Featured Article: Davies, G. F. (2009), Reconciling the geophysical and geochemical mantles: Plume flows, heterogeneities, and disequilibrium

Geochem. Geophys. Geosyst., 10, Q10008, doi:10.1029/2009GC002634.

"For several decades geophysicists and geochemists have favoured rather different and mutually incompatible models of the mantle. Geochemists have persistently favoured some form of layering, mainly because the strongly depleted MORB source cannot accomodate the Earth's budgets of many incompatible trace elements without assuming some large reservoir that is less depleted or enriched in those elements. Davies argues here that the geophysical evidence precludes not only a boundary between the upper and lower mantles, but also any layering other than the thin D" layer at the base of the mantle. The evidence comes not only from seismology but also from the topography of the sea floor, which constrains the form of convection in the mantle, including whether it is layered. If there is no major enriched layer then the MORB source itself is required to be less depleted than previously inferred. For example the U concentration needs to be 10 ng/g rather than the 3-5 ng/g previously inferred. Davies argues that the degree of MORB source depletion has been overestimated for several reasons, all related to the major-element heterogeneity of the mantle. First, geochemists have assumed the MORB source gives rise to a fictitious "normal MORB", which is relatively depleted, rather than to the full distribution of MORBs, yet the "All-MOR" data set from PetDB yields concentrations of incompatible trace elements that are 2-3 times higher than "normal MORB". Second, they have tended to exclude OIB-like signatures because they would come from plumes that do not sample the MORB source yet, surprisingly, up to 25% of mantle heterogeneities may be former plume material. Third, recent estimates rely on compositions of peridotites that may not have equilibrated with the mean composition of the heterogeneous MORB source, and thus may not record the more enriched components. - G. Davies