Mantle-derived trace element variability in olivines and their melt inclusions

• Published in: Earth and Planetary Science Letters
• Main Authors: Neave, David A., Shorttle, Oliver, Oeser, Martin, Weyer, Stefan, Kobayashi, Katsura
• Format: Article in Journal/Newspaper
• Language: English
• Published: Elsevier 2018
• Subjects: 05 - Petrology - Igneous; Metamorphic and Volcanic Studies; Reykjanes; Stapafell; Iceland
• Online Access: http://eprints.esc.cam.ac.uk/4112/; http://eprints.esc.cam.ac.uk/4112/1/1-s2.0-S0012821X17307136-main.pdf; https://doi.org/10.1016/j.epsl.2017.12.014

Trace element variability in oceanic basalts is commonly used to constrain the physics of mantle melting and the chemistry of Earth's deep interior. However, the geochemical properties of mantle melts are often overprinted by mixing and crystallisation processes during ascent and storage. Studying primitive melt inclusions offers one solution to this problem, but the fidelity of the melt-inclusion archive to bulk magma chemistry has been repeatedly questioned. To provide a novel check of the melt inclusion record, we present new major and trace element analyses from olivine macrocrysts in the products of two geographically proximal, yet compositionally distinct, primitive eruptions from the Reykjanes Peninsula of Iceland. By combining these macrocryst analyses with new and published melt inclusion analyses we demonstrate that olivines have similar patterns of incompatible trace element (ITE) variability to the inclusions they host, capturing chemical systematics on intra- and inter-eruption scales. ITE variability (element concentrations, ratios, variances and variance ratios) in olivines from the ITE-enriched Stapafell eruption is best accounted for by olivine-dominated fractional crystallisation. In contrast, ITE variability in olivines and inclusions from the ITE-depleted Háleyjabunga eruption cannot be explained by crystallisation alone, and must have originated in the mantle. Compatible trace element (CTE) variability is best described by crystallisation processes in both eruptions. Modest correlations between host and inclusion ITE contents in samples from Háleyjabunga suggest that melt inclusions can be faithful archives of melting and magmatic processes. It also indicates that degrees of ITE enrichment can be estimated from olivines directly when melt inclusion and matrix glass records of geochemical variability are poor or absent. Inter-eruption differences in olivine ITE systematics between Stapafell and Háleyjabunga mirror differences in melt inclusion suites, and confirm that the Stapafell eruption was fed by lower degree melts from greater depths within the melting region than the Háleyjabunga eruption. Although olivine macrocrysts from Stapafell are slightly richer in Ni than those from Háleyjabunga, their overall CTE systematics (e.g., Ni/(Mg/Fe), Fe/Mn and Zn/Fe) are inconsistent with being derived from olivine-free pyroxenites. However, the major element systematics of Icelandic basalts require lithological heterogeneity in their mantle source in the form of Fe-rich and hence fusible domains. We thus conclude that enriched heterogeneities in the Icelandic mantle are composed of modally enriched, yet nonetheless olivine-bearing, lithologies and that olivine CTE contents provide an incomplete record of lithological heterogeneity in the mantle. Modally enriched peridotites may therefore play a more important role in oceanic magma genesis than previously inferred.