Following the recent work of Bréger et al., 2000 in which it was shown by forward modelling that the structure in D'' can explain a large portion of anomalous PKP traveltimes, and one by Romanowicz and Bréger, 2000 in which it was shown that a combination of lateral heterogeneity in the deep mantle and outer core structure contributes significantly to splitting of modes, we inverted a PKP(AB-DF) dataset to obtain the P velocity structure in the lowermost 300 km of the mantle. It is of particular interest to accurately estimate the effects of complex structure in the deep mantle and D'' on PKP differential travel times in order to reach reliable conclusions about the physical and chemical properties of the Earth's inner and outer core. It is important to assess how much of the data can be explained by mantle structure alone. In order to achieve a robust coverage of PKP paths in the lowermost mantle, we compiled several independent, hand-picked high quality PKP(AB)-PKP(DF) differential travel time datasets (McSweeney et al., 1997, Creager, 1999, Souriau, personal communication, Wysession, personal communication). By using differential travel times, we assume that effects of upper-mantle and source mislocation are minimized, and that the dataset is mostly sensitive to the deep mantle and core where the paths of the two phases differ the most. If the D'' region is parametrized by equiangular block cells, 5 x 5 degrees, then it is visible from Figure 29.1, that the best sampled regions lay beneath South America, Caribbean Sea, southwestern Atlantic, Asia, southwestern pacific and Australia. Number of hits per block in some areas, e.g. South Sandwich Islands, exceeds one hundred.
 |
The influence of nine P and S tomographic mantle models, on PKP(AB-DF) residuals is tested in order to choose a model which achieves the best variance reduction in our dataset. The P velocity model (Karason and van der Hilst, 2000), is our preferred model for mantle correction of PKP(AB-DF) travel times due to significantly larger variance reduction in the PKP dataset, compared to the other P and S converted to P models. The map of P velocity anomalies in D" shows prominent fast features in northeastern Asia, Arabian Sea, south Atlantic, Caribbean Sea and Alaska, as well as slow features in the southwest Pacific and under southern Africa. Furthermore, we compare models obtained with and without polar paths (paths with angles of less than 35o with respect to Earth's spin axis) and find that their inclusion or exclusion does not significantly affect the D'' model. The model derived without polar paths is shown in Figure 29.2. Our results demonstrate that, with only about 1500 PKP differential travel time data, we are able to retrieve a D'' map of quality comparable to other maps derived using by far larger number of ISC data. Preliminary maps of PcP-P anomalies, plotted at the PcP bouncing points, correlate well with the PKP(AB-DF) maps of D'' under Americas in places where coverage exists for both datasets. The level of P heterogeneity in D'' that we obtain is dependent on damping factor in the inversion procedure and it determines the percentage of PKP(AB-DF) data that can be explained by mantle-only structure. If the level of P heterogeneity in D'' is assumed to be no larger than 
in absolute sense, mantle structure alone can explain more than 
of the variance in the data, including polar paths, although some unexplained large residuals remain, mostly at angles with the rotation axis between 25 and 35o, corresponding to specific anomalous paths (Figure 29.3).
 |
 |
We thank Ken Creager, Tom McSweeney, Annie Souriau and Michael Wysession for generously sharing their data, and also to Roger Hensen, Kent Lindquist, Doug Christensen and personnel of UAF Geophysical Institute, Bill Shannon, Sylvia Lehman and Luc Saumure from CNSN, for helping one of us, Hrvoje Tkalcic, collect data from the Alaska Network. This work was partially funded by a grant from the IGPP/LLNL program and by NSF.
Bréger, L., H. Tkalcic and B. Romanowicz, The effect of D'' on PKP(AB-DF) travel time residuals and possible implications for inner core structure, Earth Planet. Sci. Lett., 175, 133-143, 2000.
Creager, K. C., Large-scale variations in inner core anisotropy, J. Geophys. Res., 104, 23,127-23,139, 1999.
Engdahl, E. R., R. D. van der Hilst and R. P. Buland, Global teleseismic earthquake relocation with improved travel times and procedures for depth determination, Bull. Seismol. Soc. Am., 88, No 3, 722-743, 1998.
Karason, H. and R. van der Hilst, Improving global tomography models of P-wavespeed I: incorporation of differential times for refracted and diffracted core phases (PKP, Pdiff), submitted, 2000.
Kennett, B. L. N., E. R. Engdahl and R. Buland, Constrains on seismic velocities in the Earth from traveltimes, Geophys. J. Int., 122, 108-124, 1995.
McSweeney, T. J., K. C. Creager and R. T. Merrill, Depth extent of inner-core seismic anisotropy and implications for geomagnetism, Phys. Earth Planet. Int., 101, 131-156, 1997.
Romanowicz, B. and L. Bréger, Anomalous splitting of core sensitive modes: a revaluation of possible interpretations, J. Geophys. Res.,in press, 2000.
Tkalcic, H, B. Romanowicz and N. Houy, Modelling of D" by using high quality PKP(AB-DF) datasets, in preparation., 2000.
), therefore without paths that might be affected by inner and outer core. Unsampled regions are black. The robustness of this model should be assessed by comparing it with figure 1 which shows sampling per each particular cell. Note the fast horizontal band across the north Pacific.
are shown by diamonds. Theoretical PKP(AB-DF) traveltimes are with respect to ak135 model and corrected for ellipticity (Kennett et al., 1995), after taking relocated in space and time events from Engdahl et al., 1998 catalog. Triangles correspond to residuals predicted from the mantle model (Karason and van der Hilst, 2000) stripped of D'' in combination with the model of D'' derived in this study.