Source Process and Constitutive Relations of the 2011

Tohoku Earthquake Inferred from Near-Field

Kunikazu Yoshida, Anatoly Petukhin & Ken Miyakoshi

Koji Hada
NEWJEC, Japan

Atsushi Okazaki
The Kansai Electric Power Company, Japan

SUMMARY:
The spatial and temporal distribution of the stress on the fault plane of the Mw9.0 2011 Tohoku, Japan,
earthquake is calculated from kinematic inversion results. We have investigated the rupture process of the
earthquake using the long-period strong-ground motion data. We calculate the stress changes using the
finite-difference method to infer the constitutive relations and dynamic source parameters. The static stress drop
at the large slip area along the trench, which is identified as an asperity, is about 25 MPa. Strength excess at the
deeper margin of the asperity is about 3 MPa, although strength excesses on the most of the fault are generally
small. The slip-weakening distances (Dc) obtained from the constitutive relations around the asperity are about 7
to 10 m. Dc of the asperity is unclear from the constitutive relations due to the complexity, but Dc obtained from
peak slip velocities indicate about 15 to 22 m.

Keywords: 2011 Tohoku earthquake, Source process, constitutive relations

1. INTRODUCTION

On 11 March 2011, the Tohoku earthquake of MW9.0 (Japan Meteorological Agency) struck the
Tohoku region, Japan, and generated a huge tsunami. The 2011 Tohoku earthquake (the 2011 Off the
Pacific Coast of Tohoku earthquake named by JMA) is the first M9 earthquake that is recorded by a
lot of near-field strong-motion stations. During historical M9-class earthquakes (e.g., 2004 Sumatra,
1964 Alaska, 1960 Chile), few strong-motion seismometers near the source region were operated.
Many kinematic source models are proposed using the strong-ground motion records for this
earthquake.

A constitutive relation is a key element of dynamic descriptions of the seismic source which are based
on models that satisfy the elastodynamics equation. Some dynamic source parameters are estimated
from the constitutive relations. Many researchers investigated the constitutive relations of natural
earthquakes (e.g, Ide and Takeo, 1997; Zhang et al., 2003; Hok and Fukuyama, 2011). The
constitutive relations are estimated with combining the slip evolution and stress evolution at each
location on a fault plane. A spatio-temporal stress change can be estimated by solving the
elastodynamic equation with a kinematic slip history estimated by an inversion analysis as boundary
conditions.

The 2011 Tohoku earthquake is the first M9 event that is proposed many detailed kinematic source
models. In this study, we will use a 3-D finite-difference method (FDM) proposed by Pitarka (1999)
under appropriate boundary conditions to directly solve the elastodynamic equations and calculate the
stress spatiotemporal distribution over a fault plane from the kinematic inversion result estimated by
Yoshida et al. (2011). We apply this approach to study the dynamic source parameters of the 2011
Tohoku earthquake.
2. SOURCE MODEL

We have investigated the rupture process of the 2011 Tohoku earthquake by the multi-time-window
linear waveform inversion method using the long-period strong-ground motion data. Detail of the
analysis is denoted in Yoshida et al. (2011).

We used the strong-motion data obtained from 37 stations of F-net and KiK-net (NIED) (Aoi et al.,
2004; Okada et al., 2004), and 1 station of Hokkaido University (MYR, Sasatani et al., 2002) for the
inversion (Fig. 2.1). These stations are located on relatively hard rock. The data were windowed for
300 sec, starting at P-wave arrival time, and band-pass-filtered between periods of 200 s to 20 s
(0.005–0.05 Hz) for waveform inversion.

We assumed a single planar fault model of 468 km long in strike and 228 km wide in dip. We assume
N193°E and 10° as the strike and dip angles, respectively, referring to the JMA CMT solution. The
rupture starting point is located at the hypocenter determined by JMA: 38.103°N, 142.861°E, 23.7 km.
Theoretical Green’s functions are calculated using the discrete wavenumber method (Bouchon, 1981)
and the Reflection/Transmission coefficient matrix method (Kennett and Kerry, 1979) using a
stratified medium. The velocity structure model proposed by Wu et al. (2008, 2009) for the waveform
inversion of the 1979 and 2005 Miyagi-oki earthquake was used for calculating the theoretical Green’s
functions. We use multi-time-window linear waveform inversion procedure (e.g., Hartzell and Heaton,
1983) in which the moment-release distribution is discretized in both space and time. For
discretization in space, we divide the fault plane into 39 in the strike direction and into 19 in the
down-dip direction (making a total of 741 subfaults with an area of 12 km x 12 km). We use 8
smoothed ramp functions with duration of 16 seconds separated by 8 seconds’ interval to represent the
slip history of each subfault.

The seismic moment of this earthquake was estimated to be 4.3 x 1022 Nm (Mw 9.0). The inverted slip
distribution shows a large slip area with a maximum slip of about 47 m which is located on the
shallower part of the fault plane (Fig. 2.1). The shallow large slip area is identified as an asperity with
the criteria of Somerville et al. (1999) (Fig. 2.2A). The moment distribution is dominated by the
asperity (moment released area) in the shallower part of the fault plane. However, the peak
moment-rate distribution indicates that high peak moment-rate areas extend to the deeper part of the
fault plane (Fig. 2.2B). The feature of the ground motion suggested from the record section is well
represented by the peak moment rate distributions, rather than the slip distributions. This source model
is successfully validated by tsunami simulation (Petukhin et al., 2012).
Figure 2.1. Projection of the final slip distribution on the map. The star and squares indicate the rupture starting
point and the strong-motion stations that are used for the inversion. Dotted line shows plate boundaries (Bird,
2003).

Figure 2.2. Upper: Distributions of coseismic slip obtained for the 2011 Tohoku earthquake. The star indicates
the rupture starting point. The purple dashed box indicates the asperity identified with the criteria of Somerville
et al. (1999). Lower: Distributions of peak moment rate together with the slip rate function at each subfault.
3. CALCULATION METHOD FOR STRESS CHANGES

To determinate stress changes on the fault plane from the slip spatiotemporal distribution obtained by
kinematic inversion, we used a 3-D finite difference method (FDM) presented by Pitarka (1999) to
solve the following elastodynamic equations with appropriate boundary condition. In this method, the
fourth-order staggered grid expressions for discrete spatial differential operators and the second-order
approximate for time derivatives are used to model wave propagation in 3-D elastic media.

In the FDM coordinate system, we adopt x axis is the direction N103°E, y axis is the direction N193°E,
and z axis is direction down (Fig. 3.1). During the calculation, we used four boundary conditions. The
first set of boundary conditions is the free surface condition, which is the zero-stress formulation
(Graves, 1996). The second set of boundary conditions is the continuity condition at each layer
interface included in the medium, that is, the continuity of stresses and displacements. The third set of
boundary conditions is an absorbing boundary condition (Clayton and Engquist, 1977) for the side and
bottom boundaries that attenuates elastic waves reflected back from these artificial side and bottom
boundaries of the model space. The fourth set of boundary conditions is for the points on the fault
plane. For these points, we used the kinematic inversion results as the fault plane boundary conditions.
The kinematic inversion results gave us the slip distributions on each subfault. Each subfault is like a
point source with arbitrary focal mechanism; the moment tensor components can be represented as an
equivalent distribution of body force couples centred at the grid point corresponding to each subfault.
We used the moment tensor source formulation using stress components proposed by Pitarka (1999).

The target event has an oblique fault plane. Stress is calculated under the Cartesian coordinates of the
finite-differential calculation. Stresses along dip and fault normal directions are taken by
transformation of the stress tensor calculated by FDM on the Cartesian coordinates (Fig. 3.1).

Figure 3.1. Definition of the fault orientation parameters and slip direction.

4. CONSTITUTIVE RELATIONS AND DYNAMIC PARAMETERS OF THE 2011 TOHOKU

The kinematic inversion indicates that the fault rupture process during this earthquake is very complex.
In this study, we used the kinematic inversion result (Yoshida et al., 2011) to analyse the stress
distribution on the fault. The projection of the fault plane onto the surface is 468 km in length and
224.5 km in width. The fault is divided into 39 x 19 subfaults, and each subfault occupies a 12 km x
12 km area on the fault plane. The crustal structure is the same as the one used in the Green’s function
calculation for the kinematic inversion (Wu et al., 2008, 2009). In order to represent the fault
geometry whose dip angle is 10°, the grid sizes used in our FDM calculation were chosen as Δx = 11.8
km, Δy = 12 km, and Δz = 2.08 km. The top of the fault plane for the FD calculation is aligned to the
surface, although the top of the fault plane for the kinematic inversion is the depth of 8.1 km, because
some direct observations of the sea bottom investigate that the faulting clearly reached to the sea
bottom (Ito et al., 2011). This modification affects the stress field near the top of the fault. The static
stress drop on the fault plane used in the original kinematic inversion exceeds 40 MPa on the large slip
area on the shallower part of the plane, but differences in most area of the fault between the two
models are very little.
 Figure 4.1. Snapshots of the slip rate 𝑣𝑑 and shear stress τ𝑑𝑛 in the dip direction. The history of 𝑣𝑑 is
determined by the inversion analysis, and that of τ𝑑𝑛 is calculated by FDM. The star indicates the rupture
starting point. In the FDM calculation, we used a oblique fault plane whose top depth is aligned to the surface.

Figure 4.1 shows the spatial and temporal distribution of the slip rate and shear stress on the fault
plane. Although three components of the slip vector and six components of the stress tensor are
determined everywhere on the fault plane, we show only shear stress along the dip direction τ𝑑𝑛 and
slip rate 𝑣𝑑 since the rupture of this earthquake is dominated by dip slip. Stress increases of less than
5 MPa appears around a rupture front area.

Figure 4.2 shows the spatial distribution of static stress changes (static stress drop). The distribution of
the stress drop is very complex, and is similar to the distribution of the peak moment rate. The
maximum stress drop in τ𝑑𝑛 is about 25 MPa at the region of large slip area. This maximum stress
drop is relatively high comparing to those for other plate-boundary earthquake, e.g. about 10 MPa
(Yagi, 2004), but it seems to be reasonable due to large slip of about 47 m at the asperity. Moderate
stress drop (~ 10 MPa) of the deeper part of the initial rupture point corresponds to the high peak
moment rate area on the deeper part of the fault plane. Small stress increases of less than 10 MPa
appear in wide area of the deeper part of the fault plane. Fig. 4.3 shows the spatial distribution of
dynamic stress drop. The spatial distribution of the dynamic stress drop is very similar with that of the
static stress drop.

Figure 4.4 shows the distribution of strength excess on the fault. The deeper margin of the asperity has
rather high strength (about 3 MPa). This relatively high strength excess suggests to cause that the
rupture progression suspended before the rupture of the asperity. The stress increase just before the
rupture of the asperity (at the frame of 40 s in Fig. 4.1) is recognized. Distribution of the high strength
area corresponds to the low peak moment rate, small total slip area (Figs. 2.2 and 4.4). Variety of the
strength suggests causing the complex slip distribution and the rupture progression.
Figure 4.2. Spatial distribution of stress change (static stress drop) for the 2011 Tohoku earthquake. The star
indicates the rupture starting point. Depth of the top of the fault plane is aligned to the surface.

Figure 4.3. Spatial distribution of dynamic stress drop for the 2011 Tohoku earthquake. The star indicates the
rupture starting point. Depth of the top of the fault plane is aligned to the surface.
Figure 4.4. Spatial distribution of strength excess for the 2011 Tohoku earthquake. The star indicates the rupture
starting point. Depth of the top of the fault plane is aligned to the surface.

The distribution of slip and stress can be used to give the constitutive relation at each grid point in
space. Figure 4.5 shows the relations plotted at each time step up to 200 s after the rupture initiation.
At almost point, stress decreases with increasing slip (slip weakening) as seen in frictional
experiments of rocks. Obviously, the slip weakening rate (du𝑑 /dτ𝑑𝑛 ) in the large slip point (A) is
significantly smaller than those of the other parts.

The relations between slip velocity and stress (right panel of Fig. 4.5) indicate that velocity weakening
behaviour is not unclear. This figure also shows that slips repeat twice (B, C, D, F) or three-time (E),
except for the point A. Slip at these points stopped (slip velocity dropped to zero) once, and then
restarted. These repetitions of the slip correspond to the three-stage rupture obtained from the
kinematic inversion (Yoshida et al., 2011).

Another important parameter in the slip-weakening model is critical slip-weakening distance Dc. Dc is
defined as the slip displacement required for the local strength in the breakdown zone behind the
rupture front to degrade to a residual friction stress level. The constitutive relations indicate the
slip-weakening distances (Dc) are about 7 m for point C, and about 5 m for point D. However, these
slip-weakening distances from the constitutive relations are upper bound because of the limitation due
to the resolution of the kinematic inversion. For point A, B, E and F, Dc are not identified clearly due
to the complex behaviour of the constitutive relations.

Mikumo et al. (2003) presented an approach to estimate the critical slip-weakening distance Dc based
on the relation between the breakdown time of shear stress, 𝑇𝑏 , and the time of peak slip velocity, 𝑇𝑝𝑣 .
Figure 4.6 show the spatial distribution of Dc estimated based on peak slip velocity. The Dc for point
C and D are 9.6 m and 6.9 m, respectively. At these points both Dc obtained from the constitutive
relations and 𝑇𝑝𝑣 are similar results. The Dc obtained from 𝑇𝑝𝑣 on the asperity (points A and B) are
20.4 m and 14.8 m, respectively. These values are very large, and are limited by the inversion
resolution, so that the validity of these values should be checked.
 Figure 4.5. (Left)The constitutive relation between slip and stress on the fault plane. (Right) The relation
between slip rate and stress. Each trace is the function calculated at the corresponding location in the bottom
figure of slip distribution. Depth of the top of the fault plane is aligned to the surface.

Figure 4.6. Spatial distribution of the estimated Dc values on the fault of the 2011 Tohoku earthquake obtained
with 𝑇𝑝𝑣 (Mikumo et al., 2003). Depth of the top of the fault plane is aligned to the surface.

5. SUMMARY

Stress evolution is estimated from the kinematic inversion results on the entire plane of the 2011
Tohoku earthquake fault. The relations between stress and slip, and between stress and slip rate on the
fault, show that most locations on the fault plane followed the slip-weakening model during the
rupture. The static stress drop at the large slip area along the trench is about 25 MPa. Dynamic
strength excess are similar to the static stress drop. Strength excess at the bottom of the large slip areas
is about 3 MPa, although strength excesses on the most of the fault are generally small. The
constitutive relations indicated that the slip weakening rate in the large slip point is significantly
smaller than those of the other parts. The slip-weakening distances (Dc) around the asperity obtained
from the constitutive relations are about 7 to 10 m. Dc of the asperity is not clear from the constitutive
relations, but Dc obtained from the peak slip velocity indicates about 15 to 22 m.

AKCNOWLEDGEMENT
We thank the National Research Institute for Earth Science Disaster Prevention and Hokkaido University to
provide the strong-motion data.

REFERENCES

Aoi, S., Kunugi, T., and Fujiwara, H. (2004). Strong-motion seismograph network operated by NIED: K-NET
and KiK-net, J. Jpn. Assoc. Earthq. Eng., 4, 65–74.
Bird, P. (2003). An updated digital model of plate boundaries, Geochem. Geophys. Geosyst., 4, 1027.
Bouchon, M. (1981). A Simple Method to Calculate Green’s Functions for Elastic Layered Media, Bull. Seism.
Soc. Am., 71, 959–971.
Clayton, R. and Engquist, B. (1977). Absorbing Boundary Conditions for Acoustic and Elastic Wave Equations,
Bull. Seism. Soc. Am., 67, 1529–1540.
Graves, R. W. (1996). Simulating Seismic Wave Propagation in 3D Elastic Media Using Staggered-Grid Finite
Differences, Bull. Seism. Soc. Am., 86, 1091–1106.
Hartzell, S. H. and Heaton, T. H. (1983). Inversion of Strong Ground Motion and Teleseismic Waveform Data
for the Fault Rupture History of the 1979 Imperial Valley, California, Earthquake, Bull. Seism. Soc. Am., 73,
1553–1583.
Hok, S. and Fukuyama, E. (2011). A new BIEM for rupture dynamics in half-space and its application to the
2008 Iwate-Miyagi Nairiku earthquake, Geophys. Jour. Inter., 184, 301–324.
Ide, S. and Takeo, M. (1997). Determination of constitutive relations of fault slip based on seismic wave analysis,
J. Geophys. Res., 102, 27379–27391.
Ito, Y., Tsuji, T., Osada, Y., Kido, M., Inazu, D., Hayashi, Y., Tsushima, H., Hino, R., and Fujimoto, H. (2011).
Frontal wedge deformation near the source region of the 2011 Tohoku-Oki earthquake, Geophys. Res. Lett.,
38, L00G05.
Kennett, B. L. N. and Kerry, N. J. (1979). Seismic waves in a stratified half space, Geophys. J. R. Astr. Soc., 57,
557–583.
Mikumo, T., Olsen, K. B., Fukuyama, E., and Yagi, Y. (2003). Stress-Breakdown Time and Slip-Weakening
Distance Inferred from Slip-Velocity Functions on Earthquake Faults, Bull. Seism. Soc. Am., 93, 264–282.
Okada, Y., Kasahara, K., Hori, S., Obara, K., Sekiguchi, S., Fujiwara, H., and Yamamoto, A. (2004). Recent
progress of seismic observation networks in Japan –Hi-net, F-net, K-NET, KiKnet–, Earth Planets Space,
56, xv–xxviii.
Petukhin, A., Yoshida, K., Miyakoshi, K., and Irikura, K., (2012). Tsunami simulation for the 2011 Great
Tohoku earthquake (Mw9.0), Japan, using seismic inversion source model and fully nonlinear tsunami
model. 15th World Conference on Earthquake Engineering.
Pitarka, A. (1999). 3D Elastic Finite-Difference Modeling of Seismic Motion Using Staggered Grids with
Nonuniform Spacing, Bull. Seism. Soc. Am., 89, 54–68.
Sasatani, T., Maeda, T., Yoshida, K., Morikawa, N., Ichiyanagi, M., Motoya, Y., and Kasahara, M. (2002).
Strong-motion Observation in Hokkaido with Broadband, Velocity-type Seismometers, Geophys. Bull.
Hokkaido Univ., 65, 335–354.
Somerville, P., Irikura, K., Graves, R., Sawada, S., Wald, D., Abrahamson, N., Iwasaki, Y., Kagawa, T., Smith,
N., and Kowada, A. (1999). Characterizing Crustal Earthquake Slip Models for the Prediction of Strong
Ground Motion, Seism. Res. Lett., 70, 59–80.
Wu, C., Koketsu, K., and Miyake, H. (2008). Source processes of the 1978 and 2005 Miyagi-oki, Japan,
earthquakes: Repeated rupture of asperities over successive large earthquakes, J. Geophys. Res., 113,
B08316.
Wu, C., Koketsu, K., and Miyake, H. (2009). Correction to “Source processes of the 1978 and 2005 Miyagi-oki,
Japan, earthquakes: Repeated rupture of asperities over successive large earthquakes”, J. Geophys. Res.,
114, B04302.
Yagi, Y. (2004). Source rupture process of the 2003 Tokachi-oki earthquake determined by joint inversion of
teleseismic body wave and strong ground motion data, Earth Planets Space, 56, 311–316.
Yoshida, K., Miyakoshi, K., and Irikura, K. (2011). Source Process of the 2011 Off the Pacific Coast of Tohoku
Earthquake Inferred from Waveform Inversion with Long-Period Strong-Motion Records, Earth Planets
Space, 63.577-582.
Zhang, W., Iwata, T., Irikura, K., Sekiguchi, H., and Bouchon, M. (2003). Heterogeneous distribution of the
dynamic source parameters of the 1999 Chi-Chi, Taiwan, earthquake, J. Geophys. Res., 108, 2232.