2006TokaiSeismicity

7/31/2019 2006TokaiSeismicity

1/6

GEOPHYSICAL RESEARCH LETTERS, VOL. ???, XXXX, DOI:10.1029/,

Short-term slow slip and correlated tremor episodes in the Tokai

region, central Japan

Hitoshi Hirose and Kazushige ObaraNational Research Institute for Earth Science and Disaster Prevention, Tsukuba, Japan

We observe the related phenomenon of short-term slowslip event (SSE) and low-frequency nonvolcanic tremor onthe deeper extension of the 1944 Tonankai and the antici-pated Tokai earthquake rupture areas, central Japan. Japannation-wide seismometer and tiltmeter array, NIED Hi-nethas recorded these correlated episodes of SSE and tremor,known as episodic tremor and slip (ETS) in 2004 and 2005.The events are located on two areas in the Tokai region, oneis near the Shima peninsula, and the other is around theAichi area. Both episodes have a typical duration of 2-3 daysand released seismic moment corresponding to Mw 6.0.

Moreover, these episodes repeatedly occur with a recur-rence interval of about six months in both areas. This ev-idence suggests that the coincidence between the SSE andthe tremor is a characteristic behavior in the whole tremorbelt-like zone in the Nankai trough subduction zone.

1. Introduction

The Tokai region is located at the northern end of thePhilippine Sea plate (PH), which is subducting beneath theJapan islands along the Suruga-Nankai trough (Figure 1).At the trough, interplate thrust earthquakes of magnitude 8recur with an interval of nearly 100 years [Ando, 1975b;

Sangawa, 1993]. The source region of the latest event,1944 Tonankai earthquake, extends from southeast off Kiipeninsula to the Shima peninsula along the Suruga-Nankaitrough [e.g., Kikuchi et al., 2003]. The northern part ofthe trough remains unbroken (Expected Tokai EQ in Fig-ure 1), though the coseismic slip in the previous events arethought to extend into the northern segment [e.g., Ando,1975b]. Thus the Tokai earthquake is anticipated in thenear future in this region [Ando, 1975a].

Recent seismological and geodetic observations have re-vealed that there are more diverse natural phenomena insubduction zones than it had been previously thought.These are nonvolcanic deep low frequency tremors [Obara,2002; Katsumata and Kamaya, 2003], slow slip events (SSE)[e.g., Hirose et al., 1999; Dragert et al., 2001; Ozawa et al.,

2002; Sagiya, 2004], and very low frequency earthquakes[Obara and Ito, 2005; Ito and Obara, 2006]. Among theseevents, the Tokai SSE [Ozawa et al., 2002] is particularly in-teresting because the slip region is adjacent to the expectedsource area of the Tokai earthquake (Figure 1) and the SSEmight promote the occurrence of the earthquake.

In the Cascadia margin, Rogers and Dragert [2003] foundinteresting episodes called episodic tremor and slip (ETS),in which the SSE is accompanied by nonvolcanic tremor ac-tivity. A similar phenomenon has also been found in west-ern Shikoku, southwest Japan [Obara et al., 2004; Hirose

Copyright 2006 by the American Geophysical Union.0094-8276/06/$5.00

and Obara, 2005], but has not been identified in the otherregions within the widely distributed belt-like tremor zone[Obara, 2002]. In this paper, we report that episodes of SSEand tremors have also repeatedly occurred in the Tokai re-gion, central Japan. These episodes suggest that ETS is acharacteristic behavior of the deeper region of the subduct-ing plate boundary where nonvolcanic tremor occurs alongthe Nankai trough.

2. Observation and Data Analysis

We observe slow crustal deformation associated with thetremor activity using a tiltmeter equipped with a high-sensitivity seismometer at NIED Hi-net stations [Obara etal., 2005]. Tiltmeter signals are originally recorded with asampling frequency of 20 Hz. These records are resampledso that we obtain a one hour sampling interval. In conjunc-tion with hourly atmospheric pressure data provided by theJapan Meteorological Agency (JMA), we then estimate bothtidal components and a response to the atmospheric pres-sure contained in the tilt records by applying BAYTAP-G[Tamura et al., 1991]. Subtracting the tidal components andthe atmospheric pressure response from the original record,we can determine if any transient signal is present in thedata.

Time variations of the tilt record due to a short-termSSE are characterized by a slow rise lasting for a few days

and small amplitude typically less than 0.1 rad [Obara etal., 2004; Hirose and Obara, 2005]. In addition, adjacentstations also record a similar deformation with almost thesame time duration. In some cases the onset of the slowtilt change is slightly different at each station, indicating apropagation of the slow slip [Hirose and Obara, 2005].

The tilt record usually contains a long-term drift, which isnot dealt with in this study. To illuminate the slow transienttilt signal from the detided record with a drift, we first fittwo lines which have the same slope on the detided time se-ries, one for the time period before the slow slip episode, theother for that after the episode. Then we subtract the trendfrom the detided record. The tilt offset between the twotime periods before and after the transient episode can becalculated simultaneously in the detrend procedure. A set

of such tilt offsets recorded by several stations are invertedto estimate a slip on a plane rectangular fault using Okadas[1992] formulation. The processing of the tilt records con-ducted in this paper is the same as that in Hirose and Obara[2005].

The tremor signals are recorded by the high-sensitivityseismometer at the NIED Hi-net stations. To clarify thesesignals, original 100 Hz sampling data are converted intoenvelope waveforms, which are RMS traces of bandpass fil-tered seismograms. The tremor signal is indicated by coher-ent envelopes over several neighboring stations. Using sucha waveform similarity, the source of the tremor is locatedby the envelope correlation method [Obara, 2002]. In thisprocedure, a tremor location is estimated once per minute.We then select well-located tremors, and count the numberof events per day as a measure of the tremor activity.

1


2/6

X - 2 HIROSE AND OBARA: SLOW SLIP AND TREMOR IN TOKAI, JAPAN

3. Slow Slip Events and Tremors in theTokai Region

By the end of 2004, we detected short-term episodes ofthe tilt transients and the accompanying tremor activity intwo areas in the Tokai region: One occurs near the Shimapeninsula (northern part of the Kii peninsula) in the end ofNovember, and the other near Nagoya, Aichi area in Decem-

ber (locations are shown in Figure 1).Figure 2a shows the time series of the tiltmeter records

at Hi-net stations near the Shima peninsula. These recordsshow that the slow deformation started on 29 November2004 (JST) and seemed to last for 2-3 days. A maximumtilt change of approximately 0.1 rad, down in the WNWdirection is recorded at the station MASH (Figure 3). Onthe other hand, the record at WATH, southern part of thepeninsula, shows almost purely southward down tilt. Thesevariations in the tilt direction indicate that the maximumground uplift is centered north of WATH near MASH. Un-fortunately, a large disturbance on the tilt records betweenthe 4th and 5th of December due to the passage of a low at-mospheric pressure front prevents us from determining thetermination of this transient deformation, even after the ap-plication of the atmospheric pressure correction. We assumethis episode ends on 1 December 2004 (Table 1).

The low-frequency tremor became active on the same dayas the slow tilt changes began, and lasted for six days (Fig-ure 2a; Table 1). The results of the tremor source locationsshow that the tremors are distributed on the northwest partof the Shima p eninsula (near the station URSH and MASH)on 29 and 30 November (Figure 3) and the foci of the tremorsmigrate to the southwest throughout the episode along thestrike direction of the subducting plate surface.

The estimated slow slip fault model is shown in Figure 3with the tremor locations which occurred during the sametime period as the tilt transient. The slip region is locatedbeneath MASH and URSH, northwest part of the penin-sula with averaged slip of about 1.2 cm (the released seis-mic moment corresponds to a moment magnitude of 5.9).

The depth, strike and dip angle of the fault are consistentwith the configuration of the plate interface. The slip areacorrelates with most of the tremor locations on 29 and 30November. On 1 December a decrease in the tilt changerate is seen at most stations (Figure 2a). This correspondsto the migration of the tremors to the south as noted above.This means that the observed transient tilt changes mainlyreflect the two-day-long ETS activity in the northwest partof the Shima peninsula (Figure 3), and both the tremor andthe SSE might migrate to the south, though this southernSSE has not been identified because of an insufficient signal-to-noise ratio and inadequate tiltmeter station coverage inthis area.

In December 2004, we detected another ETS episode inthe Aichi area. Figure 2b shows the detided tilt recordsand the tremor activity in the area. In this episode, the

tremor activity and the tilt changes started on 17 Decemberand seemed to last for 4- 5 days. The observed tilt changesare very small, a maximum of 0.06 rad. The tilt vectorsshow a radial pattern in spatial distribution which is cen-tered near the station OKZH (Figure 4a). This indicatesthat the center of the uplift region is located there, whichcould be caused by reverse fault slip at depth.

The locations of the inverted fault and the tremors showthat the main activity is concentrated in the central part ofthe Aichi area, near OKZH (Figure 4a). The depth of theupper end of the fault is estimated to be 35 km with an un-certainty of several kilometers. The estimated fault locationsuggests that the slow slip event occurs on the subductingplate interface.

In July 2005, the following episodes took place in the twoareas in the Tokai region (Table 1). The tilt records show

that in the Shima peninsula area a slow transient signal sim-ilar to those in Figure 2a lasted from 10 to 12 July 2005 witha longer tremor episode from 10 to 16 July, and a tilt tran-sient and tremors began on 20 July and continued until 23July in the Aichi area [NIED, 2006]. The later episode thatoccurred in the Aichi area is also recorded by several strain-meters operated by other institutes [e.g., Kobayashi et al.,2006]. These two episodes indicate that the ETS events in

the Tokai region occur repeatedly as in the Cascadia margin[Rogers and Dragert, 2003] and in western Shikoku, south-west Japan [Obara et al., 2004; Hirose and Obara, 2005].

The estimated slip area and the tremor locations arenearly identical in both November 2004 and July 2005 Shimapeninsula episodes, while slightly different in December 2004and July 2005 Aichi episodes. Figure 4b shows the obtainedSSE fault model and the tremor epicenters for the July 2005Aichi event. A comparison of the slipped area in the 2005episode with the previous one demonstrates that that in2005 is located on the adjacent part to the east of that in2004. This result is supported by differences in the observedtilt direction. For example, the tilt vector in OKZH points inthe southeast direction in 2004 (Figure 4a), whereas almostwest in 2005 (Figure 4b). The change in the tilt direction is

also observed at STRH and THNH. These may suggest thatthere are at least two areas where ETS episode is likely tooccur, SSE patches [Hirose and Obara, 2005].

4. Discussion

The short-term SSE and the correlated tremor episodehave been detected not only in the western Shikoku region,southwest Japan [Obara et al., 2004; Hirose and Obara,2005], but also in the Tokai region, central Japan. Althoughthere are still some regions where the tremor is observed butthe SSE has not been identified, the observations reportedin this paper may suggest that the coincidence between theSSE and the tremor could be a characteristic behavior at thedeeper part of the subduction zone along the whole tremor

belt in southwest Japan [Obara, 2002].The recurrence of SSE provides us with another impor-tant source of information about the nature of SSE. As men-tioned above, SSE with tremor occurred in November 2004and July 2005 near Shima peninsula and the slip area forboth episodes is almost identical. Moreover, the tiltmeterrecord at MASH shows several recurrences of ETS with anaverage interval of six months [Obara and Hirose, 2006]. Re-call that the estimated slip in each episode is 1-2 cm and theETS takes place roughly twice a year. This means that theannual displacement released by means of SSE is 2-4 cm.Taking an uncertainty in the slip estimation into account,this amount is quite comparable to the relative plate conver-gence rate [Heki and Miyazaki, 2001]. This indicates thatthe SSEs may be the process of strain release which is accu-

mulated by the relative plate motion. This may also supportthe idea that the SSEs may occur on the subduction plateinterface which is estimated by the SSE fault inversion.

Although the frequency of ETS activity in the Aichi areais not well known, two episodes were observed with an in-terval of seven months, and they seem to occur twice a year.We found two spots where ETS could happen every oneyear with out of phase occurrence. In addition, the amountof slip in each event is almost the same as the Shima penin-sula case. This means that the Aichi ETS is half as frequentthan the Shima ETS. This variation in ETS activity mightbe affected by the reduced convergence rate at the northern-most part of the Suruga-Nankai trough in comparison withthe rest of the trough [Heki and Miyazaki, 2001]. The otherfactor which may affect the ETS activity is the Tokai long-term SSE [Ozawa et al., 2002], which has lasted for over


3/6

HIROSE AND OBARA: SLOW SLIP AND TREMOR IN TOKAI, JAPAN X - 3

five years in the southeast portion of the Aichi ETS area(Figure 1). The possible connection between the short- andthe long-term SSEs is suggested from similar activity in thewestern Shikoku area, southwest Japan, where an increasein frequency of the short-term SSEs is observed during theadjacent long-term SSE in the Bungo channel [Hirose andObara, 2005]. If this is the case, the Aichi ETS activity isexpected to decrease in its frequency after the termination

of the Tokai long-term SSE.

5. Conclusion

We detected new ETS events occurred in 2004 and 2005in the Tokai region, central Japan. These events last for2-3 days. The SSE equivalent magnitudes are of Mw 6.0.These ETS recur quasi-periodically with the recurrence in-terval of about six months. These episodes, in addition tothe previously reported events in western Shikoku, south-west Japan, suggest that the ETS is a characteristic behav-ior at the plate boundary along the tremor belt zone in theNankai trough subduction zone. The recurrence features ofETS suggest that the short-term SSE may be the process ofstrain release which is accumulated by the relative plate mo-tion. This means that the frequency of ETS episodes maybe partly controlled by a plate convergence rate.

Acknowledgments. We are grateful to K. Kasahara, F. Ya-mamizu and K. Shiomi for the operation of the observation net-work, NIED Hi-net. We would like to thank M. Ishida, M.Sakanashi and S. Sekine for providing the plate configurationdata. Meteorological data were provided by Japan MeteorologicalAgency. We wish to thank anonymous reviewers for their sugges-tions for ways to improve the manuscript. The writing work ofthis article was partly done on the first authors visiting PacificGeoscience Centre, Geological Survey of Canada.

References

Ando, M. (1975a), Possibility of a major earthquake in the Tokaidistrict, Japan and its pre-estimated seismotectonic effects,Tectonophys., 25, 6985.

Ando, M. (1975b), Source mechanisms and tectonic significanceof historical earthquakes along the Nankai trough, Japan,Tectonophys., 27, 119140.

Dragert, H., K. Wang, and T. S. James (2001), A silent slipevent on the deeper Cascadia subduction interface, Science,292, 15251528.

Heki, K., and S. Miyazaki (2001), Plate Convergence and Long-Term Crustal Deformation in Central Japan, Geophys. Res.Lett., 28(12), 23132316.

Hirose, H., and K. Obara (2005), Repeating short- and long-term slow slip events with deep tremor activity around theBungo channel region, southwest Japan, Earth Planets Space,57, 961972.

Hirose, H., K. Hirahara, F. Kimata, N. Fujii, and S. Miyazaki(1999), A slow thrust slip event following the two 1996 Hyu-ganada earthquakes beneath the Bungo Channel, southwestJapan, Geophys. Res. Lett., 26(21), 32373240.

Ishida, M., and M. Sakanashi (2003), Geometry of the PhilippineSea slab from Kanto to Kyushu district, Japan (in Japanese),Monthly Chikyu, 25, 168172.

Ito, Y., and K. Obara (2006), Dynamic deformation of the ac-cretionary prism excites very low frequency earthquakes, Geo-phys. Res. Lett., 33(2), L02311, doi:10.1029/2005GL025270.

Katsumata, A., and N. Kamaya (2003), Low-frequency continu-ous tremor around the Moho discontinuity away from volca-noes in the southwest Japan, Geophys. Res. Lett., 30(1), 1020,doi:10.1029/2002GL015981.

Kikuchi, M., M. Nakamura, and K. Yoshikawa (2003), Source rup-ture processes of the 1944 Tonankai earthquake and the 1945Mikawa earthquake derived from low-gain seismograms, EarthPlanets Space, 55, 159172.

Kobayashi A., T. Yamamoto, K. Nakamura, and K. Kimura(2006), Short-term slow slip events in Tokai region detectedby the strainmeters (submitted).

Miyazaki, S., and K. Heki (2001), Crustal velocity field of south-west Japan: Subduction and arc-arc collision, J. Geophys.Res., 106(B3), 43054326.

National Research Institute for Earth Science and Disaster Pre-vention (NIED, 2006), Short-term slow slip event with deeplow-frequency tremor activity in the Tokai area (in Japanese),Rep. Coord. Commit. Eq. Pred., 75, 375379.

Obara, K. (2002), Nonvolcanic deep tremor associated with sub-duction in southwest Japan, Science, 296, 16791681.

Obara, K., and H. Hirose (2006), Non-volcanic deep low-frequency tremors accompanying slow slips in the southwestJapan subduction zone, Tectonophys., 417, 3351.

Obara, K., and Y. Ito (2005), Very low frequency earthquakesexcited by the 2004 off the Kii peninsula earthquakes: A dy-namic deformation process in the large accretionary prism,Earth Planets Space, 57, 321326.

Obara, K., H. Hirose, F. Yamamizu, and K. Kasahara (2004),Episodic slow slip events accompanied by non-volcanic tremorsin southwest Japan subduction zone, Geophys. Res. Lett.,31 (23), L23602, doi:10.1029/2004GL020848.

Obara, K., K. Kasahara, S. Hori, and Y. Okada (2005), Adensely distributed high-sensitivity seismograph network inJapan: Hi-net by National Research Institute for Earth Sci-ence and Disaster Prevention, Rev. Sci. Instrum., 76, 021301,doi:10.1063/1.1854197.

Okada, Y. (1992), Internal deformation due to shear and tensilefaults in a half-space, Bull. Seism. Soc. Am., 82, 10181040.

Ozawa, S., M. Murakami, M. Kaidzu, T. Tada, T. Sagiya, Y.Hatanaka, H. Yarai, and T. Nishimura (2002), Detection andmonitoring of ongoing aseismic slip in the Tokai region, centralJapan, Science, 298, 10091012.

Rogers, G., and H. Dragert (2003), Episodic tremor and slip onthe Cascadia subduction zone: The chatter of silent slip, Sci-ence, 300, 19421943.

Sagiya, T. (2004), Interplate coupling in the Kanto district, cen-tral Japan, and the Boso peninsula silent earthquake in May1996, Pure Appl. Geophys., 161, 23272342.

Sangawa, A. (1993), Research on paleoearthquakes using tracesdiscovered at archaeological sites (in Japanese with Englishabstract), Quaternary Res., 32, 249256.

Tamura, Y., T. Sato, M. Ooe, and M. Ishiguro (1991), A proce-dure for tidal analysis with a Bayesian information criterion,Geophys. J. Int.

,104

, 507516.

H. Hirose and K. Obara, National Research Institute for EarthScience and Disaster Prevention, 3-1 Tenno-dai, Tsukuba, Ibaraki305-0006, Japan. ([email protected]; [email protected])


4/6


135 136 137 138 139

33

34

35

36

0 50 100

km

10

km

203

050

20

30

40

50

TsuShima pen. area

Aichi area

KP

SP

Suruga

-Nankaitro

ugh

1944Tonankai

EQ

130 140

30

40NA

PA

AM

PH

Tokai region

40

Nagoya

ExpectedTokaiEQ

Tokai SSE

Figure 1. Index map showing the tectonic settingand station locations. NA, AM, PA, PH, KP and SPrepresent North American, Amurian, Pacific, PhilippineSea plate, Kii Peninsula and Shima Peninsula, respec-tively. Crosses show NIED Hi-net station locations. Solid

squares are the locations of JMA meteorological observa-tories. Small gray circles display the epicenters of tremors[Obara and Hirose, 2006]. Dark gray line ellipses indi-cate the rupture areas of the 1944 Tonankai earthquakeand the expected Tokai earthquake. Light gray line el-lipse denotes the source area of the Tokai SSE [Ozawa etal., 2002]. The tremor activity occurring inside the rect-angle areas are plotted in Figure 2. Open arrow shows therelative plate convergence direction of PH with respectto AM [Miyazaki and Heki, 2001]. Contour lines denotethe plate boundary depth inferred from microearthquakedistribution along subducting PH [Ishida and Sakanashi,2003].


5/6

HIROSE AND OBARA: SLOW SLIP AND TREMOR IN TOKAI, JAPAN X - 5

21 28 05Nov 2004 Dec 2004

02040 Tsu precip. [mm]

-300

30 Tsu atm. pres. [hPa]0

50 Shima tremor [count/day]0.0

0.2

0.4

0.6

0.8

Tilt[

rad]

WATH E

WATH N

MGWH E

MGWH N

MASH E

MASH N

URSH E

URSH N

OYMH E

OYMH N

N, E down

(a)

N, E down

12 19 26Dec 2004

02040 Nagoya precip. [mm]

-150

15 Nagoya atm. pres. [hPa]0

50100 Aichi tremor [count/day]0.0

0.1

0.2

0.3

0.4

0.5

INSH E

INSH N

THNH E

THNH N

NUKH E

NUKH N

OKZH E

OKZH N

STRH E

STRH N

(b)

Figure 2. Time series of detided and detrended tilt-meter records for (a) November 2004 Shima peninsulaepisode; and (b) December 2004 Aichi episode. Traceswith four-character station code followed by N and Edenote the northward and eastward ground down compo-

nents, respectively. The bottom three records show thetremor activity, the atmospheric pressure change, and theprecipitation, respectively. The meteorological recordsare observed at (a) Tsu; and (b) Nagoya, respectively.Dotted lines depict the identified time period for eachSSE.

Table 1. ETS Episodes in the Tokai Region

Dates (Tilt)a Dates (Tremor) Areab Mwc

29/1101/12/2004 29/1103/12/2004 S 5.917/1220/12/2004 17/1221/12/2004 A 5.810/0712/07/2005 10/0716/07/2005 S 5.920/0723/07/2005 20/0723/07/2005 A 5.7

a Time window in which tilt change vectors are calculated; b S: Shima peninsula, A: Aichi; c Moment magnitude corre-sponding to the seismic moment released by a SSE


6/6


136 137

34

35

0 50

km

Slip3 cm

Obs.Calc.

Tilt

0.1 rad

WATHMGWH

MASHURSH

OYMH

lon. 136.75

dip 13

slip 1.2 cm

wid. 45 km

MW5.9 rake 94

lat. 34.60

strike 219.36

depth 25 km

leng. 37 km

M07.9e+17 Nm

Figure 3. The estimated SSE fault model and tremorepicenters for the November 2004 episode around theShima peninsula. Solid and open arrows with stationcodes represented by four letters show the observed andcalculated tilt change vectors (downward tilt direction),respectively. Rectangle with thick arrow denotes the slowslip area and the slip vector projected onto the groundsurface. Epicenters of tremor activity during the sametime period are also plotted (solid circles: tremors on 29and 30 November; open circles: on 1 December).

137 138

35

0 50

km

Tilt

0.1 radObs.Calc.

Slip3 cm

lon. 137.32

dip 36

slip 1.8 cm

wid. 29 km

MW5.8 rake 116

lat. 34.96

strike 246

depth 35 km

leng. 30 km

M06.3e+17 Nm

137 138

0 50

km

lon. 137.72

dip 30

slip 0.80 cm

wid. 41 km

MW5.7 rake 128

lat. 34.93

strike 258

depth 29 km

leng. 41 km

M05.3e+17 Nm

Tilt

0.1 radObs.Calc.

Slip3 cm

(a) (b)

INSH

THNH

STRH

NUKH

OKZH OKZH

STRH

THNH

Figure 4. Same as Figure 3 but for (a) December 2004;(b) July 2005; Aichi episodes. Solid circles show thetremor epicenters on the first two days of each episodeand open circles are those on the following two days.

2006TokaiSeismicity

Documents

Transcript of 2006TokaiSeismicity