Download Repeated foreshock sequences in the thrust faulting environment of

GEOPHYSICAL RESEARCH LETTERS, VOL. 31, L13601, doi:10.1029/2004GL019833, 2004
Repeated foreshock sequences in the thrust faulting environment of
eastern Taiwan
Cheng-Horng Lin
Institute of Earth Sciences, Academia Sinica, Nankang, Taipei, Taiwan
Received 26 February 2004; revised 25 May 2004; accepted 8 June 2004; published 1 July 2004.
[1] Foreshock sequences have repeatedly been observed in
the thrust faulting system of the Chengkung area in eastern
Taiwan. Examinations of earthquakes since 1992 in this area
show all six of the larger earthquakes (ML R 5) had
foreshocks at a distance of only a few kilometres within just
a few days before the main shocks. The largest earthquake
(ML = 6.5) on December 10, 2003 was preceded by two
significant foreshocks about four days and 6 minutes
respectively prior to the main shock. This earthquake
sequence shows the tectonic boundary can most probably
be described as a variable-dip reverse fault. The second
largest earthquake occurred on May 28, 1992 (ML = 5.4)
was similar with two clear foreshocks. Repeated foreshock
sequences in this reverse-faulting environment are primarily
associated with a higher degree of small-scale heterogeneity
in the crust and might very well be considered as potential
INDEX
predictors of large earthquakes in the future.
TERMS: 7209 Seismology: Earthquake dynamics and mechanics;
7223 Seismology: Seismic hazard assessment and prediction; 7230
Seismology: Seismicity and seismotectonics; 8102 Tectonophysics:
Continental contractional orogenic belts. Citation: Lin, C.-H.
(2004), Repeated foreshock sequences in the thrust faulting
environment of eastern Taiwan, Geophys. Res. Lett., 31,
L13601, doi:10.1029/2004GL019833.
[3] Since those previous studies concerning foreshock
sequences were largely obtained in the western United
States where most of the main shocks were dominated by
strike-slip faulting, it is worthwhile to determine whether or
not the occurrence patterns of foreshock sequences and the
interpretations of these patterns are also appropriate to
foreshock occurrences in other areas under such various
regimes as reverse faulting. Simply put, if increasing normal
stress does indeed inhibit the occurrence of a foreshock, it
should likely be very difficult at best to observe foreshocks
for a large earthquake with reverse faulting because normal
stress would significantly increase from normal faulting to
reverse faulting [Sibson, 1974].
[4] For the examination of foreshock sequences in a
reverse faulting environment, Taiwan serves as one of the
best candidates the world over not only because it is under
strong collision between two plates but also because a
wealth of high-quality earthquake data have been collected
from several dense seismic arrays. This study specifically
focuses on foreshock sequences in the Chengkung area of
eastern Taiwan where seismicity is very strong. All large
earthquakes (ML R 5) over the past twelve years are
examined to identify possible temporal and spatial variations
in foreshock sequences.
1. Introduction
2. General Geology and Seismic Networks
[2] Foreshocks might well be considered as one of the
most reliable short-term predictors once the factors that
control foreshock occurrence are fully understood. To
improve our understanding of possible mechanisms for
foreshock occurrences, a number of foreshock sequences
have been investigated over time [Richter, 1958; Mogi,
1963; Utsu, 1970; Jones, 1984; Abercrombie and Mori,
1996]. Among these interpretations, the analyses of the
spatial and temporal distribution of seismicity preceding
the main shocks in the San Andreas Fault system in
California have often shown that the durations of each of
the foreshock sequences decreased with the focal depths of
the main shocks and that no foreshocks occurred with
earthquakes with reverse faulting [Jones, 1984]. Similar
features of foreshock occurrence were later confirmed in a
broader area of the western United States [Abercrombie
and Mori, 1996]. These features were successfully
explained by a decrease in small-scale crustal heterogeneity
with increasing depth, suggesting that increasing normal
stress inhibits the occurrence of foreshocks.
[5] The island of Taiwan is located on a section of the
convergent boundary between the Eurasian plate (EUP) and
the Philippine Sea plate (PSP) (Figure 1). The major part of
Taiwan is a result of the strong collision between the Luzon
volcanic arc of the PSP and the continental margin of the
EUP [Lin, 2002]. The Coastal Range (COR) on the island of
Taiwan represents one part of the Luzon volcanic arc, and
the Eastern Central Range (ECR) displays the easternmost
portion of the continental margin. These two geological
provinces are unmistakably separated by the Longitudinal
Valley (LV), a clear suture zone between the two plates. Off
the island of Taiwan, the Luzon volcanic arc, striking in the
N-S direction, is distinctly marked along the Lanyu and
Lutao islands.
[6] Since the island of Taiwan is under strong plate
convergence, numerous earthquakes dominated by either
reverse or strike-slip faulting [Lin et al., 1985; Cheng, 1995]
have been recorded by several dense seismic arrays [Shin et
al., 2000]. Every earthquake in Taiwan has been routinely
located by a regional seismic network, the Central Weather
Bureau Seismic Network (CWBSN) which is composed of
79 three-component digital short-period seismic stations
covering the Taiwan area. Apart from this, all larger earthquakes (ML R 3.5) have been recorded by a broadband
seismic network, the Broadband Array in Taiwan for
Copyright 2004 by the American Geophysical Union.
0094-8276/04/2004GL019833$05.00
L13601
1 of 4
LIN: REPEATED FORESHOCK SEQUENCES IN EASTERN TAIWAN
L13601
L13601
earthquakes, possible foreshock sequences for the six larger
earthquakes (ML R 5) (Table 1) are shown below.
3.1. The 2003 Earthquake (ML = 6.5)
[8] On December 10, 2003, a large earthquake occurred
near the small town of Chengkung (Figure 2). The focal
depth of the main shock was preliminarily located at the
depth of 10 km by the CWBSN but on the basis of the
waveform inversion of the broadband seismograms, it was
further relocated at the depth of 25 km by the BATS. The
difference might be largely attributed to the waveform
saturation at the CWBSN stations near the epicenter. The
relocated hypocenter in the deep crust is surely more
consistent with an east dipping seismic zone as determined
by hundreds of aftershocks in the following days. In fact, a
variable-dip plane which smoothly decreases as depth
increases might very well be more realistic describing the
seismic zone in the crust. The dipping angle of the seismic
zone at depths shallower than 10 km is close to vertical
(70 – 80 degrees) but gradually decreases to 40– 50 degrees
at depths of 20– 30 km. This seismic zone is also consistent
with the fault plane solution of the main shock as deterFigure 1. (a) Background seismicity (small circles) and
8 larger earthquakes (large circles) in the Chengkung area of
eastern Taiwan shown respectively on the map and
projected (b) in the N-S and (c) W-E profiles. Focal
mechanisms of 8 larger earthquakes and major geological
provinces in eastern Taiwan, namely the Coastal Range
(COR), Longitudinal Valley (LV), Eastern Central Range
(ECR) and Western Central Range (WCR) are shown on the
map. (d) Generalized regional plate tectonics in the Taiwan
area. The plate boundary between the Eurasian plate (EUP)
and Philippine Sea plate (PSP) is indicated by thick lines.
Triangles along one side of the boundary indicate the
subduction direction, and the large arrow shows relative
plate motion.
Seismology (BATS) to immediately determine the CMT
solution [Kao et al., 1998].
3. Foreshock Sequences in the Collision Zone
[7] Background seismicity routinely located by the
CWBSN has validated a large number of earthquakes were
clustered in the Chengkung area (Figure 1), where the
strongest arc-continental collision takes place. Among these
Table 1. Source Parameters for the Six Larger Earthquakes
(ML R 5) and Two Foreshocks (Events 6 and 7) in the Chengkung
Area of Eastern Taiwan Since 1992
Time
Longitude Latitude Depth Magnitude
Event No. (yr/mo/dy/hr/min)
(E°)
(N°)
(km)
(ML)
1
2
3
4
5
6
7
8
1992/05/28/23/19
1995/05/27/18/11
1995/06/06/23/21
1996/12/18/11/20
2000/03/09/05/08
2003/12/06/06/34
2003/12/10/04/32
2003/12/10/04/38
121.351
121.465
121.457
121.358
121.493
121.520
121.312
121.340
23.132
23.008
23.014
22.821
23.222
23.010
23.112
23.100
13.68
19.73
23.65
16.22
33.78
42.00
16.24
25.00
5.4
5.3
5.1
5.0
5.1
5.1
4.1
6.5
Figure 2. (a) Locations of aftershocks (circles), the main
shock and two immediate foreshocks (pluses) of the
December 10, 2003 earthquake sequence in the Chengkung
area of eastern Taiwan. The focal mechanisms of the main
shock and two foreshocks are also shown. The index map
shows the study area and a regional seismic network
(CWBSN) in the Taiwan area. (b) Projections of the
aftershock hypocenters and focal mechanisms of the main
shock and two foreshocks on the back-wall hemisphere in
the A-B profile.
2 of 4
L13601
LIN: REPEATED FORESHOCK SEQUENCES IN EASTERN TAIWAN
mined by the CMT inversion of broadband waveforms
recorded at the BATS.
[9] Significantly, the detailed examination of the 2003
earthquake sequence here found two immediate foreshocks
had been reported by the CWBSN about 4 days (ML = 5.1)
and 6 minutes (ML = 4.1) before the main shock. In general,
the hypocenters and focal mechanisms of the two foreshocks were consistent with both the seismic zone identified
by aftershocks and the focal mechanism of the main shock
(Figure 2). Of interest, one immediate foreshock (ML = 4.1)
occurred 6 minutes before the main shock and was located
at a depth of 16 km along with a great many aftershocks
detected later. This foreshock had a similar focal mechanism
to that of the main shock and was highly indicative of its
occurrence on a single plane in a similar slip direction to
that of the main shock. The other foreshock (ML = 5.1)
preceded the main shock by 4 days at the depth of 42 km
where the main shock was located. Although it was deeper
than the main shock and other aftershocks, a variable-dip
plane might be connected by the clustered seismic zone and
three east-dipping planes of the main shock and two
immediate foreshocks if one considered that all of those
earthquakes took place at the same fault. The extension of
the seismic zone to the surface is clearly located within the
LV (the suture); therefore, this fault might be also considered as the major tectonic boundary between the PSP and
the EUP in the Chengkung area of eastern Taiwan.
3.2. The 1992 Earthquake (ML = 5.4)
[10] Earlier, on May 28, 1992, again in the Chengkung
area, another larger earthquake was recorded by the
CWBSN. This earthquake was the second largest since
1992. The focal mechanism determined by the CMT
[Dziewonski et al., 1981] showed occurrence to be a pure
thrust-faulting earthquake. An east-dipping fault plane was
roughly consistent with aftershock seismicity even though
only some ten aftershocks were recorded. Like the 2003
earthquake, two large immediate foreshocks near the main
shock occurred, separately, about two days (ML = 4.7) and
two hours (ML = 4.9) prior to the main shock.
3.3. Four Other Large Earthquakes
[11] In addition to the two largest earthquakes presented
above, four other large earthquakes (ML R 5) were found in
the CWBSN catalogue of the Chengkung area for the period
under consideration (Figure 1 and Table 1). To ascertain the
possibility of any foreshocks that, within a few days,
preceded those large earthquakes, all earthquakes with
magnitudes between 1.8 and 4.9 were examined. Essentially,
all six of the larger earthquakes had foreshocks within a few
days and several kilometres of the epicentres (Figure 3). In
fact, many foreshocks were detected within only a few hours
and at distances of less than 10 km between each of the main
shocks and respective foreshocks.
4. Discussion
[12] The 2003 earthquake sequence provides some clear
evidence to delineate a major tectonic fault between the PSP
and EUP in the Chengkung area of eastern Taiwan. The
seismic zone determined by clustered aftershocks shows the
major fault plane dipping to the east from the uppermost
crust down to 30 km (Figure 2). Since the extension of this
L13601
Figure 3. Temporal and spatial variations of the foreshocks for 6 larger earthquakes (ML R 5) in the Chengkung
area of eastern Taiwan. For each individual main shock,
foreshocks occurred at the times before the main shock and
are plotted against the horizontal distance from the main
shock.
seismic zone to the surface is clearly located in the LV (the
suture), the crustal boundary between the COR and ECR
might be delineated by the geometry of this seismic zone. If
the two larger immediate foreshocks are further considered
as both occurred at the same fault of the main shock, this
tectonic boundary can be extended down to at least 42 km
along a variable-dip plane. Thus temporal variations of the
2003 earthquake sequences might be shown as: (1) one
larger foreshock (ML = 5.1) first took place at the bottom of
the major tectonic fault four days before the main shock;
(2) another immediate foreshock (ML = 4.1) occurred at
mid-crust depths just 6 minutes before the main shock;
(3) the main shock occurred in the lower crust; and (4) a
numerous of aftershocks were triggered along an eastdipping plane throughout the crust. Therefore, such a
tectonic fault ought to be considered as the major plate
boundary between the volcanic arc of the PSP and the
continental margin of the EUP in the Chengkung area of
eastern Taiwan.
[13] It may also seem surprising to see that foreshocks
repeatedly occurred in a reverse faulting environment contemplated in this study along with previous studies (such as
those by Jones [1984] and Abercrombie and Mori [1996]),
as a general rule, foreshocks more apt to occur in either
normal or strike-slip faulting environments. In the western
United States, for instance, foreshock occurrence is clearly
dependent on both focal depths and slip orientation of the
main shocks [Abercrombie and Mori, 1996]. Foreshocks
which decrease with increasing focal depths of the main
shocks might reflect either small-scale crustal heterogeneity
deceasing with depth or normal stress increasing with depth.
This means foreshocks might be inhibited by either increasing normal stress or decreasing heterogeneity as the focal
depths of main shocks increase. As a result of this, the
absence of foreshocks in a reverse faulting environment
such as that in the San Andreas Fault system can be
explained simply by the characteristics of normal stress
increasing with depth. Like a foreshock sequence in Japan
[Umino et al., 2002], however, it should be kept in mind
that it is unlikely that the foreshock occurrences in the
reverse fault environment of the Chengkung area are asso-
3 of 4
L13601
LIN: REPEATED FORESHOCK SEQUENCES IN EASTERN TAIWAN
ciated with normal stress on the fault plane because it
significantly increases from normal faulting to reverse
faulting [Sibson, 1974]. Instead, a higher degree of smallscale heterogeneity in the crust should be considered to be
the major factor controlling the occurrence of foreshocks in
this reverse fault environment.
[14] A higher degree of small-scale heterogeneity in the
Chengkung area can be also reflected in the background
seismicity. It is widely accepted that small earthquakes, or
foreshocks, can be interpreted as the breaking of small
patches of fault as they reach a critical stress, and thus
leading to more small earthquakes in regions with a higher
degree small-scale heterogeneity zone [Abercrombie and
Mori, 1996]. The background seismicity in eastern Taiwan
shows a large number of small earthquakes clustered in
the Chengkung area. The higher degree of small-scale
heterogeneity represented by both the large number small
earthquakes along with the repeated foreshock sequences in
the Chengkung area of eastern Taiwan might have resulted
from a strong collision of the volcanic arc of the PSP on the
continental margin of the EUP.
[15] Therefore, the repeated foreshock sequences in the
Chengkung area of eastern Taiwan demonstrate an earthquake in the short term may be predicted in an area with a
higher degree of small-scale heterogeneity area where many
small earthquakes are clustered. The foreshock patterns in
the Chengkung area show each of the six larger earthquakes
had foreshocks during the past 12 years. In general, the
magnitude of the foreshocks is smaller than that of the main
shock by about 1 or 2 scales, except the May 28, 1992
earthquake sequence that shows the magnitude of the largest
foreshock is only 0.5 less than that of the main shock. These
results may suggest there is a certain scale length to
heterogeneity associated with foreshock occurrence.
5. Conclusion
[16] On the basis of both aftershock seismicity and the
focal mechanisms of the 2003 Chengkung earthquake
sequence, the major tectonic boundary between the volcanic
arc of the Philippine Sea plate and continental margin of the
Eurasian plate in eastern Taiwan could be very well delineated by a variable-dip plane from surface down to 42 km.
Repeated foreshock sequences in the Chengkung area of
eastern Taiwan do, in fact, indicate foreshocks could have
taken place in reverse faulting as well as in strike-slip or
normal faulting environments. The foreshocks in the reverse
L13601
faulting system might not be associated with normal
stress on the fault plane, but may largely depend on a
higher degree of heterogeneity in the crust. As a consequence, foreshock sequences can be concerned a short-term
predictor in the areas with a higher degree of small-scale
heterogeneity.
[17] Acknowledgments. The author would like to thank two anonymous reviewers for constructive comments. Gratitude is also due to the
Central Weather Bureau in Taipei for providing earthquake catalogue data
and the author’s colleagues (W. T. Liang and Y. H. Liu) in the Institute of
Earth Sciences, Academia Sinica for contributing the CMT solutions. This
research has been funded in part by the National Science Council, Taipei,
Taiwan.
References
Abercrombie, R. E., and J. Mori (1996), Occurrence patterns of foreshocks
to large earthquakes in the western United States, Nature, 381, 303 – 307.
Cheng, S. N. (1995), The study of stress distribution in and around Taiwan,
Ph.D. thesis, 225 pp., Natl. Cent. Univ., Chungli, Taiwan.
Dziewonski, A. M., T. A. Chou, and J. H. Woodhouse (1981), Determination of earthquake source parameters from waveform data for studies of
global and regional seismicity, J. Geophys. Res., 86, 2825 – 2852.
Jones, L. M. (1984), Foreshocks (1966 – 1980) in the San Andreas system,
California, Bull. Seismol. Soc. Am., 74, 1361 – 1380.
Kao, H., P. R. Jian, K. F. Ma, B. S. Huang, and C. C. Liu (1998), Momenttensor inversion for offshore earthquakes east of Taiwan and their
implications to regional collision, Geophys. Res. Lett., 25, 3619 – 3622.
Lin, C. H. (2002), Active continental subduction and crustal exhumation:
The Taiwan orogeny, Terra Nova, 14, 281 – 287.
Lin, C. H., Y. H. Yeh, and Y. B. Tsai (1985), Determination of regional
principal stress directions in Taiwan from fault plane solutions, Bull. Inst.
Earth Sci. Acad. Sin., 5, 67 – 86.
Mogi, K. (1963), Some discussions on aftershocks, foreshocks and earthquake swarms: The fracture of a semi-infinite body caused by an inner
stress origin and its relation to earthquake phenomena, Bull. Earthquake
Res. Inst. Tokyo Univ., 41, 615 – 658.
Richter, C. F. (1958), Elementary Seismology, 768 pp., W. H. Freeman,
New York.
Shin, T. C., K. W. Kou, W. H. K. Lee, T. L. Teng, and Y. B. Tsai (2000), A
preliminary report on the 1999 Chi-Chi (Taiwan) earthquake, Seismol.
Res. Lett., 71, 23 – 29.
Sibson, R. H. (1974), Frictional constraints on thrust, wrench and normal
faults, Nature, 249, 542 – 544.
Umino, N., T. Okada, and A. Hasegawa (2002), Foreshock and aftershock
sequence of the 1998 M = 5.0 Sendai, northeastern Japan, earthquake and
its implications for earthquake nucleation, Bull. Seismol. Soc. Am., 92,
2465 – 2477.
Utsu, T. (1970), Aftershocks and earthquake statistics (II)—Further investigation of aftershocks and other earthquake sequences based on a new
classification of earthquake sequences, J. Fac. Sci. Hokkaido Univ., Ser. 7,
3, 197 – 266.
C.-H. Lin, Institute of Earth Sciences, Academia Sinica, P. O. Box 1-55,
Nankang, Taipei, Taiwan. ([email protected])
4 of 4