Excursions to the base of the mantle.

Abstract

The core mantle boundary (CMB) region, consisting of the core mantle interface itself and the few hundred kms above and below it, is investigated in two different types of large-scale seismological studies. In the first study, a large set of long-period S phases is analyzed to determine large scale shear velocity structure in the CMB region beneath Alaska. Subdivision of the data by depth, source region, and source to receiver azimuth establishes that the observations exhibit stable behavior, which is consistent with extensive local stratification of D$\sp{\prime\prime}$. The preferred model, SYLO, has a 2.75% velocity increase 243 km above the core mantle boundary and a negative velocity gradient within the D$\sp{\prime\prime}$ layer. Analysis of core-penetrating phases indicates that the outermost 100 km of the core below the fast D$\sp{\prime\prime}$ region underlying Alaska must be slower than the PREM structure. In the second study, a set of short-period P-wave amplitude profiles near the core shadow zone is utilized to explore lateral variations in P velocity structure at the base of the mantle. Various radially symmetric models are tested by comparison of the data with amplitudes measured from synthetics. The observed amplitude versus distance profiles exhibit significant regional variations of the apparent shadow zone boundary, with as much as a 5$\sp\circ$ shift in onset distance, but it is possible to model the overall behavior using simple, regionally varying positive P velocity gradients in the lowermost mantle. The strong lateral variations required by the data support the presence of compositional heterogeneity in D$\sp{\prime\prime}$. The data also indicate that if there is a P velocity discontinuity corresponding to the S discontinuity beneath Alaska, it must be less than 1% in magnitude, indicating a change in Poisson's ratio across the discontinuity. The results from these two studies, as well as those from previous studies of the CMB region (which are briefly reviewed), are used to evaluate several simple schemes for the role of D$\sp{\prime\prime}$ in mantle convection.