Cratonic lithospheric mantle: Is anything subducted?

William L. Griffin*, Suzanne Y. O'Reilly

*Corresponding author for this work

    Research output: Contribution to journalArticlepeer-review

    143 Citations (Scopus)

    Abstract

    If the subcontinental lithospheric mantle (SCLM) formed through the repeated underthrusting of oceanic slabs, peridotitic SCLM should resemble oceanic peridotites, and mafic rocks (eclogites, s.l.) should be distributed throughout the SCLM. However, cratonic peridotites (both exposed massifs and xenoliths) differ markedly from oceanic and ophiolitic peridotites in their Fe-Cr-Al relationships and abundances of trace elements (Li and B) diagnostic of subduction. "Typical" cratonic peridotites have experienced extensive metasomatism; modelling of their refractory protoliths indicates high-degree melting at high P, perhaps a uniquely Archean process. Cratonic eclogites are strongly concentrated at the base of the depleted SCLM or at major layer boundaries, and are accompanied by intense melt-related metasomatism in adjacent peridotites. This distribution, and the preservation of exsolution microstructures, suggest an origin by the ponding and cooling of magmas at a compositional/rheological boundary. Compositionally, cratonic eclogites are similar to Phanerozoic garnet pyroxenites that originated as cumulates of high-Al pyroxenes and as reaction zones between melts and peridotite wall rocks. Eu anomalies in peridotitic garnets from the lithospheric mantle are unlikely to reflect plagioclase fractionation, but may be redox-related metasomatic signatures; such anomalies in eclogitic minerals or whole rocks are thus not prima facie evidence of low-P origin. Mg-isotope fractionation in high-T mantle xenoliths indicates that stable-isotope variations (including O and C) in cratonic eclogite suites may not be evidence of ocean-floor processes. Covariations between C and O isotope ratios suggest that high-T redox-related Rayleigh fractionation, and mixing process involving carbonatitic melts, can explain the ranges of δ18O and δ13C in eclogite suites. There is thus little compelling evidence that any rocks in the cratonic SCLM represent unambiguous samples of subducted oceanic plates.

    Original languageEnglish
    Pages (from-to)43-53
    Number of pages11
    JournalEpisodes
    Volume30
    Issue number1
    Publication statusPublished - Mar 2007

    Fingerprint

    Dive into the research topics of 'Cratonic lithospheric mantle: Is anything subducted?'. Together they form a unique fingerprint.

    Cite this