Which Mantle Interfaces Do Seismologists See? Peter Shearer* IGPP/SIO/U.C. San Diego * With figures...
-
Upload
mavis-garrison -
Category
Documents
-
view
216 -
download
0
Transcript of Which Mantle Interfaces Do Seismologists See? Peter Shearer* IGPP/SIO/U.C. San Diego * With figures...
Which Mantle Interfaces Do Seismologists See?
Peter Shearer*
IGPP/SIO/U.C. San Diego
* With figures from Castle, Deuss, Dueker, Fei, Flanagan, Gao, Gu, Kaneshima, Kato, Kellogg, Kosarev, Kruger, Lebedev, Li, Masters, Niu, Owens, Reif, Tackley, Tseng, van der Hilst, Vidale, Vinnik
Interface Depth vs. Publication Date
Most depths are sampled at least once
Consistency in depths greatest for 220, 410, 520, 660
Note: plot is not complete, especially in last 15 years
Three Types of Mantle Interfaces
Unobserved seismically, hypothesized by geochemists or geodynamicists (e.g., Kellogg et al. lower mantle boundary)
Routinely observed seismically, known mineral physics origin (410, 520, 660)
Intermittently observed seismically, undetermined origin (220, 900, 1200, D”)
SS precursors probe layering near the SS bounce point
Nice for global studies since SS bounce points are widely distributed
figure from Lebedev et al. (2002)
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
410 and 660 observations are consistent with mineral physics predictions for olivine phase changes
• Absolute depths agree with expected pressures
• Topography consistent with Clapeyron slopes
• Size of velocity and density jumps are about right
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
Flanagan & Shearer (1998) Lebedev et al. (2002)
Global, SS precursors Australia region, Receiver functions
• Correlation between TZ thickness and velocity anomalies• Agrees with mineral physics data for olivine phase changes• Permits calibration of dT/dv and Clapeyron slopes
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
Analysis of different discontinuity phases can resolve density, P & S velocity jumps across discontinuities
A puzzle: Where is the 660 reflector?
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
Shearer & Flanagan (1999)
SS & PP precursors
Kato & Kawakatsu (2001)
ScS reverberations
Tseng & Chen (2004)
Triplicated waveforms
Estimated S velocity and density jumps across 660 km
Global Study Northwest Pacific Philippine Sea
Dueker & Sheehan (1997)
Snake River Plane
Eastern US, MOMA Array
Li et al. (1998)
Tibet
Tanzania
Kosarev et al. (1999)
Owens et al. (2000)
Southern Africa
Gao et al. (2002)
Earthquake
Station P'P'df P'P'ab
Mantle
OuterCore
InnerCore
Figure 1
P’P’ phase: seen at short periods, good for sharpness constraints
0
0.2
0.4
0.6
0.8
1
-200 -150 -100 -50 0 50 100
Envelope stack:1/19/69 earthquake at LASA
Time relative to P'P'(ab) (sec)
P'P' onset
P'660P' P'410P'
Several minute envelope stack
0
0.01
0.02
0.03
0.04
0.05
-200 -150 -100 -50
Precursors to P'P'
Time relative to P'P' (sec)
P'660P'P'410P'
XXlong-period
reflectionamplitudes
Comparison to long-period reflections
Corrected for attenuation
0.00
0.02
0.04
0.06
0.08
0.10
2200 2240 2280 2320
LASA stacks at two frequencies
0.7 Hz stack1.0 Hz stack1.3 Hz stack
Time Figure 11
"660"
"410"
No visible 410 in P’P’ at higher frequencies
from Fei, Vidale & Earle
Conclusions from Fei, Vidale & Earle P’P’ study
410 is not so sharp — results suggest half is sharp jump, half is spread over 7 km
520 is not seen in short-period reflections — jump must occur over 20 km or more
660 is sharp enough to efficiently reflect 1 Hz P-waves — less than 2-km thick transition
S520S
Seen in global stacks but weaker than 410 and 660 reflectors
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
Deuss & Woodhouse (2001)
520-km discontinuity may be intermittent and/or split into two interfaces
SS precursors stacked in bounce point caps
Deuss & Woodhouse propose this may be phase changes in two components, olivine and garnet, whose depths don’t always coincide.
Transectsof the 520
Lateral continuity of
structure
A global map, where there is
coverage
JohnWoodhouse
Intermittently observed seismically, undetermined origin (220, 900, 1200, D”)
Three Types of Mantle Interfaces
Are there regional reflectors hidden in here, such as the 220?
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
figure from Deuss & Woodhouse (2002)
SS Precursor Stacks in Different Regions
figure from Deuss & Woodhouse (2002)
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
figure from Deuss & Woodhouse (2002)
Discontinuity depths in SS precursor bounce point caps
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
figures from Gu, Dziewonski & Ekstrom (2001)
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
But another SS precursor study does not find N. American 220
Are there regional reflectors hidden down here, such as at 900 km or 1200 km?
Strongest evidence for mid-mantle discontinuities is from S-to-P conversions from deep earthquakes
to receiver
Slide 44
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
figures from Niu & Kawakatsu (1997)
• Interface near 1000 km below Indonesia
• Seen in S-to-P conversions
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
figure from Kaneshima & Helffrich (1999)
• Interface near 1500 km below Marianas
• Seen in S-to-P conversions below deep earthquakes
Unobserved seismically, hypothesized by geochemists or geodynamicists (e.g., Kellogg et al. lower mantle boundary)
Three Types of Mantle Interfaces
Hypothesized Undulating Mid-Mantle Discontinuity
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
figures from Kellogg, Hager, van der Hilst (1999)
Low density contrast (chemical+thermal) implies large dynamic topography
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
figure from van der Hilst & Karason (1999)
Some tomography models may have change in character near 1700 km depth
But this is not direct evidence for an interface,and models have limited resolution and uniqueness….
Tackley (2002) test of Kellogg et al. hypothesized layer
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
Interface position Temperature perturbation
• 3-D numerical simulation of mantle convection
• Filtered to match seismic modeling
• Simulations predict peak in heterogeneity near interface depth
• Not seen in real tomography models – argues against hypothesis
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
Masters et al. model 10% in dense layer 30% in dense layer
figures from Tackley (2002)
Radial correlation function is measure of vertical continuity of model
Convection simulations with dense lower layer predict significant de-correlation (narrowing) near interface, which is not observed in Masters or Grand tomography models
Convection SimulationsSeismic Inversion
Thermal boundary layer at interface will cause sharp velocity change, which should cause observable triplication in seismic travel time curve*
* Provided source/receiver geometry is just right
figure from Vidale, Schubert & Earle (2001)
figures from Vidale, Schubert & Earle (2001)
• California network data, 10 earthquakes• No triplications or complicated waveforms• But approach could miss:
- interface not at 1770 km, or - dipping interface
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
figure from Castle & van der Hilst (2003)
S-to-P conversions can detect discontinuities below earthquake source regions
S1700P phase arrives between P and pP
Useful for finding interfaces below subduction zones from ~800 to ~2000 km depth
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
figure from Castle & van der Hilst (2003)
1500 - 1800 km
Systematic search for interfaces below Pacific subduction zones
Nothing
(see also Wicks & Weber, 1996; Kaneshima & Helffrich, 1999)
Nothing
Nothing
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
figure adapted from Vinnik, Kato & Kawakatsu (2001)
1850 km
1200 km
900 km
950 & 1050 km
1200 km
Yet Vinnik et al. study finds interfaces in some of same places
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
figure from Vinnik, Kato & Kawakatsu (2001)
Quake in Fiji-Tonga, recorded in Japan
S1200Ps410P S1700P
Quake in Marianas, recorded in western US
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
figure from Castle & van der Hilst (2003)
What about the lowermost mantle?
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
figures from Castle & van der Hilst (2003)QuickTime™ and a
TIFF (LZW) decompressorare needed to see this picture.
• Mid-mantle reflectors seen below many subduction zones
• Related to ancient slabs?
• Discontinuity observations lacking in other areas
• Not present? or Not easy to observe?
• Conclusion: Seismic evidence argues against continuous mid-mantle boundary