Regional body wave and surface wave studies indicate that there is a low‐Q upper mantle layer underlying a high‐Q lithosphere. Great circle surface wave attentuation is used to refine the Q structure of the upper mantle and to demonstrate that these features are consistent with the global data. Body wave results are used to constrain the average Q of various regions of the mantle and core and the Q gradient in the lower mantle. Normal mode data are used to test the hypotheses that bulk dissipation is not required in the mantle and that the inner core has low Q. Both hypotheses are consistent with the data. The data are also consistent with a smooth increase of Q with depth over most of the lower mantle and a low‐Q zone at the base of the mantle. The radial modes require bulk dissipation somewhere in the earth, probably in the inner core. A series of parametric models is presented which illustrate the sensitivity of the attentuation data to major features of the Q distribution.

No abstract

Several recent inversion studies have clearly indicated the lack of resolving power of the normal mode data set and the possible trade-offs among the various parameters. These studies have also shown that the final model is as dependent on the starting model as on the data set. It is therefore important to incorporate body wave data into any inversion scheme not only to gain resolution but also to reduce trade-offs between density and velocity. An earth model based on special studies of the structure of the mantle and core is inverted to be consistent with both body wave data and a representative set of normal mode observations (

Sn is a short‐period shear wave that propagates in the lithosphere. For long paths, the wave may travel well below the M discontinuity as it seeks a least‐time path. It cannot penetrate the partially molten asthenosphere. More than 130 Sn phases from earthquakes in the North Atlantic, recorded by stations on islands and surrounding coasts, have been analyzed by regression to obtain regionalized velocities corresponding to different age zones of the Atlantic lithosphere. The highest velocity, 4.71 ± 0.01 km/sec, was found for sea floor older than 50 m.y. The average velocity for younger sea floor, age 0 to 50 m.y., was 4.58 ± 0.02 km/sec. For continental paths, Sn velocity was 4.61 ± 0.02 km/sec. These velocities are interpreted in terms of a lithosphere evolving into a colder, denser mineral assemblage as part of the spreading process. Rocks with elastic velocity higher than that of peridotite must be present in the suboceanic lithosphere to account for the high value of 4.71 ± 0.01 km/sec found in this study.

Attenuation of seismic waves indicates that the earth is not perfectly elastic. Dispersion accompanying absorption gives frequency-dependent "elastic" moduli, a fact that must be taken into account when inverting seismic data. Normal mode data are reinverted after correcting for absorption. The correction removes the discrepancy between body wave and free oscillation interpretations of earth structure.