-25-20-15-10

-505

10152025

Ve

loci

ty (

mm

/yr)

-200 -150 -100 -50 0 50 100 150 200

Distance from fault trace (km)

3yrs

KANRAKCO UCG2PIREKAZI

SISL

SMASMURT

KTOPSEFIAGUZMEKEIGAZIUCKTEBASEYHCMLN

MHGZ ESKIAGOK

KFKT

KDER

50

60

70

80

90

100

deep

slo

w s

lip

0 5 10 15 20shallow locked slip rate

-20

-15

-10

-5

0

de

pth

(km

)

-20 0 20 40 60slip rate (mm/yr)

fault slip rate

-20

-15

-10

-5

0

5

10

15

20

Ve

loci

ty (

mm

/yr)

-200 -150 -100 -50 0 50 100 150 200

Distance from fault trace (km)

inter

SILE

GATE

YUHETUBI

SISL

SMASOLUK

MEKEIGAZIUCK

DERBHMZAKPKL

0

5

10

15

20

de

ep

lock

ed s

lip r

ate

0 5 10 15 20shallow locked slip rate

-20

-15

-10

-5

0

de

pth

(km

)

-30 -25 -20 -15 -10 -5 0slip rate (mm/yr)

fault slip rate

−1500

−1000

−500

0

500

1000

1500

Slip

(m

m)

−200 −150 −100 −50 0 50 100 150 200

Distance from fault trace (km)

cos

IZ19IZ23

IZ16

IZ11

IZ20

IZ22IZ25IZ24IZ12IZ18

IZ09IZ17IZ10

IZ15IZ14IZ13

−20

−15

−10

−5

0

dep

th (

km)

0 1000 2000 3000 4000slip (mm)

fault slip vs. depth

-40-30-20-10

010203040

Ve

loci

ty (

mm

/yr)

-200 -150 -100 -50 0 50 100 150 200

Distance from fault trace (km)

1yr

KANRAKCO

UCG2

PIREKAZISISL

SMASMURT

KTOPSEFIAGUZ

MEKEIGAZIUCKTEBASEYH

CMLNMHGZ

ESKI

AGOK

KFKT

KDER

200

220

240

260

280

300

deep s

low

slip

0 5 10 15 20shallow locked slip rate

-20

-15

-10

-5

0

depth

(km

)

0 50 100 150 200 250slip rate (mm/yr)

fault slip rate

FAULT CREEP AND PERSISTENT ASPERITIES ON THE WESTERN SECTION OF THE NORTH ANATOLIAN FAULT, TURKEYRobert Reilinger (1), Semih Ergintav (2), Hayrullah Karabulut (2), Philippe Vernant (3), Michael Floyd (1), Ozgun Konca (2), Ugur Dogan (4), Seda Cetin (4), Ziyadin Cakir (5), David Mencin (6), Roger Bilham (6), Robert King (1)

 (1) MIT, DEAPS, Cambridge, USA; (2) Bogazici Univ., KOERI, Istanbul, Turkey; (3) U. of Montpellier II, Geosciences, France; (4) Yildiz Tech. U. Geomatic Engineering, Istanbul, Turkey; (5) ITU, Dept. of Geology, Istanbul, Turkey; (6) U. of Col., Geol. Sci., Boulder, USA

-20

-15

-10

-5

0

5

10

15

20

Velo

city

(m

m/y

r)

-200 -150 -100 -50 0 50 100 150 200

Distance from fault trace (km)

inter

SILE

GATE

YUHETUBI

SISL

SMASOLUK

MEKEIGAZIUCK

DERBHMZAKPKL

0

5

10

15

20

deep lo

cked s

lip r

ate

0 5 10 15 20shallow locked slip rate

-20

-15

-10

-5

0

depth

(km

)

-30 -25 -20 -15 -10 -5 0slip rate (mm/yr)

fault slip rate

−1500

−1000

−500

0

500

1000

1500

Slip

(m

m)

−200 −150 −100 −50 0 50 100 150 200

Distance from fault trace (km)

cos

IZ19IZ23

IZ16

IZ11

IZ20

IZ22IZ25IZ24IZ12IZ18

IZ09IZ17IZ10

IZ15IZ14IZ13

−20

−15

−10

−5

0

de

pth

(km

)

0 1000 2000 3000 4000slip (mm)

fault slip vs. depth

-40-30-20-10

010203040

Ve

loci

ty (

mm

/yr)

-200 -150 -100 -50 0 50 100 150 200

Distance from fault trace (km)

1yr

KANRAKCO

UCG2

PIREKAZISISL

SMASMURT

KTOPSEFIAGUZ

MEKEIGAZIUCKTEBASEYH

CMLNMHGZ

ESKI

AGOK

KFKT

KDER

200

220

240

260

280

300

deep s

low

slip

0 5 10 15 20shallow locked slip rate

-20

-15

-10

-5

0

depth

(km

)

0 50 100 150 200 250slip rate (mm/yr)

fault slip rate

-25-20-15-10-505

10152025

Velo

city

(m

m/y

r)

-200 -150 -100 -50 0 50 100 150 200

Distance from fault trace (km)

3yrs

KANRAKCO UCG2PIREKAZI

SISL

SMASMURT

KTOPSEFIAGUZMEKEIGAZIUCKTEBASEYHCMLN

MHGZ ESKIAGOK

KFKT

KDER

50

60

70

80

90

100

deep s

low

slip

0 5 10 15 20shallow locked slip rate

-20

-15

-10

-5

0

depth

(km

)

-20 0 20 40 60slip rate (mm/yr)

fault slip rate

-20-15-10-505

101520

Ve

loci

ty (

mm

/yr)

-200 -150 -100 -50 0 50 100 150 200

Distance from fault trace (km)

6.5yrs

KANRAKCO CALTUCG2

KAZI

SMAS

SEFIAGUZ

DGCTIGAZTEBASEYH CMLNMHGZ ESKI

AGOK

KFKT

KDER

0

10

20

30

40

50

de

ep

slo

w s

lip

0 5 10 15 20shallow locked slip rate

-20

-15

-10

-5

0

de

pth

(km

)

-30 -20 -10 0 10 20 30slip rate (mm/yr)

fault slip rate

Seismic segmentation of the NAF 3-layer elastic model for Izmit earthquake cycle

Figure 12 Left column: GPS velocity profiles across the North Anatolian Fault around the Izmit earthquake rupture for (a) pre-earthquake velocities, (b) coseismic displacements, (c) velocities 1 year after the earthquake, (d) velocities 3 years after the earthquake and (e) velocities 6.5 years after the earthquake. Blue line represents the model shown in the panel to the right. Right column: Slip rate models to fit velocity profiles consist of a middle layer (2–8 km depth) fully locked interseismically and upper and lower layers allowed to creep at a rate estimated from the data (Vernant et al., in prep.).

2. 1912 Ganos M7.4 earthquake segment

3. Marmara Sea segments

Figure 2 Yellow lines are active faults. Red trace the estimated surface break of the 1912 Earthquake. Seismic activity before the 2014 North Aegean Earthquake are black circles and aftershocks blue circles. The focal mechanisms of significant earthquakes (M>4.7) since 1975 – red the 2014 event.

1. 2014 Gokceada earthquake (Konca et al., 2018)

Figure 4 (a) Slip distribution obtained from modeling of GPS data. (Bottom) Slip distribution obtained from GPS and seismic data. (b) (Top) Coseismic slip distribution and the seismicity before the main shock (projected onto the fault. (Bottom) Same for aftershocks. Seismicity and aftershocks lie outside coseismic  slip patches.

1.6

1.2

0.8

0.4

0.0

faul

t slip

, mm

11/25/17 11/29/17 12/3/17

Tepetarla Extensometer

1/1/15 1/1/16 1/1/17

Izmit Extensometer(100 mm range)

43.543.042.542.041.541.040.5

Faul

t sli

p, m

m (10 mm range, 3 µm resolution)40.72097°N, 30.0784E

40.7211°N , 29.9465°E

1 year 1 week

Figure 11 Postseismic GPS velocities at 6 years following the eq corrected for interseismic deformation attest to broad deformation, suggesting deep active fault creep (from Ergintav et al., 2009, Figure 6c).

SUMMARY AND CONCLUSIONS—New, precise geodetic and seismological observations allow mapping variations in coupling on the western North Anatolian Fault. —Variations in fault coupling have a first-order effect on the earthquake cycle.—Seismicity appears to be directly associated with areas of low coseismic slip that we infer reflect areas of interseismic fault creep.—Models with rapid, post-earthquake fault healing in mid-layer and post-earthquake creep in top and bottom layers can account for eq-cycle deformation for the western North Anatolian Fault.—Long-term, deep post-earthquake fault slip may contribute to earthquake interaction/propagation on the North Anatolian Fault, including deep afterslip from Izmit triggering the Duzce earthquake (Reilinger et al., 2000; Burgmann et al., 2002).

Distance perpendicular to North Anatolian Fault (km)

-20-15-10-505

101520

Ve

loci

ty (

mm

/yr)

-200 -150 -100 -50 0 50 100 150 200

Distance from fault trace (km)

6.5yrs

KANRAKCO CALTUCG2

KAZI

SMAS

SEFIAGUZ

DGCTIGAZTEBASEYH CMLNMHGZ ESKI

AGOK

KFKT

KDER

0

10

20

30

40

50

deep s

low

slip

0 5 10 15 20shallow locked slip rate

-20

-15

-10

-5

0

depth

(km

)

-30 -20 -10 0 10 20 30slip rate (mm/yr)

fault slip rate

Slip ratedeficit

Coseismicslip

Slip ratedeficit

Accelerated slip rate

Accelerated slip rate

Slip ratedeficit

Accelerated slip rate

Slip ratedeficit

Figure 3 Coseismic offsets (black vectors) with 1-σ error ellipses. Red vectors horizontal displacements predicted from the joint slip model shown below. The map view of depth-averaged slip is shown by the blue curve. The blue rectangle shows the surface projection of the finite-fault model.

Figure 5 Currently aseismic 1912 earthquake segment is fully locked according to geodetic data. Black bracket is GPS velocity profile at right (from Ergintav et al., 2014, Figure 2d).

Figure 8 Pre-Izmit earthquake seismicity on fault plane; aseismic mid-crust? (Bohnhoff et al., 2016, Figure 5)

Figure 6 (Top) Marmara Sea seismicity. Black brackets are GPS profiles to right using the velocity solution of Ergintav et al. (2014). (Bottom) Seismicity projected on Main Marmara Fault (Schmittbuhl et al., 2016, Figure 2).

REFERENCESBohnhoff et al. (2016), doi:10.1016/j.tecto.2016.07.029Burgmann et al., (2002), doi:10.1785/0120000833Cakir et al. (2012), doi:10.1130/G33522.1Ergintav et al. (2009), doi:10.1029/2008JB006021Ergintav et al. (2014), doi:10.1002/2014GL060985Feigl et al. (2002), doi:10.1785/0120000830Konca et al. (in revision), Geophys. J. Int.Reilinger et al. (2000), doi:10.1126/science.289.5484.1519Schmittbuhl et al. (2016), doi:10.1002/2015GC006120 Vernant et al. (in prep.)

Figure 1 Overview of significant sections of the North Anatolian Fault, with historical observations showing varying behaviors over the seismic cycle. Here we synthesize geodetic and seismic observations from the four westernmost segments.

(a) Interseismic

(b) Coseismic

2014 Gokceada eqk

Ganos segment

Marmara Sea

1999 Izmit eqk

Ismetpasa

Eastern North Anatolian Fault

Figure 5

Anatolia

B l a c k S e a

Figure 7 GPS fault-xing profiles (Top) Princes Islands segment shows strain accumulation. (Bottom) Kumburgaz Basin segment unclear, note different scales (Ergintav et al., 2014, Figures 2b and 2c).

4. 1999, M7.6 Izmit/M7.2 Duzce sequence

Figure 10 Fault creep continues on the section of the North Anatolian Fault that ruptured during the 1999 Izmit earthquake. Top: InSAR observations (2002–2009) of Izmit post-earthquake surface creep from Cakir et al. (2012, Figure 3). Circle shows creepmeter location (time series below). Bottom: 2015 (left) and 2017 surface creep events recorded at 2 locations on the Izmit coseismic fault attest to ongoing surface creep.

Figure 9 Izmit coseismic right-lateral slip from inversion of GPS, InSAR (ERS and RADARSAT), and SPOT correlation map. Maximum coseismic slip between 2-12 km depth.

-6

-12

-18

Dept

h

Abstract EGU2018-17354

(c) After 1 year

(d) After 3 years

(e) After 6.5 years

c

b

D

d

C B

Slip ratedeficit

European Geosciences UnionGeneral Assembly 2018

Vienna | Austria | 8–13 April 2018