Flow and melting of a heterogeneous mantle: 2. Implications for a chemically nonlayered mantle

Abstract

We have recently examined the importance of upper mantle flow and melting of a heterogeneous mantle on the incompatible trace element and isotope composition of ocean island basalts (OIBs) and mid-ocean ridge basalts (MORBs) [G. Ito, J. Mahoney, Flow and melting of a heterogeneous mantle 1: method and importance to the geochemistry of ocean island and mid-ocean ridge basalts, Earth Planet. Sci. Lett. 230 (2005) 29-46]. We found that the combination of thicker lithosphere and enhanced mantle upwelling deep in the melting zone beneath hotspots can lead to many geochemical characteristics that distinguish OIBs from MORBs. These results occurred in the absence of any geochemical distinction between mantle plumes, which we assumed to feed OIB volcanism, and the upper mantle, which we assumed to feed MORB volcanism. Our findings renew the possibility for OIB and MORB geochemical systematics to arise out of a well-stirred heterogeneous mantle without large-scale layering. In this study, we test this possibility by solving for the mantle source compositions that, when melted in a hot plume beneath lithosphere of appropriate thickness, can explain key geochemical correlations at specific hotspots. Both the age of the seafloor at the time of volcanism and correlations between Sr, Nd, and Pb isotopes are considered in model fits to the OIB data. We then compute the magma composition that would arise out of these same mantle sources when melted at normal temperatures beneath a mid-ocean ridge. On one hand, some predicted MORB compositions lie outside the range of compositions represented by our collection of “normal” MORB data. This result is consistent with the conventional wisdom that mantle plumes sample a deep mantle layer that is compositionally distinct from the upper mantle. On the other hand, many solutions are indistinguishable from the observed MORB compositions and thus do not require a chemically layered mantle. Focusing on these solutions, the implied range of (nonlayered) mean mantle compositions can satisfy the global Th and U budget if the upper-bound estimates for the content of these elements in the continents are used. Our mean mantle produces only a small percentage (<∼30%) of the mantle heat budget and therefore requires a large heat flux from the core. Such a condition can allow for reasonably low secular cooling rates if a thick dehydrated lithosphere has existed over the mantle's history and has retarded global heat loss, as a recent study [J. Korenaga, Energetics of mantle convection and the fate of fossil heat, Geophys. Res. Lett. 30 (8), (2003) 1437, (doi:10.1029/2003GL016982)] suggests. Future tests for a nonlayered mantle should include integrated geodynamic and geochemical studies of correlations between noble gases and other isotopes, magma compositions at ultraslow spreading centers, and geographic variations in magma composition at hotspots, including those interacting with mid-ocean ridges.