Partial melting experiments on a MORB-like pyroxenite between 2 and 3 GPa: Constraints on the presence of pyroxenite in basalt source regions from solidus location and melting rate

[1] We present partial melting experiments at 2-3 GPa on a basaltic pyroxenite (G2) similar in composition to typical oceanic crust. The 3.0 GPa solidus is located at 1310 +/- 12degreesC and the liquidus is 1500-1525degreesC. Clinopyroxene, garnet, quartz, and rutile are subsolidus phases. Garnet, quartz, and rutile are absent above 1475degreesC, 1365degreesC, and 1335degreesC, respectively. At the solidus, the garnet mode is low (18 wt.%) because clinopyroxene is unusually aluminous (13.8-15.5 wt.% Al2O3). In adiabatically upwelling mantle near 2-3 GPa, G2-like pyroxenite begins melting 35-50 km deeper than peridotite. The calculated near-solidus adiabatic productivity for G2 is similar to13%/GPa and averages similar to59%/GPa through the melting interval, suggesting substantial partial melting deep in basalt source regions: G2 is similar to60% molten at the 3 GPa peridotite solidus. Small percentages of pyroxenite in the source significantly affect oceanic crust production and composition, as the proportion of pyroxenite-derived melt contributed to oceanic crust formation is 5 to > 10 times the pyroxenite proportion in the source. Given the overall depleted isotopic character of mid-ocean ridge basalt (MORB), oversampling of fertile G2-like pyroxenite limits the abundance of such lithologies to similar to<2% of the MORB source. Owing to high extents of partial melting, the effect of modest amounts of pyroxenite on Sm/Yb ratios of aggregated basalts is limited and depends largely on the average bulk composition of the pyroxenite source. Low near-solidus adiabatic productivities could allow small (similar to 1-2%) proportions of basaltic pyroxenite to enhance (Th-230)/(U-238) in oceanic basalts without requiring marked shifts in other indicators of heterogeneity, such as Sr or Pb isotopes.