 |
| Titel |
Simultaneous inversion for anisotropic and structural crustal properties by stacking of radial and transverse receiver functions |
| VerfasserIn |
Frederik Link, Georg Rumpker, Ayoub Kaviani, Manvendra Singh |
| Konferenz |
EGU General Assembly 2016
|
| Medientyp |
Artikel
|
| Sprache |
en
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 18 (2016) |
| Datensatznummer |
250132572
|
| Publikation (Nr.) |
 EGU/EGU2016-13091.pdf |
|
|
|
|
|
| Zusammenfassung |
| The well-known H-κ-stacking method of Zhu and Kanamori (2000) has developed into a
standard tool to infer the thickness of the crust, H, and the average P to S-wave velocity ratio,
κ. The stacking approach allows for the largely automated analysis of teleseismic waveforms
recorded in the distance range between 30˚ and 95˚ . Here, we present an extension of the
method to include the inversion for anisotropic crustal properties. For a single anisotropic
crustal layer, this involves the computation of delay times and amplitudes for 20 P-to-S
converted phases and their crustal reverberations, instead of (up to) five phases in the
isotropic case (Kaviani and Rümpker, 2015). The delay times and amplitudes exhibit a
complex dependency on slowness and backazimuth. They can be calculated semi-analytically
from the eigenvalues and eigenvectors of the system matrix, as defined by Woodhouse
(1974). A comparison of the calculated delay times and amplitudes with those obtained by
similar methods (Frederiksen and Bostock, 2000) shows a very good agreement between the
results.
In our approach, the crust exhibits hexagonal anisotropy with a horizontal symmetry axis,
such that the anisotropic properties are defined by two parameters: the orientation of the
symmetry axis w.r.t. North, φ, and the percentage of anisotropy, a. The inversion, thus,
involves a grid search in a 4-dimensional parameter space (H, κ, φ, a) and the stacking of
both radial and transverse receiver functions. Known input parameters are the average P-wave
velocity of the crust, and the slowness vector (as given by the event-receiver configuration
and a global 1D-velocity model). The computations are performed by the new software
package AnStack which is based on MATLAB.
Synthetic test show that the extended anisotropic stacking has advantages compared to the
conventional H-κ stacking as it may allow for inversions at even higher noise levels. We
further test for the effect of the azimuthal distribution of events on the results. It turns out,
that the orientation of the symmetry axis is most sensitive to limitations and gaps in the
azimuthal distribution. The extended stacking method provides an average model of the
anisotropic crust below a station. Therefore, internal (vertical) variations cannot be resolved.
Complex structures, which differ from the assumed single-layer model, will also affect the
results. For example, an inclination of the layer boundary may cause an apparent anisotropic
effect. We will also show examples for the application of the method to recently obtained data
sets. |
|
|
|
|
|
|