Volatile Systematics of the Icelandic Mantle from Olivine-Hosted Melt Inclusions

Abstract:

The behaviour of volatiles within a volcanic system can tell us about melt storage depths, melt evolution, and degassing processes. It is therefore important to be able to quantify the amount of each volatile species entering the system. Olivine-hosted melt inclusions can provide compositions of the primitive mantle melt entering volcanic systems. In some cases they have been unaffected by processes such as melt mixing, fractional crystallisation and degassing; so chemical variations are source signatures. With the knowledge that volatiles behave as very incompatible elements, and with measured volatile/trace-element ratios, we can estimate volatile concentrations within the Icelandic mantle source.

Abundances of CO₂, H₂O, Cl, F, S and trace elements have been measured by SIMS for over 100 olivine-hosted (Fo_87.8-88.6) melt inclusions from Kistufell, a monogenic subglacial eruption known to have elevated ³He/⁴He (15.5 R/Ra)^[1] and situated above the inferred locus of the Icelandic mantle plume. H₂O concentrations in the melt inclusions are near uniform and similar to that of the carrier glass (0.10-0.15 wt%), likely due to diffusive equilibration through the olivine host at low pressure. CO₂ shows more variation, ranging from 1200 to 0ppm, probably from a combination of melt heterogeneity and degassing. The majority of melt inclusions are incompatible trace element-depleted (La/Yb = 1.3), indicating that they were probably generated by high fraction melting of a depleted source in the shallow mantle. The CO₂/Nb ratio of this source mantle is ~308, very similar to that of Borgarhraun, ~314^[2], which has a lower ³He/⁴He ratio (12.2 R/Ra)^[3]. This CO₂/Nb ratio, along with H₂O/Ce ~200 and F/Nd ~18, are consistent with other measured melt inclusion suites from Iceland, suggesting limited volatile/trace-element ratio variation within the Icelandic depleted melt source. By analysing more eruptions across Iceland, we hope to identify any spatial variations in volatile/trace-element ratios within the Icelandic mantle.

[1] K. Breddam (2002) J. Petrology 43, 2, 345-373; [2] E. H. Hauri (2002) Goldschmidt Abstract; [3] C. Macpherson et al (2005) EPSL 233,411-427.