition (Si/(Mg - - Fe) = 2.29) intermediate between our extreme end-members of the Helffrich et al. (1989) eclogite (1.65) and the Helffrich and Stein (1993) gabbro (2.58). 

Calculated bulk rock trace-element systematics of eclogites have wider implications for mantle recycling models and bulk silicate earth mass balance. The subchondritic Nb/Ta, Nb/La, and Ti/Zr of both continental cmst and depleted mantle require the existence of an additional reservoir with superchondritic ratios to complete the terrestrial mass balance. Rudnick et al. (2000) have shown that rutile-bearing eclogites from cratonic mantle have suitably superchondritic Nb/Ta, Nb/La, and Ti/Zr such that if this component formed 1 -6% by weight of the bulk silicate earth, this would resolve the mass imbalance. This mass fraction far exceeds the likely mass of eclogite in the continental lithosphere and so the material is proposed to reside in the lower mantle, possibly at the core-mantle boundary. 

Mass balance calculations suggest that this eclogitic reservoir is of considerable size. Kamber and Collerson (2000) estimated the mass of deeply subducted ocean crust over 4.3 Ga to be about 1.4 X 1026 g, that is, about 3% of the mass of the silicate Earth, greater than the mass of the continental crust, and equal to about 20% of the mass of oceanic crust subducted over time. Where this reservoir is located is not precisely known. However, Kamber and Collerson (2000) argued that the Nb and Ta contents of MORB worldwide are very well correlated, implying that, for these elements, the depleted mantle is a well-mixed reservoir, and unlikely therefore to contain high Nb/Ta slabs. For this reason they propose that the eclogitic slab reservoir is located in the lower mantle. 

Griffin W. L., Carswell D. A., and Nixon P. H. (1979) Lower-cmstal granulites and eclogites from Lesotho, southern Africa. In The Mantle Sample Inclusions in Kimberlites (eds. F. R. Boyd and H. O. A. Meyer). American Geophysical Union, Washington, DC, pp. 59-86. 

In summary, for a variety of reasons, it is unwarranted to use geodynamical models to rule out partial melting of eclogite facies sediment, basalt, or lower crustal gabbro beneath arcs. Meanwhile, we believe that geochemical data on primitive arc lavas are best understood if a partial melt of subducted sediment and/or basalt (and/or gabbro) is the primary medium for transport of in compatible trace elements into the mantle wedge. 

---