Restrictions on the compositions of mid-ocean ridge basalts: a fluid dynamical investigation

Abstract

The nature of the melts parental to mid-ocean ridge basalts has been vigorously debated. Opinions vary between those considering the parental magma to be magnesia-rich (picritic basalt) with 15–20% MgO and those arguing that the parental melts are compositionally related to the most primitive liquids erupted at mid-ocean ridges (MgO∼9–11%). This difference of between 5 and 11% in MgO content represents a temperature interval of 100–250 °C. The actual MgO content places constraints on the thermal conditions at the mantle source. The high-magnesia liquid theory is supported by melting experiments on mantle and basalt samples^1–3 and by petrological studies of ophiolite complexes⁴. However, this theory does not seem to be immediately consistent with the fact that no glasses or aphyric rocks of these compositions have ever been sampled from the sea floor. Indeed the most primitive glass compositions that occur as inclusions in olivine megacrysts and as glassy margins to pillow lavas have MgO contents in the range 9–11%. Models of generating low MgO partial melts from the mantle have been proposed. For example, Presnall et al.⁵ consider that melting at cusps in the peridotite mantle solidus could generate such melts. We outline here a fluid dynamical model for a basaltic mid-ocean ridge magma chamber supplied by high-magnesia melt from the mantle. The model predicts that high-magnesia extrusives will not be erupted, that ultramafic cumulates should occur at the base of the oceanic crust, that lava compositions will be restricted to MgO contents <10% and that magma mixing will be a common feature of sea-floor basalts⁶.