next up previous contents
Home: Berkeley Seismological Laboratory
Next: Direct Inversion of Mode Up: Ongoing Research Previous: Asymptotic and non-asymptotic approaches

The effects of theoretical formalism and data selection scheme on mantle models derived from waveform tomography

Charles Mégnin, Barbara Romanowicz

Subsections


Most global tomographic models to date are derived using a combination of surface wave (or normal mode) data and body wave travel time data. The travel time approach limits the number of phases available for inversion by requiring them to be isolated on the seismogram. This may ultimately result in limiting the resolution of 3D structure, at least in some depth ranges in the mantle. In previous work, we successfully derived a degree 12 whole mantle SH velocity tomographic model (SAW12D) (Li and Romanowicz, 1996) using exclusively waveform data. In this inversion, a normal mode formalism suitable for body waveforms, the nonlinear asymptotic coupling theory (NACT) (Li and Romanwoicz, 1995), was combined with a bodywave windowing scheme, referred to as the 'individual wavepacket' technique, which allows to assign individual weights to different body-wave energy packets. We here compare the relative merits of these two inversion ingredients, at different depth ranges in the mantle. Choosing as reference a model obtained using 7,500 transverse component body wave and 8,000 surface wave seismograms and the NACT and IW approaches, we discuss the relative performance of the NACT versus the path average approximation (PAVA), a zeroth order theoretical approximation appropriate for single mode surface waves, and the IW windowing scheme (IW) (figure 19.1, bottom) with a more standard 'full window' (FW) (figure 19.1, top) approach, in which a single time window is considered from the first body wave arrival to the fundamental mode surface waves. The combination PAVA/FW is often used in global tomography to supplement the travel time data. The comparison is performed by inverting the data using the four possible combinations of theoretical formalism and data selection scheme (table 19.1).


 
Table 19.1: Nomenclature of the models derived.
Data selection scheme / Theory NACT PAVA
Individual wavepacket ( IW) 1 2
Full waveform ( FW) 3 4
 

In figure 19.2, we show the correlation at depth between the NACT/IW model and the other three. The quality of the image derived under PAVA/FW formalism is seen to be adequate in the top half of the upper-mantle, where the resolution is dominated by surface waves, but deteriorates at greater depths. In the lower mantle the tomographic image is strongly sensitive to the theoretical formalism used. In contrast, resolution of structure near the core-mantle boundary depends mostly on the windowing scheme. This is because this resolution is controlled by low amplitude phases such as Sdiff, which are downweighted in the FW scheme.

Figure 19.3 shows the results from each inversion at two lower mantle depths: 1,500 km at the top and 2,800 km at the bottom. The top figure shows substantial disagreement between PAVA and NACT-derived models while the choice of windowing presents little impact. In contrast, while the image obtained in D" using the combination NACT/IW is in good agreement with images obtained by other authors using both waveforms and travel times (e.g. Bréger et al., 1998), it is apparent that, when using FW, uppermost mantle structure is mapped into D". Table 19.2 shows that model 4 is actually better correlated with the surface features (0.85) than it is with the D'' heterogeneity dderived with the IW formalism (0.68). The IW models, in contrast, show little correlation with the uppermost mantle (0.35).


  
Table 19.2: Correlation between uppermost mantle (150 km) and D'' (2,800 km) heterogeneity for each model.
\begin{table}%
\begin{center}
\epsfig{file=figs/bsl98_pepe_fig6.ps, width=8cm}\end{center}\end{table}

This result is confirmed by synthetic tests performed on a composite of the geodynamic model 3SMAC (Nataf and Ricard, 1996) for the upper mantle and PREM in the lower mantle. Figure 19.4 shows the result of the NACT/IW inversion in the lower mantle where low-amplitude scattering is observed. The extent of the scatter is similar in the PAVA/FW model (figure 19.5) at all lower mantle depths, except in the D'' where mapping of uppermost mantle structure is once again observed.

While a combination of travel times and surface wave data is adequate for resolving relatively smooth features in the mantle, our results show that, by potentially increasing the achievable sampling, the waveform approach, cast in appropriate theoretical framework, which allows more accurate mapping of the location of heterogeneity, shows great promise for future high-resolution tomographic modeling of mantle structure.


  
Figure 19.1: Waveform windowing schemes. Top: 'Full window' ( FW) scheme. The entire trace is selected. The data is biased toward the information provided by the high amplitude phases. Bottom: 'Individual wavepacket' ( IW) scheme. Each energy wavepacket is selected individually. The wavepackets are weighted as to normalize their respective contribution.
\begin{figure}%
\begin{center}
\epsfig{file=figs/bsl98_pepe_fig1.ps, width=8cm}\end{center}\end{figure}


  
Figure 19.2: Correlation at depth of each model with the NACT/IW-derived model. The numbers correspond to the nomenclature of table 19.1.
\begin{figure}%
\begin{center}
\epsfig{file=figs/bsl98_pepe_fig2.ps, width=8cm}\end{center}\end{figure}


  
Figure 19.3: top: Heterogeneity at 1,500 km depth derived from each waveform inversion technique. This region is predominantly sampled by large amplitude body waves (S, SS, ...). The quality of the image at that depth is strongly dependent on the accuracy of the waveform modeling and weakly on the type of windowing scheme. bottom: Heterogeneity at 2,800 km depth derived from each waveform inversion technique. The use of a non-selective windowing scheme (bottom graphs) causes the mapping of surface heterogeneity at the CMB.
\begin{figure}%
\begin{center}
\epsfig{file=figs/bsl98_pepe_fig3.ps, width=8cm}\end{center}\end{figure}


  
Figure 19.4: Result of the NACT/IW inversion of a composite model consiting of the model 3SMAC in the upper mantle and PREM in the lower mantle. Lower mantle depths. Small amplitude scattering is observed.
\begin{figure}%
\begin{center}
\epsfig{file=figs/bsl98_pepe_fig4.ps, width=8cm}\end{center}\end{figure}


  
Figure 19.5: Same as 19.4 but with the PAVA/FW scheme. Here again, the surface features are mapped into the D''.
\begin{figure}%
\begin{center}
\epsfig{file=figs/bsl98_pepe_fig5.ps, width=8cm}\end{center}\end{figure}

References

Bréger, L., B. Romanowicz and L. Vinnik, Test of tomographic models of D'' using differential travel time data, Geophys. Res. Lett., Vol. 25, pp. 5-8, 1998.

Li, X.D. and B. Romanowicz, Comparison of global waveform inversions with and without considering cross-branch modal coupling, Geophys. J. Int., Vol. 121, pp. 695-709, 1995.

Li, X.D. and B. Romanowicz, Global mantle shear-velocity model developed using nonlinear asymptotic coupling theory, Geophys. J. R. Astr. Soc., 101, pp. 22,245-22,272, 1996.

Nataf, H.-C. and Y. Ricard, 3SMAC: an a priori tomographic model of the upper mantle based on geopphysical modeling, Phys. Earth and Plan. Int., 95 pp. 101-122, 1996.


next up previous contents
Next: Direct Inversion of Mode Up: Ongoing Research Previous: Asymptotic and non-asymptotic approaches

The Berkeley Seismological Laboratory, 202 McCone Hall, UC Berkeley, Berkeley CA 94720
Questions and comments to www@seismo.berkeley.edu
Copyright 1999, The Regents of the University of California.