Role of H2O in subduction-zone magmatism: New insights from melt inclusions in high-Mg basalts from central Mexico
- 1Department of Geology and Geophysics, Texas A&M University, College Station, Texas 77845, USA
- 2Ocean Drilling Program, Texas A&M University, College Station, Texas 77845, USA
Abstract
Although there is a growing body of data on H2O in arc magmas, there is still considerable uncertainty about the relationship between H2O and various incompatible elements during enrichment of the mantle wedge by subduction processes. We report data for H2O, other volatiles (CO2, S, Cl), and trace elements in olivine-hosted melt inclusions from high-Mg basalts in central Mexico that exhibit varying degrees of subduction-related enrichment. Most melt inclusions were trapped at low pressure, but rare inclusions (Mg# 65–78, olivine hosts Fo85–90) trapped at upper to middle crustal pressures (1–6 kbar) contain high CO2 (250–2120 ppm). The high-pressure inclusions indicate magmatic H2O contents from 1.3 to 5.2 wt%. Enrichment of H2O relative to Nb correlates positively with K/Nb, Ba/Nb, and La/Nb, indicating a clear link between H2O and trace element enrichment of the mantle wedge. Our results show that fluxing of the wedge with an H2O-rich component from the subducted slab is important in formation of magmas that are enriched in large ion lithophile (LILE) and light rare earth (LREE) elements relative to high field strength elements (HFSE). In contrast, magmas with low LILEs and LREEs relative to HFSEs have relatively low H2O, and must have formed largely by decompression melting of unmodified mantle. Our data for volcanoes <50 km apart show evidence of significant variability in the composition of H2O-rich subduction components that are added to the mantle wedge beneath central Mexico.
Footnotes
-
↵*Corresponding author. Present address: Department of Geological Sciences, University of Oregon, Eugene, Oregon 97403, USA; pwallacedarkwing.uoregon.edu
-
↵GSA Data Repository item 2003021, Table DR1, trace element abundances, is available from Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301-9140, editinggeosociety.org, or at http://www.geosociety.org/pubs/ft2003.htm.
-
Note: Major elements (wt%) by electron microprobe. H2O (wt%) and CO2 (ppm) by Fourier Transform Infrared (FTIR) spectroscopy using band assignments and compositionally dependent absorption coefficients from Dixon et al. (1995) and Dixon and Pan (1995) as described in Cervantes (1999). CO2 (1515 and 1430 cm–1) and molecular H2O (1630 cm–1) absorbances were measured after subtraction of a reference spectrum for decarbonated basaltic glass. 1σ uncertainties in parentheses are mostly due to uncertainties in thickness of sectioned inclusions. Accuracy of the absorption coefficients is about ±20%. Trace elements (in ppm) were measured by Cameca IMS-3f ion microprobe at Wood's Hole Oceanographic Institution. Analytical procedures and uncertainties are described in Shimizu (1998). All data have been modified from analyzed values to account for post-entrapment crystallization (%PEC) of the olivine host (Sobolev and Chaussidon, 1996). Mg# is calculated using Fe3+/Fe2+ ratios from whole rock samples (Wallace and Carmichael, 1999). No post-entrapment diffusive loss of Fe from the inclusions has occurred (Danyushevsky et al., 2000).
-
- Accepted November 14, 2002.
- Received June 17, 2002.
- Revision received November 8, 2002.
- Geological Society of America
