INVESTIGATING SOLID MANTLE UPWELLING RATES BENEATH MID-OCEAN RIDGES USING U-SERIES DISEQUILIBRIA, 1: A GLOBAL APPROACH

Abstract

U-series disequilibria in MORB produce trends of (230Th)/(232Th) as a function of (238U)/(232Th) for given lengths of ridge often defined by first-order segmentation. These trends are most easily explained as mixing between melts derived from chemically heterogeneous mantle sources. We propose that the slopes of these disequilibria trends provide a method for constraining solid upwelling rates beneath mid-ocean ridges that may avoid difficulties in interpreting absolute disequilibria. Melting models based on melt ascent under both equilibrium and disequilibrium transport conditions are used to explore the effect of solid upwelling rate variations on the slope of such trends. Solid upwelling rate variations have no effect on the slope of the disequilibria trends in near-fractional melting models in which melting only begins in the presence of garnet. If the high (238U)/(232Th) melt is derived from melting spinel peridotite, then near-fractional models can produce an upwelling rate dependence. Equilibrium porous flow models have a greater dependence of slope on upwelling rate. This dependence occurs irrespective of whether or not the high (238U)/(232Th) melt comes from a garnet-bearing source. We regress all available U-series data from individual areas of ridge for both the slopes of the trends and the corresponding errors. The slopes of all trends except one decrease as spreading rate decreases, consistent with the prediction of our modeling. The sole exception is the trend of data from the northern MAR which produces a well-defined mixing trend with samples from Sao Miguel (Azores) suggesting it is influenced by plume-type melting. The dependence of disequilibria slopes on spreading rate suggests that melting rate variations are recorded by 230Th excess variation. This mutual variation is most easily explained as a direct relation between spreading rate and solid upwelling rate. The close correspondence between the observed dependence of slope on spreading rate and the modeled dependence of slope on upwelling rate is consistent with purely passive plate driven flow of the solid mantle beneath ridges.