Effects of Variable Melt Productivity and Active Mantle Upwelling on Trace-Element and Isotopic Composition of Hotspot Magmas

We examine the effects of variable melt productivity and upwelling rate on the composition of magmas generated by mantle plumes. Melting of peridotite with relatively high concentrations of highly incompatible elements and volatiles is expected to begin at greater depths than for more depleted peridotite. The deepest melting is expected to occur at a minimal rate of melt production per increment of decompression (i.e. melt productivity) and to liberate melts with maximum concentrations of incompatible elements. In addition, the rate of upwelling, and thus decompression melting, in a buoyant mantle plume is expected to be relatively high at the base of the melting zone. As melting proceeds, melt productivity increases substantially, incompatible element concentrations decrease, and mantle decompression rate decreases. Together these effects predict pooled melts with greater concentrations of incompatible elements compared to models that do not include a deep zone of low melt productivity and active mantle upwelling. We consider these effects on the mixing of melts derived from a heterogeneous mantle source in which an incompatible-element-enriched component begins melting deeper, in an expanded zone of low melt productivity, and a more depleted component begins melting shallower, in a smaller zone of low productivity. We also assume the enriched component has relatively high ⁸⁷Sr/⁸⁶Sr and low ¹⁴³Nd/¹⁴⁴Nd. Beneath young/thin lithosphere, the effects of a deep low-productivity zone and active upwelling are minimized; thus magmas are only moderately enriched in incompatible elements and are isotopically similar to the depleted source component. Beneath lithosphere of greater age, where melting stops deeper, the effects of the deep low-productivity zone and active upwelling become increasingly important. Predicted magma compositions become increasingly enriched in incompatible elements and isotopically more similar to the enriched source component. These predictions are consistent with the observations of relatively low ⁸⁷Sr/⁸⁶Sr, high ¹⁴³Nd/¹⁴⁴Nd and a smaller total range of isotopic variability at hotspots near mid-ocean ridges compared to intraplate hotspots, and with observations along the Hawaiian-Emperor Chain. We do not require mixing with depleted upper mantle to explain the compositions of near-ridge hotspots. More broadly, compared to models that ignore a deep low-productivity zone and active upwelling, our models can account for many geochemical differences among hotspots, and between hotspots and mid-ocean ridges, with a smaller range of mantle source composition.