On the Heterogeneity of the Lowermost Mantle from PcP-P and ScS-S Differential Travel Times

There has been an ongoing debate about the nature of heterogeneities in the lowermost mantle. Different P and S-wave tomography models can be compared to obtain an ad hoc picture of the nature of heterogeneities (e.g. thermal versus chemical), but because such models are normally derived from different datasets and using different assumptions, such a comparison is a nuisance. Here we use existing data from deep and large magnitude earthquakes to analyze and compare differential travel times of PcP-P and ScS-S waves. This type of analysis is the most direct probe of the nature of heterogeneities in the lowermost mantle because the PcP and ScS waves sample the core-mantle boundary (CMB) region in the same way. The predicted PcP-P and ScS-S travel time differential is subtracted from the observed travel time difference, and that residual can then be plotted onto a map to show where parts of the CMB produce slower or faster PcP or ScS. By then comparing the PcP-P and ScS-S travel time residuals as well as their lateral gradients, we can infer what may be the nature of heterogeneity in a given area of the lowermost mantle. The goal is to expand the data set by Tkalcic and Romanowicz (2002) by using seismic stations and earthquakes in areas not initially used. Because of the ten-year difference between the initial data set, we are able to utilize the new data recorded at the newer seismic stations, in particular in Australia and Antarctica. In this study, over 480 measurements were made. Although more observational work still needs to be done in finding events that produce good PcP-P and ScS-S, we are beginning to fill in the "holes" of the existing spatial coverage. Events in locations such as the Mediterranean Sea add to the velocity model particularly in the northern Atlantic Ocean. In the new data set, the largest negative PcP-P residuals are located along the coast of Antarctica and between Alaska and Russia. These larger negative residuals indicate significantly faster paths through the lowermost mantle. When compared to some whole-mantle P-wave velocity models, the acquired data appears to coincide with the models in many areas. The surface projections of the bounce points in east Asia and North America show fast patches at the CMB which fit previous models. After analyzing PcP-P, we moved onto ScS-S to compare the former residuals with the latter to presume the nature of the heterogeneities. The measurements vary in comparison to those of PcP-P, as some are very close to their PcP-P counterparts, others are very different in magnitude, and even opposite in sign. Because of the locations of the chosen events, a high percentage of the data relates to the CMB beneath Asia and the Middle East. Further analysis is necessary in order to fill in the bigger "holes" in the spatial coverage of the lowermost mantle. Large areas, mainly in the oceans, have yet to be examined. The difficulty in doing so comes from the lack of seismometer stations in these areas and the noisy data often retrieved from stations on islands. A possible solution to this problem would be the installation of borehole ocean bottom seismometers at key locations.