Inverse relationship between Sr isotope diversity and rate of oceanic volcanism has implications for mantle heterogeneity

Abstract

Cohen and O'Nions¹ noted that basalts from the Mid-Atlantic Ridge (MAR) display greater diversity of Pb, Nd and Sr isotopes than do basalts from the East Pacific Rise (EPR). They attributed this difference not to greater isotopic heterogeneity beneath the MAR, but rather to more effective mixing (to eradicate heterogeneity) beneath the EPR. Allègre et al.² reached the same conclusion on the basis of isotope data including noble gas isotopic data. To test this idea further, I compare here the average spreading rate of nine portions of the mid-ocean ridge system with the measured ⁸⁷Sr/⁸⁶Sr diversity of basalts. This shows that the maximum observed ⁸⁷Sr/⁸⁶Sr diversity is inversely proportional to spreading rate. Furthermore, globally averaged eruption rates for mantle-derived rocks of intra-oceanic island arcs, ocean islands and the mid-ocean ridge system exhibit a negative correlation with ⁸⁷Sr/⁸⁶Sr diversity. These results suggest that the upper mantle is everywhere heterogeneous on a small scale and that the extent of observed heterogeneity is a function of mixing. Independent geophysical and petrological evidence for the existence of steady-state crustal magma chambers beneath fast-spreading ridges and their absence below slow spreading ridges favours this hypothesis because mixing is enhanced by the presence of a magma chamber. Alternatively, mixing during melt production in the mantle, during melt segregation or during ascent could also result in the observed patterns if the degree of mixing is related to eruption rate. Large heterogeneities (mantle plumes?) are required to explain why isotopic variations are well correlated geographically in some places but not in others.