Redox conditions for mantle plumes

Lara E. Heister and Charles E. Lesher

The vanadium to scandium ratio (V/Sc) for basalts from mid-ocean ridge (MOR) and arc environments has been proposed as a proxy for fO2 conditions during partial melting (e.g. [1] and [2]).  Contrary to barometric measurements of the fO2 of primitive lavas, the V/Sc ratio of the upper mantle at mid-ocean ridges and arcs is similar, leading previous authors to propose that the upper mantle has uniform redox potential and is well-buffered.  We have attempted to broaden the applicability of the V/Sc parameter to plume-influenced localities (both oceanic and continental), where mantle heterogeneities associated with recycled sediments, mafic crust, and metasomatized mantle, whether of shallow or deep origin, exist.  We find that primitive basalts from the North Atlantic Igneous Province (NAIP), Hawaii (both the Loa and Kea trends), Deccan, Columbia River, and Siberian Traps show a range of V/Sc ratios that are generally higher (average ~9) than those for MOR (average ~ 6.7) or arc (average ~7) lavas.  Based on forward polybaric decompression modeling, we attribute these differences to polybaric melting and melt segregation within the garnet stability field rather than the presence of a more oxidized mantle in plume-influenced settings.  Like MORB, the V/Sc ratios for plume-influenced basalts can be accounted for by an oxidation state approximately one log unit below the Ni-NiO buffer (NNO-1).  Our analysis suggests that source heterogeneities have little, if any, resolvable influence on mantle redox conditions, although they have significant influence on the trace element and isotopic composition of mantle-derived melts.  We suggest that variations in the redox of erupted lavas is largely a function of shallow lithospheric processes rather than intrinsic to the mantle source, regardless of tectonic setting.

[ return to abstract list ]