HR: 11:05h
AN: T12A-04    [Abstracts]
TI: MORB Composition in Intra-transform Spreading Centers: A Key Test of Models of Mantle Flow and Melt Transport
AU: Saal, A E
EM: asaal@brown.edu
AF: Department of geological Sciences, Brown University, 324 Brook Street, Providence, RI 02912 United States
AU: * Forsyth, D W
EM: Donald_Forsyth@brown.edu
AF: Department of geological Sciences, Brown University, 324 Brook Street, Providence, RI 02912 United States
AB: Models of mantle flow, melt generation and melt transport predict the composition and volume of melts delivered to the base of the crust in a mid-ocean ridge system. Many models can be tuned to match normal crustal thickness and composition, but they differ in predicting how the system behaves when it is perturbed by a transform fault offset. However, this "transform edge effect" or the pattern of along-axis delivery of melt to the crust within a ridge segment can be obscured by along-axis transport of magma within the crust in dikes or long-lived, continuous magma chambers. The advantage of sampling intra-transform spreading centers is that they provide well-defined locations of upwelling and crustal formation in a perturbed part of the system that does not allow along-axis transport of melt in the crust from other parts of the system. The composition of basalts from these settings can thus provide a critical test of mantle flow and melting models. Basalts from both the Garrett and Siqueiros intra-transform settings on the East Pacific Rise (EPR) have primitive composition with very depleted trace element contents, low ratios of very incompatible to moderately incompatible elements, and (230Th/238U) activity ratios ranging from slightly higher to lower than 1 suggesting that the first incompatible-element-rich melts are missing and only the magma from subsequent, shallower melting of an already depleted mantle reaches the intra-transform spreading centers. We have constructed 3-D models of passive mantle flow and melt driven by dynamic pressure gradients and buoyancy using the specific geometry of these two transform settings to predict the resulting composition. We compare the measured trace element and isotopic composition of both the intra-transform and normal segment lavas to those calculated using our model. To minimize the influence of a heterogeneous source, we chose samples from the 9-10 N EPR and Siqueiros with very similar isotopic composition. In contrast, to evaluate the effect of mantle heterogeneities we selected basalts from the Garrett and 13-23 S EPR, where the mantle probably contains heterogeneities that melt at greater depth. This passive flow model is very successful in reproducing the composition of lavas from both the transform fault and the normal segment. Some other melt migration models can be rejected, as they fail to correctly predict the compositional variations.
DE: 3035 Midocean ridge processes
DE: 1040 Isotopic composition/chemistry
DE: 1065 Trace elements (3670)
SC: Tectonophysics [T]
MN: 2004 AGU Fall Meeting