Squigglylineland view of the Earth

1
Squiggly-line-land view of the Earth

• Whats going on in the upper mantle?
• Receiver function, powerful seismic tool
• What in the world does the structure of the inner
  core mean?
• Is it still rotating, like it was in 1996?

2
Outline of mantle discussion

• USArray
• Receiver function analysis
• MOMA
• Africa
• RISTRA
• The upper mantle discontinuities
• Water at 410-km-depth
• A double 520

(I just got a digital camera)
3
EarthScope Components

• EarthScope's facilities include the following
  four coupled components
• USArray (United States Seismic Array)
• SAFOD (San Andreas Fault Observatory at Depth)
• PBO (Plate Boundary Observatory)
• InSAR (Interferometric Synthetic Aperture Radar)

4
USArray Permanent Array
5
Big Foot Array
6
Flexible Arraysexample from recent experiments
7
Why look at the upper mantle?

• Mapping seismic structure
• P S velocity, density, anisotropy
• To deduce physical characteristics
• Chemical and thermal heterogeneity
• To deduce whats going on
• Stagnant or moving continental keels
• Dynamics of upper thermal boundary layer of the
  mantle
• Mantle circulation

8
Seismic-style study

• Reflection for crustal structure
• S-wave splitting for anisotropy
• Flow direction - aesthenosphere
• Relic fabric - lithosphere
• Surface and body wave tomography
• Absolute velocities in upper few 100 km
• Body wave tomography (deeper)
• Receiver functions
• Best resolution of radial velocity gradients

9
The receiver function

• Pioneered by seismologists including Bob Phinney
  and Chuck Langston
• Examines echoes of the P wave to determine zones
  of high radial gradient in seismic velocity
• It is proving to be a very useful companion to
  seismic tomography, providing detailed pictures
  of near-receiver structure

10
Chuck Langston, after igniting 50 pounds of
explosives in sand
11
Chuck Ammons notes
12
40-80 distance range best
13
Ray paths contributing to receiver functions
Chuck Ammon
14
Radial component of receiver function
Just useful for finding the Moho
15
Lateral variations
Adam
Alan
16
Mechanics of a receiver function

• Extract the P wave from the vertical component
• Deconvolve it from the horizontal component
• This should leave a spike at the P arrival time
  and a string of P-S conversions
• Convert the conversions (as a fcn of time and
  ground motion) to structure (impedance as a
  function of depth)
• Average together the records from many distances
  and azimuths

17
Some limitations

• Assumes no lateral variations in structure
• Migration can overcome this limitation
• Only works in a frequency pass band
• Cannot recover baseline, trends, or really much
  beyond about 100-200 km wavelength velocity
  structure
• Generally falls apart shorter than 5-10 km
  wavelengths

18
MOMA

• Missouri to Massachusetts transect
• 19 stations placed every 100 km
• Chosen for nice graphics

Mike Wysession
Keith Koper
19
(No Transcript)
20
MOMAdiscontinuity imaging
Mike Wysession
Karen Fischer
21
Stereo vision
Receiver functions from events to the north
East!?
West
Events to the south
22
Tomography plus receiver functions
?T lt 150 C
Disagreement with individual profiles
Farallon depression?
23
Steve Gao
24
Shows trend of smaller time separation with more
vertical incidence
Gao, GRL, 2002
25
Again, well-resolved reflections from near 410
and 660
Note the presence of clear 410 conversions at
short-period
26
Thicker transition zone to NE
Transition thickness near global average of 245
km, so not cold under region, 10 km of relief may
correspond to 60 temperature difference
cooler
warmer
27
Receiver function migration

• Just like migrating seismic reflection data
• Benefits from adequate spatial sampling
• Ability to image structure depends on
• Depth of structure
• Frequency of waves recorded
• Of course, more events with more back-azimuths,
  and more distances are helpful

28
Resolution with 70 km spacing
T 15s
29
Resolution with 10 km spacing
T 2s
30
Alan
31
A test model
32
Recovery of the test model
33
MOMA Array Depth Migration LP10s
MOMA migration
34
Cheyenne Belt Receiver Functions
Imbricated Moho
Mantle layered
Archean Mantle
Modified Proterozoic Mantle
Fast from tomography
From Ken Dueker
35
RISTRA Rio Grande Rift Ran from Texas into Utah
Rick Aster
Receiver functions across the 1000-km line give a
good picture of the shallow structure, and show
little topography on the 410 and 660.
moho
36
Flat discontinuities
37
Hot off the JGR press
Ken Dueker

• Hersh, Dueker, Sheehan, and Molnar, JGR
• 410 and 660 topography under western US
• 20-30 km topography, with 500 km scale length
• No relation to surface tectonics
• Sharpness not easily related to depth
• Conclusions
• Either transition zone has smaller scale
  convection than deep mantle
• Or there is a lot of compositional variation down
  there

38
Anne Sheehan
Field area
39
Average receiver function structure
Seymour Hersh
410
660
40
410 topography
/- 10 km
660 topography
/- 15 km
41
(No Transcript)
42
Science 6 June 2003
Seismic evidence for water deep in the Earths
upper mantle
Federica Marone
Mark van der Meijde
Domenico
Suzanne van der Lee
43
Science 6 June 2003 - van der Meijde et al.
1000 ppm water broadening the 410-km-discontinuity
?
44
Main points of van der Meijde

• Conversion from 410 stronger at low frequency
  than high, but conversion from 660 is steady
• So 410 must be broader, in fact very broad,
  20-40 km wide
• Subduction has been pervasive, so water might be
  common near 410-km-depth
• Entire story is consistent if about 1000 ppm
  water is present.

45
1 s period
6 s period
46
9 stations The general trend is consistent, and
statistics can be constructed to support the
significance of the trend.
47
X
48
The phase PP
Jim Whitomb
DLA
49
JGR, Fei, Vidale and Earle

• Rounded 3 good datasets of PP
• California networks
• LASA recordings
• Highly selected GSN seismograms
• Well see
• Sharp 660-km-depth discontinuity
• Somewhat less sharp 410, sometimes
• (but MUCH sharper 410 than claimed for Europe)
• No 520

50
Several minute envelope stack
51
The 660 and 410 corrected for steady noise
52
A global average
53
More 660 than 410 energy, Nothing else
Fei Xu
54
Comparison to long-period reflections
Corrected for attenuation
55
No visible 410 at higher frequencies
56
This means

• 660 sharp enough to efficiently reflect 1 Hz
  waves - less than 2 km thick transition
• 410 not so sharp - our data is fit by half a
  sharp jump, half spread over 7 km

57
SS precursors as a probe of layering near their
bounce point
Peter Shearer
58
Science, 2001. Sees 520 sometime simple,
sometimes split.
Arwen Deuss
Interprets this as the 520 having phase changes
in two components, olivine and garnet, whose
depths dont always coincide.
(Also has claims to see PKJKP and a 250)
59
Transects that indicate lateral continuity of
structure
60
Transects of the 520
Lateral continuity of structure
61
A global map, where there is coverage
John Woodhouse
62
Some high points

• 410
• Why is its brightness variable?
• Can we map the pattern globally to learn more?
• Is topography real?
• 520
• Why does it flicker?
• 660
• Is topography a thermometer?
• Other discontinuities?
• Better images on the way from USArray

63
The enigmatic inner core

• Layering
• Anisotropy
• Rotation
• Possible origins of structure
• Combined my slides with those of Ken Creager and
  Shun Karato

Some slides lent by Ken Creager and Shun Karato
64
Seismic characteristics of the inner core

• A large Poissons ratio, close to that of a
  liquid
• High attenuation (Qs100-200)
• Strong anisotropy

65
A current working model
Upper Inner Core Isotropic, finely
heterogeneous West 0.8 slower 250 km thick Q
600 East thicker Q 250 in east Middle Inner
Core Strong anisotropy Isotropic Voigt average
is homogeneous Innermost Core Different
anisotropy?
Isotropic Upper Inner Core
Transition Region
IMIC
Anisotropic Lower inner Core
66
Niu and Wen, 2001
Red - western hemisphere Black - eastern
hemisphere
67

• Comparing polar and equatorial data

Ouzounis and Creager, GRL, 2001
68
Beghein and Trampert Science, 2003
Adam and Miaki Ishii
69
Summed slant stack
(Vidale Earle)
70
Proposed mechanisms of inner core anisotropy
Convective flow due to high Rayleigh number
aligns crystals (most effective near surface)
Jeanloz Wenk, GRL, 1988
71
Inhomogeneous growth of inner core drives
convective flow that restores isostatic
equilibrium
Yoshida et al., JGR, 1996
72
Dendritic growth of crystals aligns a-axes
radially with heat flow direction (assumes c-axis
is fast)
Michael Bergman, Science, 1997 (modified by
Michael Wysession)
73
Strong heterogeneities, various crystal alignment
orientations
Modified from Annie Souriau, Science, 1998
74
Rotationally wrapped magnetic field around inner
core causes Maxwell stresses that align crystals
(c-axes cylindrically radially out)
Bruce Buffett, Nature, 2001
75
Lorentz forces produce axisymmetric, sustained
flow that aligns crystals
Modified from Shun-Ichiro Karato, Nature, 1999
76
Hemispherical asymmetry
Sumita and Olson (1999)
Hemispherical asymmetry might be due to
heterogeneous thermal boundary conditions at the
inner-core boundary caused by core-mantle
interaction. Time-scale for anisotropic
structure formation must be comparable to or
shorter than the time scale for changes in mantle
structure.
77
Does the inner core rotate with respect to the
mantle?
Song and Richards, 1996 yes 1.1 deg/yr
Creager, 1997, yes 0.2-0.3 deg/yr
Vidale et al., 2000, yes 0.15 deg/yr
Song, 2002, yes 0.5-1.0 deg/yr
Laske and Masters , 2002, maybe 0.130.11 deg/yr
Souriau, 2001, no, at least not very fast, lt0.1 -
0.2 deg/yr
78
Why do we care?

• I think its interesting
• Would mean the core has either
• Quite low viscosity
• Can deform fast enough to keep moving
• Quite low viscosity
• Deforms so little that there is little viscous
  drag
• Would prevent association of IC structure with
  mantle structure

79
25 years of data
Xiao-Dong Song and Paul Richards
80
Li and Richards, submittedSouth Sandwich Islands
Doublet
81
Song, AGU Monograph, 2002
More Sandwich doublets
82
Laske and MastersNormal mode analysisAGU
Monograph, 2002
83
Geometry
84
PKKP comparison
85
PKiKP waveform correlation
86
P660P correlation
87
Bottom lineInner core maylap Earth every2000
years
Wild card - Does the outer core change over time?
88
Quick Review

• Mantle discontinuities still remain interesting
  after 40 years
• Inner core is being mapped but not yet understood
• Inner core is likely turning slowly
• Seismology and mineral physics must progress
  together