GEOPHYSICAL RESEARCH LETTERS, VOL. 30, NO. 1, 1020, doi:10.1029/2002GL015981, 2003
Low-frequency continuous tremor around the
Moho discontinuity away from volcanoes in
the southwest Japan
Akio Katsumata
Meteorological College, Japan Meteorological Agency, Kashiwa, Chiba, Japan
Noriko Kamaya
Seismological and Volcanological Department, Japan Meteorological Agency, Chiyoda-ku, Tokyo, Japan
Received 26 July 2002; revised 23 October 2002; accepted 15 November 2002; published 11 January 2003.
[1] Recent enhancement of seismic networks in the Japan
Islands revealed occurrence of low-frequency continuous
tremors of a beltlike distribution in the southwest Japan,
where the subducting Philippine Sea plate reaches depths of
30– 40 km. Source depth of the tremor is estimated by
selecting tremor segments with relatively clear P-wave
onsets. The source region of the tremor is assumed to
correspond to lowermost parts of crust close to a triple
boundary of the crust, mantle wedge, and the subducting
slab. The long duration of the phenomenon indicates
existence of fluid relating to the generation of the tremor.
The most probable fluid is considered to be water produced
by dehydration process of chlorite and amphibole in the slab
on the basis of data from high pressure and temperature
experiments. The northern border of the beltlike distribution
is possibly rimmed by the edge of the mantle wedge because
INDEX TERMS:
serpentine formation absorbs fluid water.
3699 Mineralogy and Petrology: General or miscellaneous; 7220
Seismology: Oceanic crust; 7230 Seismology: Seismicity and
seismotectonics; 8130 Tectonophysics: Evolution of the Earth:
Heat generation and transport. Citation: Katsumata, A., and
N. Kamaya, Low-frequency continuous tremor around the Moho
discontinuity away from volcanoes in the southwest Japan,
Geophys. Res. Lett., 30(1), 1020, doi:10.1029/2002GL0159812,
2003.
1. Introduction
[2] Since Oct. 1997, short-period seismic records observed
at most of stations in the Japan Islands have been transmitted
to the Japan Meteorological Agency (JMA). The JMA has
processed those records to make a comprehensive seismic
catalog in Japan. The short-period instruments include those
installed by universities, the National Research Institute for
Earth Science and Disaster Prevention (NIED), the JMA,
other national institutes, and local governments. This integrated processing lowered the detectable magnitude of earthquakes. Moreover, deployment of the Hi-net [e.g., Okada et
al., 2000], which is a dense network of highly sensitive shortperiod instruments installed by the NIED, improved the
detectivity remarkably. Hi-net data has been included to
produce the seismic catalog since Oct. 2000.
[3] The enhanced detection ability revealed occurrence of
low-frequency events in the Japan Islands. The JMA has
Copyright 2003 by the American Geophysical Union.
0094-8276/03/2002GL015981$05.00
distinguished low-frequency earthquakes (LFE) in the seismic catalog since 1999. Occurrence of low-frequency
events around the Moho discontinuity in the areas away
from volcanoes became recognized based on the seismic
catalog [Nishide et al., 2000]. Although seismicity of lowfrequency events beneath volcanoes has been well-known
[e.g., Ukawa and Obara, 1993], such seismic activity away
from volcanic areas had not been noticed. Obara [2002]
found the nonvolcanic deep low-frequency continuous
tremors (LFT in this article) of a beltlike distribution along
the strike of the subducting Philippine Sea plate in the
southwest Japan using data from the Hi-net.
[4] The beltlike distribution is also seen in the comprehensive seismic catalog (Figure 1). Activities of eastern end
of the beltlike distribution in the southwest Japan were
detected by Nishide et al. [2000], but they did not recognize
the activities as continuous tremors because the JMA
identifies enlarging parts of the LFT as onsets of earthquakes (LFE). In Figure 1, solid circles show LFE occurring
within 10 km (epicentral distance) from the centers of
volcanoes which have been active in the Quaternary [The
Research Group for Active Faults of Japan, 1991], and open
circles show LFE away from volcanoes.
[5] Averaged source depth of LFT is about 30 km
[Obara, 2002], but the estimated source depths often scatter.
In this article, source depth of LFT is investigated, and its
causes are discussed.
2. Source Depth of LFT
[6] Figure 2 shows examples of LFT waveforms of
horizontal (N/S) component. The peak amplitudes in the
figure indicate that the magnitudes [Katsumata, 2001] are
no more than 0.7. Due to the small amplitude and noise-like
appearance, the tremor had not been recognized until recent
years.
[7] The JMA estimates the source locations by an ordinary hypocenter determination method based on the onset
times of enlarging parts of the tremors. The calculated travel
times in Figure 2 are based on locations estimated by the
JMA. Onsets of the tremor are close to the calculated S
travel times. Corresponding onsets of P-waves are seldom
seen even on the vertical component. Lack of P-wave
arrival data makes accuracy of source depths poor.
[8] Obara [2002] estimated the locations of LFT by
applying S-wave velocity model to envelopes of tremor
waveforms. The method of Obara [2002] may have
20 - 1
20 - 2
KATSUMATA AND KAMAYA: DEEP TREMOR IN THE SOUTHWEST JAPAN
Figure 2. Example of LFT observed on June 3, 2001.
Waveforms are arranged in epicentral distance order from an
estimated source location. Broken lines show calculated P
and S travel times for enlarging parts of the tremor. Station
code are denoted on the right side. ND, NU, and LG signify
the NIED, Nagoya University, and local government,
respectively. YASUOK and TAKASA are stations of the
JMA. Station and source locations are plotted in Figure 3.
Figure 1. Seismicity of low-frequency earthquakes (LFE)
in Japan identified by the JMA (Sep. 1999– Dec. 2001).
Solid circles show LFE occurring close to active volcanoes
in the Quaternary (filled gray triangles [The Research Group
for Active Faults of Japan, 1991]), and open circles show
LFE away from volcanoes.
produced rather poor depth estimation due to lack of Pwave information and errors in cross-correlation between
envelopes.
[9] Although most of LFT does not show onsets of Pwave, a small number of LFT segments does show onsets of
P-wave. We selected forty-seven enlarging parts of the
tremor (LFE) which show P-wave onsets and estimated
source depths by using S-P times at stations close to
epicenters with the velocity structure used in the JMA
routine work [Ueno et al., 2002]. Estimated source locations
are shown with solid circles in Figure 3. The estimated
depths of the most of events were distributed from 25 to 35
km (±1 4 km), and the average depth is 31 ± 2 km. There
are 2 events located deeper than 40 km, but their P-wave
onsets are not so clear. In Figure 3, contours shows the top
of the Philippine Sea plate which were estimated as upper
limits of seismicity in the plate using the seismic catalogue
since 1997. The beltlike distribution lies almost between the
contours of 30 and 40 km.
[10] The distribution indicates that this activity is related
to the subduction of the Philippine Sea plate. The beltlike
distribution has eastern and western ends [Obara, 2002].
There is no tremor activity in the central Honshu in the east
of 138°E and Kyushu island, though the Philippine Sea
plate subducts beneath both areas. Moreover, seismicity gap
is recognized around the Kii channel [Obara, 2002]. In
addition to the beltlike area along the Nankai Trough, some
spots of non-volcanic LFE activities are seen. Beltlike
activity is not seen in the subduction zone along the Japan
trench where the Pacific plate subducts (Figure 1).
[11] Zhao et al. [1992] estimated depths of the Moho
discontinuity beneath the Japan Islands by Pn and Sn arrival
times of natural earthquakes, and their estimated depths are
31– 38 km (±2 km) beneath the beltlike region. Kurashimo
et al. [2002] estimated crustal and upper mantle structure
beneath eastern Shikoku by seismic refraction/reflection
Figure 3. Distribution of LFE (open circles) in the
southwest Japan (Sep. 1999 – Dec. 2001). Solid circles,
crosses, a solid square, and filled gray triangles show
revised source locations of selected LFE in this study,
stations of which records are shown in Figure 2, the
estimated source location for the records in Figure 2, and
active volcanoes in the Quaternary, respectively. The top
depth contours (km) of the subducting Philippine Sea plate
are also shown. The line segments A and B indicate
Tonankai and Nankaido profiles for which Hyndman et al.
[1995] estimated geotherms, respectively.
KATSUMATA AND KAMAYA: DEEP TREMOR IN THE SOUTHWEST JAPAN
20 - 3
method. Their profile (almost along a meridian of about
134.3°E) was located inside the activity gap of LFT.
According to their structure model, the Philippine Sea plate
contacts with inland crust with a depth range of 25– 33 km
beneath interpolated beltlike region of LFT. The source
depths of about 30 km of LFE are considered to be in the
lower crust and very close to the triple boundary of the
crust, the subducting Philippine Sea plate, and mantle
wedge (Figure 4).
3. Fluid Generating the Tremory
[12] The activities in the beltlike region along Nankai
trough should be related to the subducting Philippine Sea
plate from the epicenter distribution.
[13] The continuous tremors remind us of volcanic tremors, which are caused by some kinds of fluid [e.g., Chouet,
1996]. Since a large quantity of water is conveyed by
subduction of the Philippine Sea plate, the most plausible
liquid causing the tremors is assumed to be water. We assume
some possible sources of the water which causes the LFT.
1. Extraction of interstitial water from oceanic crust.
2. Dehydration from minerals in the oceanic crust.
3. Dehydration from minerals in the mantle of the
Philippine Sea plate.
4. Dehydration from minerals in the mantle wedge.
[14] Although extraction of interstitial water from the
oceanic crust might be a possible source, it would be
difficult to explain why the extraction occurs at the
restricted depth range of 30– 40 km. Since extraction of
interstitial water should occur gradually, the limited depth
range of water supply is considered to be related to limited
pressure and temperature condition.
[15] Figure 5 shows the phase relationships for watersaturated mid-ocean ridge basalt obtained by Schmidt and
Poli [1998]. This phase relationships are applicable to
minerals in a oceanic crust in a subducting slab. The broken
curves A and B in the figure show geotherms at the top of
the subducting Philippine Sea plate estimated by Hyndman
et al. [1995] for Tonankai and Nankaido profiles, respectively (Figure 3). Depth contours in Figure 3 were used to
convert distance from the trench axis into depth at the top of
the slab. The Philippine Sea plate subducts with a steeper
incline at the profile A than at the profile B. The geotherms
cross the phase boundary shown by the heavy solid curve.
While crossing the boundary, chlorite decomposes for
forming mainly garnet and water is released. Amphibole
continuously dehydrates around the boundary, too [Schmidt
Figure 4. Schematic chart of source region of the LFT.
Figure 5. Phase relationships for water-saturated midocean ridge basalt, retouch of a figure by Schmidt and Poli
[1998]: amph, amphibole; chl, chlorite; cld, chloritoid; cpx,
jadeitic or omphacitic clinopyroxene; epi, epidote; gar,
garnet; law, lawsonite; zo, zoisite. The broken curves A and
B show geotherms at the top of the subducting Philippine
Sea slab based on the estimation by Hyndman et al. [1995]
for Tonankai and Nankaido profiles, respectively. The
geotherms cross phase boundaries for chlorite (the heavy
solid curve) and amphibole decomposition at depths of
40– 50 km.
and Poli, 1998]. The crossing points of the geotherms do
not coincide with the depth range of 30– 40 km in the
figure. But, inside of the slab has higher pressure and
temperature than the top of the slab, and those conditions
promote the dehydration process.
[16] According to another phase diagram by Schmidt and
Poli [1998], serpentine and some other hydrous minerals
including chlorite and amphibole in the mantle peridotite
also dehydrate under the conditions of 600– 700°C and 30–
50 km. This indicates a possibility that hydrous minerals in
the descending mantle wedge along with the slab would
also release water. Though, temperature at the top of the
mantle wedge beneath Shikoku was estimated at about
450– 500°C by Hyndman et al. [1995]. The temperature is
considered too low for serpentine in the mantle wedge to
release much of water.
[17] LFT is observed in the limited range of slab top
depth, 30– 40 km. Since the amount of water from dehydration processes in the oceanic crust would increase as
temperature and pressure increase, a wider beltlike area
would be expected. According to the structure shown by
Kurashimo et al. [2002], descending oceanic crust starts
contact with the mantle wedge under the northern rim of the
beltlike distribution. Fluid water relating to the tremor
would not be supplied to the inland crust in the north of
the rim because temperature in the mantle wedge would be
low enough for mantle peridotite to absorb ascending water
forming hydrous minerals. There would be another possibility that tremor would not occur in the boundary region
between the inland crust and the mantle wedge due to a
different mechanical condition from that between the inland
crust and the oceanic crust.
[18] The origin of the water should be in the subducting
slab. But the tremor itself is considered occurring in the
lower crust, and not in the slab. This is considered to be
related to differences of mechanical properties of minerals
in the lower crust and the slab.
20 - 4
KATSUMATA AND KAMAYA: DEEP TREMOR IN THE SOUTHWEST JAPAN
[19] Beltlike LFT activities are not recognized in the
east of 138°E line and in Kyushu Island. In addition, a
gap is recognized around the Kii channel. Since the
activity is considered to be related to dehydration process
of hydrous minerals such as chlorite and amphibole, the
inactive areas would indicate absence of dehydration
process due to the following reasons. The Philippine Sea
plate moves northwestward in this area [e.g., Seno et al.,
1993], and the plate moves almost along the iso-depth
curves of the plate under the channel. If the dehydration
process has almost done around the Kii Peninsula before
reaching this area, water would not be supplied. On the
other hand, in the area east of 138°E, there are some
volcanoes, Mt. Fuji, Mt. Hakone, and so on. Volcanic
activities would affect thermal condition and dehydration
process of the subducting plate.
[20] The tremor have features of no large P-wave sources
and no large frequency components lower than 1 Hz
[Obara, 2002]. If the seismic source is pressure variation
in the ground, clear P-wave should be observed. It is
assumed that the tremor was produced by some kind of
shear forces between the rock and fluid. If the viscosity of
the fluid is large or if the surface of the rock is rough, flow
of fluid would act on the rock with shear force.
4. Conclusions
[21] Source depth of low-frequency tremor (LFT) along
the subduction zone in the southwest Japan where the
Philippine Sea plate subducts was investigated, and the
cause of the tremor was discussed. The depth estimation
was based on selected tremor segments with recognizable
onsets of P-wave. The estimated source depths of LFT in the
beltlike area ranged from 25 to 35 km. It is considered that
the source region of the LFT is close to the boundaries
among lower crust, the Philippine Sea plate, and mantle
wedge. It is assumed that the tremor is caused by ascending
water which is produced by dehydration process of chlorite
and amphibole in the subducting slab. These hydrous
minerals would release water at a temperature of about
600°C in a depth range of 30 –50 km. The northern rim of
the beltlike region seems to correspond to the edge of the
mantle wedge. In the north of the edge, fluid water would be
depleted to form hydrous minerals such as serpentine in the
mantle wedge. This would be the cause of the relatively
narrow width of the beltlike distribution.
[22] Acknowledgments. We used seismic data from the National
Research Institute for Earth Science and Disaster Prevention, Hokkaido
University, Hirosaki University, Tohoku University, University of Tokyo,
Nagoya University, Kyoto University, Kochi University, Kyushu University, Kagoshima University, the National Institute of Advanced Industrial
Science and Technology, Tokyo metropolitan government, Shizuoka prefectural government, Kanagawa prefectural government, the City of Yokohama, the Japan Marine Science and Technology Center, and the Japan
Meteorological Agency.
References
Chouet, B. A., Long-period volcano seismicity: Its source and use in eruption forecasting, Nature, 380, 309 – 316, 1996.
Hyndman, R. D., K. Wang, and M. Yamano, Thermal constraints on the
seismogenic portion of the southwestern Japan subduction thrust, J. Geophys. Res., 100, 15,373 – 15,392, 1995.
Katsumata, A., Magnitude determination of deep-focus earthquakes in and
around Japan with regional velocity-amplitude data, Earth Planets Space,
53, 333 – 346, 2001.
Kurashimo, E., M. Tokunaga, N. Hirata, T. Iwasaki, S. Kodaira, Y. Kaneda,
K. Ito, R. Nishida, S. Kimura, and T. Ikawa, Geometry of the subducting
Philippine Sea plate and the crustal and upper mantle structure beneath
eastern Shikoku Island revealed by seismic refraction/wide-angle reflection profiling, Zisin 2, 54, 489 – 505, 2002 (in Japanese with English
abstract).
Nishide, N., T. Hashimoto, J. Funasaki, H. Nakazawa, M. Oka, H. Ueno,
N. Yamada, I. Sasakawa, K. Maeda, K. Sugimoto, and T. Takashima,
Nationwide activity of low-frequency earthquakes in the lower crust in
Japan, Abstr. Jpn. Earth and Planet. Sci. Joint Meeting, sk-p002, 2000.
Obara, K., Nonvolcanic deep tremor associated with subduction in southwest Japan, Science, 296, 1679 – 1681, 2002.
Okada, Y., K. Kasahara, S. Hori, K. Obara, and S. Aoi, Hi-net (1): outline,
Prog. Abstr. Fall Meeting Seismol. Soc. Jpn., P004, 2000. (in Japanese)
Schmidt, M. W., and S. Poli, Experimentally based water budgets for
dehydrating slabs and consequences for arc magma generation, Earth
Planet. Sci. Lett., 163, 361 – 379, 1998.
Seno, T., S. Stein, and A. E. Gripp, A model for the motion of the Philippine Sea plate consistent with NUVEL-1 and geological data, J. Geophys.
Res., 98, 17,941 – 17,948, 1993.
The Research Group for Active Faults of Japan, Active faults in Japan,
University of Tokyo Press, Tokyo, pp. 437, 1991 (in Japanese).
Ueno, H., S. Hatakeyama, T. Aketagawa, J. Funasaki, and N. Hamada,
Improvement of hypocenter determination procedures in the Japan Meteorological Agency, Quart. J. Seism., 65, 123 – 134, 2002 (in Japanese
with English abstract).
Ukawa, M., and K. Obara, Low frequency earthquakes around Moho beneath the volcanic front in the Kanto district, central Japan, Bull. Volcanol. Soc. Japan, 38, 187 – 197, 1993 (in Japanese with English abstract).
Zhao, D., S. Horiuchi, and A. Hasegawa, Seismic velocity structure of the
crust beneath the Japan Islands, Tectonophysics, 212, 289 – 301, 1992.
A. Katsumata, Meteorological College, Asahi-cho 7-4-81, Kashiwa,
Chiba 277-0852, Japan. ([email protected])
N. Kamaya, Seismological and Volcanological Department, Japan
Meteorological Agency, Ohtemachi 1-3-4, Chiyoda-ku, Tokyo 100-8122,
Japan. ([email protected])