REGIONALLY DISTINCTIVE SOURCES OF DEPLETED MORB - EVIDENCE FROM TRACE-ELEMENTS AND H2O

The sources of depleted mid-ocean ridge basalts (N-MORB) worldwide are regionally distinctive in that they share incompatible element ratios with spatially associated, enriched mid-ocean ridge basalts (E-MORB). The ratio of H2O/Ce is uniform for N-MORB and E-MORB within a region, suggesting that the order of incompatibility during the evolution of MORB sources is La > H2O approximate to Ce > Nd. However, there are significant regional variations in H2O/Ce. N-MORB and E-MORB from the American-Antarctic Ridge (AAR), Southwest Indian Ridge (SWIR), southern Mid-Atlantic Ridge (SMAR), Pacific-Nazca Ridge 27-34 degrees S (PNR), East Pacific Rise 10-12 degrees N (EPR), Explorer Ridge, Mid-Cayman Rise Spreading Center (MCR) and Galapagos Spreading Center (GSC), as well as basalts from Loihi Seamount, have H2O/Ce ratios that average about 155-213 (+/-40 for each region). N-MORB through E-MORB from the Mid-Atlantic Ridge north of about 22 degrees N (NMAR) have higher H2O/Ce ratios, averaging 240-280 (+/-50 for each region). There are no correlations of H2O/Ce with spreading rate or extent or depth of melting, indicating that variations in H2O/Ce are not related to MORB melting but are a characteristic of the source. K/Nb ratios are also regionally variable but interpretation is complicated by a slight dependence of K/Nb on source enrichment. H2O/Ce is not correlated with K/Nb, He-3/He-4 or Pb isotopic parameters but may be correlated with high Sr-87/Sr-86 at a given Nd-143/Nd-144. Data from other regions are needed before a correlation between H2O/Ce and Sr-87/Sr-86 can be established. The poor correlation of H2O/Ce with He-3/He-4 makes it unlikely that H2O/Ce variations are related to variations in juvenile H2O in the source. It is more likely that H2O in MORB is derived from recycled, subducted, altered oceanic crust. High H2O/Ce in MORB from the NMAR might be related to a period of rapid subduction in the past that resulted in depressed isotherms and less dehydration in the slab. The constancy of H2O/Ce within regions despite differences between regions indicates that N-MORB and E-MORB sources may share a common heritage. This constraint on the evolution of the depleted mantle is not easily reconciled with most conventional models of mantle evolution. A model in which the sources of N-MORB have been influenced by inputs of regionally distinctive plume material that has been previously depleted by small extents of melting could account for the trace element variations but is physically implausible. It is possible that high H2O/Ce may be a regional characteristic of the mantle that is unrelated to plumes.