Hawaii, Boundary Layers and Ambient Mantle—Geophysical Constraints

Recent high-resolution seismic observations and geodynamic calculations suggest that mid-plate swells and volcanoes are plausibly controlled by processes and materials entirely in the upper boundary layer (<220 km depth) of the mantle rather than by deep-seated thermal instabilities. The upper boundary layer (BL) of the mantle is fertile enough, hot enough and variable enough to provide the observed range of temperatures and compositions of mid-plate magmas, plus it is conveniently located to easily supply these. Seismic data show that the outer ∼220 km of the mantle is heterogeneous, anisotropic and has a substantially superadiabatic vertical temperature gradient. This is the shear, and thermal, BL of the upper mantle. It is usually referred to as the ‘asthenosphere’ and erroneously thought of as simply part of the well-mixed ‘convecting mantle’. Because it supports both a shear and a thermal gradient, the lower portions are hot and move slowly with respect to the surface and can be levitated and exposed by normal plate tectonic processes, even if not buoyant. The nature of BL anisotropy is consistent with a shear-induced laminated structure with aligned melt-rich lenses. The two polarizations of shear waves travel at different velocities, VSV and VSH, and they vary differently with depth. VSH is mainly sensitive to the temperature gradient and indicates a high thermal gradient to 220 km depth. VSV is mainly sensitive to melt content. The depth of the minimum isotropic shear velocity, Vs, under young plates occurs near 60 km and this rapidly increases to 150 km under older oceanic plates, including Hawaii; 150 km may represent the depth of isostatic compensation for swells and the source of tholeiiic basalt magmas. The high-velocity seismic lid thickens as the square-root of age across the entire Pacific, but the underlying mantle is not isothermal; average sub-ridge mantle is colder, by various measures, than mid-plate mantle. Ambient mantle potential temperature at depth under the central Pacific may be ∼200°C higher, without deep mantle plume input, than near spreading ridges. This is consistent with bathymetry and seismic velocities and the temperature range of non-ridge magmas. Some of the thinnest and, in terms of traditional interpretations, hottest transition zones (TZ; ∼410–650 km depth) are under hotspot-free areas of western North America, Greenland, Europe, Russia, Brazil and India. The lowest seismic velocity regions in the upper mantle BL are under young oceanic plates, back-arc basins and hotspot-free areas of California and the Pacific and Indian oceans. Cold slabs may displace hotter material out of the TZ but geophysical data, and geodynamic simulations, do not require deeper sources. Magmas extracted from deep in a thick conduction layer are expected to be hotter than shallower oceanic ridge magmas and more variable in temperature. Mid-plate magmas appear to represent normal ambient mantle at depths of ∼150 km, rather than very localized very deep upwellings. Shear-driven upwellings from the base of the BL explain mid-plate magmatism and its association with fracture zones and anomalous anisotropy, and the persistence of some volcanic chains and the short duration of others. The hotter deeper part of the surface BL is moving at a fraction of the plate velocity and is sampled only where sheared or displaced upwards by tectonic structures and processes that upset the usual stable laminar flow. If mid-plate volcanoes are sourced in the lower half of the BL, between 100 and 220 km depth, or below, then they will appear to define a relatively fixed reference system and the associated temperatures will increase with depth of magma extraction. Lithospheric architecture and stress control the locations of volcanoes, not localized thermal anomalies or deep mantle plumes.