Wysession, M. E., C. R. Bina, and E. A. Okal, Constraints on the temperature and composition of the base of the mantle, in Dynamics of the Earth's Deep Interior and Earth Rotation, Geophysical Monograph Series, edited by J.-L. LeMouël et al., American Geophysical Union, Washington, D.C., pp. 181-190, doi:10.1029/gm072p0181, 1993.
Abstract. P and S waves diffracted around the core-mantle boundary (CMB) are examined to obtain measurements of long-wavelength average velocities in D'', the base of the mantle. Observations are made of profiles of diffracted waves (Sd and Pd) from WWSSN and Canadian stations, and are compared to synthetic seismograms generated with the reflectivity method. The apparent ray parameters of the data and synthetic profiles serve as the basis of comparisons, which suggest significant lateral heterogeneity on the order of about 4% for both P and S velocities at the base of the mantle. While most of the D'' average velocity anomalies are on the order of ~1% relative to PREM, our range in seismic heterogeneity is largely the result of a 3% S and P velocity low in D'' beneath Indonesia, which is made even more unusual by the fact that it is adjacent to a regional fast velocity anomaly beneath Southeast Asia. This velocity low is over a major rising plume in the outer core, as calculated by Voorhies , and if this plume has been held in place over time through core-mantle coupling then the low velocity would be expected due to the increased mantle influx of heat and iron. We undertake a calculation of the variations in D'' seismic velocities due to changes in temperature and composition using a third-order Birch-Murnaghan equation of state with current available thermoelastic data on perovskite and magnesiowüstite. Using this model, small seismic velocity anomalies in D'' could be the result of temperature variations, though small fluctuations in relative amounts of magnesium and iron would have a greater effect on the velocities. For example, the Indonesian anomaly cannot be explained by only a thermal anomaly, but requires only a 20% increase in the Fe/(Mg+Fe) ratio (and even less if accompanied by a raise in temperature). In some regions of D'' the P and S velocities do not vary in tandem, as under Northern North America, where shear velocities are fast but P velocities are slightly slow. The implied lateral change in Poisson ratio could be the result of variations in the relative amounts of silicates and oxides, exacerbated by the high thermal gradients that are expected to exist in D''.Copyright © 1993 American Geophysical Union