How do finitefrequency waves sense mantle anisotropy

1
How do finite-frequency waves sense mantle
anisotropy ?
Anne Sieminski Qinya Liu, Hanneke Paulssen,
Jeannot Trampert, Jeroen Tromp
Utrecht University and Caltech Seismological
Laboratory
Utrecht, June 20th 2008
2
Sensitivity kernel

• Definition
• Constructed with adjoint spectral-element method

time (s)
Tape et al. (2007)
3
Anisotropy

• Asymptotic propagation is controlled by
• A C F L N
• Jc Js Kc Ks Mc Ms
• Bc Bs Hc Hs Gc Gs
• Dc Ds
• Ec Es
• Each parameter is ? aIJ CIJ
• Montagner Nataf (1986) Chen Tromp (2007)

dependence with i and ?
radial
East
South
4
Anisotropy

• Asymptotic propagation is controlled by
• A C F L N
• Jc Js Kc Ks Mc Ms
• Bc Bs Hc Hs Gc Gs
• Dc Ds
• Ec Es
• Each parameter is ? aIJ CIJ
• Montagner Nataf (1986) Chen Tromp (2007)

dependence with i and ?
radial
East
azimuthal anisotropy
South
5
Overview

• Rayleigh-wave fundamental mode (100s lt T lt 180s
  ? 120º)

L
Gc
A
Bc
150 km
150 km
0 km
0 km
N
Ec
Dc
etc (16 "non-zero" kernels)
150 km
0 km
0 km
s GPa-1 km-3
6
Directional dependence

• Rayleigh-wave fundamental mode (100s lt T lt 180s
  ? 120º)

Gc (cos 2?)
dependence with the azimuth ?
s GPa-1 km-3
R
?
S
South
at 150 km depth
7
Overview

• Love-wave fundamental mode (100s lt T lt 180s ?
  120º)

L
Gc
A
Bc
150 km
150 km
0 km
0 km
N
Ec
Dc
etc (3 "non-zero" kernels)
150 km
0 km
0 km
s GPa-1 km-3
8
Overview

• P wave ( T 15 s ? 75º)

Jc
Kc
Dc
L
A
C
etc (18 "non-zero" kernels)
s3 km-5
9
Directional dependence

• P wave ( T 15 s ? 75º)

Jc (cos 1?)
dependence with the azimuth ? and the incidence i
radial
i
R
S
s3 km-5
along the source-receiver plane
10
Anisotropy vs isotropy

• P wave ( T 15 s ? 75º)

Jc (cos 1?)
A ?2 (isotropic P-wave speed)
s3 km-5
along the source-receiver plane
11
Overview

• SKS splitting (T 8 s ? 105º)

Gs
Ks
Ds
Ms
s3 km-5
SKS-path geometry
mantle
core
12
Overview

• SKS splitting (T 8 s ? 105º)

Gs
Ks
Ds
Ms
s3 km-5
L
N
etc (8 "non-zero" kernels)
13
Conclusion

• The sensitivity to anisotropy cannot be
  neglected ? We need anisotropy for
  mantle imaging
• The sensitivity to anisotropy is complicated
  ? We need 3D full-wave
  propagation
• The "asymptotic" description of anisotropy is
  not optimal ? How to reduce the number of
  significant parameters ? ? By working with the
  "principal parameters" ?

14
Composite arrivals

• SS wave ( T 18 s ? 90º)

? (isotropic S-wave speed)
x 1e7 s km-3
Liu Tromp (2008)
along the source-receiver plane
15
In 3D models

• Rayleigh-wave fundamental mode (40s lt T lt 100s
  ? 32º)

L
L
50 km
50 km
1D
model 3D model
s GPa-1 km-3
16
Conclusion

• The sensitivity to anisotropy cannot be
  neglected ? We need anisotropy for
  mantle imaging
• The sensitivity to anisotropy is complicated
  ? We need 3D full-wave
  propagation
• The "asymptotic" description of anisotropy is
  not optimal ? How to reduce the number of
  significant parameters ? ? By working with the
  "principal parameters" ?

17
Principal parameters

• Rayleigh-wave fundamental mode (40s lt T lt 100s
  ? 50º)

100 km
s GPa-1 km-3
18
(No Transcript)
19
Directional dependence

• Rayleigh-wave fundamental mode (100s lt T lt 180s
  ? 120º)

Gc (cos 2?)
In Born scattering theory K cos (?1 ?2)
Ray
Gc
Rayleigh, 0
Rayleigh, 0
?2
Gc
?1
South
at 150 km depth