Hashima, A.1, H. Sato 1, T. Ishiyama 1,A.M. Freed 2, T.W. Becker 3

1Earthquake Research Institute, University of Tokyo, 2Purdue University, 3University of Texas at Austin

EGU2020-18393Fault stressing in the overriding plate due to megathrust coupling along the Nankai trough, Japan

TS5.1Seismic analysis and geodetic modelling:multi-disciplinary approach to problem-solving

EGU 2020 Online4 May 2020, 16:15-18:00

Recent inland earthquakes around SW Japan

2020/5/4 Intra-plate stressing due to megathrust, SW Japan 2

Tectonics around Philippine Sea (Wu et al., 2016, JGR) 126˚ 128˚ 130˚ 132˚ 134˚ 136˚

28˚

30˚

32˚

34˚

36˚

2016 Mw7.1

1995

2000

2005

2015

Korea

Kyushu

Ryuk

yu

Trenc

h

Nankai Trough

Okinawa Trough

1946 M8.0

Active seismic activity of ~M7, especially around Kyushu since 1995

Inland earthquakes due to Nankai megathrust

2020/5/4 Intra-plate stressing due to megathrust, SW Japan 3

6.5

7.0

7.5

8.0

8.5

1700 1800 1900 2000

Nankai Yr−M

1707M8.6

1854M8.4

1944 M7.91946 M8.0Megathrust

Inland

Nankai Map

M<78>M>=7M>=8

128˚ 132˚ 136˚ 140˚

32˚

36˚

128˚ 132˚ 136˚ 140˚

32˚

36˚1605 ~ 1707

1620 1650 1680

M<78>M>=7M>=8

128˚ 132˚ 136˚ 140˚

32˚

36˚

128˚ 132˚ 136˚ 140˚

32˚

36˚1707 ~ 1854

1710 1740 1770 1800 1830

M<78>M>=7M>=8

128˚ 132˚ 136˚ 140˚

32˚

36˚

128˚ 132˚ 136˚ 140˚

32˚

36˚1854 ~ 1946

1860 1890 1920

M<78>M>=7M>=8

128˚ 132˚ 136˚ 140˚

32˚

36˚

128˚ 132˚ 136˚ 140˚

32˚

36˚1946 ~ 2020

1950 1980 2010Year

Nankai Map

M<78>M>=7M>=8

128˚ 132˚ 136˚ 140˚

32˚

36˚

128˚ 132˚ 136˚ 140˚

32˚

36˚1605 ~ 1707

1620 1650 1680

M<78>M>=7M>=8

128˚ 132˚ 136˚ 140˚

32˚

36˚

128˚ 132˚ 136˚ 140˚

32˚

36˚1707 ~ 1854

1710 1740 1770 1800 1830

M<78>M>=7M>=8

128˚ 132˚ 136˚ 140˚

32˚

36˚

128˚ 132˚ 136˚ 140˚

32˚

36˚1854 ~ 1946

1860 1890 1920

M<78>M>=7M>=8

128˚ 132˚ 136˚ 140˚

32˚

36˚

128˚ 132˚ 136˚ 140˚

32˚

36˚1946 ~ 2020

1950 1980 2010Year

Inland earthquakes tend to occur before megathrust earthquakes

1944, 1946 ???

Data from historical documents (Utsu, 1990, 2002, 2004) http://iisee.kenken.go.jp/utsu/index_eng.html

2016 Mw7.1

M-T diagram

Map View

Time in earthquake cycle

Motivation of this study

• Inland source faults are activated by the Nankai megathrust coupling? How the Ryukyu trench contributes?

Methods– Calculate the stressing rate on source faults by the

interplate coupling using 3-D finite element model

– interplate coupling under the Ryukyu-Nankai trenches is constrained by the GPS data

2020/5/4 Intra-plate stressing due to megathrust, SW Japan 4

This method can be used to determine the inland fault going close to rupture

4700 km

4600 km

700 km

PAC PHSEUR

PHSPAC

Tohoku-okiNankai

3D FEM around the Japanese islands

5

(Hashima et al., 2016, EPS; Freed et al., 2017, EPSL)

2020/5/4 Intra-plate stressing due to megathrust, SW Japan

View from NWEU removed

• Plate interface geometry based on Nakajima & Hasegawa (2006), Hayes et al. (2012), etc.

• 1000,000 Elements with the size of 5–100 km

Viscoelastic effect of asthenosphere (Li et al., 2015, 2018; Noda et al., 2018)

1019 Pa s 1019 Pa s

70 kmTe

*We use solution at t = 100 yr (perfectly relaxed)

Elastic thickness Te is critical to deformationWe examine cases of Te = 30, 50, and 80 km

(Flat layer)

125˚ 130˚ 135˚

25˚

30˚

35˚

5 cm/yr

data/disp_oct.dat

1998/1/1-2010/12/31, 404 stations

Reference

GPS Data and inland fault model

6

• Plate interface divided into 8×27 Patches• Slip response is computed for each patch for inversion

2020/5/4 Intra-plate stressing due to megathrust, SW Japan

Source fault modelGPS data (from GSI of Japan)

Slip angle is estimated to be the optimum direction by regional stress field (Terakawa & Matsu’ura, 2010)

Landward

Trenchward

Rotation

Results of slip-rate inversion for Te = 30 km

2020/5/4 Intra-plate stressing due to megathrust, SW Japan 7

Figure 4

120˚

125˚

125˚

130˚

130˚

135˚

135˚

140˚

25˚

35˚

40˚

Observed

Calculated

5 cm/yr

−10 −5 0 5 10

Slip (cm/yr)

Figure 4

120˚

125˚

125˚

130˚

130˚

135˚

135˚

140˚

25˚

35˚

40˚

Residual

2 cm/yr

−10 −5 0 5 10

Slip (cm/yr)

Te = 30 km (Comparison)

Te = 30 km (Residual)

(obs. - cal.)

Widespread residuals

Deficit Excess

Deficit Excess

5 cm/yr (Slip rates)

Results of slip-rate inversion for Te = 50 & 80 km

2020/5/4 Intra-plate stressing due to megathrust, SW Japan 8

Figure 4

120˚

125˚

125˚

130˚

130˚

135˚

135˚

140˚

25˚

35˚

40˚

Residual

2 cm/yr

−10 −5 0 5 10

Slip (cm/yr)

Figure 4

120˚

125˚

125˚

130˚

130˚

135˚

135˚

140˚

25˚

35˚

40˚

Residual

2 cm/yr

−10 −5 0 5 10

Slip (cm/yr)

Fitting too much!

Te = 50 km (Residual)

Te = 80 km (Residual)

Deficit Excess

Deficit Excess

Opposite slip at depth

5 cm/yr (Slip rates)

5 cm/yr (Slip rates)

The optimum model (Te = 50 km)

2020/5/4 Intra-plate stressing due to megathrust, SW Japan 9

Figure 4

120˚

125˚

125˚

130˚

130˚

135˚

135˚

140˚

25˚

35˚

40˚

Residual

2 cm/yr

−10 −5 0 5 10

Slip (cm/yr)

Residuals might be due to structural anomalies

Edge effect

Slip-rate deficit ~ 8 cm/yr(Interplate locking)

Consistent with previous studies (e.g. Ito et al., 1999; Yokota et al., 2016; Noda et al., 2018)

Slip-rate excess 3~ 6 cm/yr

Direction of slip-rate excess is deviated from the EUR-PHS relative motion, which represents different characteristics from relative plate motion

Related to slab rollback?

Te = 50 km (Residual)

Deficit Excess

5 cm/yr (Slip rates)

102020/5/4 Intra-plate stressing due to megathrust, SW Japan

(extension)(contraction)(σ11+ σ22 +σ33)/3

NS extension

NW-SE compression

Te = 50 km

Intra-plate stressing rate for optimum modelStress at depth of 10 km

Coulomb stress rates on inland faults

112020/5/4 Intra-plate stressing due to megathrust, SW Japan

promotionsuppression

2005 Mw6.6

2015 Mw6.7

2016 Mw7.1

Coulomb StressΔCFS = σn+ µσs (µ = 0.4)

Three recent ~M7 earthquakes occurred on the faults of positive ΔCFF

Summary

• Using 3-D FEM, we inverted GPS data in the Ryukyu-SW Japan arcs to obtain slip-rate excess/deficit under the Ryukyu-Nankai trenches– Slip-rate excess under the Ryukyu trench might be due to

slab rollback– Slip rates under the Ryukyu trench is critical to the

deformation on the Kyushu island, junction of the Ryukyu-SW Japan arcs

• Using slip-rate distribution, we computed stressing rate on the inland faults. In particular, source faults of the recent ~M7 earthquakes shows positive Coulomb stressing rates

122020/5/4 Intra-plate stressing due to megathrust, SW Japan

For more details, please mail me to hashima@eri.u-tokyo.ac.jp