Generation of arc and back-arc magmas

Together with John Eiler, Katherine Kelley (Carnegie Institution of Washington), Sally Newman (Caltech), and Terry Plank (Boston University), I have been involved in several studies related to magma genesis at arcs and back-arc spreading centers. I initially got into this through some measurements we made of water and carbon dioxide contents of submarine glasses from the Mariana trough in the western Pacific (see figure). The water contents showed some unusual correlations with other aspects of the chemical compositions of the glasses; in particular, water contents were positively correlated with the concentrations of some incompatible elements (e.g., Na, K, U, Th) but negatively correlated with others (e.g., Ti, Y). We developed a model to explain these contrasting trends that required that the degree of melting of mantle sources is correlated with the H2O content of the source. These results also indicate that the integrated degree of melting of mantle sources increases from typical mid-ocean ridge basalts to the parental liquids of island arc lavas. Finally, these results were used to estimate the compositions of mobile fluids in the mantle wedge overlying the Mariana subduction zone and to speculate on the nature of interactions between fluids generated by dehydration of the subducting slab and the overlying mantle.

Our conclusion that the amount of melting was correlated with the water content of the source was not unexpected (i.e., it is well-known that water will, like other incompatible elements, flux melting in the mantle), but our modelling presented the first quantitative evaluation of the relationship between the amount of water and the amount of melting in mantle peridotite (see figure). As shown in the figure, we concluded that this relationship is roughly linear, a result in agreement with MELTS calculations (the green curves). Note based on the MELTS calculations that the effect of water on melt production is strongly temperature dependent, and this suggested that quantification of this relationship for natural suites could provide a crude measure of mantle temperature in back-arc regions.


CO2 and H2O contents of glasses from the Mariana trough. The calculated isobars show the water and carbon dioxide contents of liquids coexisting with H2O-CO2 vapor at the indicated pressures.

Comparison of % melting vs. water content relationship for peridotite based on mixing/melting model for Mariana Trough basalts and direct thermodynamic modeling using "MELTS".

Following up on these studies of volatiles related to magmagenesis in the Marianas, we have been involved in a collaborative study with Terry Plank and Katherine Kelley extending this work to a larger number of arc and back-arc systems (e.g., the Mariana Trough, Sumisu Rift, Manus Basin, Lau Basin, East Scotia Ridge, and North Fiji Basin). Many of the samples included in this study are glassy melt inclusions in olivine phenocrysts, ensuring that the glasses preserve their volatile contents on eruption (see figure below).

These photomicrographs show analyzed inclusions, indicated by the white arrows, in two olivine crystals from (a) Pagan and (b) Guguan islands (from K.A. Kelley, Ph.D. dissertation).

These studies have demonstrated that the relationship between water content of the source and degree of melting deduced for the Mariana trough is a general feature of back-arc basins (see figure), but that the slope of the water content vs. degree of melting relationship varies from one back-arc basin to another. However, as anticipated based on our earlier analysis, this slope is relatively simply related to mantle temperature (see figure).

In another study of arc volcanism, I have worked with John Eiler on oxygen-isotope ratios of phenocrysts in basalts and basaltic andesites from the Central American arc. As shown in the figure shown below, the oxygen-isotope ratios vary systematically with location, from a minimum δ18O olivine value of 4.6 (below the range typical of terrestrial basalts) in Nicaragua near the center of the arc, to a maximum δ18O olivine value of 5.7 (above the typical range) in Guatemala near the northwestern end of the arc. Moreover, these oxygen isotope variations correlate with major- and trace-element abundances and with Sr and Nd isotope compositions of host lavas, defining trends that suggest variations in δ18O reflect slab contributions to the mantle sources of these lavas. These trends can be explained by a model in which both a low-δ18O, water-rich component and a high-δ18O, water-poor component are extracted from the subducting Cocos slab and flux melting in the overlying mantle wedge. The first of these components dominates slab fluxes beneath the center of the arc and is the principal control on the extent of melting of the mantle wedge, which is highest in the center of the arc; the second component dominates slab fluxes beneath the northwestern margin of the arc; fluxes of both components are small or negligible beneath the southeastern margin of the arc. A key feature of this model is that it incorporates the relationship between water content of the source and degree of melting described above based on our studies of arc and back-arc basin lavas. We suggested that the low-δ18O component is a solute-rich aqueous fluid produced by dehydration of hydrothermally altered rocks deep within the Cocos slab (perhaps serpentinites produced in deep normal faults off-shore of Nicaragua), and that the high-δ18O component is a partial melt of subducted sediment on top of the plate.