The seismic mid-lithosphere discontinuity

Seismic S-wave receiver functions (SRF) are a uniquely powerful tool for imaging velocity discontinuities within the upper mantle. SRF data frequently contain negative phases at depths between similar to 80 and 100 km within the continental lithosphere, indicative of large and sharp velocity drops at these depths. In young, actively tectonic areas with thin lithosphere, this feature is generally interpreted as the lithosphere-asthenosphere boundary. However, in tectonically stable areas it occurs within the continental lithospheric mantle and has been termed the mid-lithosphere discontinuity (MLD). A significant velocity drop at such depths is unexpected and its cause is unknown. In this manuscript, we summarise the current observations and assess the main mechanisms that could produce such a feature. We find that changes in mantle iron content (Mg#) and elastically-accommodated grainboundary sliding are unlikely to result in sufficiently large velocity decreases to produce an observable SRF response, while partial melt will generally only exist at greater depths within stable lithosphere. Radial and azimuthal seismic anisotropy are both capable of producing negative SRF phases. However, azimuthal anisotropy will not produce consistently negative phases independent of back-azimuth. Some geometries of radial anisotropy can produce consistent negative phases but such geometries are not observed universally and are hard to explain tectonically. Low-velocity minerals can cause sharp and large decreases in seismic velocity. Amphibole-rich layers are likely to form at MLD depths in metasomatised regions, making amphibole a possible cause for the MLD. However, some xenolith sections contain no amphibole, suggesting this may not be a universal explanation. A careful assessment of SRFs shows that the continental lithospheric mantle generally contains numerous positive and negative velocity discontinuities and is spatially heterogeneous. Long-period band-pass filtering can combine smaller features and may lead to the appearance of a larger and more coherent velocity decrease at the MLD than actually exists. We propose that many of the assessed mechanisms may be acting at different depths in different locations to produce numerous velocity discontinuities. The large MLD phase is likely to be commonly associated with amphibole but on current evidence there is no universal cause for the MLD. (C) 2014 Elsevier B.V. All rights reserved.