Search

High-velocity keels are present beneath stable continental interiors, but the maximum depth to which they extend, the degree to which they cool the surrounding mantle and the role in mantle flow, remains uncertain. Thermal anomalies throughout the eastern margin of the North American continental keel were investigated using images of seismic discontinuities. It is thought that the cold keel material and sub-keel down-wellings are confined largely to the upper mantle and can impinge on the transition zone only in localized areas.

A differential technique compares the arrival times of waves refracted across and diffracted around the core-mantle boundary to map the large-scale compressional-wave velocity variations within the D' thermal and chemical boundary layer. This map depicts the continent-sized variations at the core-mantle boundary that complements the shear-wave data. The diffracted waves provide an image of the D' layer, and the differential technique negates the effect of ray paths that are outside of D'. The results support a strong coupling between surface plate tectonics and the mantle base.

Seismic anisotropy occurs in the thermal boundary layer known as the D' which lies 200 km above the core of the earth. The cause of seismic anisotropy in the crust may be due to layered rock fabric caused by subducted crustal slabs. Similar activity in the upper mantle and the inner core possibly occurs due to mineralogical anisotropic nature. Properties of D' such as increase in seismic velocity and waves help in explaining anisotropy. Lack of knowledge about melting points and physical behaviors of the lower mantle minerals creates difficulty in selecting a particular model of D'.

Geophysical research using data recorded at US and Canadian regional seismographic networks following the May 1992 Chinese nuclear test at Lop Nor reveals a seismically fast layer 130 kilometers (km) thick and 300 km wide at the boundary of the core and mantle. The resolution needed to observe such a mass was obtained by organizing data from 14 regional networks of short-period vertical-component seismometers. The data suggest that the boundary region between mantle and core cannot be viewed as a single homogeneous layer and may be as structurally complex as the crust.