HR: 11:35h
AN: V41D-06 [PDF]
TI: Mantle Melting as a Function of Water Content in Arcs
AU: * Kelley, K A
EM: kelley@dtm.ciw.edu
AF: Dept. of Earth Sciences, Boston University, Boston, MA 02215 United States
AU: Plank, T
EM: tplank@bu.edu
AF: Dept. of Earth Sciences, Boston University, Boston, MA 02215 United States
AU: Newman, S
EM: sally@gps.caltech.edu
AF: Div. of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125 United States
AU: Stolper, E
EM: ems@gps.caltech.edu
AF: Div. of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125 United States
AU: Grove, T L
EM: tlgrove@MIT.EDU
AF: Dept. of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology, Cambridge,
MA 02139 United States
AU: Parman, S
EM: parman@MIT.EDU
AF: Dept. of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology, Cambridge,
MA 02139 United States
AU: Hauri, E
EM: hauri@dtm.ciw.edu
AF: Dept. of Terrestrial Magnetism, Carnegie Institution of Washington, Washington, DC 20015 United States
AB:
Subduction zone magmas are characterized by high concentrations of dissolved H$_{2}$O, presumably derived from the subducted
plate and ultimately responsible for melt generation in this tectonic setting. Almost ten years ago, Stolper and Newman
(EPSL, 1994) illustrated a linear relationship between the concentration of water (H$_{2}$O$_{o}$) and the fraction of
melting (F) in the mantle beneath the Mariana back-arc. Here we report new major element and volatile data for
olivine-hosted melt inclusions from the Mariana Islands to test this relationship for melting beneath an arc. Basaltic melt
inclusions from the Mariana arc have water contents (2.3-6.1 wt% H$_{2}$O) significantly higher than all basaltic glasses or
melt inclusions from the Mariana back-arc (0.2-2.2 wt% H$_{2}$O). We use TiO$_{2}$ as a proxy for F, after correcting for
crystal fractionation, and evaluate the Ti source composition with a model based on Ti/Y variations in mid-ocean ridge
basalts (MORBs). Each calculated F thus represents the amount of mantle melting for a single melting episode. Even after
accounting for mantle depletion, the TiO$_{2}$ concentrations in Mariana arc magmas record higher extents of mantle melting
(F = 10-30%) than recorded in back-arc magmas (F = 5-24%). As a whole, the Mariana arc broadly extends the linear
H$_{2}$O$_{o}$-F array defined by the back-arc, although in detail the islands show important differences. Two islands from
the Mariana arc (Guguan and Pagan) define a H$_{2}$O$_{o}$-F slope similar to the Mariana back-arc, suggesting similar mantle
potential temperature beneath the arc and back-arc ($\sim$1360 $\pm$ $20\deg$C). Melts from Agrigan island, however,
indicate a steeper slope suggestive both of cooler mantle beneath Agrigan and of along-strike thermal variations beneath the
Mariana Islands. Both the arc and back-arc arrays project to finite F at zero water in the mantle, providing evidence for
decompression melting in both settings. These relationships may be extended globally to other back-arc and arc systems, using
published data. For back-arc basins (Mariana, Lau, Scotia, Manus), the main factor contributing to regionally distinct
trends in H$_{2}$O$_{o}$ vs. F is mantle potential temperature, which may be determined in dry back-arc melts using the same
geochemical criteria as MORBs, and spans a similar range ($\sim$1300-$1500\deg$C; see Plank and Kelley, this meeting).
Increasing mantle temperature leads to a progressive change in slope and intercept in the back-arc H2Oo-F arrays. In the
Cascadian, Central American, and Mexican arcs, the wet melting functions indicate overall colder mantle potential
temperatures in arcs ($\sim$1100-$1300\deg$C) than in global back-arcs, presumably due to proximity to the cold subducting
plate. The simple relationships between water and melt fraction recorded in arc magmas thus allow for quantitative separation
of the factors affecting mantle melt production in subduction zones, namely water addition from the slab, decompression, and
mantle temperature.
DE: 1025 Composition of the mantle
DE: 3640 Igneous petrology
DE: 3670 Minor and trace element composition
DE: 8439 Physics and chemistry of magma bodies
SC: Volcanology, Geochemistry, Petrology [V]
MN: 2003 Fall Meeting