The Mantle Transition Zone: Seismic Properties, Deep Subduction, Earthquakes, and Petrology Wang-Ping Chen University of Illinois, Urbana-Champaign [Green,

The Mantle Transition Zone: Seismic Properties, Deep Subduction, Earthquakes, and Petrology

Wang-Ping Chen

University of Illinois, Urbana-Champaign

[Green, ’01]

Outline• Why the mixture of topics?

– Forego comprehensive review– Work toward an integration, relating these topics

• Preview of overall theme– Intricate interplays between temperature and

petrology

• Current issues on deep slab-penetration and mantle convection/mixing

• Compare Tonga-Fiji and Izu-Bonin: – Counter-intuitive– Extremely fast subduction of cold slab results in

impounding of buoyant slab in transition zone

Introspection/Re-interpretation

• Low VP in the lower mantle

– High temperature?– High density?

• van der Hilst [’04]• Perspective on Trampert et al. [’04]

• Scale of VP not reported

A picture is worth a thousand words – If the image is valid, clear, and

correctly interpreted

• Deep slab-penetration advanced by travel-time tomography

• Issues and assumptions– High seismic wave speed implies low temperature?

• Ignores petrologic effects (mineralogy, composition, volatiles, kinetics)

– Amplitude of anomalies (and color-scale)• ± 0.5% in lower mantle (except near the base)• ± 2-3% in upper mantle and transition zone

– Watch out for color-saturation

– Inconsistency among images (with similar data and techniques)

2 pix

[Grand et al., 1997]VS

Tonga-Fiji: Examples of Travel-Time Tomography

[Deal et al., 1999]

Some Fundamental Issues

• Mass imbalance– Fast subduction of cold slab but only modest

anomalies in the lower mantle• Tonga subduction zone

– Rate of convergence > 200 mm/yr (from GPS)– Slab older than 100 Ma– 65% of world’s deep seismicity

– Best example of lower mantle anomaly• Caribbean anomaly (remnant of Farallon plate?)• Thermal window for deep slab penetration?

[Grand et al., 1997]

Some Fundamental Issues (Cont.)

• Incomplete mantle mixing– Primordial component of mantle

• Transition zone “water filter” hypothesis [Bercovici and Karato, ‘03]

• Tolstikhin-Hofmann hypothesis

• Origin of deep earthquakes– Double couple

• Sudden slip over faults• Kinematics similar to shallow earthquakes

– Brittle failure requires unrealistic shear stress at depth• Strength is limited by ductile flow at high temperatures

– Potential mechanisms• “Embrittlement” (fluid/fluid-like materials)• Transformational faulting (phase change)

V p3%

Natural Laboratory: Seismicity

Tonga-Fiji Region: Seismicity

• Wadati-Benioff zone • The usual inclined

zone of seismicity– Consistency in fault

plane solutions

• The “toe”– Complexity in fault

plane solutions

• Outboard earthquakes

• A separate zone• Lack of pattern in

fault plane solutions

Experimental Configuration(Simplified)

Seismic Profiles– Triplicate waveforms

• Basis for recognizing major discontinuities• Principles

– Aperture of profiles• Variations in focal depth (50 - 670 km)• Variations in epicentral distance (1200 - 2500 km)

3000

20

1000

50

Waveform Modeling– Match absolute timing, relative timing, amplitudes– Isotropic medium: WKBJ method– Anisotropic medium: ANRAY and reflectivity

Distance (km)

Red

uce

d T

ime

(s)

410-km disc660-km disc

Fore-Arc– Geologic baseline before subduction– Similar to average Earth models (i.e., iasp91)– Laterally homogeneous and isotropic over 1000 km– Anomalies under back-arc must be a consequence of

subduction

Seismic Anisotropy• Definition

– Variations of seismic wave speeds along different directions of propagation or polarization

– Shear-wave birefringence (splitting)

• Anisotropy in the Mantle– Usually concentrates above 200 km– Typically null anisotropy in transition zone– New observations in the TZ: Split-time up to 3 s

Lateral Variations of Anisotropy

Sector I: Strong Birefringence

Focal Depth: ~ 150 km ~ 590 km ~ 660 km

Modeling Birefringence

Sectors II & III: Null Birefringence

Tonga-Fiji Region: Transition Zone Structures

• 3D rendition of seismicity• Radial seismic anisotropy

– Associated only with zone of outboard earthquakes– Petrofabric (VSH > VSV)

Lateral Variations in Wave Speeds

• Compare fore-arc, source zone of outboard earthquakes, and the rest of the back-arc

Waveforms: Sector I

• Similar to fore-arc• NO anomalies associated with outboard earthquakes

Waveforms: Sector II

• High VP and VS in the TZ adjacent to outboard earthquakes

Magnitude of Anomalies

Tonga-Fiji Region: Transition Zone Structures (II)

• Lack of high VP and VS within zone of outboard earthquakes– Seismicity indicates low temperature

• Global patterns of deep seismicity and zones of recent convergence• Low strength of rocks at high temperatures

– Otherwise mantle earthquakes away from subduction zones and sluggish mantle convection

• Limiting temperature for seismicity in the mantle – 600 to 800C [Chen and Molnar, 1983; Molnar et al., 1979; Wiens and Stein, 1983]

Tonga-Fiji Region: Transition Zone Structures (III)

– Effect of low temperature to raise V counteracted by petrologic anomalies (volatiles, meta-stable phases, melts?)

• Contrast in anisotropy• Sharp boundary for VP and VS (3% over 25 km)

• High VP and VS surrounding outboard earthquakes– Gradual decrease in magnitude outward– Cold “aureole”, yet too warm to maintain petrologic anomaly

Region I:

Coldest Core

Region II: Cold AureoleRegion III: Assimilated Slab

Metastable Olivine Hydros Phases Partial Melt

Stability in Transition Zone

Favorable Favorable Induced bydehydration

Trigger for Earthquakes

Transformationalfaulting

Embrittlement Embrittlement

3% reduction in VP 60% of olivine-polymorph in -phase

~2.2 wt % water ~2% melt (~0.04 wt% water)

Corresponding reduction in VSH

Consistent Consistent Too much (~6%)

SH-SV splitting Plausible, fossil fabric

Minor effect by themselves

Oriented magmapockets (SPO)

No pattern in Focal

Mechanisms

Self-stress Inconsistent, need other localized source of stress

Inconsistent, need regional stress for SPO

Buoyancy of slab 2%, but only in TZ 1%, buoyant in the upper mantle

~0% (depending on compressibility)

The Check List

Andean Seduction Zones

• N3: Moderate-dipping slab– Classic down-dip

extension• N2: Sub-horizontal slab

– Incoherent P/T-axis– No regional stress field

Global Study of Seismic Strain• Two populations• Down-dip compression or

extension, if– Slab dip > 20 – Down-dip comp. > sin(20)

of total slab-pull• No clear pattern of strain, if

– Slab dip < 20 – Vanishing down-dip comp.

< 1/3 of total slab-pull

Observations Regional Stress (Slab-Pull) Localized Stress

Numerous seismicity in sub-horizontal slabs

Small to not present, thus not necessary

Only alternative, thus both necessary and sufficient

Great pressure at depth Slab pull too small to be sufficient,~0.5% negative buoyancy

Only alternative,thus necessary and sufficient

Down-dip compression from depths of 100 to 700 km, but gap in seismicity near 300 km

Insufficient Not present?

• Localized, self-stress is both necessary and sufficient for generating deep earthquakes

Tonga-Fiji Region: Synthesis

• Detached remnant of slab juxtaposed over active Wadati-Benioff zone• Buoyancy

– Amount of reduction in VP and VS (2-3%) requires lots of meta-stable olivine (60% of olivine) or volatiles

– Both are buoyant– Self-limiting buoyancy for meta-stable olivine

• Petrologic anomalies are likely trigger of deep earthquakes• Large-scale remnant of slabs alleviate problems in mass imbalance and mantle

mixing

Evolution of Buoyant Slab

• Toe of WBZ as the predecessor of detached slab– The missing link between WBZ and detached slab

• Detached slab–

Anchored by dense leading edge where petrologic anomaly dissipates with rising temperature

• Partially assimilated slab remnant– Aseismic– Petrologic anomaly dissipated– Attenuated thermal anomaly remains– Rests directly above the 660-km discontinuity (e.g., N. Philippine Sea anomaly)

[Modifying Green, ’01]

Stage 1: “Toe”

Stage 2b: “Anchor”

Stage 2a: Buoyant remnant

N. Philippine Sea Anomaly

[Tseng and Chen, ‘04]

[van der Hilst and Seno, ‘93]

• Null Birefringence

• Moderate increase in VP and VS (~1.5%)

• Aseismic• Resting on 660-km discontinuity

Conclusions• Extremely fast subduction of

cold slab results in impounding of buoyant slab in transition zone– Unified interpretation for

observed seismic anisotropy, lateral variations in seismic wave speeds, and outboard earthquakes

– Help resolving paradox in mass imbalance and perplexing patterns of seismicity

– Slab penetration into the lower mantle may require special thermal window

Proposed Experiments in Tonga-Fiji• CAVASCOPE Array• Deep earthquakes

– How deep is the deepest?– Still in the TZ?

• The nature of the 660-km discontinuity– Topography

• Compare with lab. data on pure Mg2SiO4?

– Density contrast • High contrast (Japan Sea):

8.1±0.8%

– Contrast in seismic anisotropy

• Detached slab in the TZ– Toe of Wadati-Benioff Zone –

detached slab in the making?– Complete mapping of cold

aureole surrounding outboard earthquakes

– Connection between two detached slab-remnants? [Tseng and Chen, 2004]

Collaborators• Michael R. Brudzinski (Miami Univ., Ohio)• Tai-Lin (Ellen) Tseng (UIUC)• Robert L. Nowack (Purdue)• Robert Pillet (Institut de Recherche pour le

Développement, New Caledonia)• Bor-Shouh Huang (Institute of Earth Sciences,

Taiwan)• Special thanks to

– Harry Green, Steve Kirby– Jay Bass, Chu-Yung Chen, Jennie Jackson, Jie Li,

Holger Hellwig, Stas Sinogeikin

Sources of materials not explicitly cited are mainly from the following articles:

Brudzinski, M. R., W.-P. Chen, R. L. Nowack, and B.-S. Huang, Variations of P-wave speeds in the mantle transition zone beneath the Northern Philippine Sea, J. Geophys. Res., 102, 11,815–11,827, 1997.

Nowack, R. L., E. Ay, W.-P. Chen, and B.-S. Huang, A seismic profile of the upper mantle along the southwestern edge of the Philippine Sea plate using short-period array data, Geophys. J. Int., 136, 171–179, 1999.

Wu, L.-R., and W.-P. Chen, Anomalous aftershocks of deep earthquakes, Geophys. Res. Lett., 26, 1977–1980, 1999.

Brudzinski, M. R., and W.-P. Chen, Variations of P-wave speeds and outboard earthquakes: Evidence for a petrologic anomaly in the mantle transition zone, J. Geophys. Res., 105, 21,661–21,682, 2000.

Wu, L.-R., and W.-P. Chen, Rupture of the large (MW 7.8), deep earthquake of 1973 beneath the Japan Sea with implications for seismogenesis, Bull. Seismol. Soc. Am., 91, 102–111, 2001.

Chen, W.-P., and M. R. Brudsinzski, Evidence for a large-scale remnant of subducted lithosphere beneath Fiji, Science, 292, 2475-2479, 2001.

Brudzinski, M. R., and W.-P. Chen, A petrologic anomaly accompanying outboard earthquakes beneath Fiji-Tonga: Corresponding evidence from broadband P and S waveforms, J. Geophys. Res., 108(B6), 2299 (19 pp.), doi:10.1029/2002JB002012, 2003.

Chen, W.-P. and Brudzinski, M.R., Seismic anisotropy in the mantle transition zone beneath Fiji-Tonga, Geophys. Res. Lett., 30, 1682 (4 pp.), doi:10.1029/2002GL016330, 2003.

Tseng, T.-L., and W.-P. Chen, Contrasts in seismic waves speeds and density across the 660-km discontinuity beneath the Philippine and the Japan Seas, J. Geophys. Res., 109, (12 pp.), B04302, doi:10.1029/2003JB002613, 2004.

Chen, W.-P., and Z.-H. Yang, Earthquakes beneath the Himalayas and Tibet: Evidence for strong lithospheric mantle, Science, 304, 1949-1952, 2004.

Brudzinski, M.R., and W.-P. Chen, Earthquakes and strain in sub-horizontal slabs, J. Geophys. Res., in press, 2005.