Water, partial melting and the origin of the seismic low velocity and high attenuation zone in the upper mantle

Abstract

The common belief that the seismic low velocity and high attenuation zone (the asthenosphere) is caused by the presence of a small amount of melt is not supported by recent mineral physics and seismological observations. A review of recent mineral physics observations suggests that water significantly reduces seismic wave velocities through anelastic relaxation and hence, at a small melt fraction expected in most of the Earth's upper mantle, partial melting will increase seismic wave velocities through the removal of water from minerals such as olivine. Therefore the asthenosphere, in this model, is a layer where no significant partial melting occurs and hence a high water content is retained. We apply this model to calculate seismic wave velocities and attenuation in the upper mantle with a range of water contents. The seismic structures calculated from this model depend on geotherm, the mode of partial melting (batch or fractional melting) and the geometry of upwelling flow (passive flow or dynamic upwelling). The sharp velocity change around 60–80 km (the Gutenberg discontinuity) can be attributed to a sharp change in water content due to partial melting, if the temperature there is relatively high as implied by the plate model and if melting occurs as fractional melting but not by batch melting. However, the significant increase in seismic wave velocity with age in young oceanic upper mantle suggests rapid cooling as predicted by a cooling half-space model. Thus, the present model suggests fast cooling in the early stage but slow cooling in the later stage of evolution of the oceanic upper mantle, the latter being caused presumably by some additional heat in the old oceanic upper mantle. The seismic structures of typical oceanic upper mantle with a fast spreading rate (e.g., the Pacific) is consistent with passive spreading, whereas the greater depth of the G-discontinuity and the weaker seismic anisotropy in back-arc regions (e.g., the Philippine Sea) suggest dynamic upwelling caused presumably by a higher degree of melting due to a larger amount of water.