Weak ductile shear zone beneath the western North ... · (a) after the earthquakes (Ergintav et al....
1
: Elastic model : LWZ (ΩG ' = 0.1) : LWZ (ΩG ' = 1.0) : LWZ (ΩG ' = 0.6) : LWZ (ΩG ' = 0.2) : LWZ (ΩG ' = 2.0) η w ' = 0.01 η ' = 1.0 ΘG ' = 0.3 Weak ductile shear zone beneath the western North Anatolian Fault Zone: inferences from earthquake cycle model constrained by geodetic observations SUMMARY 1. AIM OF THIS STUDY Tadashi Yamasaki, Tim J. Wright and Gregory A. Houseman COMET+, School of Earth and Environment, University of Leeds, UK 3. THE EARTHQUAKE CYCLE MODEL Profiles of GPS velocities across the North Anatolian Fault, before and after the 1999 İzmit and Düzce earthquakes, require a localised weak zone in the mid crust beneath the western North Anatolian Fault Zone. The thickness and width of this low viscosity zone are ~ 12 km and ~ 20 - 40 km, respectively, and the effective viscosities of the weakened and the non-weakened domains are ~ 2 × 10 18 and ~ ≥ 2 × 10 20 Pa s, respectively. Using a 3D finite element model, we examine the linear Maxwell visco-elastic response to a repeated strike-slip faulting event under the condition of a constant far-field loading rate for three basic crustal viscosity models beneath an upper elastic layer: uniform viscosity (UNV), depth-dependent viscosity (DDV) and localised weak zone (LWZ). We compare predicted surface velocity profiles with the GPS velocity profiles across the North Anatolian Fault Zone (NAFZ), before and after the 1999 İzmit and Düzce earthquakes, in order to constrain the actual viscosity variation in the crust. 6. CONCLUSIONS 5. THE LWZ MODEL FOR THE NAFZ CASE ACKNOWLEDGEMENTS: REFERENCES: This study is supported by the UK Natural Environment Research Council (NERC) under grant number NE/I028017/1, and with the support of the National Centre for Earth Observation (NCEO-COMET+). Cluster computing was provided by NCEO and by the University of Leeds Advanced Research Computing Node 1 (arc1). Bürgmann et al., 2002, BSSA 92, 126-137; Ergintav et al. 2009, JGR 114, B07403, doi:10.1029/2008JB006021; Feigl et al., 2002, BSSA 92, 138-160; Klinger et al., 2003, BSSA 93, 2317-2332; Kozacı et al., 2009, JGR 114, B01405, doi:10.1029/2008JB005760; McClusky et al., 2000, JGR 105, 5695-5719; Reilinger et al., 2000, Science 289, 1519-1524; Reilinger et al., 2006, JGR 111, B05411, doi:10.1029/2005JB004051; Rockwell et al., 2009, Geol. Soc. Lon. Spec. Pub. 316, 31-54; Stein et al., 1997, GJI 128, 594-604; Vanacore et al., 2013, GJI, 193, 329-337; Tatar et al., 2012. Tectonophysics 518-521, 55-62; Wright et al., 2001, GRL 28, 2117-2120; Yamasaki & Houseman, 2012, GJI 190, 769-784. 4.1 Uniform viscosity (UNV) model 4.2 Depth-dependent viscosity (DDV) model -12.2 -12.0 -11.8 -11.6 -11.4 -11.2 Displacement (U x ' - t ' ) Time (t ' ) -1.5 -1.0 -0.5 0.0 -1.5 -1.0 -0.5 0.0 -2.0 -1.5 -1.0 -0.5 0.0 -14 -12 -10 -8 -6 -4 -2 0 Displacement (U x ' - t ' ) Displacement (U x ' - t ' ) Displacement (U x ' - t ' ) Time (t ' ) Displacement (U x ' - t ' ) Time (t ' ) (a) UNV (η ' = 0.01) (c) UNV (η ' = 1.0) (b) UNV (η ' = 0.1) (d) UNV (η ' = 10.0) (x ' , y ' , z ' ) = (0.0,-0.5,1.0) (x ' , y ' , z ' ) = (0.0,-0.5,1.0) (x ' , y ' , z ' ) = (0.0,-0.5,1.0) (x ' , y ' , z ' ) = (0.0,-0.5,1.0) 0 30 2 6 10 14 18 22 26 4 8 12 16 20 24 28 0 200 20 60 100 140 180 40 80 120 160 0 10 1 2 3 4 5 6 7 8 9 0 10 1 2 3 4 5 6 7 8 9 180 200 182 184 194 192 190 186 188 196 198 4.3 Localised weak zone (LWZ) model 4.1.1 The surface displacement with time 4.1.2 The surface velocity profiles before and after earthquake -30 -20 -10 0 10 20 30 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 -8 -6 -4 -2 0 2 4 6 8 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 Velocity (V x ' ) Velocity (V x ' ) Velocity (V x ' ) Velocity (V x ' ) (a) (c) (b) (d) DDV ( η b ' = 0.001) After earthquake DDV ( η b ' = 0.001) Before earthquake DDV ( η b ' = 0.01) After earthquake DDV ( η b ' = 0.01) Before earthquake : Elastic model : η h ' = 1000 η b ' : η h ' = 10000 η b ' : η h ' = 100000 η b ' Distance (y ' ) Distance (y ' ) Velocity (V x ' ) Velocity (V x ' ) (e) (f) DDV ( η b ' = 0.1) After earthquake DDV ( η b ' = 0.1) Before earthquake -5 5 4 3 2 1 0 -1 -2 -3 -4 -5 5 4 3 2 1 0 -1 -2 -3 -4 -5 5 4 3 2 1 0 -1 -2 -3 -4 -5 5 4 3 2 1 0 -1 -2 -3 -4 -5 5 4 3 2 1 0 -1 -2 -3 -4 -5 5 4 3 2 1 0 -1 -2 -3 -4 : Elastic model : η h ' = 10 η b ' : η h ' = 100 η b ' : η h ' = 1000 η b ' (a ' ) Cross-section view Elastic layer H ' = 0.3, D ' = 0.4 Visco-elastic layer z ' y ' V xb ' -V xb ' Θ ' Ω ' S x ' + S x ' - (a '' ) Map view y ' x ' Right-lateral strike-slip fault S x ' + S x ' - -V xb ' V xb ' V xb ' = S x ' / ∆ t ' = 1 S x ' = 1 ∆ t ' = 1 Displacement Fault Boundary Time (t ' ) (b) Prescribed displacements 2. THE GEODETIC OBSERVATIONS 25˚ 25˚ 26˚ 26˚ 27˚ 27˚ 28˚ 28˚ 29˚ 29˚ 30˚ 30˚ 31˚ 31˚ 32˚ 32˚ 33˚ 33˚ 34˚ 34˚ 35˚ 35˚ 39˚ 39˚ 40˚ 40˚ 41˚ 41˚ 42˚ 42˚ 43˚ 43˚ Black Sea Anatolia North Anatolian Fault Eurasia Marmara Sea Gulf of Saros Agean Sea The İzmit earthquake (Mw = 7.5, 17 August 1999) The Düzce earthquake (Mw = 7.2, 12 November 1999) : GPS velocity pre. 1999 : GPS velocity Feb. 2000 The GPS velocity profiles across the North Anatolian Fault Zone near the epicentres of the 1999 İzmit and Düzce earthquakes show: (a) after the earthquakes (Ergintav et al. 2009), the peak velocities either side of the fault differ by as much as 150 mm/yr, 5 or 10 times greater than the long-term relative displacement rate of ~ 15 to 30 mm/yr (Tatar et al., 2012; Reilinger et al., 2006). (b) before the earthquakes (McClusky et al., 2000), displacement rates vary monotonically across fault zone, but the velocity gradient is ~ 3 times higher within a zone of about 30 km width, centred on the fault. We use a parallelized 3-D finite element code, oregano_ve (Yamasaki and Houseman, 2012), to solve the linear Maxwell visco-elastic problem in which an instantaneous strike-slip fault occurs at a periodic interval, releasing elastic strain accumulated by regional loading that occurs at a constant rate. -16 -12 -8 -4 0 4 8 12 16 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 Velocity (V x ' ) Velocity (V x ' ) Velocity (V x ' ) Distance (y ' ) Velocity (V x ' ) Distance (y ' ) (a) UNV (η ' = 0.01) (c) UNV (η ' = 1.0) (b) UNV (η ' = 0.1) (d) UNV (η ' = 10.0) : Elastic model : After earthquake : Before earthquake -5 5 4 3 2 1 0 -1 -2 -3 -4 -5 5 4 3 2 1 0 -1 -2 -3 -4 -5 5 4 3 2 1 0 -1 -2 -3 -4 -5 5 4 3 2 1 0 -1 -2 -3 -4 : Elastic model : After earthquake : Before earthquake : Elastic model : After earthquake : Before earthquake : Elastic model : After earthquake : Before earthquake For the models with η ’ = 0.01 and 0.1, a steady-state periodicity is established rapidly, implying that the stress relaxation is virtually completed by the end of the inter-seismic period. For the models with η ’ = 1.0 and 10.0, only a smaller proportion of the deviatoric stress can be relaxed between seismic events, and the un-relaxed stresses evolve until the steady-state is established. Some DDV models predict the general observed features. However, the pre-seismic velocity gradient across the fault is not adequately high, and the wavelength of the post-seismic velocity profile is inconsistent with the data. A DDV model alone cannot explain the NAFZ velocity profiles. LWZ model predicts the general observed features of the both post- and pre-seismic velocities with consistent wavelength of the profiles. Note that the Gaussian-function type of LWZ results in similar behaviour. An LWZ model can explain the NAFZ velocity profiles. Symbol Meaning Value TABLE 1: Parameter values used for the western North Anatolian Fault Zone (NAFZ) *: Stein et al. (1997), Klinger et al. (2003) and Rockwell et al. (2009) #: Reilinger et al. (2000), Bürgmann et al. (2002) and Feigl et al. (2002) $: Reilinger et al. (2000; 2006), Kozacı et al. (2007; 2009), Wright et al. (2001) and Tatar et al. (2012) Shear modulus Minimum viscosity in localised weak zone (LWZ) Background viscosity Earthquake cycle period Depth-extent of fault Elastic layer thickness Poisson’s ratio Layer thickness 3 x 10 10 0.25 12 km * 16 km * ~ 200 yrs # ~ ≥ 2 x 10 20 Pa s ~ 1 - 2 x 10 18 Pa s 40 km µ ν ∆t η η w Boundary velocity Slip rate (Effective) thickness of LWZ (Effective) width of LWZ ~ 24 - 40 km ~ 12 km ~ 15 - 30 mm/yr $ ~ 7.5 - 15 mm/yr Ω H , Ω G Θ H , Θ G 2S x / ∆t V xb H D L 0 : Elastic model : η h ' = 1000 η b ' : η h ' = 10000 η b ' : η h ' = 100000 η b ' : Elastic model : η h ' = 10 η b ' : η h ' = 100 η b ' : η h ' = 1000 η b ' : Elastic model : η h ' = 10 η b ' : η h ' = 100 η b ' : η h ' = 1000 η b ' : Elastic model : η h ' = 10 η b ' : η h ' = 100 η b ' : η h ' = 1000 η b ' The high post-seismic displacement rate is predicted for UNV model with η ’ = 0.01 or 0.1, but such a model does not predict the higher pre-seismic velocity gradient across the fault. A UNV model cannot explain the NAFZ velocity profiles. The LWZ model prediction shows a good fit to both pre- and post-seismic data for the LWZ configuration shown in Table 1. Asymmetry of the observed profiles is not included in these LWZ models. Comparison with observations favours the block-type LWZ rather than the Gaussian-type LWZ. Distance (y km) (a) After the earthquakes N S -100 -75 -50 -25 0 25 50 75 100 Velocity (V x mm/yr) Slip rate = ~ 15 - 30 mm/yr Slip rate = ~ 15 - 30 mm/yr -100 100 75 50 25 0 -75 -50 -25 -20 -15 -10 -5 0 5 10 15 20 Velocity (V x mm/yr) Distance (y km) (b) Before the earthquakes N S Slip rate = ~ 15 - 30 mm/yr Slip rate = ~ 15 - 30 mm/yr -100 100 75 50 25 0 -75 -50 -25 5.1 The Block-type of LWZ 5.2 The Gaussian-type of LWZ Velocity (V x mm/yr) Distance (y km) (b) Before the earthquakes N S -20 -15 -10 -5 0 5 10 15 20 -100 100 75 50 25 0 -75 -50 -25 Velocity (V x mm/yr) Distance (y km) N S -120 -90 -60 -30 0 30 60 90 120 Slip rate = 22.5 ± 7.5 mm/yr ∆ t = ~ 200 yrs η = ~ 2 × 10 20 Pa s Θ H = 12 km H = 12 km, D = 16 km -100 100 75 50 25 0 -75 -50 -25 -100 100 75 50 25 0 -75 -50 -25 Velocity (V x mm/yr) Distance (y km) (a) After the earthquakes Slip rate = 22.5 ± 7.5 mm/yr ∆ t = ~ 200 yrs η = ~ 2 × 10 20 Pa s η w = ~ 2 × 10 18 Pa s N S -120 -90 -60 -30 0 30 60 90 120 Θ G = 12 km Ω G = 40 km H = 12 km, D = 16 km Velocity (V x mm/yr) Distance (y km) (b) Before the earthquakes Slip rate = 22.5 ± 7.5 mm/yr ∆ t = ~ 200 yrs η = ~ 2 × 10 20 Pa s η w = ~ 2 × 10 18 Pa s N S -20 -15 -10 -5 0 5 10 15 20 Θ G = 12 km Ω G = 40 km H = 12 km, D = 16 km -100 100 75 50 25 0 -75 -50 -25 (a) After the earthquakes : η w = ~ 2 × 10 18 Pa s Ω H = 40 km : η w = ~ 10 18 Pa s Ω H = 24 km : η w = ~ 2 × 10 18 Pa s Ω H = 40 km : η w = ~ 10 18 Pa s Ω H = 24 km Slip rate = 22.5 ± 7.5 mm/yr ∆ t = ~ 200 yrs η = ~ 2 × 10 20 Pa s Θ H = 12 km H = 12 km, D = 16 km X L ' = X L ' /L 0 = 25.0 Y L ' = Y L /L 0 = 10.0 x ' z ' y ' Right-lateral strike-slip fault S x ' + S x ' - O 1.0 (Elastic layer: η ' = 10 20 ) (Visco-elastic layer: η ' ) - V xb ' Weak Zone ( η w ' ) V xb ' H ' = H /L 0 = 0.3 Ω ' Θ ' (a) D ' = D /L 0 = 0.4 L 0 = 40 km, the thickness of the crust beneath the western North Anatolian Fault (Vanacore et al., 2013) (1) The ratio τ / ∆ t of the Maxwell relaxation time τ to the earthquake cycle period ∆ t controls the surface velocity variation during the earthquake cycle. (2) Post-seismic velocities near a major fault that are much greater than the long-term relative displacement rate indicate a relatively low Maxwell time constant (low viscosities), and strong localization of the pre-seismic strain rate near the fault requires a relatively high Maxwell time constant (high viscosities). (3) Of the models we tested, only the localized weak zone (LWZ) model is able to explain at once the GPS velocities before and after the 1999 İzmit and Düzce earthquakes. We infer a zone of low viscosity (~ 2 x 10 18 Pa s) centred below the fault, relative to a general background crustal viscosity of at least 2 x 10 20 Pa s. The low viscosity zone is ~ 12 km thick (below a 12 km thick elastic layer), and its width is between ~ 20 and 40 km. (4) A likely explanation for the block-type low viscosity zone is a slice of allocthonous crust about 20 km wide between northern and southern strands of the North Anatolian Fault in the İzmit region, though differences in stress, grain-size, water content, or temeperature could also be significant. 4. THE MODEL BEHAVIOUR: INFERENCES FOR THE NAFZ CASE Velocity (V x ' ) Velocity (V x ' ) Distance (y ' ) (d) Gaussian-type LWZ Before earthquake (c) Gaussian-type LWZ After earthquake -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 -5 5 4 3 2 1 0 -1 -2 -3 -4 -5 5 4 3 2 1 0 -1 -2 -3 -4 : Elastic model : LWZ (ΩG ' = 0.1) : LWZ (ΩG ' = 1.0) : LWZ (ΩG ' = 0.6) : LWZ (ΩG ' = 0.2) : LWZ (ΩG ' = 2.0) η w ' = 0.01 η ' = 1.0 ΘG ' = 0.3 (c) Viscosity in UNV and DDV models UNV DDV Viscosity (log 10 ( η ' )) η b ' 0.01 η b ' 10 η b ' 100 η b ' 1000 η b ' Elastic layer ( η ' = 10 20 ) Depth (z ' ) 0.0 0.2 1.0 0.1 0.4 0.3 0.6 0.5 0.8 0.7 0.9 0.1 η b ' Distance from fault (y ' ) 5.0 1.0 0.0 3.0 2.0 4.0 0.0 1.0 0.4 0.6 0.8 0.2 Depth (z ' ) -2 0 Log 10 ( η ' ) Elastic layer ( η ' = 10 20 ) 5.0 1.0 0.0 3.0 2.0 4.0 -2 0 Log 10 ( η ' ) 0.0 1.0 0.4 0.6 0.8 0.2 Depth (z ' ) Elastic layer ( η ' = 10 20 ) (d) Viscosity in block-type LWZ model Θ B ' Ω B ' Θ G ' Ω G ' Distance from fault (y ' ) (e) Viscosity in Gaussian-type LWZ model Velocity (V x ' ) Velocity (V x ' ) Distance (y ' ) (b) Block-type LWZ Before earthquake After earthquake -8 -6 -4 -2 0 2 4 6 8 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 -5 5 4 3 2 1 0 -1 -2 -3 -4 -5 5 4 3 2 1 0 -1 -2 -3 -4 : Elastic model : LWZ (Ω B ' = 0.2) : LWZ (Ω B ' = 1.0) : LWZ (Ω B ' = 0.6) : LWZ (Ω B ' = 0.4) : LWZ (Ω B ' = 2.0) : Elastic model : LWZ (Ω B ' = 0.2) : LWZ (Ω B ' = 1.0) : LWZ (Ω B ' = 0.6) : LWZ (Ω B ' = 0.4) : LWZ (Ω B ' = 2.0) η w ' = 0.01 η ' = 1.0 Θ H ' = 0.3 η w ' = 0.01 η ' = 1.0 Θ H ' = 0.3 -8 -6 -4 -2 0 2 4 6 8 (a) Block-type LWZ T23E-2634 Fall AGU 2013
Transcript of Weak ductile shear zone beneath the western North ... · (a) after the earthquakes (Ergintav et al....
: Elastic model: LWZ (ΩG' = 0.1)
: LWZ (ΩG' = 1.0): LWZ (ΩG' = 0.6): LWZ (ΩG' = 0.2)
: LWZ (ΩG' = 2.0)
ηw' = 0.01η' = 1.0ΘG' = 0.3
Weak ductile shear zone beneath the western North Anatolian Fault Zone: inferences from earthquake cycle model constrained by geodetic observations
SUMMARY
1. AIM OF THIS STUDY
Tadashi Yamasaki, Tim J. Wright and Gregory A. HousemanCOMET+, School of Earth and Environment, University of Leeds, UK
3. THE EARTHQUAKE CYCLE MODEL
Proles of GPS velocities across the North Anatolian Fault, before and after the 1999 İzmit and Düzce earthquakes, require a localised weak zone in the mid crust beneath the western North Anatolian Fault Zone. The thickness and width of this low viscosity zone are ~ 12 km and ~ 20 - 40 km, respectively, and the eective viscosities of the weakened and the non-weakened domains are ~ 2 × 1018 and ~ ≥ 2 × 1020 Pa s, respectively.
Using a 3D nite element model, we examine the linear Maxwell visco-elastic response to a repeated strike-slip faulting event under the condition of a constant far-eld loading rate for three basic crustal viscosity models beneath an upper elastic layer: uniform viscosity (UNV), depth-dependent viscosity (DDV) and localised weak zone (LWZ). We compare predicted surface velocity proles with the GPS velocity proles across the North Anatolian Fault Zone (NAFZ), before and after the 1999 İzmit and Düzce earthquakes, in order to constrain the actual viscosity variation in the crust.
6. CONCLUSIONS
5. THE LWZ MODEL FOR THE NAFZ CASE
ACKNOWLEDGEMENTS:
REFERENCES:
This study is supported by the UK Natural Environment Research Council (NERC) under grant number NE/I028017/1, and with the support of the National Centre for Earth Observation (NCEO-COMET+). Cluster computing was provided by NCEO and by the University of Leeds Advanced Research Computing Node 1 (arc1).
Bürgmann et al., 2002, BSSA 92, 126-137; Ergintav et al. 2009, JGR 114, B07403, doi:10.1029/2008JB006021; Feigl et al., 2002, BSSA 92, 138-160; Klinger et al., 2003, BSSA 93, 2317-2332; Kozacı et al., 2009, JGR 114, B01405, doi:10.1029/2008JB005760; McClusky et al., 2000, JGR 105, 5695-5719; Reilinger et al., 2000, Science 289, 1519-1524; Reilinger et al., 2006, JGR 111, B05411, doi:10.1029/2005JB004051; Rockwell et al., 2009, Geol. Soc. Lon. Spec. Pub. 316, 31-54; Stein et al., 1997, GJI 128, 594-604; Vanacore et al., 2013, GJI, 193, 329-337; Tatar et al., 2012. Tectonophysics 518-521, 55-62; Wright et al., 2001, GRL 28, 2117-2120; Yamasaki & Houseman, 2012, GJI 190, 769-784.
4.1 Uniform viscosity (UNV) model
4.2 Depth-dependent viscosity (DDV) model
−12.2
−12.0
−11.8
−11.6
−11.4
−11.2
Dis
plac
emen
t (U
x' - t'
)
Time (t' )
−1.5
−1.0
−0.5
0.0
−1.5
−1.0
−0.5
0.0
−2.0
−1.5
−1.0
−0.5
0.0
−14
−12
−10
−8
−6
−4
−2
0
Dis
plac
emen
t (U
x' - t'
)
Dis
plac
emen
t (U
x' - t'
)
Dis
plac
emen
t (U
x' - t'
)
Time (t' )
Dis
plac
emen
t (U
x' - t'
)
Time (t' )
(a) UNV (η' = 0.01)
(c) UNV (η' = 1.0)
(b) UNV (η' = 0.1)
(d) UNV (η' = 10.0)
(x', y', z' ) = (0.0,-0.5,1.0)
(x', y', z' ) = (0.0,-0.5,1.0)
(x', y', z' ) = (0.0,-0.5,1.0)
(x', y', z' ) = (0.0,-0.5,1.0)
0 302 6 10 14 18 22 264 8 12 16 20 24 28 0 20020 60 100 140 18040 80 120 160
0 101 2 3 4 5 6 7 8 9 0 101 2 3 4 5 6 7 8 9
180 200182 184 194192190186 188 196 198
4.3 Localised weak zone (LWZ) model
4.1.1 The surface displacement with time
4.1.2 The surface velocity proles before and after earthquake
−30
−20
−10
0
10
20
30
−1.0−0.8−0.6−0.4−0.2
0.00.20.40.60.81.0
−8
−6
−4
−2
0
2
4
6
8
−1.0−0.8−0.6−0.4−0.2
0.00.20.40.60.81.0
−1.2−1.0−0.8−0.6−0.4−0.2
0.00.20.40.60.81.01.2
−1.0−0.8−0.6−0.4−0.2
0.00.20.40.60.81.0
Vel
ocit
y (V
x' )
Vel
ocit
y (V
x' )
Vel
ocit
y (V
x' )
Vel
ocit
y (V
x' )
(a)
(c)
(b)
(d)
DDV (ηb' = 0.001)After earthquake DDV (ηb' = 0.001)
Before earthquake
DDV (ηb' = 0.01)After earthquake DDV (ηb' = 0.01)
Before earthquake
: Elastic model: ηh' = 1000ηb'
: ηh' = 10000ηb'
: ηh' = 100000ηb'
Distance (y' ) Distance (y' )
Vel
ocit
y (V
x' )
Vel
ocit
y (V
x' )
(e) (f)
DDV (ηb' = 0.1)After earthquake
DDV (ηb' = 0.1)Before earthquake
-5 543210-1-2-3-4 -5 543210-1-2-3-4
-5 543210-1-2-3-4-5 543210-1-2-3-4
-5 543210-1-2-3-4 -5 543210-1-2-3-4
: Elastic model: ηh' = 10ηb'
: ηh' = 100ηb'
: ηh' = 1000ηb'
(a' ) Cross-section viewElastic layerH' = 0.3, D' = 0.4
Visco-elastic layer
z'
y'Vxb'-Vxb'
Θ'
Ω'
Sx'+Sx'-
(a'' ) Map viewy'
x'
Right-lateralstrike-slip fault
Sx'+
Sx'-
-Vxb'
Vxb'
Vxb' = Sx' / ∆t' = 1Sx' = 1
∆t' = 1
Dis
plac
emen
t
Fault
Boundary
Time (t' )
(b) Prescribed displacements
2. THE GEODETIC OBSERVATIONS
25˚
25˚
26˚
26˚
27˚
27˚
28˚
28˚
29˚
29˚
30˚
30˚
31˚
31˚
32˚
32˚
33˚
33˚
34˚
34˚
35˚
35˚
39˚ 39˚
40˚ 40˚
41˚ 41˚
42˚ 42˚
43˚ 43˚
Black Sea
Anatolia
North Anatolian Fault
Eurasia
Marmara Sea
Gulf of Saros
Agea
n Se
a
The İzmit earthquake(Mw = 7.5, 17 August 1999)
The Düzce earthquake(Mw = 7.2, 12 November 1999)
: GPS velocity pre. 1999
: GPS velocity Feb. 2000
The GPS velocity proles across the North Anatolian Fault Zone near the epicentres of the 1999 İzmit and Düzce earthquakes show:(a) after the earthquakes (Ergintav et al. 2009), the peak velocities either side of the fault dier by as much as 150 mm/yr, 5 or 10 times greater than the long-term relative displacement rate of ~ 15 to 30 mm/yr (Tatar et al., 2012; Reilinger et al., 2006).(b) before the earthquakes (McClusky et al., 2000), displacement rates vary monotonically across fault zone, but the velocity gradient is ~ 3 times higher within a zone of about 30 km width, centred on the fault.
We use a parallelized 3-D nite element code, oregano_ve (Yamasaki and Houseman, 2012), to solve the linear Maxwell visco-elastic problem in which an instantaneous strike-slip fault occurs at a periodic interval, releasing elastic strain accumulated by regional loading that occurs at a constant rate.
−16
−12
−8
−4
0
4
8
12
16
−2.0
−1.5
−1.0
−0.5
0.0
0.5
1.0
1.5
2.0
−1.0−0.8−0.6−0.4−0.2
0.00.20.40.60.81.0
−1.0−0.8−0.6−0.4−0.2
0.00.20.40.60.81.0
Vel
ocit
y (V
x' )
Vel
ocit
y (V
x' )
Vel
ocit
y (V
x' )
Distance (y' )
Vel
ocit
y (V
x' )
Distance (y' )
(a) UNV (η' = 0.01)
(c) UNV (η' = 1.0)
(b) UNV (η' = 0.1)
(d) UNV (η' = 10.0)
: Elastic model: After earthquake: Before earthquake
-5 543210-1-2-3-4
-5 543210-1-2-3-4
-5 543210-1-2-3-4
-5 543210-1-2-3-4
: Elastic model: After earthquake: Before earthquake
: Elastic model: After earthquake: Before earthquake
: Elastic model: After earthquake: Before earthquake
For the models with η’ = 0.01 and 0.1, a steady-state periodicity is established rapidly, implying that the stress relaxation is virtually completed by the end of the inter-seismic period. For the models with η’ = 1.0 and 10.0, only a smaller proportion of the deviatoric stress can be relaxed between seismic events, and the un-relaxed stresses evolve until the steady-state is established.
Some DDV models predict the general observed features. However, the pre-seismic velocity gradient across the fault is not adequately high, and the wavelength of the post-seismic velocity prole is inconsistent with the data.
A DDV model alone cannot explain the NAFZ velocity proles.
LWZ model predicts the general observed features of the both post- and pre-seismic velocities with consistent wavelength of the proles. Note that the Gaussian-function type of LWZ results in similar behaviour.
An LWZ model can explain the NAFZ velocity proles.
Symbol Meaning Value
TABLE 1: Parameter values used for the western North Anatolian Fault Zone (NAFZ)
*: Stein et al. (1997), Klinger et al. (2003) and Rockwell et al. (2009)#: Reilinger et al. (2000), Bürgmann et al. (2002) and Feigl et al. (2002)
$: Reilinger et al. (2000; 2006), Kozacı et al. (2007; 2009), Wright et al. (2001) and Tatar et al. (2012)
Shear modulus
Minimum viscosity in localised weak zone (LWZ)Background viscosity
Earthquake cycle periodDepth-extent of faultElastic layer thickness
Poisson’s ratio
Layer thickness3 x 1010
0.25
12 km*
16 km*
~ 200 yrs#
~ ≥ 2 x 1020 Pa s~ 1 - 2 x 1018 Pa s
40 kmµν
∆t
ηηw
Boundary velocitySlip rate
(Eective) thickness of LWZ(Eective) width of LWZ ~ 24 - 40 km
~ 12 km
~ 15 - 30 mm/yr$
~ 7.5 - 15 mm/yr
ΩH , ΩG
ΘH , ΘG
2Sx / ∆t
Vxb
H
D
L0
: Elastic model: ηh' = 1000ηb'
: ηh' = 10000ηb'
: ηh' = 100000ηb'
: Elastic model: ηh' = 10ηb'
: ηh' = 100ηb'
: ηh' = 1000ηb'
: Elastic model: ηh' = 10ηb'
: ηh' = 100ηb'
: ηh' = 1000ηb'
: Elastic model: ηh' = 10ηb'
: ηh' = 100ηb'
: ηh' = 1000ηb'
The high post-seismic displacement rate is predicted for UNV model with η’ = 0.01 or 0.1, but such a model does not predict the higher pre-seismic velocity gradient across the fault.
A UNV model cannot explain the NAFZ velocity proles.
The LWZ model prediction shows a good t to both pre- and post-seismic data for the LWZ conguration shown in Table 1.
Asymmetry of the observed proles is not included in these LWZ models.
Comparison with observations favours the block-type LWZ rather than the Gaussian-type LWZ.
Distance (y km)
(a) After the earthquakes
N S−100
−75
−50
−25
0
25
50
75
100
Vel
ocit
y (V
x mm
/yr)
Slip rate = ~ 15 - 30 mm/yr
Slip rate = ~ 15 - 30 mm/yr
-100 1007550250-75 -50 -25−20
−15
−10
−5
0
5
10
15
20
Vel
ocit
y (V
x mm
/yr)
Distance (y km)
(b) Before the earthquakes
N S
Slip rate = ~ 15 - 30 mm/yr
Slip rate = ~ 15 - 30 mm/yr
-100 1007550250-75 -50 -25
5.1 The Block-type of LWZ
5.2 The Gaussian-type of LWZ
Vel
ocit
y (V
x mm
/yr)
Distance (y km)
(b) Before the earthquakes
N S−20
−15
−10
−5
0
5
10
15
20
-100 1007550250-75 -50 -25
Vel
ocit
y (V
x mm
/yr)
Distance (y km)N S−120
−90
−60
−30
0
30
60
90
120
Slip rate = 22.5 ± 7.5 mm/yr
∆t = ~ 200 yrsη = ~ 2 × 1020 Pa s
ΘH = 12 kmH = 12 km, D = 16 km
-100 1007550250-75 -50 -25
-100 1007550250-75 -50 -25
Vel
ocit
y (V
x mm
/yr)
Distance (y km)
(a) After the earthquakes
Slip rate = 22.5 ± 7.5 mm/yr
∆t = ~ 200 yrs
η = ~ 2 × 1020 Pa sηw = ~ 2 × 1018 Pa s
N S−120
−90
−60
−30
0
30
60
90
120
ΘG = 12 kmΩG = 40 km
H = 12 km, D = 16 km
Vel
ocit
y (V
x mm
/yr)
Distance (y km)
(b) Before the earthquakes
Slip rate = 22.5 ± 7.5 mm/yr
∆t = ~ 200 yrs
η = ~ 2 × 1020 Pa sηw = ~ 2 × 1018 Pa s
N S−20
−15
−10
−5
0
5
10
15
20
ΘG = 12 kmΩG = 40 km
H = 12 km, D = 16 km
-100 1007550250-75 -50 -25
(a) After the earthquakes : ηw = ~ 2 × 1018 Pa s ΩH = 40 km: ηw = ~ 1018 Pa s ΩH = 24 km
: ηw = ~ 2 × 1018 Pa s ΩH = 40 km: ηw = ~ 1018 Pa s ΩH = 24 km
Slip rate = 22.5 ± 7.5 mm/yr
∆t = ~ 200 yrsη = ~ 2 × 1020 Pa s
ΘH = 12 kmH = 12 km, D = 16 km
XL' = XL'/L0 = 25.0
YL' = YL/L
0 = 10.0
x'
z'y'Right-lateralstrike-slip fault
Sx'+
Sx'-
O
1.0
(Elastic layer: η' = 1020)
(Visco-elastic layer: η' )-Vxb'
Weak Zone
(η w' )
Vxb'
H' = H/L0 = 0.3Ω
'
Θ'
(a)
D' = D/L0 = 0.4
L0 = 40 km, the thickness of the crust beneath the western North Anatolian Fault(Vanacore et al., 2013)
(1) The ratio τ/∆t of the Maxwell relaxation time τ to the earthquake cycle period ∆t controls the surface velocity variation during the earthquake cycle.(2) Post-seismic velocities near a major fault that are much greater than the long-term relative displacement rate indicate a relatively low Maxwell time constant (low viscosities), and strong localization of the pre-seismic strain rate near the fault requires a relatively high Maxwell time constant (high viscosities). (3) Of the models we tested, only the localized weak zone (LWZ) model is able to explain at once the GPS velocities before and after the 1999 İzmit and Düzce earthquakes. We infer a zone of low viscosity (~ 2 x 1018 Pa s) centred below the fault, relative to a general background crustal viscosity of at least 2 x 1020 Pa s. The low viscosity zone is ~ 12 km thick (below a 12 km thick elastic layer), and its width is between ~ 20 and 40 km.(4) A likely explanation for the block-type low viscosity zone is a slice of allocthonous crust about 20 km wide between northern and southern strands of the North Anatolian Fault in the İzmit region, though dierences in stress, grain-size, water content, or temeperature could also be signicant.
4. THE MODEL BEHAVIOUR: INFERENCES FOR THE NAFZ CASE
Vel
ocit
y (V
x' )V
eloc
ity
(Vx' )
Distance (y' )
(d) Gaussian-type LWZ Before earthquake
(c) Gaussian-type LWZ After earthquake
−1.0−0.8−0.6−0.4−0.2
0.00.20.40.60.81.0
-5 543210-1-2-3-4
-5 543210-1-2-3-4
: Elastic model: LWZ (ΩG' = 0.1)
: LWZ (ΩG' = 1.0): LWZ (ΩG' = 0.6): LWZ (ΩG' = 0.2)
: LWZ (ΩG' = 2.0)
ηw' = 0.01η' = 1.0ΘG' = 0.3
(c) Viscosity in UNV and DDV models
UN
V DDV
Viscosity (log10(η' ))
ηb'0.01ηb' 10ηb' 100ηb' 1000ηb'
Elastic layer (η' = 1020 )
Dep
th (z
' )
0.0
0.2
1.0
0.1
0.4
0.3
0.6
0.5
0.8
0.7
0.9
0.1ηb'
Distance from fault (y' ) 5.01.00.0 3.02.0 4.0
0.0
1.0
0.4
0.6
0.8
0.2
Dep
th (z
' )
−2
0
Log10 (η')
Elastic layer (η' = 1020 )
5.01.00.0 3.02.0 4.0
−2
0
Log10 (η')
0.0
1.0
0.4
0.6
0.8
0.2
Dep
th (z
' ) Elastic layer (η' = 1020 )
(d) Viscosity in block-type LWZ model
ΘB'
ΩB'
ΘG'
ΩG'
Distance from fault (y' )
(e) Viscosity in Gaussian-type LWZ model
Vel
ocit
y (V
x' )V
eloc
ity
(Vx' )
Distance (y' )
(b) Block-type LWZ Before earthquake
After earthquake
−8
−6
−4
−2
0
2
4
6
8
−1.0−0.8−0.6−0.4−0.2
0.00.20.40.60.81.0
-5 543210-1-2-3-4
-5 543210-1-2-3-4
: Elastic model: LWZ (ΩB' = 0.2)
: LWZ (ΩB' = 1.0): LWZ (ΩB' = 0.6): LWZ (ΩB' = 0.4)
: LWZ (ΩB' = 2.0)
: Elastic model: LWZ (ΩB' = 0.2)
: LWZ (ΩB' = 1.0): LWZ (ΩB' = 0.6): LWZ (ΩB' = 0.4)
: LWZ (ΩB' = 2.0)
ηw' = 0.01η' = 1.0ΘH' = 0.3
ηw' = 0.01η' = 1.0ΘH' = 0.3
−8
−6
−4
−2
0
2
4
6
8(a) Block-type LWZ
T23E-2634 Fall AGU 2013
![> r 45?H;CHIECI E0F04CG;0>OA35LIF;OAH8G>CA;HG0G r ... · [29 31] as well as experience data from sitting of nuclear reactor in-stallations where near- eld earthquakes of magnitude](https://static.fdocument.org/doc/165x107/5e7d18784afd9f356f6ddbc6/-r-45hchieci-e0f04cg0oa35lifoah8gcahg0g-r-29-31-as-well-as.jpg)
