Consequences of diffuse and channelled porous melt migration on uranium series disequilibria

Magmas erupted at mid-ocean ridges (MORB) result from decompression melting of upwelling mantle. However, the mechanism of melt transport from the source region to the surface is poorly understood. It is debated whether melt is transported through melt-filled conduits or cracks on short time scales ( 10 yrs), or whether there is a significant component of slow, equilibrium porous flow on much longer time scales ( 10–10 yrs). Radiogenic excess Ra in MORB indicates that melt is transported from the melting region on time scales less than the half life of Ra ( 1600 yrs), and has been used to argue for fast melt transport from the base of the melting column. However, excess Ra can be generated at the bottom of the melt column, during the onset of melting, and at the top of the melt column by reactive porous flow. Determining the depth at which Ra is generated is critical to interpreting the rate and mechanism of magma migration. A recent compilation of high quality U-series isotope data show that in many young basalts, Ra excess in MORB is negatively correlated with Th excess. The data suggest that Ra excess is generated independently of Th excess, and cannot be explained by “dynamic” or fractional melting, where observed radiogenic excesses are all generated at the base of the melt column. One explanation is that the negative correlation of activity ratios is a result of mixing of slow moving melt that has travelled through reactive, low-porosity pathways and relatively fast moving melt that has been transported in unreactive high-porosity channels. We investigate this possibility by calculating U-series disequilibria in a melting column in which high-porosity, unreactive channels form within a low-porosity matrix that is undergoing melting. The results show that the negative correlation of Ra and Th excesses observed in MORB can be produced if 60% of the total melt flux travels through the low-porosity matrix. This melt maintains Ra excesses via chromatographic fractionation of Ra and Th during equilibrium transport. Melt that travels through the unreactive, high-porosity channels is not able to maintain significant Ra excesses because Ra and Th are not fractionated from each other during transport and the transport time for melt in the channels to reach the top of the melt column is longer than the time scale for Ra excesses to decay. Mixing of melt from the high porosity channels with melt from the low-porosity matrix at the top of the melting column can produce a negative correlation of Ra and Th excesses with the slope and magnitude observed in MORB. This transport process can also account for other aspects of the geochemistry of MORB, such as correlations between La/Yb, Sm/Nd, and Th/U and Ra and Th excess. Copyright © 2002 Elsevier Science Ltd