Moment Tensor Determination of the September 2, 2009 Tasikmalaya, West Java EarthquakeMoment Tensor Determination of the September 2, 2009 Tasikmalaya, West Java Earthquake

Using the Waveform Inversion Method of Near Field Data* (T2-P45)Using the Waveform Inversion Method of Near Field Data* (T2-P45)

I. Suardi1, Y. Yagi2, Afnimar3, S. Widiyantoro3

1 Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia

INTRODUCTION

On Wednesday, September 2, 2009, a strong earthquake struck the southern coast of West Java, i.e. about 142 km 1 Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia2 Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Japan3 Faculty of Mining and Petroleum Engineering, Institut Teknologi Bandung, Bandung, Indonesia

On Wednesday, September 2, 2009, a strong earthquake struck the southern coast of West Java, i.e. about 142 km

southwest of the Tasikmalaya District (Figure 1). Earthquake parameters released by the MCGA are as follows: origin time

= 02/09/2009 07:55:00 UTC; epicenter = 8.24OS, 107.32OE; depth = 30 km; and moment magnitude MW = 7.3.

DATA

The source mechanism of the September 2, 2009 Tasikmalaya, West Java earthquake is not consistent with the

characteristic of the tectonic stress in this region in which the strike direction in general parallels to the present

day trench. In fact, the strike of the September 2, 2009 Tasikmalaya earthquake is nearly perpendicular to the

DATA

We used 18 components of the near field data from 5 stations of the MCGA broadband seismometer network and 1

station of the Incorporated Research Institution for Seismology of the Data Management Center (IRIS-DMC) broadband

seismometer network (Figure 2).

To calculate the Green’s function, we used several velocity models, i.e. the Jeffreys-Bullen (Jeffreys and Bullen, 1940),

Figure 6. Comparison of observed

waveform (black curves) withday trench. In fact, the strike of the September 2, 2009 Tasikmalaya earthquake is nearly perpendicular to the

trench. We determined the moment tensor using the near field data from the regional network of the

Meteorological, Climatological and Geophysical Agency (MCGA) of Indonesia, and from the IRIS-DMC seismic

network. The frequency of band pass filter and the velocity structure model are determined by referring to

To calculate the Green’s function, we used several velocity models, i.e. the Jeffreys-Bullen (Jeffreys and Bullen, 1940),

AK135 (Kennett et al., 1995), Koulakov et al. (Koulakov et al., 2007), and modified Koulakov et al. models (Laksono, 2011),

respectively (Table 2).

For each pair of the range of band pass filters and the velocity structure model we applied a grid search method for the

waveform (black curves) with

theoretical waveform (gray curves).

Numbers below the station codes

depict the maximum amplitude of

the observed waveform.network. The frequency of band pass filter and the velocity structure model are determined by referring to

previous study results, as well as by trial and error. The band pass filter and the velocity structure model that

produce the smallest variance of 0.2402 is 0.01 to 0.03 Hz and the Jeffreys-Bullen model, respectively. The

Green's functions were calculated using the extended reflectivity method for the near field data. Our inversion

results show that the earthquake is an interplate earthquake type, which is located at the border around the

For each pair of the range of band pass filters and the velocity structure model we applied a grid search method for the

hypocenter depth determination. The range of centroid hypocenter depth used in the grid search started from a depth of

10 km to 100 km with an increment of 1 km.

RESULTS AND DISCUSSION

Figure 4 and Table 3. The table shows the smallest variance of 0.2402, that was obtained by using the band pass filter with

the observed waveform.

results show that the earthquake is an interplate earthquake type, which is located at the border around the

plate interface at a depth of 44 km. The strike is almost perpendicular to the trench, which may be related to a

strong slab pull beneath the region.

Figure 4 and Table 3. The table shows the smallest variance of 0.2402, that was obtained by using the band pass filter with

cut off frequencies of 0.01 Hz and 0.03 Hz (this study), and using the Jeffreys-Bullen velocity model. The source

parameters of the September 2, 2009 Tasikmalaya earthquake are as follows:

Moment = 0.2788x1020 (Nm), Mw = 6.9Moment = 0.2788x10 (Nm), Mw = 6.9

Depth = 44 km, variance = 0.2402.

Figure 1. Tectonic setting and seismicity around Java. Seismicity isFigure 1. Tectonic setting and seismicity around Java. Seismicity is

marked with small gray circles. The yellow star depicts the MCGA

epicenter of the September 2, 2009 event, and the aftershocks for

one week after the mainshock are depicted by red circles. The

average rate of plate motion is marked by arrows (modified from

Figure 3. Example of near field waveform data from

the MCGA regional network used in the moment

tensor inversion.

Figure 4. Results of the grid search process to determine the

best solution of focal mechanism as a function of ∆t andTable 4: Comparison of source mechanisms of the

September 2, 2009 Tasikmalaya earthquakeaverage rate of plate motion is marked by arrows (modified from

Bock et al., 2003).

tensor inversion. best solution of focal mechanism as a function of ∆t and

variance for each velocity model (JB Model, AK135, Koulakov et

al., and Modified Koulakov et al.) and the range of band pass

filter as shown in Table 1. The solutions at ∆t = 1 second (the

lowest focal mechanism in each square) are the best and

September 2, 2009 Tasikmalaya earthquake

Figure 2. Configuration of lowest focal mechanism in each square) are the best and

displayed in Table 3.

CONCLUDING REMARKS

Figure 2. Configuration of

the MCGA broadband

seismic network. The

yellow star depicts the

epicenter of the September

Figure 5. Map view (upper panel) and

vertical cross section (lower panel) with the

CONCLUDING REMARKS

We have analyzed the moment tensor of the September 2, 2009 Tasikmalaya

earthquake by employing the waveform inversion method of near field data in

detail. The best result with the smallest variance of 0.2402 was obtained by using

the band pass filter with cut off frequencies of 0.01 Hz and 0.03 Hz, and the

epicenter of the September

2, 2009 Tasikmalaya

earthquake. The red

triangles and the blue vertical cross section (lower panel) with the

source mechanism of the September 2,

2009 Tasikmalaya earthquake. The star and

open circles, indicate the relocated

mainshock hypocenter and the relocated

the band pass filter with cut off frequencies of 0.01 Hz and 0.03 Hz, and the

Jeffreys-Bullen velocity model. The focal mechanism resulting from this study is

in excellent agreement with those from the Global CMT, USGS and MCGA, as

shown in Table 4. The waveform fitting for the near field data used in this study

triangles and the blue

triangle in the circle depict

the MCGA near field

stations and the IRIS-DMC

station, respectively mainshock hypocenter and the relocated

aftershocks, respectively (Laksono, 2011).

Small gray circles indicate the seismicity

relocated by PDE. The curved thick line

shown in Table 4. The waveform fitting for the near field data used in this study

showed a good result. From the waveform data inversion, the centroid depth of

the September 2, 2009 Tasikmalaya earthquake is shallow i.e. 44 km. This

suggests that the earthquake type was interplate. The strike that is almost

perpendicular to the trench is interpreted to be related to a strong slab pull

stations and the IRIS-DMC

station, respectively

relocated by PDE. The curved thick line

depicts the USGS global slab 1.0 model.

Notice that the aftershocks distribution

plotted in the lower panel, to some extent,

depicts a splay fault.

perpendicular to the trench is interpreted to be related to a strong slab pull

beneath Java.

depicts a splay fault.

Description: The values listed above are the cut off low frequency, *in press.

Table 3: The grid search results of source mechanisms

Description: The values listed above are the cut off low frequency,

the cut off high frequency, and the stop band frequency in Hz,

respectively