Subduction Zone Geodynamics: Walking the Maze of Coupling and Decoupling Kelin Wang Pacific Geoscience Centre, Geological Survey of Canada.

Subduction Zone Geodynamics:

Walking the Maze of Coupling and Decoupling

Kelin Wang

Pacific Geoscience Centre, Geological Survey of Canada

Subduction Zone Geodynamics:

Walking the Maze of Coupling and Decoupling

Coupling or decoupling

• Velocity continuous or discontinuous (long-term)

• Seismic or aseismic

• Locked or creeping (short-term)

• Strong or weak interface

~ 10

0 km

• Low seismic attenuation• Low Vp/Vs• Serpentinization• Stagnant

• High attenuation• High Vp/Vs• Melting• Vigorous wedge flow

Cold Forearc Hot Arc, Back Arc

70 ~

80 km

Blue: Basaltic crust

Purple:Serpentine stability

Basalt to eclogite ~ 40-50 km depthFeeble arc volcanismSerpentinized mantle wedge cornerIntraslab earthquakes to ~90 km depth

Basalt to eclogite ~ 100-140 kmActive arc volcanismHigh-velocity wedge cornerEarthquakes to hundreds of km

N Cascadia NE Japan

Kirby et al., 1996; van Keken et al., 2002; Wada and Wang, 2009; Syracuse et al., 2010

End-member warm-slab and cold-slab subduction zones

Survival depth of basaltic crust (blue diamonds)

anddepth range of intraslab

earthquakes (purple)

Model-predicted peak dehydration depth (blue)

andserpentine stability in

subducting slab (purple)

Wada and Wang, 2009

Warm Cold

Questions:

• What controls the abrupt transition from decoupling to coupling?• What is the role of petrology, fluids, and rheology?

?

~ 10

0 km

70 ~

80 km

Seismogenic zone(stick-slip,

velocity-weakening, “seismically coupled”)

For coseismic deformation (a few minutes), this is all elastic.

Blue: Basaltic crust

Purple:Serpentine stability

N Cascadia NE Japan

• Temperature plays a role but perhaps not via a single critical value• Continental Moho seems to be a limit, but there are counter examples:

Many events in NE Japan, 2004 Sumatra (Klingelhoefer et al., 2010)

End-member warm-slab and cold-slab subduction zones

Seismogenic zone

?

Questions:

• What determines the downdip limit of seismic rupture?• What is the frictional behavior of the slab – mantle wedge interface?

Locked zone

• Interseismic deformation changes with time and therefore is not a mirror image of coseismic deformation.

• Locked zone (future rupture zone) cannot be determined by inverting interseismic geodetic data using an elastic model.

coastline coastline

Interseismic Coseismic

Locked zone

For interseismic deformation (decades to centuries), this is viscoelastic.

Rupture

Stress relaxation

Stress relaxation

Afterslip

Locking

Three primary processes after an earthquake: afterslip, viscoelastic stress relaxation, and fault locking.

Sumatra: A few years after a great earthquake:

All sites move seaward

Courtesy Kelly Grijalva and Roland Burgmann

Alaska and Chile: ~ 40 years after a great earthquake:Opposing motion of coastal and inland sites

M = 9.2 M = 9.2 19641964

Freymueller et al. (2009)

M = 9.5 M = 9.5 19601960

Wang et al. (2007)

Cascadia: ~ 300 years after a great earthquake:

Wells and Simpson (2001)

All sites move landward

Inter-seismic 2 (Cascadia)

Inter-seismic 1(Alaska, Chile)

Co-seismic

Coast line

Coast line

Post-seismic (Sumatra)

Questions:• What do interseismic deformation observations tell us about

fault friction, rock rheology, and state of locking? • What can we learn by observing subduction zones presently at

different stages of the earthquake cycle?

Locked zone

?

ETS

Seismogenic zone

GPS displacements and slip distribution on subduction interface determined by inverting the GPS data.

Northern Cascadia ETS event of May 2008

Tremor located by Kao (white) and Wech (gray)

Comparison with a worst-case scenario of megathrust rupture

(Ichinose et al., 2003)

(Baba and Cummins, 2005)

(Kikuchi and Yamanaka, 2001)

(Sagiya and Thatcher, 1999)

Including afterslip

19461944

1944

1944

19461944

Non-volcanic tremor

Survival depth of basaltic oceanic crust (blue)

anddepth range of intraslab

earthquakes (purple)

Model-predicted peak dehydration depth (blue)

andserpentine stability in

subducting slab (purple)

Wada and Wang, 2009

Nan

kai

Mex

ico

Ala

ska

Cos

ta R

ica

Warm Cold

Blue: Basaltic crust

Purple:Serpentine stability

N Cascadia NE Japan

ETS at mantle wedge corner

No ETS has been reported

End-member warm-slab and cold-slab subduction zones

In addition:• Other types of slow slip events: long- and short-duration slow slip

without tremor, very-low-frequency earthquakes in ETS zone … …• Vp/Vs anomaly associated with ETS (fluid?)• Mike Brudzinski will provide other details this afternoon

ETS

Seismogenic zone

Questions:• What is the relation between the earthquake cycle, ETS, and

other slow slip phenomena? • What are the thermally controlled petrologic and hydrologic

conditions of ETS?

Seismogenic zone

b 0.04

b -0.01

Average stress

~ 15 MPa

Stress dropa few MPa

coastlineInterseismic

coastlineCoseismic

b > 0Stress

increasea few MPa

Co-seismic

Post-seismic

Stress increase; resisting slip

Rupture;Stress drop

Stress decrease

Locked;Stress increase

ClassicalCoulomb Wedge

DynamicCoulomb Wedge

Updip zone Seismogenic zone

Understanding how the prism is made … …

Nankai

Moore et al., 2007

(based on von Huene et al. (2004)

Costa Rica

Ranero et al., 2007

Relation between seismogenic zone and prism structure … …

Prism stress from NanTroSeize boreholes

Byrne et al., 2009

Inter-seismic 2 (Cascadia)

Inter-seismic 1(Alaska, Chile)

Co-seismic

Coast line

Post-seismic (Sumatra)

?

?

?

?

How the leading edge behaves in earthquake cycles … …

Hsu et al. (2006)

2005 Nias-Simeulue earthquake:1-yr postseismic slip (color)

Updip segment off Peru:Not slipping. Fully relaxed?

Gagnon et al. (2005)

Very-low-frequency earthquakes possibly in Nankai accretionary prism

CORK fluid pressure transients associated with Nankai VLF

(Ito and Obara, 2006)

(Davis et al., 2006)

Fluid transients have also been observed at prism toe, Costa Rica, using flowmeters and also interpreted to indicate transient fault slip (Brown et al., 2005; Labonte et al., 2009).

Seismogenic zone

Questions:• What stops coseismic rupture at accretionary and erosional margins? • How does the updip segment move during the interseismic period? • How do stress and fluid in the wedge evolve throughout earthquake

cycles at accretionary and erosional margins?• What can we learn by observing subduction zones presently at

different stages of the earthquake cycle?

?

Seismogenic zone

Bilek (2007)

smoothrough

very rough

Two aspects of the megathrust: 1) Fault zone material and its frictional behaviour2) Fault zone morphology and its scale variability

Frictional contact Uneven fault zone

Slip can break velocity-strengthening barrier, allowing large displacement in earthquakes. – localized shear

Large displacement requires modification of fault geometry, involving complex deformation of the fault-zone volume. – distributed cataclastic shear

Bilek (2007)

smoothrough

very rough

Two aspects of the megathrust: 1) Fault zone material and its frictional behaviour2) Fault zone morphology and its scale variability

Smoothly coupled fault;Rate-state friction;Giant earthquakes possible

Roughly coupled fault;Friction and complex deformation;Earthquakes and creep

Very roughly coupled fault;Complex fault zone deformation;Creep and small earthquakes

Built on Ruff (1985)

Seismogenic zone• Thermal state

• Evolution throughout earthquake cycles

• Comparison between subduction zones