20061010WS-TAIWANinAIST

Transport properties and its implication of pore pressure change due to frictional heating during 1999 Taiwan Chi-Chi earthquake

Wataru Tanikawa (Kochi core research center / JAMSTEC） Toshihiko Shimamoto (Kyoto University) 1999 Taiwan Chi-Chi Earthquake 20061010WS-TAIWANinAIST

Acceleration Velocity Displacement ‹September 20, 1999 - Mw 7.6 ‹Rupture of Chelungpu Fault ‹Propagation from South to North ‹Remarkable difference between N – S

North South

Displacement Large -10m Small Ma et al. 2003

Velocity Large – 4.5m/s Small

Acceleration Small Large

High freq. Low level radiation Stress Drop Large Small

Zhang et al. 2003 Question to be Solved 20061010WS-TAIWANinAIST

What made the contrast between North and South? What caused such a large displacement at Northern portion?

Variation of fault rock property and Dynamic fault weakening mechanism ・Melting (Hirose and Shimamoto, 2005) – Psuedotachylyte is rare in fault zones. ・(Elast) hydrodynamic lubrication (Ma et al., 2003) – Fault rocks behaves as viscous? ・Acoustic fluidization (Melosh, 1996) – Difficult how to identify – injection vein? ・Thermal pressurization (Lachenbruch 1980)

Current researches related to the Chelungpu Fault Borehole temperature observation (Kano et al. 2006) - Low friction during slip event → Dynamic weakening mechanism effectively occurred? Core observation (Hirono et al. 2006) - Temperature doesn’t rise to melting point → Melt weakening is ineffective? Concept of Thermal Pressurization 20061010WS-TAIWANinAIST

Fault zone Fault core (fault gouge) ⇒Heat source （fault breccia） Frictional Heating (during earthquake) ↓ Thermal expansion of pore water ↓(Undrained condition) Pore pressure generation ↓ Velocity Reduction of effective pressure Thickness Strain distribution ↓ Dynamic fault weakening -Stress change at the fault zone- ↓ Unstable → Large slip? Stress drops due to pore pressure Critical Parameters for TP generation ・Diffusion parameter - Permeability, Specific storage

Shear stress (MPa) ・Heat source parameter - Shear strength, Thickness of fault core Pore pressure generation (MPa) Slip Displacement (m) Research Area 20061010WS-TAIWANinAIST Eurasian Plate 0204060 Taipei Northern Hole km STUDY Taichung AREA Chelungpu Shuichangliu 1)Depth Variation Ryukyu Tainan Taitung Trench Kaoshung 8 .2 cm Chelungpu Fault / Manila yr Trench Philippine Shuangtung Fault Sea Plate Shuichangliu Fault (Stratigraphic cross section, Vitrinite reflectance)

2)Along-Fault Variation (borehole sample) Northern site (Fengyuan 400m) Southern hole Southern site (Nantou 200m) TCDP - (Dakeng A:2000m, B:1350m)

Shuangtung Chelungpu Fault-Northern borehole 20061010WS-TAIWANinAIST 3 possible candidates for slip zone - fault zones are developed within siltstone →Still under discussion which is the best choice!? Candidate1(329 - 330 m) 広い温度異常が確認280m～340m 10m random oriented fault breccia Siltstone/Sandstone

Slip plane? Candidate2 (224.55 – 224.75 m)

Siltstone Fault breccia Siltstone

No strong shear deformation zone

Candidate3（405m) Thin clayey fault gouge (7 mm)

Siltstone Thick fault gouge（50cm） Deformed sandstone

Very thin hard black material (ultra cataclasite?) Shuangtung Fault-outcrop 20061010WS-TAIWANinAIST

Fault breccia and fractured hostrock

Black Layers

zBoundary between Pleistocene and late Miocene sediment. z8 m thickness of the clay-rich foliated fault gouge. zBlack layers（ultra cataclasite） are developed in the both boundary of the thick foliated fault gouge (23 mm thick).

70

) 60 23±13 mm 50 40 30 20 10 thickness (mm 0 10 90 250 450 650 10 m-thickness clayey foliated fault gouge distance (cm) Transport Property Test 20061010WS-TAIWANinAIST

Experimental condition Pore Fluid - N2 gas （low viscosity） Temperature - room temperature Confining pressure - 0 ~ 200 MPa (12km) Fluid pressure - 0 ~ 2 MPa Sample size – φ20mm ×Length 10 - 40 mm Cylindrical samples Flow meter Method for measurements Permeability - steady state gas flow method using accurate gas flow meter (ADM2000, Alicat flowmeter) Gas permeability is arranged to water permeability using the Klinkenberg equation. Porosity - calculated by the pore pressure change under undrained condition Specific storage- approximated by drained pore compressibility that is estimated from porosity test

（ -1 Ss = β + φβ Ss: Specific storage Pa ) φ f -1 βf: Fluid compressibility(Pa ） 1 ∂φ φ : Porosity β = − φ Pc: Confining pressure 1− φ ∂Pc p=0 p: Pore pressure

Pressure vessel Example of Experiment – permeability & porosity20061010WS-TAIWANinAIST Northern Shallow borehole for Chelungpu Fault

30

-13 10 25 -14 10 ） 2

） 20 m -15 （ 10 % （ -16 10 15

-17

10 Porosity 10

Permeability -18 φ2 10 k1 -19 5 φ1 10 -20 k2 10 0 0 50 100 150 0 50 100 150 200 Effective Pressure （MPa） Effective Pressure （MPa） Applied for Thermal Pressurization analysis

1.Strong sensitivity for effective pressure Proper description of parameters as 2.Elast-plastic behaviors a function of pressure is important! Permeability StructureⅠ- Along Fault Variation 20061010WS-TAIWANinAIST

Northern borehole （329m depth） Southern borehole ） ）

2 (305.5m) 2 m m （ （ (326.5m depth)

(conglomerate) Permeability Permeability

(fault breccia)

Distance from the fault (m) Depth (m)

◆Permeability for fault rock is larger in Southern site (one order). ◆Permeability for wall rock is larger in Southern site. Permeability StructureⅡ- Depth Variation 20061010WS-TAIWANinAIST

Shallow? Deep? Chelungpu Fault Shuangtung Fault Shuichangliu Fault ） 2 m

浸透係数（ fault zone

（ ） Distance from fault core（m） Distance from fault core（m） Distance from fault core m

‹Chelungpu Fault < Shuangtung Fault • Shuichangliu Fault ‹Permeability variation within fault zone is small Specific Storage- Depth Variation 20061010WS-TAIWANinAIST Chelungpu Fault ) -1 Small difference of specific storage among the faults, within a fault zone, and between fault and host rocks. Most of them are around 10-9 Pa-1. Specific storage (Pa

Distance from the slip surface（m） Shuangtung Fault Shuichangliu Fault ) ) -1 -1 Specific storage (Pa Specific storage (Pa

Distance from the slip surface（m） Distance from the slip surface（m） Frictional Property Test 20061010WS-TAIWANinAIST Bi-axial typical frictional test High shear velocity machine

Shear stress Gabbro blocks

Vertical stress Designed by Shimamoto at Kyoto Univ.

Assembly of high frictional test for fault gouge

Cylindrical Gouge material (1.5~2.0 g) host rock High Velocity Low Velocity

Slip velocity 0.1~1 m/s 0.1~100μm/s Vertical Rotation Stress Vertical stress 1 MPa 0 ~60MPa

Slip distance ∞ 20mm Gouge material（2g） Teflon-sleeve Summery of Low Velocity Friction 20061010WS-TAIWANinAIST 0 .4 20

・Stable friction achieved from 10 mm 0 .3 15Vertical stress(MPa)

・Chelungpu > Shuangtung > Shuichangliu efficient ・Wet gouge < Dry gouge （except South) 0 .2 Velocity step test Hold-slide- 10 holdconvention test ・South - Velocity weakening velocity step Shear cycles at high al H.S.H 0 .1 test 5 ⇔ North - Velocity hardening Frictional co vertical stress test BAF099 0 shear load cycles 0 0 5 1 0 1 5 2 0 Slip displacement (mm)

Chelungpu1 Dry Velocity Wet hardening ‐South

Chelungpu2 ‐North

Shangtung3

Shuichangliu4 Velocity weakening

0.2 0.3 0.4 0.5 0.6 0.7 0.8 -0.01 -0.005 0 0.005 0.01 Frictional coefficient a – b High Velocity Frictional Behavior for FG 20061010WS-TAIWANinAIST

→video image Chelungpu Fault Peak friction – Southern Shallow core (Black Gouge, 176.75 ~ 176.83m) µp slip velocity：1 m/s 0.6 MPa 0.9 MPa dry condition

Temperature rise < 300oC Steady state frictional coefficient µ (Mizoguchi 2005) ⇒ Gouge d frictional coefficient does not melt ⇒ What weakening mechanism? Dc:Slip weakening distance

slip distance（m）

１）Large peak friction～１ Low velocity frictional test ２）Rapid reduction with slip ⇒ stable 1)Dc 10 µm～1 mm ３）Large weakening distance Dc～10m 2)µd 0.5 ~ 0.85 ４）Low steady state friction µd～0.2 Fault Gouge Variation for H-V Friction 20061010WS-TAIWANinAIST

CONDITION Chelungpu Fault - North Chelungpu Fault - South ・Slip velocity:1 m/s Grayish Gouge（285m） Black Gouge (176.75-176.83m) ent ent ・Vertical stress:1MPa i i ・Room Temperature 0.6 MPa ・Dry gouge 0.6 MPa 0.6 MPa 0.9 MPa 0.9 MPa

・All gouges show similar behaviors Frictional Coeffic Frictional Coeffic ・Dc － 6 ~ 13m． ・South is largest Slip Distance （m） Slip Distance （m）

Shuangtung Fault Shuichangliu Fault

Brownish Gouge ent Black Gouge i ent However weakening i mechanism is not 0.6 MPa 0.6 MPa researched in detail! 0.9 MPa 0.9 MPa →Hydrodynamic lubrication? Frictional Coeffic

→Tribo-chemical Frictional Coeffic reaction with heating?

Slip Distance （m） Slip Distance （m） Thermal Pressurization Analysis 20061010WS-TAIWANinAIST

1D heat and fluid diffusion equation ＋ High velocity frictional behavior

Lachenbruch (1980)

Mase and Smith (1987) µ p

Heat generation and diffusion efficient

∂T 1 ⎛ ∂2T ⎞ = ⎜A +κ ⎟ µss ⎜ 2 ⎟

∂t ρc ∂x Frictional co ⎝ ⎠ dc V Heat production rate A = µ ()σ −P d W Slip distance（m） Fluid flow and Pp generation ⎛ ln(0.05)⋅ d ⎞ µd = µss + ()µ p − µss exp⎜ ⎟ ∂P 1 ⎛ ∂T ∂ ⎛ k ∂P ⎞⎞ ⎝ Dc ⎠ = ⎜φ()γ − γ + ⎜ ⎟⎟ ⎜ f ⎜ ⎟⎟ ∂t ⎛ ⎛ ∂φ ⎞ ⎞ ⎝ ∂t ∂x ⎝ µ ∂x ⎠⎠ ⎜φβ + ⎜ ⎟ ⎟ ⎜ f ∂P ⎟ ⎝ ⎝ ⎠ T ⎠ Shear Stress Change Laboratory results (Effective pressure functions) τ= (Pc-Pp)×μd TP Result Ⅰ- Pp Generation Ⅰ 20061010WS-TAIWANinAIST Slip Velocity－1 m/s Thickness of fault－20 mm

Chelungpu Fault – North Chelungpu Fault – South 深度 (km) 1 Without TP 2 l shear strength l shear strength 3 4 Initia Initia Fault weakening due to TP 5 ／ ／ 6 TnoPなし TP Shear strength Shear strength

Small Dc at depth Slip distance (m) Slip distance (m)

TP is effective!! ⇒ Weakening TP is ineffective!! ⇒ Weakening is due is accelerated with depths． to high velocity (mechanical?) weakening TP Result Ⅱ-Pp Generation Ⅱ 20061010WS-TAIWANinAIST

Slip Velocity－1 m/s Thickness of fault－20 mm

Shuangtung Fault Shuichangliu Fault 深度 (km) 1 2 3 l shear strength l shear strength 4 5 Initia Initia 6 ／ ／ TPなし Shear strength Shear strength

Slip displacement (m) Slip displacement (m)

TP is much effective !! TP is relatively effective (Similarity to Northern Chelungpu Fault) TP Result Ⅲ - Temperature Rise 20061010WS-TAIWANinAIST Chelungpu Fault – North Chelungpu Fault – South 1000 1000 C) TP effective regime 800 C) 800 o o Reducing shear strength of 600 600 e rise ( e rise ( fault. Low, and relatively depth (km) atur 1 atur low temperature rise in depth (km) 400 2 400 1 Shuangtung, and Northern 3 2 4 3 Temper Temper Chelungpu Faults . 5 4 200 6 200 5 slip rate =1 m/s 6 deformationzone=20mm 0 0 0 2 4 6 8 10 0 2 4 6810 Slip displacement (m) Slip displacement (m) Shuangtung Fault Shuichangliu Fault 1000 1000 depth (km) 1 TP ineffective regime 2 C) C) 800 3 800 o o 4 Rapid temperature 5 6 600 e rise ( rise easily reaches e rise ( 600 atur to melting point. atur depth (km) 400 400 1 2 3 Temper Temper 4 200 200 5 6

0 0 0 2 4 6 8 10 0 2 4 6 8 10 Slip displacement (m) Slip displacement (m) TP Result Ⅳ - Importance of Width 20061010WS-TAIWANinAIST TCDP-Hole B 1194m Width of the fault core is directly related to heating rate, and identification of the shear zone is important (though it is difficult).

BM disk（20mm thickness） 1 1000 width (mm) width (mm) slip rate =1 m/s 10 slip rate =1 m/s 10 depth = 5 km 0.8 depth = 5 km 20 800 20 C) 30 o 30 40 40 l shear strength 0.6 100 600 100 200 200 Initia

／ 0.4 400

0.2 Temperature rise ( 200

Shear strength 0 0 0 1 2 3 4 5 0 1 2 3 4 5 Slip displacement (m) Slip displacement (m) Summery Ⅰ- fault variation 20061010WS-TAIWANinAIST

North South Depth Remark

Permeability 10-16~10-17 10-15~10-16 10-16~10-18 North < South （m2） Specific storage Small difference among faults （10-9 Pa-1） （Pa-1）

High velocity Similar behavior （exponential decay curve，Low Relatively large frinction in friction steady state friction，Large initial friction） Southern Chelungpu Fault

TP analysis Relatively Ineffective Effective Possible existence of effective overpressure at depth might be negative influence for TP.

Low μdry 0.7 0.7 0.5 ~0.6 Reduction in 50% at wet velocity condition friction a-b + － +

Permeability variation ⇒ TP variation ⇒ Explain the difference between N-S? TP might be effective at depths（Overpressure can not be neglected）. South is unstable ⇒ Consistent with initiation of EQ from the south? Summery Ⅱ-TP vs Seismic Data 20061010WS-TAIWANinAIST

1.Weakening distances,Dc, High velocity friction TP analysis evaluated from TP analysis are similar order Chelungpu Fault North to that evaluated from Dc=6 to 13 m

seismic inversion analysis. t n 2. Without thermal Dc=1 to 5 m pressurization, we can account for the Dc from High velocity behavior. normalized shear strength 3.Stress drop between TP tional coefficie

and seismic data has gap. fric slip distance (m) slip distance (m)

After Jim Mori (2005) Dc = 2.6 m After Ma et al. (2005) Inversion analysis （several cm to 10 Dc = 5.06m m） Trac. Drop = (After Zheng et al. 2003) 4.5 MPa Dc = 6.26 m

slip (m)