Adv Advanced anced Worksho shop p on n Ea Earthquake Fa Fault - - PowerPoint PPT Presentation

adv advanced anced worksho shop p on n ea earthquake fa
SMART_READER_LITE
LIVE PREVIEW

Adv Advanced anced Worksho shop p on n Ea Earthquake Fa Fault - - PowerPoint PPT Presentation

Adv Advanced anced Worksho shop p on n Ea Earthquake Fa Fault Mechanics: The Theory, , Simulation on and Observation ons ICTP, Trieste, Sept 2-14 2019 Lecture 3: fault friction Jean Paul Ampuero (IRD/UCA Geoazur) Lecture 3:


slide-1
SLIDE 1

Adv Advanced anced Worksho shop p on n Ea Earthquake Fa Fault Mechanics: The Theory, , Simulation

  • n and Observation
  • ns

ICTP, Trieste, Sept 2-14 2019 Lecture 3: fault friction Jean Paul Ampuero (IRD/UCA Geoazur)

slide-2
SLIDE 2

Lecture 3: earthquake dynamics from the standpoint of fault friction

  • Zoom on the process zone
  • Laboratory-based friction laws
  • Rupture pulses
  • Stress drop scaling
  • J. P. Ampuero - Earthquake dynamics

2

slide-3
SLIDE 3

Rock strength is finite

  • J. P. Ampuero - Earthquake dynamics

3

Byerlee’s law ! ∼ 0.6&

slide-4
SLIDE 4

The process zone

Stress singularities are unphysical: inelastic processes occur at the small scale (damage, weakening)

Process zone imaged by acoustic emissions in laboratory fracture of intact rock Secondary micro-cracks generated by a dynamic rupture

(mode II, numerical simulation by Yamashita 2000) Orientation Density

  • J. P. Ampuero - Earthquake dynamics

4

slide-5
SLIDE 5

Fault zone thickness

[Chester and Chester, 1998] Fault zone thickness and maturity (Savage and Brodsky, 2011)

  • J. P. Ampuero - Earthquake dynamics

5

slide-6
SLIDE 6

Cohesive zone models

Assumption: dissipative processes are mapped onto the fault plane, represented by a distribution of cohesive stresses near the crack tip Usual cohesive models:

  • constant (Dugdale, Barenblatt)
  • linearly dependent on distance to

crack tip (Palmer and Rice, Ida)

  • linearly dependent on slip (Ida,

Andrews) Slip and stress along a shear crack

(only half crack shown, Andrews 1976) Slip Stress Singular crack Slip weakening crack Process zone

ts td

  • J. P. Ampuero - Earthquake dynamics

6

slide-7
SLIDE 7

Cohesive zone size

  • Cohesive stresses t(x) generate a negative stress intensity factor

Kc = - that cancels the singularity : K + Kc = 0

  • That condition determines the size of the cohesive zone

with C1≈1 (for a linear distribution: C1=9p/32) From last lecture (mode III):

  • J. P. Ampuero - Earthquake dynamics

7

ts td

Λ = 1 − %&/(& Λ) where Λ) = *+2-.// 01 − 02

&

Slip Stress

Cohesive zone size

slide-8
SLIDE 8

Cohesive zone size

  • J. P. Ampuero - Earthquake dynamics

8

Λ = 1 − %&/(& Λ)

where Λ) = *+2-.// 01 − 02 & t45 = Λ/% Increasing rupture velocity à contraction of the process zone à higher frequency radiation à larger ground acceleration ∼ 1/ 7

slide-9
SLIDE 9

Tohoku: high-frequency radiation deeper than low-freq slip

Ito et al (2007)

Brownish symbols: 1Hz radiators extracted from back- projection movies Colored contours: static slip from GPS & tsunami data Spatial complementarity of high- and low-frequency slip:

HF radiation is deeper than static slip HF radiation occurs even where the rupture is slow

slide-10
SLIDE 10

Ito et al (2007)

Huang, Ampuero, Kanamori (2013) |<- Bottom Trench ->| |<- Bottom Trench ->|

Tohoku: high-frequency radiation deeper than low-freq slip

slide-11
SLIDE 11

Friction

More lateral force is needed to slide a taller, heavier object The resisting force is friction at the base of the object Friction force is proportional to the compressive force

Fault resistance is classically described by friction

  • J. P. Ampuero - Earthquake dynamics

11

Byerlee’s law ! ~ 0.6& !

slide-12
SLIDE 12

A brief history of fault friction

  • Coulomb friction: strength

SLIP S T R E S S STRENGTH

slide-13
SLIDE 13

A brief history of fault friction

  • Coulomb friction: strength
  • Static/dynamic friction:

stress drop

SLIP S T R E S S Static Dynamic Stress drop

slide-14
SLIDE 14

A brief history of fault friction

  • Coulomb friction: strength
  • Static/dynamic friction:

stress drop

  • Cohesion models: fracture

energy Gc Nucleation size

Lc ~ µGc/Dt2

SLIP S T R E S S Fracture energy

Gc

slide-15
SLIDE 15

A brief history of fault friction

  • Coulomb friction: strength
  • Static/dynamic friction: stress

drop

  • Cohesion models: fracture

energy Gc

  • Slip weakening friction: critical

slip Dc, weakening rate W

Lc ~ µ / W

SLIP S T R E S S W = weakening

rate

Dc

slide-16
SLIDE 16

A brief history of fault friction

  • Coulomb friction: strength
  • Static/dynamic friction:

stress drop

  • Cohesion models: fracture

energy Gc

  • Slip weakening friction:

critical slip Dc, weakening rate W

  • Rate-and-state friction:

healing, velocity weakening (a,b)

SLIP RATE S T R E S S STATE

slide-17
SLIDE 17

Laboratory-derived friction laws

Requirements :

  • High normal stress (100 MPa)
  • High slip rate (1 m/s)
  • Large displacements (>1 m)
  • Large sample (>Lc) and high resolution
  • Gouge + fluids

Only partially met by current experiments

Sandwich configuration (Ohnaka and Shen 1999) Rotary configuration (Chambon et al 2002)

  • J. P. Ampuero - Earthquake dynamics

17

slide-18
SLIDE 18

Laboratory-derived friction laws

Low resolution experiments (≈ spring+block ) record the average stress and slip à macroscopic friction

S = stress D = slip Large scale experiment Dieterich (1980)

High resolution experiments are densely instrumented à local friction + rupture nucleation and propagation

  • J. P. Ampuero - Earthquake dynamics

18

slide-19
SLIDE 19

Slip weakening friction

Slip weakening occurs during fast dynamic rupture. Linear slip weakening is a usual simplified model. Important parameters:

  • Dc = characteristic slip, associated to micro-contact evolution
  • r grain rearrangement.

Without gouge Dc ≈ 0.1 mm. With gouge Dc >10 cm

  • Strength drop: ts – td

Usually a small fraction of normal stress ≈ 0.1 s

  • Fracture energy of a linear slip weakening model :

Gc= ½ (ts - td) Dc

Slip (cm) Chambon et al (2000)

  • J. P. Ampuero - Earthquake dynamics

19

slide-20
SLIDE 20

Slip-weakening friction model

Primary parameters: dynamic friction coefficient !" and fracture energy Gc They control stress drop, rupture speed and rupture arrest Secondary parameters: critical slip distance Dc and strength drop !# − !" They control nucleation, supershear transition and peak slip velocity

slide-21
SLIDE 21

Dynamic Rupture Simulation

  • J. P. Ampuero - Earthquake dynamics

21

slide-22
SLIDE 22

Nucleation size

Nucleation size: !" =

$%& '()'*

Uenishi and Rice (2003)

slide-23
SLIDE 23

Exponential initiation

Linear slip-weakening: Δ" = "$ − "& '/') If there is some viscosity in the fault behavior: Δ" = * ̇ ' Equating both: ̇ ' = ,' Hence ' - ∼ exp ,- where , = "$ − "& /*') One form of viscosity is radiation damping, * = 2/24$ Kobe earthquake M7.2

slide-24
SLIDE 24

Seismological observations

A Mw3.9 earthquake in Alaska triggered by Love waves from the April 11, 2012 Mw 8.6 Sumatra earthquake Tape et al (2013)

slide-25
SLIDE 25

Seismological observations

Nucleation phase of the Mw3.9 Alaska triggered earthquake Tape et al (2013)

slide-26
SLIDE 26

Exponential initiation

!" = 2%& '& − ') /+,-

Simulations Ripperger et al (2007) Observations Tape et al (2013)

! = '& − ') /.,-

slide-27
SLIDE 27

Seismological constraints

Ide and Takeo (1997) Finite source inversion àslip + FDM àstress

Guatteri and Spudich (2000) Dynamic friction parameters suffer from strong trade-off

Same Gc à same strong motion <1Hz

A B

slide-28
SLIDE 28

Faults operating at low stress

How large is stress drop Δ" compared to strength drop "# − "% ? From seismological observations: Δ" = 1 − 10 Mpa From friction and lithostatic overburden: "# − "% = ) *# − *% ~100 ,-. à Δ" ≪ "# − "% Why so small?

slide-29
SLIDE 29

Faults operating at low stress

Fault loaded by deep creep à stress concentration at the base of the seismogenic zone

Interseismic slip Interseismic stress z z Seismogenic zone Creeping zone

slide-30
SLIDE 30

Faults operating at low stress

Interseismic stress z

Static strength !" Dynamic strength !#

Stress drop Strength drop $% − $'

Seismogenic depth W

slide-31
SLIDE 31

Faults operating at low stress

Interseismic stress z

Static strength !" Dynamic strength !#

Average stress drop $% Strength drop %& − %(

Seismogenic depth W

slide-32
SLIDE 32

Faults operating at low stress

Fracture energy balance: !" = $%

&' ∼ )*%+ &'

à Δ- ∼ 2/!"/1 Uenishi and Rice’s nucleation size: 2" =

'34 *56*7

à

)* *56*7 ∼ 84 + ≪ 1

slide-33
SLIDE 33

Rate-and-state friction

Second order effects: logarithmic healing (micro-contact creep) and velocity- weakening àPhenomenological rate-and-state friction law introduced by Dieterich and Ruina in the early 1980s Essential ingredients:

  • non-linear viscosity
  • evolution effect

Most important during slow slip (nucleation and post-seismic) During fast dynamic rupture, an equivalent Dc can be estimated: Dc ≈ 20 L

  • J. P. Ampuero - Earthquake dynamics

33

! = !∗ + % ln ( (∗ + ) ln (∗* + ̇ * = 1 − (* + ( = slip velocity, * = state variable

slide-34
SLIDE 34

Rate-and-state friction at high speed?

  • J. P. Ampuero - Earthquake dynamics

34

Most important during slow slip (nucleation and post- seismic) Rate-and-state behaves as slip-weakening during fast dynamic rupture Equivalent : !" = $ ln ' '∗ ≈ 20 $ ," ≈ 1 2 ./$ ln ' '∗

Kaneko et al (2008)

slide-35
SLIDE 35

Dramatic velocity-weakening at high speed

  • J. P. Ampuero - Earthquake dynamics

35

vc

Goldsby and Tullis (2011)

When sliding at high velocity: ! ∼ 1/%

Di Toro et al

slide-36
SLIDE 36

Dramatic velocity-weakening at high speed

  • J. P. Ampuero - Earthquake dynamics

36

At high velocity: ! ∼ 1/% Thermal weakening effects Predicted by flash heating (Rice, 2005)

Sutter and Ranc (2010)

slide-37
SLIDE 37

Rupture styles: cracks and pulses

In a previous lecture we focused on cracks.

Depth Along strike

Slip rate snapshots Crack : slip continues behind the rupture front, long rise time Pulse : slip heals soon behind the rupture front, short rise time

  • J. P. Ampuero - Earthquake dynamics

37

slide-38
SLIDE 38

Pulses: observations

Heaton (1990)

  • bserved that rise

times are usually short ≈10% of total earthquake duration

Source models from kinematic inversions. Contours = slip Shaded = snapshot of active slip

  • J. P. Ampuero - Earthquake dynamics

38

slide-39
SLIDE 39

Cracks and pulses

Self-healing pulses require fast strength recovery à velocity dependent friction

slide-40
SLIDE 40

Non-planar, rough faults

Power spectrum of fault surface geometry wavenumber wavelength Candela et al (2012)

slide-41
SLIDE 41

Slip and stress on rough faults

Residual off-fault stresses Dieterich and Smith (2009) Flattening of slip profiles

slide-42
SLIDE 42

Roughness drag

Fang and Dunham (2013): where Δ=slip, " =rms-amplitude-to-wavelength ratio (0.1~1 %), $%&' =small cutoff length Ignoring friction, fault opening and off-fault inelasticity.

Rougher faults need higher stresses to sustain earthquakes

slide-43
SLIDE 43

Mesoscopic model

Meso-scale representation (~homogenization) of roughness effects: Fault strength = friction (slip-weakening) + roughness drag (slip-strengthening) Key parameters: W=slip-weakening rate !=slip-strengthening rate Scope: determine overall rupture features (rupture stability, speed, slip scaling) without resolving details of high-frequency radiation

slide-44
SLIDE 44

Slip-strengthening

slide-45
SLIDE 45

Numerical results (with Franklin Koch)

Stable Unstable Steady Initial shear stress Steady pulses = boundary between decaying and sustained ruptures:

slide-46
SLIDE 46

Implications for stress in the crust

Background stress = sqrt(roughness/weakening) Rougher faults can operate seismically at higher stresses à Relation between fault maturity, geometry and strength

slide-47
SLIDE 47

Pulses: other possible origins

Pulses controlled by the depth of the seismogenic region (Day 1982) Pulses on very heterogeneous faults (Beroza and Mikumo 1996)

  • J. P. Ampuero - Earthquake dynamics

47

slide-48
SLIDE 48

Pulses: possible origins

NW SE Rupture on a bimaterial interface (between two different materials) like in the San Andreas Fault NW NW SE SE Stress at the rupture edges Pulse

  • J. P. Ampuero - Earthquake dynamics

48

slide-49
SLIDE 49

Origins of pulses: fault zone waves

[Chester and Chester, 1998]

  • J. P. Ampuero - Earthquake dynamics

49

slide-50
SLIDE 50

Origins of pulses: fault zone waves

[Ellsworth and Malin, 2012] San Andreas Fault Nojima Fault [Huang and Ampuero, 2011]

~ 200 m wide

[Chester and Chester (1998)

  • J. P. Ampuero - Earthquake dynamics

50

slide-51
SLIDE 51

Fault parallel distance (m) Fault normal distance (m) Without fault zone With fault zone

Origins of pulses: fault zone waves

Huang and Ampuero (2011) Huang, Ampuero and Helmberger (2014)

  • J. P. Ampuero - Earthquake dynamics

51

slide-52
SLIDE 52

Fault parallel distance (m) Fault normal distance (m) Without fault zone With fault zone

Origins of pulses: fault zone waves

Huang and Ampuero (2011) Huang, Ampuero and Helmberger (2014)

  • J. P. Ampuero - Earthquake dynamics

52

slide-53
SLIDE 53

Summary

  • Friction laws:
  • slip-weakening: most basic
  • rate-and-state: low speed
  • velocity-weakening: high speed
  • Fracture mechanics concepts (Gc) still useful to

rationalize results of frictional rupture models:

rupture arrest, acceleration at fault kinks

  • Features require modeling with friction laws:

nucleation, pulses (healing), supershear ruptures

  • J. P. Ampuero - Earthquake dynamics

53