High Resolution Anisotropic Structure of the North American
Upper Mantle from Inversion of Body and Surface Waveform Data

Federica Marone, Yuancheng Gung and Barbara Romanowicz


Seismic anisotropy is required for a correct interpretation of the retrieved S-velocity structure in tomographic studies at least in the first 400 km of the upper mantle (Gung et al., 2003). A detailed knowledge of the seismic anisotropic structure of the earth's mantle also provides insight into debated geophysical issues, such as the nature and strength of lithosphere/asthenosphere coupling, the depth extent of continental sub-regions and the relation of imaged seismic anisotropy to present-day asthenospheric flow and/or past tectonic events recorded in the lithosphere.

To date, our knowledge of the North American anisotropic structure arises mainly from global tomographic models (e.g. Ritsema et al., 1999; Gung et al., 2003) or SKS splitting studies (e.g. Fouch et al., 2000; Savage and Sheehan, 2000), which lack horizontal and vertical resolution respectively, and are limited to either radial or azimuthal anisotropy.

Our goal is a new high resolution model for the North American upper mantle incorporating both radial and azimuthal anisotropy. We aim at unprecedented lateral and depth resolution by improving both data coverage and methodology.


Figure 29.1: Data coverage - Paths of teleseismic events with $M_{w} \ge 6$ recorded at North American stations (indicated by white triangles), for which high quality fundamental mode Rayleigh wavepackets have been selected.
\epsfig{file=federica03_1_1.eps, width=8cm}\end{center}\end{figure}

We have collected and processed 3 component body, fundamental and higher mode surface waveforms to complement the BSL database and improve the data coverage for North America. In particular, we focused our attention to broad band seismograms recorded at the numerous permanent seismic stations deployed throughout North America (Figure 29.1), from events at teleseismic and far regional distances. From each deconvolved and filtered seismogram, individual body and surface wave energy packets have been extracted using an automated selection algorithm and subsequently checked by hand, to ensure a high quality dataset. The compiled data collection consists of more than 100,000 3 component body, fundamental and higher mode surface waveforms and provides a fairly homogeneous path (Figure 29.1) and azimuthal coverage.

We plan to use independent information from SKS splitting measurements as additional constraints on the anisotropic model.

Methodology Improvements

We invert seismic long period waveform data in the framework of normal mode asymptotic theory (NACT - Li and Romanowicz, 1996). The resulting broad band sensitivity kernels allow us to exploit the information contained in long period seismograms for body, fundamental and higher mode surface waves simultaneously.

Until now, this approach has only been applied at the global scale with lateral parametrization in terms of spherical harmonics (Li and Romanowicz, 1996; Mégnin and Romanowicz, 2000). We have adapted the procedure to the regional case by implementing a lateral parametrization in terms of spherical splines on an inhomogeneous triangular grid of knots (e.g. Wang and Dahlen, 1995), with the finest mesh for the region of interest, where the data coverage is densest, and a coarser grid outside the studied region (Figure 29.2).

Figure 29.2: Example of irregular grid with a finer mesh for North America
\epsfig{file=federica03_1_2.eps, width=8cm}\end{center}\end{figure}

Body and surface wave datasets used in mantle seismic tomography are sensitive to crustal structure, but cannot resolve details within the crust. Accurate crustal corrections are therefore essential for the quality of high resolution regional tomographic studies. The effect of shallow-layer features is often removed from the data by assuming an a priori crustal model (e.g. CRUST2.0) and applying linear perturbation corrections. However, lateral variations in Moho depth can be fairly large even over short distances, as for instance at ocean/continent transitions and the adequacy of linear corrections is questionable. In fact, Montagner and Jobert (1988) showed that the non-linearity of shallow-layer corrections is often non negligible even at long periods. In high resolution upper mantle regional tomographic studies, it is therefore important to take the crustal structure into account in a more accurate way. Going beyond the linear perturbation approximation, we follow the approach proposed by Montagner and Jobert (1988) and split the correction into a linear and non-linear part. At each point along a path, we assign a 1D reference model according to the local crustal structure (e.g. extended crust, orogen, ocean, ...). We then correct for the difference between the discontinuities in the chosen a priori crustal model (e.g. CRUST2.0) and the selected 1D local reference model assuming a linear perturbation, and exactly for the difference, if any, between the local reference model and PREM (our global reference model).

Future of the Project

Preliminary inversions of the compiled dataset using the improved NACT algorithm resulted in a radial anisotropic upper mantle structure reproducing the major features shared by all recent models (e.g. Ritsema et al., 1999; Gung et al., 2003), such as high velocities beneath cratons between 100 and 250 km depth and negative velocity anomalies in back arc regions. While refining this preliminary model, we will work on the next step and implement a more complete anisotropic parametrization. Our final goal is a model incorporating both radial and azimuthal anisotropy. Such a model can be parametrized in terms of radial anisotropy with a symmetry axis of arbitrary orientation, corresponding to the 5 Love parameters plus two angles defining the axis orientation. The backbone permanent network component of USArray, complemented by temporary Big Foot deployments, will provide an unprecedented density of recordings. This unique dataset will guarantee the resolution of this increased number of parameters.


This work has been financially supported by the Swiss foundation ``Stefano Franscini''.

We are grateful to the IRIS Data Management Center (DMC) as well as to the Geological Survey of Canada for providing waveform data used in this study.


Fouch, M.J., K.M. Fischer, E.M. Parmentier, M.E. Wysession and T.J. Clarke, Shear-wave splitting, continental keels, and patterns of mantle flow, J. Geophys. Res., 105, 6255-6275, 2000.

Gung, Y., M. Panning and B. Romanowicz, Global anisotropy and the thickness of continents, Nature, 422, 707-711, 2003.

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

Mégnin, C. and B. Romanowicz, The 3D shear velocity structure of the mantle from the inversion of body, surface and higher mode waveforms, Geophys. J. Int., 143, 709-728, 2000.

Montagner, J.-P. and N. Jobert, Vectorial tomography; II. Application to the Indian Ocean, Geophys. J., 94, 309-344, 1988.

Ritsema, J., H. van Heijst and J. Woodhouse, Complex shear wave velocity structure imaged beneath Africa and Iceland, Science, 286, 1925-1928, 1999.

Savage, M.K. and A.F. Sheehan, Seismic anisotropy and mantle flow from the Great Basin to the Great Plains, western United States, J. Geophys. Res., 105, 13,715-13,734, 2000.

Wang, Z. and F. Dahlen, Spherical-spline parametrization of three-dimensional Earth models, Geophys. Res. Lett., 22, 3099-3102, 1995.

Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail: www@seismo.berkeley.edu
© 2004, The Regents of the University of California