Role of crustal heterogeneity beneath Andaman–Nicobar Islands and its implications for coastal hazard

ORI GIN AL PA PER

Role of crustal heterogeneity beneath Andaman–NicobarIslands and its implications for coastal hazard

O. P. Mishra • D. Zhao • Chandan Ghosh • Z. Wang • O. P. Singh •

Biman Ghosh • K. K. Mukherjee • D. K. Saha • G. K. Chakrabortty •

S. G. Gaonkar

Received: 15 January 2010 / Accepted: 24 November 2010 / Published online: 8 December 2010� Springer Science+Business Media B.V. 2010

Abstract The Andaman–Nicobar (A–N) Islands region has attracted many geo-scientists

because of its unique location and complex geotectonic settings. The recent occurrence of

tsunamis due to the megathrust tsunamigenic north Sumatra earthquake (Mw 9.3) with a

series of aftershocks in the A–N region caused severe damage to the coastal regions of

India and Indonesia. Several pieces of evidence suggest that the occurrence of earthquakes

in the A–N region is related to its complex geodynamical processes. In this study, it has

been inferred that deep-seated structural heterogeneities related to dehydration of the

subducting Indian plate beneath the Island could have induced the process of brittle failure

through crustal weakening to contribute immensely to the coastal hazard in the region. The

present study based on 3-D P-wave tomography of the entire rupture zone of the A–N

region using the aftershocks of the 2004 Sumatra–Andaman earthquake (Mw 9.3) clearly

demonstrates the role of crustal heterogeneity in seismogenesis and in causing the strong

shakings and tsunamis. The nature and extent of the imaged crustal heterogeneity beneath

A paper submitted to a special issue of Natural Hazard: ‘‘Tectonics of Bengal Fan and Arabian Sea withspecial emphasis on Coastal and Marine Geohazard’’.

O. P. Mishra (&) � O. P. Singh � K. K. Mukherjee � D. K. Saha � G. K. Chakrabortty � S. G. GaonkarGeo-Seismology Division, CGD, Geological Survey of India (CHQ), Kolkata, Indiae-mail: [email protected]; [email protected];[email protected]

D. ZhaoDepartment of Geophysics, Tohoku University, Sendai, Japan

C. GhoshNational Institute of Disaster Management, New Delhi, India

Z. WangChengdu Institute of Technology, Chengdu, China

B. GhoshGeodata and Database Division, Geological Survey of India, CHQ, Kolkata, India

O. P. MishraSAARC Disaster Management Centre, New Delhi, India

123

Nat Hazards (2011) 57:51–64DOI 10.1007/s11069-010-9678-3

the A–N region may have facilitated the degree of damage and extent of coastal hazards in

the region. The 3-D velocity heterogeneities reflect asperities that manifest what type of

seismogenic layers exist beneath the region to dictate the size of earthquakes and thereby

they help to assess the extent of earthquake vulnerability in the coastal regions. The

inference of this study may be used as one of the potential inputs for assessment of seismic

vulnerability to the region, which may be considered for evolving earthquake hazard

mitigation model for the coastal areas of the Andaman–Nicobar Islands region.

Keywords Crustal heterogeneity � Tsunamis � Seismic imaging � Coastal hazard �Andaman–Nicobar Islands

1 Introduction

The damage survey reported that the Andaman–Nicobar (A–N) Islands is the worst

affected state of India because of strong shaking and tsunami caused by the 2004 Sumatra–

Andaman earthquake (Mw 9.3) (Mishra et al. 2005, 2007a, b). The mainshock also caused

a huge damage to both property and people of most coastal countries of Southeast Asia

(Bilham 2005). The mainshock was felt in the different parts of the world (e.g., Bangla-

desh, India, Indonesia, Malaysia, Maldives, Myanmar, Sri Lanka and Thailand) among

which Indonesia and India were worst affected countries during the tsunamis (Martin 2005;

Mishra et al. 2005, 2007a). In India, the earthquake was severely felt in the entire A–N

Islands and coastal belts of Tamil Nadu and Andhra Pradesh. The majority of damage in

these regions of India was caused by severe tsunamis, which witnessed a huge economic

loss of property and human lives at Carnicobar, Campbell, Camorta and Indira points in the

Andaman–Nicobar region. The earthquake shaking caused damages to buildings and

structures at Diglipur (north Andaman) and Myabandar (middle Andaman) despite their

distant locations of about 1,200 km from the north Sumatra mainshock epicenter. The

coastal people and islanders of Andaman–Nicobar were in shock and were afraid of further

imminent strong quakes and tsunamis in the region (Mishra et al. 2005). There was nothing

but chaos, panic, fear and despair among people in this region after this killer earthquake

(Mishra et al. 2007b), the second largest earthquake in the history of seismology after the

greatest 1960 Chilean earthquake (Mw 9.5) (Stein and Okal 2005; Mishra et al. 2007b).

The A–N region falls in the highest seismic hazard zone-V, which witnessed several past

damaging earthquakes that were akin to the Himalayan collision zone.

Tectonically, the A–N region (Fig. 1a) belongs to a very complex subduction system,

where the Burma–Andaman arc marks the eastern margin of the Indian plate along which

an oblique convergence is suggested by many authors (e.g., Fitch 1970; Curray 2005). The

Burma micro-plate (BMP), often referred to as an overriding sliver plate above the sub-

ducting Indian plate (IP) or subduction fault zone, forms the western segment of the main

Sunda plate (Curray 2005). The trench is a buried feature off Andaman–Nicobar fan

sediments. At the Andaman Sea, the western domain is composed of Andaman–Nicobar

Trench and accretionary subduction complex ridges (Curray 2005). The west Andaman

fault (WAF) with N–S-oriented dextral strike-slip is found to be a prominent fault all along

the A–N Islands that appears to be continuous from west off the northern Sumatra to where

it is lost beneath the terrigeneous fill of the Irrawaddy–Martaban shelf (Fig. 1a). Further

east, the line of the WAF is marked by the proximity of volcanic islands (Barren and

Narcondum).

52 Nat Hazards (2011) 57:51–64

123

Several studies, mainly based on old data recorded by the regional seismic networks,

were carried out to achieve the regional interpretation of the subduction processes in the

A–N region (Le Dain et al. 1984; Mukhopadhyay 1984; Dasgupta and Mukhopadhyay

1993). Moreover, Kayal et al. (2004) made a local aftershock investigation into the 13

September 2002 Diglipur earthquake (Mw 6.5), which, however, mainly confined to the

north Andaman, and they reported a transverse seismogenic structure to the north of the

Andaman Islands. Recently, Kennett and Cummins (2005) conducted a tomographic study

using regional data reported to the International Seismological Center (ISC) to investigate

the physical properties of the subducted slab along the Sumatra–Andaman arc, and they

found that changing morphology of the subduction zone is associated with changes in

physical properties, mainly with bulk modulus variations. However, this study lacks a

detailed insight into the geodynamical processes beneath the A–N region.

Prior to the 26th December 2004 tsunamis, the A–N Islands of India was a poorly

instrumented region without seismographic network except for a solitary permanent analog

seismograph station ascribed to the India Meteorological Department at the Port Blair,

Fig. 1 a Tectonic map of the study area with the 26 December 2004 Sumatra–Andaman mainshock (Mw9.3) shown by the pink star and four big aftershocks (Mw C 6.3) denoted by yellow stars and theircorresponding CMT solutions (peach balls) marked with 1–4. The black and white parts in each peach balldenote compressive and tensional zones, respectively. The cyan triangles denote seismograph locationsinstalled by Geological Survey of India following the mainshock. Volcanic zones (BV Barren volcano, NVNarcondum, and BT Burtang mud volcano) are denoted by golden triangles. Abbreviations denote as IPIndian Plate, BMP Burma micro-plate, GM Gulf of Martban, WAF West Andaman fault, SF Sumatra fault,WR Wharton ridge. The study region is demarcated by red rectangle on the inset map of India. b Epicenterdistribution of aftershocks located by this study. The variation in size of circles denotes the variation inmagnitude of aftershocks (see the magnitude scale at the bottom), and variation in color of the circle denotesvariations in depths of aftershocks (see the depth scale at the bottom). The red star shows the mainshockepicenter. The white triangle denotes the Baratang mud volcano. The black triangles denote Barren (BV)and Narcondum (NV) volcanoes. The white parentheses around volcanic zones denote the aftershock clustersor swarms around the volcanic zones. c Epicenter distribution of the selected aftershocks relocated by usingthe double-difference method. The symbols and tectonic features correspond to the same as those in (a)

Nat Hazards (2011) 57:51–64 53

123

a capital city of Andaman–Nicobar Islands. There was no ocean bottom seismometer

(OBS) in the Andaman Sea to record the suboceanic earthquakes. During January 6–March

16, 2005, six temporary seismic stations were installed at Carnic, Hutbay, Port Blair,

Rangat, Diglipur, and Narcondum (Fig. 1a) by the Geological Survey of India in the

Andaman–Nicobar Islands following the great tsunamigenic earthquake (Mw 9.3), which

recorded a large amount of aftershock data. Analyzing this precious data set may provide a

better insight into the seismotectonics of the A–N Islands. Recently Mishra et al. (2007a)

analyzed about 18,000 aftershocks (M C 3.5) recorded by these six seismograph stations.

Based on erratic and unsteady decay in aftershocks, a north–south-trending aftershock

cluster in an area of about 800 9 300 km2, drastic variations in b-values, and varied

faulting patterns beneath the A–N region, they reported the existence of strong hetero-

geneities in the entire A–N region. Several pieces of evidence suggest that occurrence of

the small and large earthquakes in Andaman–Nicobar region is related to complex seis-

motectonic settings and geodynamical processes in the region. Seismogenesis is generally

dictated by structural heterogeneities that may be revealed by seismic tomography (Zhao

et al. 2002, 2004; Mishra and Zhao 2004; Lei and Zhao 2009; Wang et al. 2009a, b). It is

well documented that earthquake genesis is also influenced by magma and fluids (Zhao

et al. 2002). Deep-seated crustal heterogeneities related to dehydration of the subducting

Indian plate beneath the Island arc also facilitate the process of brittle failure through

crustal weakening and contributing immensely to the coastal hazard in the region (Mishra

and Zhao 2004; Zhao et al. 2004; Mishra et al. 2007a, b). The Andaman–Nicobar region

depicts all of tectonic intricacies which may facilitate the process of seismogenesis in the

region that have caused a huge loss of both flora and faunas due to strong shaking and

tsunamis in the coastal areas of most of South Asian countries, including several coastal

states of India (e.g., Andaman–Nicobar Islands; Tamil Nadu; Andhra Pradesh) during the

huge 2004 Sumatra earthquake (Mw 9.3).

There was no much work done earlier in the region to address the disaster scenario of

Andaman–Nicobar Islands despite the region belongs to seismically active and volcani-

cally prone regime of the geotectonic settings of India. No 3-D tomography study was

carried out earlier on local-scale to know how subsurface heterogeneities were related to

the scale of damage and geological disasters of the region. In the present study, we show

how the nature of seismogenesis may bear a close relationship with nature and extent of

our imaged structural heterogeneities in the region, which in turn may help to understand

how the degree of hazards in the coastal belts is dictated by the seismogenic strength of the

subsurface layers. Eruption of the Baratang mud volcanoes and the Barren lava volcano

during onset of aftershock cycle of the 2004 Sumatra–Andaman earthquake created severe

panic among the local people (Mishra et al. 2007b). Our detailed analyses may provide

significant information to understand the role of structural heterogeneities and their bearing

on stress pattern, which may have facilitated the continuous ejection of mud and slurry

materials during aftershock cycle from the mud crater at the Baratang, middle Andaman.

During earthquake chaos due to killer tsunamis in the year 2004–2005, people had doubt

on premature ejection of lavas from the erstwhile active Barren volcano in the region. We

address the issue by providing a plausible explanation derived from our assimilated 3-D Vp

tomograms of the crust and upper mantle under the region.

In the present study, 3-D seismic structure beneath the Andaman–Nicobar (A–N)

Islands of India has been assimilated, for the first time, on local-scale using arrival time

data from aftershocks recorded by a local temporary seismic network of Geological Survey

of India following the 26 December 2004 Sumatra mainshock (Mw 9.3). Our 3-D tomo-

graphic results may offer some potential information on the seismotectonics and its bearing

54 Nat Hazards (2011) 57:51–64

123

on structural heterogeneity that exist beneath the Andaman–Nicobar region to understand

what caused the crustal and subcrustal earthquakes and how these contributed to coastal

hazards.

2 Data and method

Based on uniform distribution of events along the rupture zone beneath the A–N region,

and smaller horizontal and vertical errors in locations, we selected 2091 aftershocks from

the catalog containing over 18,000 aftershocks (M C 3.5) (Mishra et al. 2007a) (Fig. 1b).

All of the 2091 events were recorded by at least 5 stations and have more than 5 P- and 4 S-

arrivals. In this study, we have used a total of 12,433 P- and 9,691 S-wave arrival time data

from 2091 selected aftershocks. The epicenter distribution of these aftershocks is shown in

Fig. 1b–c. The reading accuracy varies from 0.05 to 0.20 s for P-wave, whereas for

S-wave, the picking accuracy is larger and found to be 0.1–0.35 s. We also used 1,137

P-arrival times from 147 earthquakes recorded by 37 ISC stations as other group of our

dataset for enhancing the ray density in the vicinity of the A–N Islands (Fig. 2). Each of the

ISC earthquakes was recorded by at least 7 stations.

We have applied the tomographic method of Zhao et al. (1992) to the two groups of

arrival time data to determine the 3-D P-wave velocity structure beneath the region. Based

on the study of deep seismic soundings (DSS), gravity data, and the geological informa-

tion, we used a modified 1-D velocity model of Kieckhefer et al. (1981) and Kayal et al.

(2004) as our starting model (Table 1). This model incorporates a P-wave velocity of

4.50 km/s for the upper 3-km thick sediments, 5.7 km/s at depths of 3–10 km, 6.80 km/s at

depths of 10–25 km, 8.0 km/s between 25 and 50 km, and 8.25 km/s at depths greater than

50 km. The corresponding S-wave velocity for the said layers is incorporated in the 1-D

model from the relation Vp/Vs = 1.74 (Mishra et al. 2007a). The grid spacing used in the

3-D tomographic inversion is 50–100 km in the horizontal direction and 3–10 km in depth.

The 2091 selected aftershocks from the originally located data have root-mean-square

(RMS) residuals between 0.63 and 0.95 s with errors in the horizontal (2–6.5 km) and

Fig. 2 Spatial distribution ofseismic rays used for thetomographic inversion. The redsquares denote seismic stations

Nat Hazards (2011) 57:51–64 55

123

vertical (7.0–20.0 km) locations (Mishra et al. 2007a). These aftershocks were relocated by

using the double-difference location method (Wang and Zhao 2006a, b), which showed a

significant reduction in the horizontal and vertical location errors to 1.5–3.5 and

2.5–8.0 km, respectively (Fig. 1c). Both P- and S-arrival times from the 2091 aftershocks

were used in the event relocation, while the 147 events reported to the ISC stations were

not relocated and they were kept fixed in the tomographic inversion.

3 Results and discussion

This endeavor may help us to understand the nature and extent of crustal heterogeneities in

P-wave velocity structure at varying depths. In addition to that, it offers how crustal and

subcrustal heterogeneities are associated with the subducting Indian plate beneath Myan-

mar micro-plate and responsible for inducing moderate to large earthquakes. The 3-D

tomographic image is extended toward volcanic (Barren and Narcondum) zones to

understand geodynamical implications of the imaged heterogeneities in the vicinity of the

volcanic region. Figure 3 demonstrates Vp images at different depths ranging from 3 to

50 km with a distribution of aftershocks that occurred in the region. Cross-sectional images

for the same simulated Vp structure nearly perpendicular to the Andaman Trench axis are

shown in Fig. 4a–d. The results (Figs. 3, 4) depict a very heterogeneous and intriguing

seismic structure beneath the A–N Islands, showing a close correspondence to the local and

regional tectonics. We performed several synthetic tests for various simulated anomalous

zones that appeared in the obtained tomographic image to ascertain the degrees of recovery

and resolvability of our obtained images. The synthetic tests were carried out with syn-

thetic arrival time data that were made by incorporating a random noise of Gaussian

distribution to the original dataset, and then we inverted the synthetic data with the same

model parameterization using the same algorithm as that in the real dataset (Zhao et al.

1992). An example of the synthetic test is shown in Fig. 4 e–f, which shows that the input

anomalies at different locations (Fig. 4e) are well recovered in the output model (Fig. 4f)

for the given grid setup used in the 3-D inversion.

The imaging of high-Vp and low-Vp anomalies (Figs. 3, 4) at shallow layers may

correspond, respectively, to highly dense crystalline rocks and water-bearing sediments or

accreted materials, where the possibility of brittle failure becomes higher. Involvement of

fluids in seismogeneis is well documented for different plate boundary and intraplate

earthquakes elsewhere in the world (e.g., Zhao et al. 2002; Mishra et al. 2008; Lei et al.

2009a; Lei and Zhao 2009). The low-Vp near the Andaman Trench may represent the

deposited terrigeneous sediments or Bengal–Nicobar fan sediments. The low-Vp in the

upper crust (3–8 km) may manifest the presence of accreted sedimentary wedge materials

above the subducting Indian plate and the overriding Burma micro-plate, which is in good

Table 1 The starting 1-Dvelocity model used for the 3-Dtomographic inversion

Number oflayers

Depth oflayers (km)

Vp (km/sec)

1 0.0–3.0 4.50

2 3.0–10.0 5.70

3 10.0–25.0 6.80

4 25.0–50.0 8.00

5 [50.0 8.25

56 Nat Hazards (2011) 57:51–64

123

agreement with the previous studies made for other subduction zones elsewhere in the

world (e.g., Pacheco et al. 1993; Suyehiro and Nishizawa 1993; Zhao et al. 2002; Wang

and Zhao 2006a). Dewey et al. (2007) and Mishra et al. (2007a) reported the occurrence of

aftershocks by a reverse fault near the trench region, which may be suggestive of rupture of

relatively weak crustal rocks at shallow layers than the rupture of competent lithosphere

below the neutral plane of a bent subducting Indian plate (Dewey et al. 2007). This

scenario is well reflected in our tomographic images that rocks in the vicinity of the

Andaman Trench are associated with low-Vp, indicating relatively weak crustal rocks

(Figs. 3, 4). The subducting Indian plate has been imaged prominently by high-Vp

anomalies, indicating a colder and thicker Indian slab beneath the A–N Islands. However, a

drastic variation of structural heterogeneity all along the subducting Indian plate may

suggest changes in the physical properties of the Indian slab that might have influenced the

amount of slip along the entire rupture zone beneath the A–N Islands during the 2004

earthquake cycle. A similar observation has also been made by Kennett and Cummins

(2005) using regional estimates of shear and bulk velocity ratio.

There was a massive eruption of mud and slurry materials to the surface at the Baratang

just within 24 h of the great Sumatra mainshock and four big aftershocks beneath the A–N

region, indicating the existence of a high compressive force that resulted in reverse to dip-

slip fault mechanisms beneath the Baratang volcano (Mishra et al. 2007b), which might

have compelled the ejection of mud and slurry materials to the surface. This hypothesis is

Fig. 3 Map views of P-wave velocity images (in %). Shown at the upper-left corner of each map are thelayer depth and the depth range (numbers in the parentheses) of the aftershocks (white dots). The black starsand black triangles denote the big aftershocks and volcanic zones (Baratang, Barren and Narcondum),respectively

Nat Hazards (2011) 57:51–64 57

123

supported by the tomographic imaging that showed a prominent low-Vp at a very shallow

layer (2–3 km) that may correspond to sandy–clayey–slurry materials (Figs. 3, 4). A

prominently high-Vp anomaly beyond this depth range ([3 km) is imaged as the competent

and compact basement beneath the Baratang mud volcano that may have accumulated and

exerted enough pressure to steer the ejection process to the surface (Fig. 4b). It is intriguing

to note that the variation in low-Vp anomaly maintains a very consistent trend beneath the

Barren and Narcondum volcanic zones, and the anomaly is found to be oriented toward the

Andaman Trench as depth increases from 3 to 15 km (Fig. 4b–d), indicating the source of

magma chambers of these lava volcanoes is possibly related to the process of dehydration of

the subducting Indian plate rather than the influence of the Sumatra–Andaman spreading

ridge (Fig. 1). The clustering of aftershocks around the Barren and Narcondum volcanoes or

occurrence of volcanic tremors during January to March 2005 (Fig. 1b–c; Mishra et al.

2007b) suggested the weakening of crustal layers through the process of underheating

(Tatsumi 1989; Zhao et al. 2002). The source of magma chambers of these lava volcanoes is

possibly related to the process of dehydration of the subducting Indian plate, which is well

supported by a recent study around the Tengchong volcano in southwest China using both

local and teleseismic data (Lei et al. 2009b).

The underheating of the crust beneath the Barren and Narcondum volcanoes associated

with tensional forces and a swarm type of earthquakes supports the concept of impending

Fig. 4 (a–d) Vertical cross-sections of P-wave velocity image along the profiles shown on the inset map.Small white dots denote the aftershocks within 100 km width of each cross-section. The deeper earthquakesare taken from the ISC catalog (1998–2002). Black arrows indicate the plausible direction of hot upwellingfrom deeper to shallow layers. The velocity perturbation scale is shown beside (b). A synthetic test for Vpimage represented by the input model (e), and the corresponding output result (f), in a vertical cross-sectionalong 12 N, indicates the recovery of the prominent low-V and high-V anomalies revealed by this study

58 Nat Hazards (2011) 57:51–64

123

eruption (Linde and Sacks 1998), which finally erupted in June–July 2005 after a gap of

about 3–4 months from the 2004 Sumatra mainshock (Mw 9.3) as per our expectation

(Mishra et al. 2005, 2007a, b). Four big aftershocks (Mw [ 6.2) occurred beneath the

Islands that produced enough seismic energy to disturb the volcanic chambers through

shaking in the region (Linde and Sacks 1998). All these observations related to volcanic

activity are supported by our conspicuous imaging of low-Vp beneath Barren and Nar-

condum volcanic zones in the region (Figs. 3, 4b–d).

It is worth to point that the Andaman–Nicobar region of India suffered severely during

the megathrust Sumatra earthquake where public panic prevailed in every part of the

Andaman and Nicobar Islands following the mainshock on 26 December 2004. The present

result vindicates our earlier inference based on our continuous analyses of three-month

(January–March 2005) seismological data monitoring and recording of aftershocks fol-

lowing the mainshock to show how this endeavor reduced public panic and constituted an

important ingredient to a disaster management program for the Andaman–Nicobar region.

The 3-D seismic imaging of the A–N Islands region helped us to understand the nature and

extent of crustal heterogeneities in the region. Crustal and subcrustal heterogeneities

associated with the subducting Indian plate beneath Myanmar micro-plate are responsible

for inducing moderate to large earthquakes. The puzzle among Islanders due to continuous

shaking during the 2004 tsunamigenic aftershock cycle was very prevalent on two

important issues to know: whether the continuous shaking of the region led to eruption of

Barren volcano in the region? and whether any possibility of conversion of Baratang mud

volcano into lava volcano was there? We extended 3-D tomographic imaging toward

Baratang mud volcano and past erupted lava volcanic (Barren and Narcondum) zones to

see the geodynamical implications of the imaged heterogeneities in the vicinity of volcanic

regions. Our 3-D seismograms (Figs. 3, 4) provide a convincing explanation to both of the

issues. As explained before, the eruption of mud and slurry materials at Baratang may have

happened due to compressive forces exerting to the upper crustal layer from the deeper

layers (Mishra et al. 2007b), which is well reflected by low-Vp at shallow depths in our 3-

D seismic cross-section (Fig. 4). There is no evidence for any interconnectivity of Bara-

tang mud volcano with the Barren lava volcano beneath the A–N Island region.

Our detailed 3-D Vp image clearly explains how mud and lava volcanoes erupted after

megathrust Sumatra earthquake (Mishra et al. 2007b). The present results clearly explain

the puzzle related to mud and lava volcanoes following the 2004 Sumatra mainshock and a

series of its aftershocks; 3-D tomographic imaging provides a very convincing mechanism

to show how mud and lava volcanoes erupted after megathrust Sumatra earthquake (Mishra

et al. 2007b), which constitutes another ingredient to a disaster management program to

reduce future public panic during earthquake chaos once inhabitants of the region would be

timely educated by scientists on these facts. In addition, it provides insight into how crustal

and subcrustal heterogeneities are associated with the subducting Indian plate beneath the

Myanmar micro-plate and responsible for inducing moderate to large earthquakes.

Since almost all recorded aftershocks confined to depths of 30–45 km, and a very few

aftershocks located below this depth range, which may suggest the existence of complex

mechanical, geophysical, and geochemical processes occurring beneath the complicated

subduction system, which might have resisted the occurrence of the intermediate-depth or

deeper earthquakes beneath the A-R region, it is also possible that deep-sea faulting or

tearing of the Indian lithosphere during episodic large to moderate earthquakes beneath the

Andaman Sea (Curray 2005) might have infiltrated enough water into deeper layers to

facilitate the process of dehydration to cause excessive serpentinization of the fore-

arc mantle (Hyndman and Peacock 2003; Mishra et al. 2003; Mishra and Zhao 2004;

Nat Hazards (2011) 57:51–64 59

123

Zhao et al. 2009), which may corroborate to prominently imaged low-Vp at deeper layers

([45 km) (Figs. 3, 4), which in turn might have hindered the occurrence of the interme-

diate-depth or deeper earthquakes during the 2004 Sumatra–Andaman aftershock

sequence. There may be other possibilities for the low-Vp anomalies in the deeper layers

([60 km), such as upwelled hot magmatic materials or dry and ductile partial melts, where

none of aftershocks occurred in the forearc mantle (Figs. 3, 4). In addition, the vertical

propulsion of mantle materials from deeper layers in the forearc mantle might have exerted

enough stress on overlying crustal rocks to bring them into a brittle failure (Fig. 4),

indicating the influence of deep structure and regime on shallower crustal earthquakes

(Zhao et al. 2004). However, we need other pieces of supportive evidence from geophysics

(e.g., gravity, magnetic, and magnetotelluric), experimental petrology and mineral physics

to arrive at a better interpretation of our tomographic results. Thus, we can infer that deep-

seated crustal heterogeneities related to dehydration of the subducting Indian plate beneath

the Island facilitate the process of brittle failure through crustal weakening and contrib-

uting immensely to the coastal hazard in the region. Our 3-D tomographic images of the

region are resolvable for layers not beyond 80 km because we used only local earthquake

data. The 3-D seismic structure beyond 80 km depth along with estimate of the geometry

of the subducting Indian plate can be achieved in future by using data from teleseismic

events (e.g., Lei and Zhao 2005; Zhao et al. 2006; Gupta et al. 2009).

Thus, the extent of coastal hazard due to suboceanic earthquakes can be related to

nature and extent of crustal heterogeneities beneath the sea. The structural heterogeneities

in the crust and upper mantle under the Andaman–Nicobar region make the region vul-

nerable for frequent tremors and shaking, and they also facilitate the magmatic eruption

process for the Barren lava volcano due to continuous shaking through occurrence of large

aftershocks in the region, which in turn may facilitate coastal hazard. The subsurface layer

is associated with high-V anomalies that infer the compact and competent parts of layers,

which may apt to accumulate more strain energy for a longer duration, and once brittle

failures occur in the layer, they may cause strong shaking and tsunamis through release of

high amount of seismic energy during strong to great earthquakes, and such scenario can

bring a huge loss of flora and faunas of the region (Table 2). We found that the occurrence

of earthquakes in the A–N Islands is associated with all sorts of anomalous layers having

low-Vp, intermediate-Vp, and high-Vp anomalies (Figs. 3, 4). We infer that the competent

and compact upper layers with very high-Vp anomalies beneath the A–N Islands may

result in moderate to strong earthquakes through weakening of strengths of rock materials

due to deep-seated extraneous agencies, exerting in form of mantle upwelling or magma

injection or injection of fluids to the upper crustal layers through the process of dehydration

of the subducting Indian plate. Thus, the extent of coastal hazards can be dictated by the

nature of structural heterogeneities and geological asperity of the seismogenic subsurface

layers. Table 2 shows the qualitative correlation between our tomographic image and the

extent of coastal hazards, keeping other conditions (e.g., mode and type of faulting, at-

tenuative agencies and amount of seismic wave attenuation) constant.

Nature of 3-D Vp structural heterogeneities may correspond to asperities that manifest

what type of seismogenic layers exists beneath the region to dictate the size of earthquake

occurrence and thereby the extent of earthquake vulnerability to coastal regions may be

estimated. Estimates of shear wave velocity and Poisson’s ratio structures in future may

provide a better constraint to our interpretation. Use of reanalyzed aftershocks of the 2004

great Sumatra earthquake in the A–N Islands using temporary local network and global

arrival time data by Engdahl et al. (2007) and precise relocations of the suboceanic

60 Nat Hazards (2011) 57:51–64

123

Ta

ble

2C

orr

elat

ion

bet

wee

nth

e3

-DV

pst

ruct

ure

and

the

exte

nt

of

coas

tal

haz

ard

s

Nu

mb

ero

fsu

bsu

rfac

ecr

ust

alla

yer

s

Nat

ure

of

3-D

Vp

ano

mal

ies

Nat

ure

of

sub

surf

ace

lay

ers

wit

hv

ary

ing

dep

ths

Ty

pes

of

eart

hq

uak

eo

ccu

rren

ceP

oss

ibil

ity

of

tsun

amis

occ

urr

ence

Ex

ten

to

fco

asta

lh

azar

dk

eep

ing

oth

erco

nd

itio

ns

con

stan

ts

Rem

ark

s

1H

igh

(Fro

mth

eav

erag

ev

eloci

tym

od

elu

sed

inth

e3

-Din

ver

sio

n)

Co

mp

eten

t/co

mp

act

lay

er/h

igh

lyd

ense

cryst

alli

ne

rock

s/p

late

boundar

yor

inte

rfac

e/la

yer

sw

ith

sup

ersa

tura

ted

flu

ids

wit

hco

mple

tece

men

tati

on

Str

on

gto

gre

atin

the

hig

has

per

ity

/st

ruct

ura

lh

eter

og

enei

tyzo

ne

Ver

yh

igh

Ver

yh

igh

(1)

Ty

pes

of

fault

ing

,m

od

eo

ffa

ult

ing,

atte

nuat

ive

agen

cies

,an

dam

ou

nt

of

seis

mic

wav

eat

tenuat

ion

rem

ain

const

ant.

(2)

Str

uct

ura

lh

eter

ogen

eiti

esco

rres

po

nd

toF

igs.

3an

d4.

(3)

Bu

ild

ing

typ

olo

gy

and

des

ign

cod

ear

eu

nif

orm

2In

term

edia

teor

com

pli

ant

(Fro

mth

eav

erag

ev

eloci

tym

od

elu

sed

inth

e3

-Din

ver

sio

n)

Rel

ativ

ely

less

com

pac

tla

yer

sas

soci

ated

wit

hm

ater

ials

of

inte

rmed

iate

den

sity

,su

bsu

rfac

ela

yer

so

ffo

rear

can

db

ack

arc

reg

ions

Mo

der

ate

tost

ron

gin

the

com

pli

ant

asp

erit

y/s

tru

ctu

ral

het

ero

gen

eity

zon

e

Unce

rtai

nR

elat

ivel

ylo

wer

than

that

of

lay

er1

3L

ow

(Fro

mth

eav

erag

ev

eloci

tym

od

elu

sed

inth

e3

-Din

ver

sio

n)

Wea

ker

lay

ers/

dec

ou

ple

dp

late

bo

und

ary

/pla

tein

terf

ace/

det

achm

ent

layer

san

dla

yer

sas

soci

ated

wit

hfl

uid

-rel

ated

low

den

sity

intr

usi

ve

or

ther

mal

anom

alie

s/ac

cret

edse

dim

enta

ryw

edg

e/v

olc

anic

zon

es,

zon

eso

fsu

bdu

ctin

gp

late

deh

ydra

tio

nan

dm

antl

eu

pw

elli

ng

Mic

roto

mod

erat

ein

the

low

asp

erit

yzo

ne/

stru

ctura

lh

eter

og

enei

tyzo

ne

Neg

lig

ible

Rel

ativ

ely

much

low

erth

anth

ose

of

lay

ers

1an

d2

.

Nat Hazards (2011) 57:51–64 61

123

aftershocks with well-detected sP-depth phase (Mishra et al. 2003; Zhao et al. 2002, 2009)

in future may improve the tomographic images of the Andaman–Nicobar forearc region.

4 Concluding remarks

Our investigation of the 3-D seismic structure beneath the Andaman–Nicobar (A–N)

Islands of India by inverting a large number of high-quality arrival time data from the 2091

selected aftershocks recorded by a local temporary seismic network revealed strong lateral

P-wave velocity variations (up to 6%) beneath the entire region of the A–N Islands,

indicating a close correspondence between the structural heterogeneities and failure of rock

materials. The subducting Indian plate and the volcanic zones (Barren and Narcondum) are

imaged as high-Vp and low-Vp anomalies, respectively. We infer that a prominent low-Vp

beneath the A–N Islands near the Andaman Trench may be associated with a deep-sea

faulting, or weak crustal rocks, or tearing of the Indian lithosphere during historical

damaging earthquakes. The low-Vp in the upper mantle beneath the subducting Indian

plate may correspond to hot upwelling materials from the deep mantle. A low-Vp anomaly

is revealed in the uppermost mantle above the Indian slab, which may reflect the dry and

ductile partial melts or excessively serpentinized forearc mantle that possibly resisted the

occurrence of intermediate-depth aftershocks ([45 km) beneath the A–N region. The

occurrence of shallow earthquakes in the A–N region may be related to the deep structure

and regime rather than the crustal heterogeneity alone. The extent of coastal hazard due to

suboceanic earthquakes can be related to the nature and extent of crustal heterogeneities

beneath the sea. Our present result vindicates the earlier inference based on our continuous

analyses of three-month (January–March 2005) seismological data monitoring and

recording of aftershocks following the 2004 mainshock (Mishra et al. 2007a, b) to show

how this endeavor reduced public panic and constituted an important ingredient to a

disaster management program for the Andaman–Nicobar region. The present results

clearly explain the puzzle related to mud and lava volcanoes following the 2004 Sumatra

mainshock and a series of its aftershocks; 3-D tomographic imaging provides a very

convincing mechanism to show how mud and lava volcanoes erupted after megathrust

Sumatra earthquake (Mishra et al. 2007b), which constitutes another ingredient to a

disaster management program to reduce future public panic during earthquake chaos once

inhabitants of the region would be timely educated with these facts by scientists.

However, in the same context, other pieces of supportive evidence from geophysics,

experimental petrology, and mineral physics coupled with shear wave velocity and Pois-

son’s ratio structures may provide a better constraint to our interpretation in future.

Incorporation of the reanalyzed aftershocks of the 2004 great Sumatra earthquake in the

Andaman–Nicobar Islands using the temporary local network and global arrival time data

and relocating the suboceanic aftershocks with well-detected sP-depth phase may improve

our tomographic results.

Acknowledgments OPM is grateful to Director General, Geological Survey of India, for his kindsupport and permission for carrying out this research. Infrastructural supports provided by localauthorities of Andaman–Nicobar administration, India, during our aftershock monitoring are gratefullyacknowledged. Critical reviews of two anonymous referees improved the original version of the MS. Wethank the Editor of the special issue for his continuous cooperation to process the revision and sub-mission of the MS.

62 Nat Hazards (2011) 57:51–64

123

References

Bilham R (2005) A flying start, then a slow slip. Science 308:1126–1127Curray JR (2005) Tectonics and history of the Andaman Sea region. J Asian Earth Sci 25:187–232Dasgupta S, Mukhopadhyay M (1993) Seismicity and plate deformation below Andaman arc, northeastern

Indian Ocean. Tectonophysics 225:529–542Dewey JW, Choy G, Presgrave B, Sipkin S, Tarr AC, Benz H, Earle P, Wald D (2007) Seismicity associated

with the Sumatra–Andaman Islands earthquake of December 26, 2004. Bull Seismol Soc Am97(1A):S34–S49

Engdahl ER, Villasenor A, DeShon HR, Thurber CH (2007) Teleseismic relocation and assessment ofseismicity (1918–2005) in the region of the 2004 Mw 9.0 Sumatra-Andaman and 2005 Mw 8.6 NiasIsland great earthquakes. Bull Seismol Soc Am 97(1A):S86–S105

Fitch TJ (1970) Earthquake mechanisms in the Himalayan, Burmese, and Andaman regions and continentaltectonics in central. Asia J Geophys Res 75:2699–2709

Gupta S, Zhao D, Rai S (2009) Seismic imaging of the upper mantle under the Erebus hotspot in Antarctica.Gondwana Res 16:109–118

Hyndman RD, Peacock SM (2003) Serpentinization of the forearc mantle. Earth Planet Sci Lett212:417–432

Kayal JR, Gaonkar SG, Chakrabortty GK, Singh OP (2004) Aftershocks and seismotectonic implications ofthe 13 September 2002 earthquake (Mw 6.5) in the Andaman Sea basin. Bull Seismol Soc Am94:326–333

Kennett BLN, Cummins PR (2005) The relationship of the seismic source and subduction zone structure forthe 2004 December 26 Sumatra–Andaman earthquake. Earth Planet Sci Lett 239:1–8

Kieckhefer RM, Moore GF, Emmel FJ (1981) Crustal structure of the Sunda forearc region west of centralSumatra from gravity data. J Geophys Res 86:7003–7012

Le Dain AY, Tapponier P, Molnar P (1984) Active faulting and tectonics of Burma and surrounding regions.J Geophysics Res 89:453–472

Lei J, Zhao D (2005) P-wave tomography and origin of the Changbai intraplate volcano in northeast Asia.Tectonophysics 397:281–295

Lei J, Zhao D (2009) Structural heterogeneity of the Longmenshan fault zone and the mechanism of the2008 Wenchuan earthquake (Ms 8.0). Geochem Geophys Geosyst 10:Q10010

Lei J, Zhao D, Steinburger B, Shen F, Wu B, Li Z (2009a) New seismic constraints on the upper mantlestructure of the Hainan plume. Phys Earth Planet Inter 173:33–50

Lei J, Zhao D, Su Y (2009b) Insight into the origin of the Tengchong intraplate volcano and seismotectonicsin southwest China from local and teleseismic data. J Geophys Res 114:B05302

Linde AT, Sacks IS (1998) Triggering of volcanic eruptions. Nature 371:408–410Martin SS (2005) Intensity distribution from the 2004 M 9.0 Sumatra–Andaman earthquakes. Seism Res

Lett 76:321–333Mishra OP, Zhao D (2004) Seismic evidence for dehydration embrittlement of the subducting Pacific slab.

Geophys Res Lett 31:L09610Mishra OP, Zhao D, Wang Z (2008) The genesis of the 2001 Bhuj, India, earthquake (Mw 7.6): a puzzle for

peninsular india. J Indian Miner (Special Issue) 61(3–4) & 62(1–4):149–170Mishra OP, Zhao D, Umino N, Hasegawa A (2003) Tomography of northeast Japan forearc and its

implications for interpolate seismic coupling. Geophys Res Lett 30. doi:10.1029/2003GL017736Mishra OP, Chakrabortty GK, Singh OP (2005) Aftershock investigation in Andaman–Nicobar Islands. Geol

Surv Ind Spcl Publ 89:115–163Mishra OP, Kayal JR, Chakrabortty GK, Singh OP, Ghosh D (2007a) Aftershock investigation in Andaman–

Nicobar Islands of India and its seismotectonic implications. Bull Seismol Soc Am 97(1A):S71–S85Mishra OP, Chakrabortty GK, Singh OP, Kayal JR, Ghosh D (2007b) Aftershock investigation in the

Andaman–Nicobar Islands: an antidote to public panic. Seismol Res Lett 78:591–600Mukhopadhyay M (1984) Seismotectonics of subduction and backarc rifting under the Andaman Sea.

Tectonophysics 108:229–239Pacheco JF, Sykes LR, Scholz CH (1993) Nature of seismic coupling along simple plate boundaries of the

subduction type. J Geophys Res 98:14,133–14159Stein S, Okal EA (2005) Speed of the Sumatra earthquake. Nature 434:581–582Suyehiro K, Nishizawa A (1993) Crustal structure and seismicity beneath the forearc off northeastern Japan.

J Geophys Res 99:22331–22347Tatsumi Y (1989) Migration of fluid phases and genesis of basalt magmas in subduction zones. J Geophys

Res 94:4697–4707

Nat Hazards (2011) 57:51–64 63

123

http://dx.doi.org/10.1029/2003GL017736

Wang Z, Zhao D (2006a) Seismic imaging of the entire arc of Tohoku and Hokkaido in Japan using P-wave,S-wave and sP depth-phase data. Phys Earth Planet Inter 152:144–162

Wang Z, Zhao D (2006b) Seismic images of the source area at the 2004 mid-Niigata prefecture earthquakein northeast Japan. Earth Planet Sci Lett 244:16–31

Wang Z, Zhao D, Huang R, Tang X, Mishra OP (2009a) Structural heterogeneity in Northeast Japan and itsimplications for the genesis of the 2004 and 2007 Niigata earthquakes. Bull Seism Soc Am99:3355–3373

Wang Z, Fukao Y, Zhao D, Kodaira S, Mishra OP, Yamada A (2009b) Structural heterogeneities in the crustand upper mantle beneath Taiwan. Tectonophysics 476:460–477

Zhao D, Hasegawa A, Horiuchi S (1992) Tomographic imaging of P and S wave velocity structure beneathnortheastern Japan. J Geophys Res 97:19909–19928

Zhao D, Mishra OP, Sanda R (2002) Influence of fluids and magma on earthquakes: seismological evidence.Phys Earth Planet Inter 132:249–267

Zhao D, Tani H, Mishra OP (2004) Crustal heterogeneity in the 2000 western Tottori earthquake region:effects of fluids from slab dehydration. Phys Earth Planet Inter 145:161–177

Zhao D, Lei J, Inoue T, Yamada A, Gao S (2006) Deep structure and origin of the Baikal rift zone. EarthPlanet Sci Lett 243:687–691

Zhao D, Wang Z, Umino N, Hasegawa A (2009) Mapping the mantle wedge and interplate thrust zone of thenortheast Japan arc. Tectonophysics 467:89–106

64 Nat Hazards (2011) 57:51–64

123

Role of crustal heterogeneity beneath Andaman–Nicobar Islands and its implications for coastal hazard

Documents