 |
| Titel |
How important is mode-coupling in global surface wave tomography? |
| VerfasserIn |
Dylan Mikesell, Guust Nolet, Sergey Voronin, Jeroen Ritsema, Hendrik-Jan van Heijst |
| Konferenz |
EGU General Assembly 2016
|
| Medientyp |
Artikel
|
| Sprache |
en
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 18 (2016) |
| Datensatznummer |
250124672
|
| Publikation (Nr.) |
 EGU/EGU2016-4148.pdf |
|
|
|
|
|
| Zusammenfassung |
| To investigate the influence of mode coupling for fundamental mode Rayleigh waves with
periods between 64 and 174s, we analysed 3,505,902 phase measurements obtained
along minor arc trajectories as well as 2,163,474 phases along major arcs. This is a
selection of five frequency bands from the data set of Van Heijst and Woodhouse,
extended with more recent earthquakes, that served to define upper mantle S velocity in
model S40RTS. Since accurate estimation of the misfits (as represented by χ2)
is essential, we used the method of Voronin et al. (GJI 199:276, 2014) to obtain
objective estimates of the standard errors in this data set. We adapted Voronin’s
method slightly to avoid that systematic errors along clusters of raypaths can be
accommodated by source corrections. This was done by simultaneously analysing
multiple clusters of raypaths originating from the same group of earthquakes but
traveling in different directions. For the minor arc data, phase errors at the one
sigma level range from 0.26 rad at a period of 174s to 0.89 rad at 64s. For the major
arcs, these errors are roughly twice as high (0.40 and 2.09 rad, respectively). In
the subsequent inversion we removed any outliers that could not be fitted at the 3
sigma level in an almost undamped inversion. Using these error estimates and the
theory of finite-frequency tomography to include the effects of scattering, we solved
for models with χ2 = N (the number of data) both including and excluding the
effect of mode coupling between Love and Rayleigh waves. We shall present some
dramatic differences between the two models, notably near ocean-continent boundaries
(e.g. California) where mode conversions are likely to be largest. But a sharpening
of other features, such as cratons and high-velocity blobs in the oceanic domain,
is also observed when mode coupling is taken into account. An investigation of
the influence of coupling on azimuthal anisotropy is still under way at the time of
writing of this abstract, but the results of this will be included in the presentation. |
|
|
|
|
|
|