Acta Geophysica

vol. 56, no. 4, pp. 1000-1014

DOI: 10.2478/s11600-008-0039-5

Swarms in Andaman Sea, India

– a Seismotectonic Analysis
Basab MUKHOPADHYAY and Sujit DASGUPTA

py
Geological Survey of India, Kolkata, India
e-mail: basabmukhopadhyay@yahoo.com (corresponding author)
co
Abstract
The seismotectonic characteristics of 1983-1984, 1993 and 2005
swarms in Andaman Sea are analysed. These swarms are characterised
by their typical pulsating nature, oval shaped geometry and higher b val-
or

ues. The migration path of the swarms from north to south along the An-
daman Spreading Ridge is documented. While the first two swarms are
located along existing mapped rift segments, the 2005 swarm appears to
have generated a new rift basin along 8oN. The analysis and supporting
h

evidences suggest that these swarms were generated by intruding mag-

matic dyke along the weak zones in the crust, followed by rifting, spread-
ut

ing and collapse of rift walls. CMT solutions for 2005 swarm activity
indicate that intrusion of magmatic dyke in the crustal weak zone is do-
cumented by earthquakes showing strike slip solution. Subsequent events
A

with normal fault mechanism corroborate the rift formation, collapse and
its spreading.
Key words: earthquake swarm, Andaman Sea, plume, dyke intrusion, rift.

1. INTRODUCTION
Earthquake swarms are sequences of earthquakes closely clustered in space
and time, in which no single earthquake dominates in size (Mogi 1963,
Scholz 2002). Swarms originate in the crust from ambient stress generated
by volcanic or tectonic activity. In many situations there is an overlapping
signature of both volcanic and tectonic causes. From teleseismic and local
network seismic data, different workers (Sykes 1970, Brocher 1983, Huang
et al. 1986, Ukawa and Tsukahara 1996, Chouet 1996, Dieterich et al. 2000,
________________________________________________
© 2008 Institute of Geophysics, Polish Academy of Sciences
 SWARMS IN ANDAMAN SEA, INDIA 1001

Toda et al. 2002, Hayashi and Morita 2003) have postulated a close relation-
ship between earthquake swarms and magmatic processes. Generation of
swarm in the spreading ridge/rift system includes interaction of intruding
magma with groundwater, crustal heterogeneity, intrusion of dyke etc., and
has a tendency to tract along the moving front of magma intrusion (Hill
1977, Rubin and Pollard 1988). Other authors (e.g., Julian and Sipkin 1985)
suggest that swarm is the result of tensile failure under high fluid pressure,
while Bergmann and Solomon (1990) postulate that propagation of dyke
swarm along the median valley of narrow segmented spreading ridge/rift
zone produces swarms of small earthquakes. Earthquake swarms have also
been produced by extraordinarily high stress release for a relatively short pe-
riod of time by crustal deformation accompanied by magma intrusion (Toda

py
et al. 2002).
The Andaman Sea, a back-arc extensional ridge/rift-transform marginal
basin in the Northeast Indian Ocean has experienced three distinct phases of
swarm activities in recent times, during the years 1983-1984, 1993 and 2005,
co
the last one following the 26 December 2004 Mw 9.3 Sumatra-Andaman
earthquake. In this note, we study the nature and characteristics of these
swarm phases and analyze the seismotectonic implications including redefin-
ing of the rift segment. Earthquake data and parameters for the swarms of
or

1983-1984 and 1993 are taken from ISC database, whereas those for 2005
are from USGS (NEIC) source. CMT solutions of important earthquakes are
taken from Harvard website (www.seismology.harvard.edu). The tectonic
base map utilized for the study is compiled from Curray (2005).
h
ut

2. TECTONIC FRAMEWORK OF THE ANDAMAN SEA

The tectonics of Andaman Sea is complex. Several workers (Curray et al.
1982, Mukhopadhyay 1984, 1988, Curray 1989, Mukhopadhyay and Krish-
A

na 1991, Halder et al. 1992, Roy 1992, Dasgupta and Mukhopadhyay 1993,
1997, Raju et al. 2004, Curray 2005, Khan and Chakraborty 2005) have de-
scribed it as a convergent margin pull-apart basin rather than a typical exten-
sional back-arc basin. Physiography of the Andaman Sea includes the Alcock
and Sewell seamounts, the central Andaman Basin between the two raises,
and smaller topographic features of unknown character. The main spreading
occurs in between the Alcock and the Sewell Rise along the Andaman
Spreading Ridge (ASR) (Fig. 1) and by several leaky transforms up to the tip
of the Sumatra Island. This basin was formed due to oblique subduction of
the Indian Plate from the west beneath the Southeast Asian Plate to the east,
with variable speed and direction of convergence in different parts of the arc
system. The inhomogeneous distributions of movement vectors in different
parts of the arc have caused strike slip faulting parallel to trench (West Andaman
1002 B. MUKHOPADHYAY and S. DASGUPTA

py
co
h or

Fig. 1. Location map showing seismic events of three swarms 1983-1984 (open cir-
cles), 1993 (solid squares) and 2005 (open triangles) along with active tectonic
ut

o
planes in Andaman Sea. Dashed lines are rifts interpreted by Currey 2005 around 8
latitude. WAF = West Andaman Fault, ASR = Andaman Spreading Ridge, B = Bar-
ren Volcano, N = Narcondam.
A

Fault and Sumatra Fault System), generation of sliver plate, trans-tension

and formation of back arc pull-apart extension basin as Andaman Sea. The
integrated study of swath bathymetry, palaeomagnetic and seismological da-
ta indicates that the Andaman Sea was created by rifting over the past 11 Ma
with active seafloor spreading only for the last 4-5 Ma (Raju et al. 2004).

3. CHARACTERISTICS OF SWARM ACTIVITY

The three distinct earthquake swarm clusters in the Andaman Sea are de-
scribed below:
‰ The 1983-1984 swarm activity is defined by 174 seismic events
(magnitude range 3.7 to 5.4) that occurred during the period from 16 De-
cember 1983 to 13 August 1984. Month/day wise detail of seismic burst is
 SWARMS IN ANDAMAN SEA, INDIA 1003

summarised (Table 1). During December 1983 to March 1984 the swarm
clustered around 13°N (Fig. 1) along the northern NE-SW rift segment with
maximum activity recorded on 17 December 1983. The swarm migrated to
the south and clustered around 11°N along the central spreading ridge segment
between Alcock and Sewell ridges during the period April-August 1984. The

Tabl e 1
Statistics of the 1983-1984 swarm

Number of Number of Depth

Year Month Day Mb
events in month events in day [km]
1983 December 49 16 1 4.9 19

py
17*) 40 3.9-5.3 10-44
18 3 4.5-5.4 33-46
20 1 4.7 44
21 1 4.7 10
co 23 2 4.5-4.8 10-34
25 1 4.7 14
1984 January 10 20 10 3.7-5.2 33-41
1984 February 1 1 1 5 37
or

1984 March 1 11 1 4.6 40

1984 April 9 2, 10-13, 21 9 4.5-5.3 20-33
1984 June 1 29 1 4.2 33
h

1984 July 98 5 3 5-5.4 18-40

8**) 42 4.3-5.2 12-158
ut

9 10 4.4-5.3 11-47
10 12 4.2-5.2 13-107
11 12 4.2-5.1 16-73
A

12 1 4.8 32
14 2 4.7-4.8 17-31
15 3 4.4-5.2 14-112
17 3 4.5-4.8 17-78
19 2 4.4-5 13-64
20 2 4.7-5.1 18-27
23 2 4.7-5.1 16-64
24 3 3.9-4.4 33-82
29 1 4.8 13
1984 August 5 1, 13 5 4.4-4.9 31-45
Pulsating nature of the swarm activity: *) one pulse between 00h and 04h, and another
at 06h; **) one pulse at 02h and 03h, another at 05h, followed by pulses at 08h, 14h
and 20h.
1004 B. MUKHOPADHYAY and S. DASGUPTA

maximum number of earthquakes in this segment was recorded on 8 July.

The northern cluster has an aspect ratio of 5 (length 60 km /width 12 km)
while the southern one is 4 (length 60 km /width 15 km) indicating an oval
shaped geometry for both. Although focal depth estimates for small earth-
quakes as given by ISC could have large errors, nevertheless this indicates
that deep sub-crustal events (focal depth from pP-P phase considered)
appear from 8 July onwards and eventually the swarm activity ceases by
mid-August 1984. The b-value for the swarm calculated through least square
method of magnitude Mb (∆m = 0.1) against log ΣN is 1.97.
‰ The 1993 swarm activity commences on 23 August and ends on
29 October that includes 34 seismic events (magnitude range 3.7 to 4.9) and
clusters around 10°N (Fig. 1) at the southwest segment of the central Anda-

py
man Spreading Ridge (ASR). The major activity pulse was on 23-24 August
(Table 2) at the beginning of the swarm. Interestingly, this swarm originates
from where the 1984 swarm ceased and proceeds towards southwest to ter-
co
minate along the West Andaman Fault (WAF). This swarm is also oval shaped
with aspect ratio of 4.3 (length 30 km/width 7 km). The swarm activity ends
in October 1993 after the occurrence of a few upper mantle earthquakes dur-
ing September-October 1993. The b-value for the swarm cluster is 2.01. It may
be noted that this swarm in the back-arc spreading ridge occurred within the
or

March 1991-May 1995 arc volcanism episode from the Barren volcano.

Tabl e 2
h

Statistics of the 1993 swarm

No of events Number of Depth
ut

Year Month Day Mb

in month events in day [km]
1993 August 25 23*) 11 4.4-4.8 14-33
24**) 9 4.2-4.9 16-33
A

25 2 4.2-4.5 14-33
26 2 4.3 33
27 1 4.4 33
1993 September 5 4, 11, 13 5 3.8-4.4 21-115
1993 October 4 4, 5, 29 4 3.7-4.6 33-135
Pulsating nature of the swarm activity: *) one pulse at 19h; **) one pulse at 23h.

‰ The 2005 swarm activity following the great earthquake of 26 De-

cember 2004 is centered near 8oN (Fig. 1) and occurred within a short time
span of only 6 days between 26 and 31 January 2005. This swarm is re-
garded as the most energetic swarm ever observed globally (Lay et al. 2005).
A total of 651 crustal earthquakes (NEIC catalog) are recorded within the
magnitude range from 3.9 to 5.9. The major pulse was on 27-28 January with
 SWARMS IN ANDAMAN SEA, INDIA 1005

more than 200 events/day (Table 3). The shape of the swarm cluster is oval
with aspect ratio of 2.75 (length 110 km/ width 40 km) and the b-value is of
the order of 1.83. The eruption of Barren Volcano (May 2005) takes place
immediately after this swarm activity. It may be noted that unlike the 1983-
1984 and 1993 swarms which occur along the mapped rift segments, the
2005 swarm locates in a segment across strands of WAF without previously
recognized spreading rift segment, though with the same NE-SW trend.

Tabl e 3
Statistics of the 2005 swarm
Numbers of Number of Depth
Year Month Day Mb

py
events in month events in day [km]
2005 January 651 26*) 41 3.9-5.6 13-49
27**) 234 3.9-5.9 16-38
28***) 243 4.0-5.7 13-43
co 29 89 3.9-5.6 16-48
30 25 4.1-5.6 11-30
31 19 4.2-5.2 15-30
Pulsating nature of the swarm activity: *) one pulse at 22h; **) pulses came at 08h,
or

12h -15h, 20h -21h; ***) pulses at 01h -02h, 04h -06h, 08h, 10h, 16h -17h.

The frequency−magnitude distribution of earthquake events for the three

swarms (Fig. 2) indicates that each swarm has one primary and another mi-
h

nor secondary frequency peaks; for the 1983-1984 swarm it is centered

ut

around magnitude 4.4 and 4.8, for 1993 swarm it is 4.3 and 4.6, while for the
2005 swarm it is centered around magnitudes 4.5 and 5.2.
A

Fig. 2. The frequency–magnitude plot of the three swarms.

1006 B. MUKHOPADHYAY and S. DASGUPTA

4. FAULT TYPE AND STRESS PATTERN

Fault type and stress pattern from the 1983-1984 and 2005 swarm sequences
has been analyzed from Harvard Best Double Couple solutions data. No such
solutions are, however, available for the 1993 swarm. The nature of fault
movement associated with swarm generating earthquake has been classified
into normal and strike slip, depending on the maximum plunge of P and N
axes and the direction of slip along fault planes (after Frohlich and Apperson
1992). Out of 14 available solutions for the 1983-1984 swarm, 12 are pure
normal fault solutions while the remaining two are with an appreciable strike-
slip component. The beach ball diagrams of these earthquakes are plotted in
the map (Fig. 3). Analysis of stress tensor data indicates a major compressive
stress (σ1) plunging 66o towards 69o and an extension (σ3) at shallow angle of

py
20o directed towards 159o (Fig. 4). Out of a total of 116 available solutions for
the 2005 swarm, 72 are strike slip earthquakes, 18 pure normal faults while the
remaining 26 events indicate normal faulting with a significant strike-slip
component. The major compressive stress (σ1) plunges 50o towards 6o with a
co
sub-horizontal (16o) extensional axis (σ3) directed towards 115o (Fig. 5a, b).
h or
ut
A

Fig. 3. Map showing seismic events of 1983-1984 (open circles) and 1993 (solid
squares) swarms along with the beach ball diagrams. The active tectonic elements in
Andaman Sea are shown. Arrowhead indicates direction of spreading. WAF = West
Andaman Fault, ASR = Andaman Spreading Ridge, MR = Margui Ridge, B = Bar-
ren Volcano, N = Narcondam.
 SWARMS IN ANDAMAN SEA, INDIA 1007

py
Fig. 4. Stereo-plot of P and T axes of available CMT data of the 1983-1984 swarm.
co
Solid circles indicate P axes and open circles indicate T axes.

From plot of poles of all the 232 nodal planes, two average fault planes are
deciphered, one with strike 58o and dip 84o towards southeast and another
with strike 146o and dip 62o southwest (Fig. 5c).
or

The stress pattern deduced shows a change in orientation spatially from

north to south along ASR. The extension (σ3) direction has changed from
SSE-NNW to ESE-WNW with sub-horizontal plunge.
h
ut

5. MAGMATISM, RIFT TECTONICS AND SWARM

The earthquakes of 2005 swarm have occurred within the aftershock sequence
following the great Sumatra-Andaman earthquake (Mw 9.3) of 26 December
A

2004. The epicenters form a cloud that is aligned across the trend of West
Andaman Fault (WAF) and Seulimeum Strand of Sumatra Fault System
(SEU). For constraining the spatial extent and orientation of this swarm clus-
ter, the better-located earthquake events (with CMT solutions), having magni-
tude (Mb ≥ 4.8), are plotted with large triangles and the rest (Mb < 4.8) with
smaller triangle (Fig. 6). The former earthquake events show an elongated
NE-SW trending zone with a strike length of 110 km and average width 40 km
across WAF and SEU. To understand the seismotectonics of the swarm, the
known rift systems in the vicinity of 8oN latitude is plotted as dashed lines
(Fig. 6). Earthquakes with available CMT solutions, in the number of 116, are
plotted with suitable symbols (Fig. 7), stars (18 in number) denote events with
pure normal fault plane solutions; triangles (26) denote events with normal
fault having appreciable amount of strike slip component and solid circles (72)
1008 B. MUKHOPADHYAY and S. DASGUPTA

py
co
or

Fig. 5. Lower hemisphere equal area net plot of available CMT data of the 2005
swarm: (a) Plot of P axes with solid circles (contour interval 2% per 1% area), mean
vector plunges 50° towards 6°; (b) Plot of T axes with solid circles (contour interval
2% per 1% area), mean vector plunges 16° towards 115°; (c) Plot of poles of 232 fault
h

planes in solid circles with contour interval 2% per 1% area, two inferred average fault
planes (one with strike 58° and dip 84°, the other with strike 146° and dip 62°).
ut

depict events with strike slip fault plane solutions. It is interesting to note
that earthquake events at the beginning of the swarm have fault plane solu-
A

tions with either pure strike slip or normal fault with appreciable strike slip
motion. The first earthquake event with a pure normal fault plane solution
appears only on 27 January at 17h40m48.90s, while the swarm has started
more than a day earlier on 26 January at 03h38m13s. After this, the swarm ac-
tivity is documented by alternate episodes of earthquake events with pure
normal and events with strike slip/normal with appreciable strike slip motion.
The beach ball diagrams of some important events are drawn to show the na-
ture of crustal movements in different parts of this tectonic zone. The direc-
tion of extensional stress is nearly horizontal (16o) along ESE-WNW while
the compressional stress is at moderately high angle (50o) towards the north
(Fig. 5a, b). To our opinion, the formation of this swarm is in response to the
crustal deformation generated by intrusion of magmatic dyke documented by
the initial events showing strike slip and normal solution with appreciable
 SWARMS IN ANDAMAN SEA, INDIA 1009

py
co
or

Fig. 6. Plot of seismic events of the 2005 swarm (open triangles) along with active
tectonic elements and inferred new rift system in Andaman Sea. The dashed line in-
h

dicates interpreted rift system by Currey (2005). The earthquake density in numbers
is high in the center (black) and low (gray) in the periphery. The solid black lines in-
ut

dicate boundary of newly formed rift across WAF and SEU, arrowhead indicates di-
rection of spreading. WAF = West Andaman Fault, SEU = Seulimeum Strand of
Sumatra Fault System.
A

strike slip component. Subsequent events show predominate normal fault me-
chanism, which indicates collapse of the rift walls and subsequent spreading.
This period with alternate phases of normal and strike-slip fault plane solu-
tions represents a cyclic episode of dyke intrusion, spreading and rift forma-
tion. The boundary of the newly formed rift zone is constrained on the basis
of the result obtained by point density spatial statistical function calculated
on earthquake numbers. Point density is calculated as the total numbers of
earthquake points that fall within a circular neighbourhood divided by the area
of neighbourhood (area = π×102 = 314.28 km2). The black-gray to white
zones indicate a high to low earthquake density values. The extremity of the
gray zone is taken as the boundary of the newly developed rift zone and
drawn as solid black lines (Fig. 6). It is interesting to note that epicenters
1010 B. MUKHOPADHYAY and S. DASGUPTA

present in the white zone have little contribution towards formation of the
NE-SW cluster. The orientation of this rift zone is parallel to the average
orientation of one of the nodal fault plane with strike 58o and dip 84o towards
southeast (Fig. 5c).

py
co
h or
ut

Fig. 7. Plot of 116 earthquakes with CMT data of the 2005 swarm, pure normal fault
plane solutions (stars); normal fault with appreciable amount of strike slip motion
(triangles), strike slip fault plane solutions (solid circles). Beach ball diagrams of
A

some important event are plotted. The solid black lines indicate boundary of newly
formed rift across WAF and SEU. WAF = West Andaman Fault, SEU = Seulimeum
Strand of Sumatra Fault System.

Similarly, the 1983-1984 and 1993 swarm is documented along the exist-
ing rift system of ASR. The CMT solutions of 1983-1984 swarm suggest that
earthquakes with appreciable strike slip motion indicate intrusion of dyke from
magma chamber to the weak crust, whereas subsequent earthquakes with
normal fault plane solutions mark the rift formation and spreading.
Our observations in both swarms support the views of Hill’s magma in-
jection model (Hill 1977) which states that the injection of shallow, vertical,
en-echelon dykes extending along a narrow rift zone may account for strike
slip motion initially. This is followed by the collapse of inner rift wall into
closely spaced normal faults (see also Bergmann and Solomon 1990). The
 SWARMS IN ANDAMAN SEA, INDIA 1011

orientation of epicenter, clustering of earthquake with normal fault plane

solution along an elongated NE-SW zone, orientation of nodal planes and
P-T-N axes in CMT solutions and episodic nature of seismicity prompts us
to infer that opening up of a new segment of rift system (boundaries of
which are shown in Figs. 6 and 7 as black lines) takes place perpendicular to
the extension (σ3) and parallel to the average nodal fault plane with strike
58o and dip 84o towards the southeast. This segment makes a high angle with
the existing rift/leaky transform fault system of WAF and SEU. The rift zone
may have many parallel narrow closely spaced rift segments.
The above magmatic model of swarm generation is further supported by
b-value analysis. The b-value is a measure of ratio of smaller earthquakes to
the number of larger earthquakes. The value of b increases where effective

py
stress decreases by interaction of magma with ground water and is also a
sign of magma accumulation and injection (Wiemer and McNutt 1997). The
high b-value obtained in our study (1.97 for 1983-1984 swarm, 2.01 for 1993
swarm and 1.83 for 2005 swarm) indicates that the swarms are generated
co
from magmatic source. The high pore pressure, heterogeneity and tempera-
ture gradient within the magmatic chamber pose buffer to the effective stress
and increase the b-value.
or

6. DISCUSSION AND CONCLUSIONS

Swarms in Andaman Sea are common. The hypocentral distribution and mi-
gration path of the swarms (1983-1984, 1993 and 2005) from north to south
h

along ASR is documented (Fig. 1). The upper mantle beneath the Andaman
Sea is warm and seismic velocities are very slow due to a highly buoyant
ut

warm crust (Shapiro et al. 2007). The activity of the Kerguelen Plume (Weis
and Frey 1996) from the upper mantle can be felt all along the extensional
basin of Andaman Sea up to the junction of ASR and WAF around 10o lati-
A

tude and down south to 8o latitude (see Shapiro et al. 2007). The activity of
the hot spot is further supported by (i) high heat flow, 220-240 mW/m2
(World Heat Flow Database), (ii) presence of high temperature thin crust
with low seismic velocity zone all along the ASR, WAF and SEU up to 5o
latitude inferred from surface wave diffraction tomography (Shapiro et al.
2007), and (iii) weaker seismic coupling between subducting plate with back
arc lithosphere. This low seismic velocity zone physically represents a weak
section of the seismogenic crust (Zhao et al. 2002) amenable for producing
new fracture system. GPS data indicate a non-negligible east-west conver-
gence (Paul et al. 2001); hence, an extension along the length of the plate
boundary. The extensional force along NNW-SSE direction has facilitated
the opening of rift system making high angle with the trench and generation
of swarm in agreement with the Hill’s magma injection model (Hill 1977).
1012 B. MUKHOPADHYAY and S. DASGUPTA

Further, the contrasting temperature gradient in the lower and upper surfaces
of this weak and thin crust makes it vulnerable for rupture. The 1983-1984
and 1993 swarm activities took place along such rupture zone of ASR by
magmatic dyke intrusion from underlying migrating plume. The volcanism
of Barren during March 1991 - May 1995 also confirms such plume activity.
Along the zone around 8oN, the interplate motion partitioning (Lay et al.
2005) along up-thrusted SE Asian plate (Nettles et al. 2005) during the rup-
ture propagation of Sumatra-Andaman earthquake has produced a new weak
zone. The injection of magmatic dyke from the plume below exerts vertical
pressure along the weak zone, which eventually ruptured and opened to form a
rift segment. Swarm releases the accumulated strain generated by this process.
The low-density magma with high amount of exsolved gas and water fills

py
the tensile fault system. Most of the water would exsolve from the magma at
depths shallower than 1 km (Wood 1995) and strength of the uppermost crust
will prevent the vesiculated magma to erupt (see Yoshida et al. 1999 for de-
tails). The swarm activities cease when expelling of exsolved gas reduces the
co
volume of fluid magma or the excess pressure in the magma is balanced with
the pressure from dyke walls that offers viscous resistance. The activity of
plume continues further and makes Barren to erupt in May 2005.
Thus, it may be concluded from the above discussion that the intrusion
or

of magmatic dyke in the thin crust from the plume below followed by
spreading and collapse of rift wall along the rift system may be responsible
for generation of all the swarms (1983-1984, 1993 and 2005) in the Anda-
man Sea. The 1983-1984 and 1993 swarms have generated by rerupturing
h

the existing rift system of ASR, whereas the opening of a new rift segment
ut

along NE-SW direction induced by the activity of Kerguelen plume may

have generated the 2005 swarm. This new rift system proposed makes a high
angle with the existing rift/leaky transform fault system of WAF and SEU.
A

A c k n o w l e d g e m e n t . The authors would like to thank Dr. Will Gos-

nold, Chester Fritz Distinguished Professor, Chair of Geology and Geologi-
cal Engineering, University of North Dakota, Grand Forks, ND 58202 for
providing the heat flow data of Indian Ocean.

References

Bergmann, E.A., and S.C. Solomon (1990), Earthquake swarms on the Mid Atlantic
Ridge: products of magmatic or extensional tectonics? J. Geophys. Res. 95,
4943-4965.
Brocher, T.M. (1983), T-phases from an earthquake swarm on the Mid Atlantic
Ridge at 31.5oN., Marine Geophys. Res. 6, 39-49.
Chouet, A.B. (1996), Long period volcano seismicity: its source and use in eruption
forecasting, Nature 380, 309-316, DOI: 10.1038/380309a0.
 SWARMS IN ANDAMAN SEA, INDIA 1013

Curray, J.R. (1989), The Sunda Arc: a mode for oblique plate convergences, Neth. J.
Sea Res. 24, 131-140.
Curray, J.R. (2005), Tectonics and history of the Andaman Sea region, J. Asian
Earth Sc. 25, 187-232.
Curray, J.R., F.J. Emmel, D.G. Moore, and R.W. Raitt (1982), Structure, tectonics
and geological history of the northeast Andaman Ocean. In: A.E.M. Nairn,
F.G. Stehli, and S. Udeya (eds.), The Ocean Basins and Margins 6, The In-
dian Ocean, Plenum, New York, 399-450.
Dasgupta, S., and M. Mukhopadhyay (1993), Seismicity and plate deformation be-
low the Andaman Arc, Northeast Indian Ocean, Tectonophysics 225, 529-
542.
Dasgupta, S., and M. Mukhopadhyay (1997), Aseismicity of the Andaman subduc-
tion zone and recent volcanism, J. Geol. Soc. Ind. 49, 513-521.

py
Dieterich, J., V. Cayol, and P. Okubo (2000), The use of earthquake rate changes as
a stress parameter at Kilauea volcano, Nature 408, 457-460.
Frohlich, C., and K.D. Apperson (1992), Earthquake focal mechanisms, moment
tensors, and the consistency of seismic activity near plate boundaries, Tec-
tonics 11, 279-296.
co
Halder, D., T. Laskar, P.C. Bandopadhyay, N.K. Sarkar, and J.K. Biswas (1992),
Volcanic eruption of Barren Island volcano, Andaman Sea, J. Geol. Soc.
Ind. 39, 411-419.
or

Hayashi, Y., and Y. Morita (2003), An image of a magma intrusion process inferred
from precise hypocentral migrations of earthquake swarm east of Izu Pen-
insula, Geophys. J. Int. 153, 159-174.
Hill, D.P. (1977), A model of earthquake swarms, J. Geophys. Res. 82, 1347-1352.
h

Huang, P.Y., S.C. Solomon, E.A. Bergman, and J.L. Nabelek (1986), Focal depths
and mechanisms of Mid Atlantic Ridge earthquakes from body waveform
ut

inversion, J. Geophys. Res. 91, 579-598.

Julian, B.R., and S.A. Sipkin (1985), Earthquake process in the Long Valley Caldera
area, California, J. Geophys. Res. 90, 11155-11169.
A

Khan, P.K., and P.P. Chakraborty (2005), Two phase of opening of Andaman Sea:
a new seismotectonic insight, Earth Planet. Sci. Lett. 229, 259-271.
Lay, T., H. Kanamori, C.J. Ammon, M. Nettles, S.N. Ward, R.C. Aster, S.L. Beck,
S.L. Bilek, R. Bredzinski, R. Butler, H.R. Deshon, G. Ekstrom, K. Satake,
and S. Sipkin (2005), The great Sumatra-Andaman earthquake of 26 De-
cember 2004, Science 308, 1127-1132.
Mogi, K. (1963), Some discussions on aftershocks, foreshocks and earthquake
swarms – the fracture of semi-infinite body caused by inner stress origin
and its relation to earthquake phenomena, Bull. Earthq. Res. Inst. Univ. To-
kyo 41, 615-658.
Mukhopadhyay, M. (1984), Seismotectonics of subduction and backarc rifting under
the Andaman Sea, Tectonophysics 108, 229-239.
1014 B. MUKHOPADHYAY and S. DASGUPTA

Mukhopadhyay, M. (1988), Gravity anomalies and deep structure of Andaman arc,

Marine Geophys. Res. 9, 197-210.
Mukhopadhyay, M., and M.R. Krishna (1991), Gravity field and deep structure of
the Bengal fan and its surrounding continental margins, northeast Indian
Ocean, Tectonophysics 186, 365-386.
Nettles, M., A.M. Ekstrom, A.M Dziewonski., N.N. Maternovskaya, and V.C. Tsai
(2005), Source characteristics of the great Sumatra earthquake and its after-
shocks, EOS Trans. AGU 86, Jt. Assem. Suppl., Abstract U43A-01.
Paul, J., R. Burgmann, V.K. Gaur, R. Bilham, K.M. Larson, M.B. Ananda, S. Jade,
M. Mukul, T.S. Anupama, G. Satyal, and D. Kumar (2001), The motion
and active deformation of India, Geophys. Res. Lett. 28, 647-651.
Raju, K.A.K, T. Ramprasad, P.S. Rao, and Juby Varghese (2004), New insights into
the tectonic evolution of the Andaman basin, northeast Indian Ocean, Earth

py
Planet. Sci. Lett. 221, 145-162.
Roy, S.K. (1992), Accretionary prism in Andaman forearc, Geol. Surv. India Special
Publication 29, 273-278. co
Rubin, A.M., and D.D. Pollard (1988), Dike induced faulting in rift zones of Iceland
and Afar, Geology 16, 413-417.
Scholz, C.H. (2002), The Mechanics of Earthquakes and Faulting, 2nd edition,
Cambridge University Press, Cambridge, UK, 471 pp.
Shapiro, N.M., M.H. Ritzwoller, and E.R. Engdahl (2007), Structural context of the
or

great Sumatra-Andaman Island Earthquake, Submitted to Geophys. Res.

Lett., July 4, 2007 (in press).
Sykes, L.R. (1970), Earthquake swarms and sea floor spreading, J. Geophs. Res. 75,
h

6598-6611.
Toda, S., R.S. Stein, and T. Sagiya (2002), Evidence from the AD 2000 Izu Islands
ut

earthquake swarm that stressing rate govern seismicity, Nature 419, 58-61.
Ukawa, M., and H. Tsukahara (1996), Earthquake swarm and dyke intrusion off the
east coast of IZU Peninsula, Central Japan, Tectonophysics 253, 285-303.
A

Weis, D., and F.A. Frey (1996), Role of the Kerguelen plume in generating the east-
ern Indian seafloor, J. Geophys. Res. 101, 13831-13849.
Wiemer, S., and S. McNutt (1997), Variation in the frequency magnitude distribu-
tion in two volcanic regions: Mt. St. Helens (Washington) and Mt. Spurr
(Alaska), Geophys. Res. Lett. 24, 189-192.
Wood, A.W. (1995), The dynamics of explosive volcanic eruptions, Rev. Geophys.
33, 495-530.
Yoshida, S., G. Seta, S. Okubo, and S. Kobayashi (1999), Absolute gravity change
associated with the March 1997 swarm in Izu Peninsula, Japan, Earth Plan-
ets Space 51, 3-12.
Zhao, D., O.P. Mishra, and R. Sanda (2002), Influence of fluids and magma on earth-
quakes: seismological evidence, Phys. Earth Planet. Inter. 132, 249-267.
Received 27 November 2007
Accepted 10 March 2008