next up previous contents
Home: Berkeley Seismological Laboratory
Next: Three-dimensional waveform calculations Up: Ongoing Research Projects Previous: On the origin of   Contents

Lateral Heterogeneity in D'': Constraints From PKP and PcP Travel Time Data

Hrvoje Tkalcic and Barbara Romanowicz

Introduction and Motivation

In the last two decades, there have been large number of studies on the inner and outer core using travel times of core sensitive phases like PKP. The observation that PKIKP (or PKP(DF)) travels faster if propagating more parallel to Earth's rotation axis (polar paths) than if forming larger angles with it (equatorial paths), along with the observation that the core sensitive normal modes are anomalously split, led to the hypothesis that the inner core is anisotropic (Morelli et al., 1986). Ten years later, another PKP travel time observation, that the inner core might be super-rotating with respect to the rest of the planet, shook the broadest scientific community (Song and Richards, 1996).

On the other hand, while very important in studying Earth's core, PKP waves also pass through the mantle - PKP(AB) phase, in particular, being very sensitive to the structure in the mantle's lowermost layer, historically called D''. Thus, on their way through Earth's interior, PKP waves can be significantly affected by the influence of various length-scale heterogeneities in the mantle. Using forward modeling approach, we previously addressed this issue, successfully explaining a lot of the variance in PKP(AB-DF) travel time residuals by increasing the amplitude in lowermost layers of existing tomographic mantle models (Bréger et al., 2000).

However, the spatial distribution of anomalies of our recently collected PcP-P dataset when plotted at the bouncing PcP points, as well as some other studies, indicate that global mantle tomographic models don't have enough resolution to explain trends and fine variations in various travel time data. In particular, it is very difficult to explain the most anomalous PKP(BC-DF) data, which, having spatially much closer sampling near core-mantle boundary than PKP(AB-DF), require more refined models. We cannot exclude the possibility that inner core anisotropy, if it exists, affects PKP(DF), but our recent analysis of short period data recorded at Alaskan stations, indicate that there exist similar trends in absolute travel times in both PKP(DF) and PKP(BC), so at least one part of the signal must come from the mantle (Romanowicz et al., 2001).

Thus, in order to obtain the model of the lowermost layer that would be compatible with high quality measured data, and that would allow us to correct core sensitive travel times prior to drawing conclusions about inner and outer core, we invert for P velocity structure in single layered D'' (Tkalcic et al., 2001). The input is the PKP(AB-DF) dataset, combined with PcP-P dataset to help resolve source-receiver side ambiguity in the inversion. The parametrization is rather refined ($5^o$x$5^o$) in order to address above mentioned short-scale spatial variations. PKP(BC-DF) dataset is used not directly in the inversion but only as a test to put constraints on damping.

At the same time, our high quality datasets of PKP(AB-DF) and PcP-P, give us an opportunity to test the behavior of the ratio $R=dlnVs/dlnVp$ in the lower mantle, a parameter of a great geodynamical importance. It has been argued recently that a direct comparison of P and S tomographic models in order to determine $R$ might be an ill-posed problem, because of the incompatibility in spatial coverage and sensitivity of data used to derive them. To avoid this problem, we scale several recent S tomographic models to obtain virtual P models, investigating the fit in PcP-P and PKP(AB-DF) data, simultaneously finding the best $R$.

Results

Our resulting model, derived from complete PKP(AB-DF) and PcP-P datasets is illustrated in Figure 1a. It is characterized by intermediate scale features and some prominent changes from fast to slow anomalies, for instance between South Sandwich Islands region and south America, as well as central and north America. Other prominent features are a slow anomaly under Africa and a fast anomaly under east Asia. This model remains stable even after exclusion of paths from south Atlantic to north America prior to inversion, which indicates that we successfully used PcP-P data to introduce more crossing paths and reduce source-receiver uncertainty. With this model, we can fit 90% of PKP(AB-DF) data almost 60% of PcP-P data and about 27% of PKP(BC-DF) data

Furthermore, we corrected data prior to inversion for spherically symmetric radially varying inner core anisotropy model by Tromp (1995). The model we obtain this way is shown in Figure 1b. Although weaker in amplitude, this model is consistent with one for which we don't take inner core anisotropy into account. This helps improve the variance reduction in the BC-DF data from 27% to 55%. However, the variance reduction in PKP(AB-DF) is slightly degraded, and this model is also unable to explain more than 2-3 sec PKP(BC-DF) travel time residuals (out of 5-6 sec) on the anomalous south Atlantic to Alaska bundle of paths. To explain these, short-scale, strong lateral heterogeneity is required in D", or, alternatively, as was proposed previously (e.g. Creager, 1997), complex heterogeneity in the inner core.

We found that the best fitting $R$ in the lower mantle varies spatially. In some regions (for example, under Central America), it can be established as a constant of proportionality indicating thermal origin of anomaly, and it is in excess of 2.5. In other regions (for example, under Africa), there is a localized transition between fast and slow PcP, indicating chemical origin to heterogeneity.

Figure 31.1: Models (a) TRH_KC and (b) TRH_KCa, for the optimal damping. These one layered models represent P velocity perturbations with respect to model ak135 (Kennett et al., 1995) in the bottom 300 km of the mantle. TRH_KC was obtained by inverting PKP(AB-DF) and PcP-P data simultaneously, without inner core anisotropy. For TRH_KCa, the same dataset was inverted, but first corrected for the transverse isotropy model of Tromp (1995). Blackened areas represent non-sampled regions.
\begin{figure}\begin{center}
\epsfig{file=hrvoje01.eps, width=8cm}\end{center}\end{figure}

Acknowledgements

Special thanks to Nicolas Houy, who collected and measured the PcP-P travel times.

References

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.

Kennett, B. L. N., Engdahl, E. R. and Buland, R., 1995. Constrains on seismic velocities in the earth from traveltimes, Geophys. J. Int., 122, 108-124.

Morelli, A., Dziewonski, A. M. and Woodhouse, J. H., 1986. Anisotropy of the core inferred from PKIKP travel times, Geophys. Res. Lett., 13, 1,545-1,548.

Romanowicz, B., Tkalcic, H. and Bréger, L., On the Origin of Complexity in PKP travel time data from broadband records, submitted to AGU Volume on Inner Core and Lower Mantle, AGU Geodynamic Series, V. Dehant, Editor, 2001.

Song, X. and Richards, P. G., Seismological evidence for differential rotation of the Earth's inner core, Nature, 382, 221-224, 1996.

Tkalcic, H., B. Romanowicz and N. Houy, Constraints on D" structure using PKP(AB-DF), PKP(BC-DF) and PcP-P travel time data from broadband records, Geophys. J. Int., in revision, 2001.


next up previous contents
Next: Three-dimensional waveform calculations Up: Ongoing Research Projects Previous: On the origin of   Contents



Berkeley Seismological Laboratory, 202 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail: www@seismo.berkeley.edu
© 2001, The Regents of the University of California.