Attenuation characteristics in eastern Himalaya and southern Tibetan Plateau: An understanding of the physical state of the medium

Accepted Manuscript Attenuation characteristics in eastern Himalaya and southern Tibetan Plateau: an understanding of the physical state of the medium Sagar Singh, Chandrani Singh, Rahul Biswas, Sagarika Mukhopadhyay, Himanshu Sahu PII: DOI: Reference:

S0031-9201(16)30063-2 http://dx.doi.org/10.1016/j.pepi.2016.05.005 PEPI 5917

To appear in:

Physics of the Earth and Planetary Interiors

Please cite this article as: Singh, S., Singh, C., Biswas, R., Mukhopadhyay, S., Sahu, H., Attenuation characteristics in eastern Himalaya and southern Tibetan Plateau: an understanding of the physical state of the medium, Physics of the Earth and Planetary Interiors (2016), doi: http://dx.doi.org/10.1016/j.pepi.2016.05.005

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

3

Attenuation characteristics in eastern Himalaya and southern Tibetan Plateau: an understanding of the physical state of the medium

4

Sagar Singh

5

Department of Geology and Geophysics, Indian Institute of Technology Kharagpur, India

6

Chandrani Singh

7

Department of Geology and Geophysics, Indian Institute of Technology Kharagpur, India

8

Rahul Biswas

9

Department of Geology and Geophysics, Indian Institute of Technology Kharagpur, India

10

Sagarika Mukhopadhyay

11

Department of Earth Sciences, Indian Institute of Technology Roorkee, India

12

Himanshu Sahu

13

Department of Geology and Geophysics, Indian Institute of Technology Kharagpur, India

1

2

14

Abstract

15

Attenuation characteristics of the crust in the eastern Himalaya and the south-

16

ern Tibetan Plateau are investigated using high quality data recorded by Hi-

17

malayan Nepal Tibet Seismic Experiment (HIMNT) during 2001-2003. The

18

present study aims to provide an attenuation model that can address the physi-

19

cal mechanism governing the attenuation characteristics in the underlying medium.

20

We have studied the Coda wave attenuation (Qc ) in the single isotropic scatter-

21

ing model hypothesis, S wave attenuation (Qs ) by using the coda normalization

22

−1 method and intrinsic (Q−1 i ) and scattering (Qsc ) quality factors by the multi-

23

ple Lapse Time Window Analysis (MLTWA) method under the assumption of

24

multiple isotropic scattering in a 3-D half space within the frequency range 2-12

25

Hz. All the values of Q exhibit frequency dependent nature for a seismically

26

active area. At all the frequencies intrinsic absorption is predominant compared

27

to scattering attenuation and seismic albedo (B0 ) are found to be lower than ∗ corresponding author Preprint submitted to Physics of the earth and planetary interiors Email address: [email protected] (Chandrani Singh )

May 13, 2016

28

0.5. The observed discrepancies between the observed and theoretical models

29

can be corroborated by the depth-dependent velocity and attenuation structure

30

as well as the assumption of a uniform distribution of scatterers. Our results

31

correlate well with the existing geo-tectonic model of the area, which may sug-

32

gest the possible existence of trapped fluids in the crust or its thermal nature.

33

Surprisingly the underlying cause of high attenuation in the crust of eastern Hi-

34

malaya and southern Tibet makes this region distinct from its adjacent western

35

Himalayan segment. The results are comparable with the other regions reported

36

globally.

37

Key words: Seismic attenuation; Q; coda normalization method; MLTWA;

38

eastern Himalaya and southern Tibet.

3

39

1. Introduction

40

The knowledge of regional values of the attenuation factor Q and its spatial

41

variation attracts considerable interest in relation to tectonics and seismicity, be-

42

ing a crucial subject in seismic risk analysis and engineering seismology (Singh

43

and Herrmann (1983); Jin and Aki (1988)). The attenuation takes place ei-

44

ther from scattering effects due to heterogeneities present in the medium or

45

from intrinsic phenomena caused due to the anelastic behavior in the medium.

46

Coda wave generates from a local event due to the randomly distributed het-

47

erogeneities present in the medium and its rate of amplitude decay is quantified

48

as Qc . By assuming multiple isotropic scattering, theoretical developments sug-

49

gest that intrinsic attenuation should be the predominant mechanism for coda

50

wave attenuation while observational data have shown that the relative contri-

51

−1 bution of both intrinsic (Q−1 i ) and scattering (Qsc ) attenuation vary in different

52

regions in the case of nonisotropic scattering (Hatzidimitriou (1994); Hoshiba

53

(1995); Sato and Fehler (1995); DelPezzo et al. (2006); Mukhopadhyay et al.

54

(2006); Badi et al. (2009)). However the important aspect is to separate out

55

both the intrinsic and scattering attenuation effect for direct application to the

56

seismic hazard study of a region (DelPezzo and Bianco (2010)). It is also im-

57

portant for the identification of subsurface material, tectonic illustrations and

58

the estimation of the ground motion (Hoshiba (1993); DelPezzo et al. (1995);

59

Bianco et al. (1999); Bianco et al. (2002); Mukhopadhyay et al. (2014)). Knowl-

60

edge of the Q−1 and Q−1 sc provides the clues to the amount of fractures and i

61

liquid present in a certain area. The Multiple Lapse Time Window Analysis

62

(MLTWA) method (Hoshiba et al. (1991); Fehler et al. (1992); Mayeda et al.

63

(1992)) that assumes multiple scattering, has become a successful tool to sep-

64

arate out the intrinsic and scattering attenuation from seismic data by using

65

the energy in several consecutive windows starting from the S-wave as a func-

66

tion of the hypocentral distance. The method is based on radiative transfer

67

theory which assumes isotropic scatters uniformly distributed throughout the

68

medium. The MLTWA technique has been widely used by several researchers

4

69

in the world (e.g. Mayeda et al. (1992); Hoshiba (1993); Pujades et al. (1997);

70

Bianco et al. (2002); Tripathi and Ugalde (2004); Bianco et al. (2005); Ugalde

71

et al. (2007); Mukhopadhyay et al. (2010); DelPezzo et al. (2011); Mukhopadhyay

72

et al. (2014)).

73

Our area of interest belongs to the zone of Nepal Himalaya and southern

74

Tibet which is a seismically very active and tectonically very complex area

75

(Figure 1). About 60 m.y.a since the time (LeFort (1996); Yin and Harrison

76

(2000)) of its collision with Asia half of the present day convergence of India

77

(36-40 mm/yr) is absorbed by the Himalaya that results in contraction of litho-

78

sphere, high-grade metamorphosis, thickening of the crust, and material escape

79

toward regions of low resistance. The Himalaya has formed due to the crustal

80

thickening and mostly consists of buried and exhumed Indian crust (Molnar

81

(1984)). In this collision zone numerous tectono-stratigraphic units such as the

82

Main Central Thrust (MCT), Main Boundary Thrust (MBT) and Main Frontal

83

Thrust (MFT) are present (Hodges (2000)) and serve as northern and southern

84

boundaries of the different Himalayas e.g. In southern Nepal, MFT limits the

85

northern boundary of the Indo-Gangetic plains; Sub-Himalayan Zone is located

86

between MFT and MBT; Lesser Himalaya lies between north of the MBT and

87

its northern boundary is limited by the MCT; north of MCT is the Greater Hi-

88

malayan zone; South Tibetan Fault Zone separates the Greater Himalaya from

89

the Tethyan Himalaya at north; Indus-Tsangpo Suture Zone (ITSZ) represents

90

the northern limit of the Tethyan Himalaya.

91

In this work we initially aim to estimate the frequency dependent quality

92

factor of S (Qs ) and coda (Qc ) waves and to investigate the relative contribution

93

−1 of intrinsic (Q−1 i ) and scattering attenuation (Qsc ) to the total attenuation

94

(Q−1 t ) as a function of frequency, which will help us to provide an attenuation

95

model for the area. Seismic attenuation study has revealed unusually high

96

seismic wave attenuation in the crust and upper mantle of the Tibetan Plateau

97

(Singh et al. (2012)), mostly along its northern part that may indicate high

98

temperatures and partial melting (Zhao et al. (2013)). It is also used to explain

99

the thickening of the Plateau and formation of the Himalaya (Nelson et al. 5

100

(1996); Royden (1996)). Inefficient transmission of shear waves is also evident

101

while crossing the Himalaya and Tibetan Plateau (McNamara et al. (1996);

102

Xie (2002)). Surface wave analysis indicates low velocities in the crust and

103

upper mantle beneath the north-central part of the plateau and higher mantle

104

velocities below the southern part of the plateau (Brandon and Romanowicz

105

(1986); Rodgers and Schwartz (1998)). All these studies point out the existence

106

of partial melting. Recent frequency independent body wave attenuation study

107

has reported a broad zone of low Q values in the southern Tibetan crust (Sheehan

108

et al. (2014)). They have found high Q in eastern Himalaya and low Q in the

109

crust beneath southern Tibet. They have interpreted the low Q to be thermal in

110

origin. However the present study will help us to verify the source of attenuation

111

in the medium below eastern Nepal and the southern Tibetan plateau. In order

112

to understand the physical state of the medium, emphasis is given to comparing

113

our results with the other published results of other adjacent parts of Himalayas

114

as well as reported worldwide.

115

2. Data

116

The Himalayan Nepal Tibet Seismic Experiment (HIMNT) was in opera-

117

tion between October 2001 and March 2003 in Eastern Nepal and Southern

118

Tibet (Monsalve et al. (2006)). The network comprises of 27 three component

119

broadband seismic stations whose sampling rates are 40-50 samples per sec-

120

ond. We have selected 123 crustal events based on high S/N ratio (> 2) out

121

of 1100 well located events (Monsalve et al. (2006)) to ensure reliable results.

122

The crustal thickness in this region varies from south to north and earthquakes

123

are selected based on the crustal thickness as reported by Hetenyi et al. (2006).

124

The magnitude (ML ) of the events are > 2.0. Figure 1 represents distribution of

125

earthquakes and stations considered for the present analysis. We benefit from

126

the use of the high quality data to investigate the attenuation characteristics

127

of the medium below eastern Himalaya and southern Tibet in a more detailed

128

manner. A typical example of a seismogram is shown in Figure 2.

6

129

3. Methodology

130

3.1. Single Isotropic scattering model

131

We have used the single scattering model of Sato (1977), which is the exten-

132

sion and modification of the method given by Aki and Chouet (1975) in case of

133

non-collocated source and receiver. We have used the following equation where

134

amplitude of coda wave has been estimated as a root mean square (rms) of

135

amplitude at a frequency f

Ac (r, f, t) = C(f )K(a, r)exp(−πf t/Q(f ))

(1)

136

Where Ac (r, f, t) is the coda amplitude at time t, C(f ) is the coda source

137

factor at frequency f , r is the distance between source and receiver, a is the

138

geometrical spreading factor defined as a= tts , ts is the S-wave arrival time and p K(a, r) = | K(a)|/r

(2)

139

a+1 where K(a) = (1/a)ln[ a−1 ] , (a > 1). Qc is derived from the slope of the

140

best-fit line while plotting ln[Ac (r, f, t)/K(a, r)] versus t (Figure 3). We have

141

estimated Qc from the band-pass filtered seismograms centered at 2Hz, 4Hz,

142

8Hz and 12Hz. The rms amplitude of the coda has been considered after 1.2

143

times the S-wave arrival time.

144

3.2. Coda Normalization Method

145

The QS for eastern Himalaya and southern Tibet is estimated by applying

146

the coda normalization method (Aki (1980)). The method assumes that coda

147

waves are composed of scattered S-waves generated from random heterogeneities

148

in the Earth (Aki (1969)). The method eliminates site effects, source power and

149

instrument response from direct S-wave by normalizing S-wave amplitude to

150

coda amplitude at a fixed lapse time tc . Details can be found in Aki (1980).

151

152

By using seismograms from earthquakes of different hypocentral distances, we calculate QS as follows

7



 AS (f, r)r πf ln =− r + const(f ) AC (f, tC ) QS (f )VS

(3)

153

Where AS (f, r) is the amplitude spectra of the direct S-wave at the hypocen-

154

tral distance r(km), AC (f, tC ) is coda spectral amplitude, tc is the reference

155

lapse time measured from the source origin time, f is the frequency and VS is

156

the S-wave velocity. The S-wave analysis is based on both the N-S and E-W

157

components of seismograms. The seismograms have been filtered at four dif-

158

ferent central frequencies such as 2Hz, 4Hz, 8Hz and 12Hz. On the filtered

159

seismograms, we measured the maximum peak-to-peak amplitude of the direct

160

S-wave in a 6-s time window starting from the onset of S wave. AC (f, tC ) is

161

calculated for a 6-s time window at tC =60 s where the lapse time tC is taken as

162

twice that of the S-wave travel time (Aki and Chouet (1975); Rautian and Khal-

163

turin (1978). We take VS =3.48 km/s (Monsalve et al. (2008)). Applying the

164

least–square method to plot of the left-hand side of (3) against the hypocentral

165

distance for different earthquakes, we can estimate QS from linear regression

166

lines (Figure 4).

167

3.3. MLTWA

168

MLTWA is probably the most potential and stable technique to calculate

169

Qi and Qsc from the seismogram energy envelopes (Hoshiba et al. (1991) and

170

Mayeda et al. (1992)). The method is based on the radiative transfer theory,

171

which is used to quantify the response of the intensity of a signal that passes

172

through a medium having an indefinite number of scatterers that are isotropic

173

in nature. Under these assumptions the seismic energy density (E(r, t)) as a

174

function of distance and lapse time can be well approximated by the analytical

175

expression (Paasschens (1997)):

( ) " ! 34 # 2 1 r (1 − vr2 t2 ) 8 r2 E(r, t) = S+W H(t− ) exp[−vt(ηi +ηs )] ×P a vtηs 1− 2 2 4πvt 32 v v t [ (3η )] s

(4)

8

176

where S = W exp[−vt(ηi + ηs )]

177

δ(t − ( vr )) 4πr2 v

(5)

and r 2.026 ∼ P a(x) = exp(x) 1 + x

(6)

178

W is the impulsive spherical source of energy, v is the average velocity, ηi

179

and ηs are the intrinsic and scattering attenuation coefficients respectively and

180

H is the Heaviside step function. Further a practical method was proposed

181

simultaneously by Hoshiba et al. (1991), Fehler et al. (1992) and Mayeda et al.

182

(1992) to estimate ηi and ηs which is called MLTWA. It measures the energy

183

envelope of the waveforms to the seismic energy envelope theoretically predicted

184

by the radiative transfer theory. All the seismograms have been filtered at the

185

central frequency 2Hz, 4Hz, 8Hz and 12Hz. The three time integrals of the

186

energy envelopes are estimated by taking the rms of the signal (A2obs (f, r, t)) at

187

different frequencies. The first integral is calculated between the arrival time

188

of S-wave (ts ) and ts + ∆t where ∆t=15 s; second between ts + ∆t and ts

189

+ 2 ∆t and third between ts + 2 ∆t and ts + 3 ∆t. We have taken the

190

average of rms amplitude over three components of seismograms to perform

191

the analysis. The normalization window of 15 s is taken at 60 s from origin

192

time to avoid any overlapping of this window over other three windows. Coda

193

energy is calculated for this window by integration of the squared rms amplitude

194

A2cobs (f, r, t). To remove the source, site and instrument effects the values of

195

A2obs (f, r, t) are normalized by A2cobs (f, r, t) (Aki, 1980). In the present study

196

we have used the average shear wave velocity 3.48 km/s for the whole area

197

as reported by Monsalve et al. (2008). The values of the three normalized

198

energy integrals are multiplied by 4πr2 at a given central frequency to correct

199

for the geometrical spreading effect. The logarithm of this factor is then plotted

200

with respect to the hypocentral distance for a given frequency (Figure 5). The

201

observed energy is finally compared with the synthetic one to obtain the best

202

pair of seismic albedo (B0 ), which is a dimensionless ratio defined as the ratio of

9

203

scattering loss to total attenuation i.e. B0 =ηs /(ηs +ηi ) and the extinction length

204

Le−1 , inverse of distance (in km) over which the S wave energy is decreased by

205

e−1 . Le−1 =1/(ηi +ηs ) (Wu (1985); Fehler et al. (1992)). The fit was carried

206

out using a Genetic algorithm (GA), a nonlinear global optimization technique

207

(Holland (1975)). We have used a series of M atlab tools included in its Global

208

Optimization T oolbox for the present analysis. The misfit function between the

209

observed and the predicted is calculated in L2 norm sense as

M (L−1 e , B0 ) =

3 NX data X

[Ek,th (ri ) − Ek,obs (ri )]2

(7)

k=1 i=1 210

where Ek,th (ri ) and Ek,obs (ri ) are the predicted/theoretical energy inte-

211

grated within the lapse time interval and integrated energy estimated from the

212

observed data respectively. The best-fitting values of B0 and L−1 e are associated

213

with the minimum of misfit function M (L−1 e , B0 ) (Figure 6).

214

215

If Qt is defined as total attenuation, we can write the following equations (Wu (1985); Fehler et al. (1992)) Q−1 t =

(ηs + ηi )V0 L−1 V0 = e ω ω

(8)

ηs V 0 = Q−1 t B0 ω

(9)

ηi V0 = Q−1 t (1 − B0 ) ω

(10)

Q−1 sc =

Q−1 = i 216

where ω is the angular frequency.

217

4. Results and Discussion

218

4.1. Seismic attenuation characteristics

219

An estimate of seismic wave attenuation is made by using S and coda waves

220

for the same dataset. The values of Qs and Qc for the region increase with

221

increasing frequency and indicate the frequency-dependent nature of Q (Figure

222

7). The estimated values of Q suggest that the attenuation is greater in high 10

223

tectonic areas. They increase from about 409 and 230 at 2 Hz to 1646.8 and

224

1661 at 12 Hz for Qs and Qc respectively. To obtain the frequency–dependent

225

relations, the estimated average Q values as a function of frequency are fitted

226

by a power law as of frequency aforementioned in the form Q = Q0 f n (where

227

Q0 is Qs or Qc at 1 Hz and n is the frequency relation parameter) as shown in

228

Figure 7. Both the Q values are found to be frequency-dependent which is a

229

feature of a tectonically active area with complex structures. Values of Q0 also

230

suggest that the area is a tectonically very active area. Coda Q0 obtained in

231

this study is comparable with the value of Q0 estimated (Singh et al. (2016))

232

for the area by using single back-scattering model of Aki and Chouet (1975).

233

Knowledge of relative contribution of scattering and intrinsic attenuation is

234

crucial for subsurface medium properties, tectonic implications and quantifica-

235

tion of the ground motion (Hoshiba (1993); DelPezzo et al. (1995); Mukhopad-

236

hyay et al. (2014)). We have estimated independent measurements of intrinsic

237

−1 (Q−1 i ) and scattering attenuation (Qsc ) using multiple scattering models (Table

238

1) based on radiative transfer theory initially introduced by Wu (1985). Sub-

239

sequent applications of this method were carried out on different regions of the

240

world such as Central California, Long Valley and Hawaii (Mayeda et al. (1992));

241

Kanto-Tokai region Japan (Fehler et al. (1992)); northern Chile (Hoshiba et al.

242

(2001)); Italy (Bianco et al. (2005)); north central Italy (DelPezzo et al. (2011));

243

Garhwal-Kumaun Himalaya (Mukhopadhyay and Sharma (2010)); source zone

244

of the 1999 Chamoli earthquake (Mukhopadhyay et al. (2014)). In the present

245

analysis Q−1 and Q−1 sc are found to be highly frequency dependent in nature i

246

and it is apparent that intrinsic dissipation predominates over scattering atten-

247

uation at all the frequencies (Table 1). The extinction length Le ranges from

248

36 km to about 67 km in the study region (Table 1). At all frequencies seismic

249

albedo B0 < 0.5 indicates the prevalence of intrinsic attenuation. We have also

250

estimated the expected coda Q (Q−1 Cexp ) by using the following expression (e.g.

251

Hoshiba et al. (1991); Mayeda et al. (1992)),

11

−1 Q−1 Cexp = Qsc [1 −

C2 + 2C3 (gvt) + 3C4 (gvt)2 + ... ] + Q−1 i 1 + C2 (gvt) + C3 (gvt)2

(11)

252

where Cn is the coefficient for the nth order scattering (Hoshiba et al., 1991)

253

and v is the S-wave velocity. For the present analysis we have taken upto 10th

254

order. Henceforth we will denote Qc as QCobs to differentiate between Q−1 Cobs

255

and Q−1 Cexp .

256

−1 −1 −1 −1 −1 The variations of Q−1 sc , Qi , Qt , QCexp , QCobs along with the Qs obtained

257

from the coda normalization method as a function of frequency are shown as

258

−1 Figure 8. We have observed that Q−1 is close to Q−1 s Cexp . In general, QCobs

259

−1 is found to be larger than Q−1 but less than Q−1 is t . It is found that Qt i

260

−1 closer to Q−1 compare to Q−1 < Q−1 t . This is sc . At all the frequencies Qs i

261

because the coda normalization method provides only the value of Qpath and

262

removes the near surface attenuation in the process of normalization of direct

263

S wave by the coda wave (Martnez-Arvalo et al. (2003)), which causes the

264

−1 overestimation of Qs . Q−1 lies between Q−1 i Cexp and QCobs but at frequency 4

265

−1 Hz they overlap. The discrepancy between Q−1 Cobs and QCexp is observed at all

266

frequencies except at 4 Hz which can be corroborated by the depth-dependent

267

Q−1 i (Hoshiba et al. (1991);Bianco et al. (2005)). For our study area, Singh et al.

268

(2016) have reported that Qc increases with lapse time, which can be explained

269

in terms of a non-uniform medium with depth-dependent intrinsic attenuation.

270

Moreover, the assumption of uniform distribution of scatterers considered in this

271

−1 study may cause the inconsistency between Q−1 Cobs and QCexp . This has been

272

observed also by others such as Mayeda et al. (1992), Pujades et al. (1997),

273

Bianco et al. (2005) and Padhy and Subhadra (2013).

274

4.2. Attenuation model

275

Attenuation of seismic waves is generally attributed to both intrinsic and

276

scattering mechanisms.

The intrinsic attenuation is associated with small-

277

scale crystal dislocations, frictional heating, and movement of interstitial fluids

278

whereas the scattering attenuation is the loss of energy of a direct wave caused

279

by reflection, refraction and conversion due to existence of heterogeneities (Sato 12

280

and Fehler (1998)). Our study area encompasses Sub-Himalaya in the south

281

up to Tethyan Himalaya in the north. The present study demonstrates that

282

the crust is highly attenuative and intrinsic absorption controls the attenua-

283

tion characteristics of the region that specifies the physical properties of the

284

crust. From the geological point of view, it is evident that our study area has

285

heterogeneities on many scales and an active zone of the ongoing convergence

286

between the Indian plate and southern Tibet. Recently Singh et al. (2015) have

287

observed two low Lg Q0 pockets that correlate well with the low Qp and Qs

288

below lesser Himalaya and southern Tibet (Sheehan et al. (2014)) and coincide

289

with low

290

These variations in the attenuation characteristics of Lg wave have been in-

291

terpreted in terms of both the intrinsic and scattering contributions caused by

292

thermal effects, presence of aqueous fluids as well as heterogeneities present

293

below these seismically active regions (Singh et al. (2015)). Interestingly re-

294

gionalization of Qc for the whole area reported by Singh et al. (2016), exhibits

295

some segments of Lesser and Sub Himalaya are characterized by very low Q0

296

values while the whole Tethyan Himalaya and some parts of Greater Himalaya

297

show low Q0 values. High electrical conductivity is also reported beneath the

298

Lesser Himalaya (Unsworth et al. (2005)). All the observations may enlighten

299

about the possible presence of significant amounts of fluids trapped in the crust,

300

since seismic waves are strongly attenuated by the presence of fluid. Xie et al.

301

(2004) also have suggested that the aqueous fluid trapped in the upper crust

302

and a mid crustal partial melting zone caused due to the underthrusting of the

303

Indian lithosphere probably explains the higher Lg attenuation observed in the

304

southern Tibet area. Sheehan et al. (2014) have observed that P and S waves

305

attenuate 2-5 times more through the crust beneath the High Himalaya and the

306

southern Tibetan Plateau than through the crust beneath the Lesser Himalaya

307

and Ganges Plain. They have interpreted that middle and upper crust of the

308

Tethyan Himalaya represents warm, felsic rock, similar to the other parts of

309

Himalaya (Yin and Harrison (2000)). The intrinsic signature below our study

310

area could also be the cause of the existence of high heat flow (Hu et al. (2000)),

Vp Vs

zones observed from a tomography study (Monsalve et al. (2008)).

13

311

which is also corroborated by the presence of plenty of hot springs (Searle (2013)

312

in southern Tibet. Thus it may be possible that all these factors collectively

313

have made the intrinsic attenuation very effective at this part of our study area.

314

In future it will be interesting to see the variations of Q−1 and Q−1 sc spatially i

315

that may provide more insight into the mechanism of attenuation for the area.

316

4.3. Comparison of Result with Global Observations

317

We have compared values of Q for different areas with our results. Figure 9

318

represents a comparison among the estimated attenuation parameters such as

319

Q−1 and Q−1 sc along with the reported parameters in several regions worldwide. i

320

It is observed that both Q−1 and Q−1 sc are comparable to other regions. At lower i

321

frequencies both the parameters are well comparable with Western India (Ugalde

322

et al. (2007)). Surprisingly intrinsic and scattering attenuation for the eastern

323

Himalaya and southern Tibet area are in general higher and lower than for

324

adjacent Himalayan segments such as Garhwal-Kumaun Himalayas (Mukhopad-

325

hyay and Sharma (2010)) and the source zone of 1999 Chamoli earthquake

326

(Mukhopadhyay et al. (2014)) respectively. This indicates that even though both

327

the areas are tectonically very active and belong to the low Q zones (Mukhopad-

328

hyay and Sharma (2010); Mukhopadhyay et al. (2014)) and form active parts of

329

ongoing continent–continent collision between Indian and Eurasian plate, the

330

governing underlying mechanisms causing high attenuation are significantly dif-

331

ferent from one segment to another. Globally the values of Q−1 and Q−1 sc for i

332

our study area are in general lower than Hawaii (Mayeda et al. (1992)), Cen-

333

tral California (Mayeda et al. (1992)), Southern Italy (Tuve et al. (2006)), Mt.

334

Etna (DelPezzo et al. (1995)) and Central Egypt (Morsy and Abed (2013)) at

335

all frequencies.

336

5. Conclusions

337

The attenuation characteristics in eastern Himalaya and southern Tibet are

338

investigated. These regions are a complex tectonically active zone of the ongo-

339

ing convergence between the Indian plate and southern Tibet. We have tried 14

340

to understand the physical mechanisms governing the attenuation character-

341

istics in the underlying medium. The attenuation factors are estimated by

342

using MLTWA, the coda normalization method and a single isotropic scattering

343

model. The values of Q represent a tectonically active area and are found to

344

be frequency dependent. MLTWA analysis provides evidence that intrinsic at-

345

tenuation dominates over scattering attenuation, which correlates well with its

346

thermal nature and possible existence of trapped fluids in the crust suggested

347

by other studies. Q−1 is observed to be closer to Q−1 compared to Q−1 t sc . We i

348

have found a good agreement among the Q−1 and Q−1 s Cexp . It is observed that

349

−1 at all frequencies Q−1 < Q−1 as The inconsistency between Q−1 t s Cobs and QCexp

350

can be explained by the depth-dependent Q−1 model and the assumption of i

351

uniform distribution of scatterers.

352

A comparative study of attenuation characteristics of eastern Himalaya and

353

southern Tibet with the adjacent segments of Himalaya along with the other di-

354

verse seismotectonic areas of the world is also performed. Interestingly it reveals

355

that the controlling mechanism for high attenuation at this part of Himalayas

356

is different than its adjacent western parts. The results are in general agree-

357

ment with the other regions reported globally. The study fills a crucial gap in

358

knowledge about the governing mechanism effective in the medium below our

359

study region. However a more realistic model with depth dependent variations

360

of Q−1 and Q−1 sc for the area will be of future interest. i

15

361

Acknowledgements: We acknowledge IRIS DMC (http://ds.iris.edu/ds/nodes/dmc)

362

and Project Team of HIMNT experiment for making seismic data available. We

363

thank Editor, Prof Vernon Cormier and both reveiwers for their constructive

364

comments.

365

References

366

Aki, K., Analysis of the seismic coda of local earthquakes as scattered waves,

367

Journal of Geophysical Research, 74 , 615–631, 1969.

368

Aki, K., Attenuation of shear-waves in the lithosphere for frequencies from 0.05

369

to 25 Hz, Physics of the Earth and Planetary Interiors, 21 , 50 – 60, 1980.

370

Aki, K., and B. Chouet, Origin of coda waves: Source, attenuation, and scat-

371

tering effects, Journal of Geophysical Research, 80 , 3322–3342, 1975.

372

Badi, G., E. Del Pezzo, J. M. Ibanez, F. Bianco, N. Sabbione, and M. Araujo,

373

Depth dependent seismic scattering attenuation in the Nuevo Cuyo region

374

southern central Andes, Geophysical Research Letters, 36 , 2009.

375

Bianco, F., M. Castellano, E. Del Pezzo, and J. M. Ibanez, Attenuation of

376

short-period seismic waves at Mt Vesuvius, Italy, Geophysical Journal Inter-

377

national , 138 , 67–76, 1999.

378

Bianco, F., E. Del Pezzo, M. Castellano, J. Ibanez, and F. di Luccio, Separation

379

of intrinsic and scattering seismic attenuation in the Southern Apennine zone,

380

Italy, Geophysical Journal International , 150 , 10–22, 2002.

381

Bianco, F., E. D. Pezzo, L. Malagnini, F. D. Luccio, and A. Akinci, Separation

382

of depth-dependent intrinsic and scattering seismic attenuation in the north-

383

eastern sector of the Italian Peninsula, Geophysical Journal International ,

384

161 , 130–142, 2005.

385

Brandon, C., and B. Romanowicz, A no-lid zone in the central Chang-Thang

386

platform of Tibet: Evidence from pure path phase velocity measurements of

16

387

long period Rayleigh waves, Journal of Geophysical Research: Solid Earth,

388

91 , 6547–6564, 1986.

389

DelPezzo, E., and F. Bianco, Mathltwa: Multiple lapse time window analysis

390

using Wolfram Mathematica 7, Computers and Geosciences, 36 , 1388 – 1392,

391

2010.

392

DelPezzo, E., J. Ibanez, J. Morales, A. Akinci, and R. Maresca, Measurements

393

of intrinsic and scattering seismic attenuation in the crust, Bulletin of the

394

Seismological Society of America, 85 , 1373–1380, 1995.

395

DelPezzo, E., F. Bianco, L. D. Siena, and A. Zollo, Small scale shallow atten-

396

uation structure at Mt. Vesuvius, Italy, Physics of the Earth and Planetary

397

Interiors, 157 , 257 – 268, 2006.

398

DelPezzo, E., F. Bianco, S. Marzorati, P. Augliera, E. D’Alema, and M. Massa,

399

Depth-dependent intrinsic and scattering seismic attenuation in north central

400

Italy, Geophysical Journal International , 186 , 373–381, 2011.

401

Fehler, M., M. Hoshiba, H. Sato, and K. Obara, Separation of scattering and

402

intrinsic attenuation for the Kanto-Tokai region, Japan, using measurements

403

of S-wave energy versus hypocentral distance, Geophysical Journal Interna-

404

tional , 108 , 787–800, 1992.

405

406

Hatzidimitriou, P., Scattering and anelastic attenuation of seismic energy in northern Greece, Pure and Applied Geophysics, 143 , 587–601, 1994.

407

Hetenyi, G., R. Cattin, J. Vergne, and J. L. Nblek, The effective elastic thickness

408

of the india plate from receiver function imaging, gravity anomalies and ther-

409

momechanical modelling, Geophysical Journal International , 167 , 1106–1118,

410

2006.

411

412

Hodges, K. V., Tectonics of the Himalaya and southern Tibet from two perspectives, Geological Society of America Bulletin, 112 , 324–350, 2000.

17

413

414

Holland, J. H., Adaptation in Natural and Artificial Systems, University of Michigan Press, Ann Arbor, MI, USA, 1975.

415

Hoshiba, M., Separation of scattering attenuation and intrinsic absorption in

416

Japan using the multiple lapse time window analysis of full seismogram enve-

417

lope, Journal of Geophysical Research: Solid Earth, 98 , 15,809–15,824, 1993.

418

Hoshiba, M., Estimation of nonisotropic scattering in western Japan using coda

419

wave envelopes: Application of a multiple nonisotropic scattering model,

420

Journal of Geophysical Research: Solid Earth, 100 , 645–657, 1995.

421

Hoshiba, M., H. Sato, and M. Fehler, Numerical Basis of the Separation of

422

Scattering and Intrinsic Absorption from Full Seismogram Envelope, Papers

423

in Meteorology and Geophysics, 42 , 65–91, 1991.

424

Hoshiba, M., A. Rietbrock, F. Scherbaum, H. Nakahara, and C. Haberland,

425

Scattering attenuation and intrinsic absorption using uniform and depth de-

426

pendent model Application to full seismogram envelope recorded in Northern

427

Chile, Journal of Seismology, 5 , 157–179, 2001.

428

429

Hu, S., L. He, and J. Wang, Heat flow in the continental area of China: a new data set, Earth and Planetary Science Letters, 179 , 407 – 419, 2000.

430

Jin, A., and K. Aki, Spatial and temporal correlation between coda Q and

431

seismicity in China, Bulletin of the Seismological Society of America, 78 ,

432

741–769, 1988.

433

434

LeFort, P., Evolution of the Himalaya, in The Tectonic Evolution of Asia, Cambridge U Press, Cambridge, p. 95109, 1996.

435

Martnez-Arvalo, C., F. Bianco, J. M. Ibez, and E. D. Pezzo, Shallow seismic

436

attenuation and shear-wave splitting in the short period range of deception

437

island volcano (antarctica), Journal of Volcanology and Geothermal Research,

438

128 , 89 – 113, 2003, putting Volcano Seismology in a Physical Context. In

439

memory of Bruno Martinelli.

18

440

Mayeda, K., S. Koyanagi, M. Hoshiba, K. Aki, and Y. Zeng, A comparative

441

study of scattering, intrinsic, and coda Q for Hawaii, Long Valley, and central

442

California between 1.5 and 15.0 Hz, Journal of Geophysical Research: Solid

443

Earth, 97 , 6643–6659, 1992.

444

McNamara, D. E., T. J. Owens, and W. R. Walter, Propagation characteristics

445

of Lg across the Tibetan Plateau, Bulletin of the Seismological Society of

446

America, 86 , 457–469, 1996.

447

Meirova, T., and V. Pinsky, Seismic wave attenuation in Israel region estimated

448

from the multiple lapse time window analysis and S-wave coda decay rate,

449

GJI , 2014.

450

Molnar, P., Structure and Tectonics of the Himalaya: Constraints and Implica-

451

tions of Geophysical Data, Annual Review of Earth and Planetary Sciences,

452

12 , 489–516, 1984.

453

Monsalve, G., A. Sheehan, V. Schulte-Pelkum, S. Rajaure, M. R. Pandey, and

454

F. Wu, Seismicity and one-dimensional velocity structure of the Himalayan

455

collision zone: Earthquakes in the crust and upper mantle, Journal of Geo-

456

physical Research: Solid Earth, 111 , 2006.

457

Monsalve, G., A. Sheehan, C. Rowe, and S. Rajaure, Seismic structure of the

458

crust and the upper mantle beneath the Himalayas: Evidence for eclogi-

459

tization of lower crustal rocks in the Indian Plate, Journal of Geophysical

460

Research: Solid Earth, 113 , 2008.

461

462

Morsy, M. A., and A. M. Abed, Attenuation of seismic waves in Central Egypt, {NRIAG} Journal of Astronomy and Geophysics, 2 , 8 – 17, 2013.

463

Mukhopadhyay, S., and J. Sharma, Attenuation characteristics of Garwhal–

464

Kumaun Himalayas from analysis of coda of local earthquakes, Journal of

465

Seismology, 14 , 693–713, 2010.

19

466

Mukhopadhyay, S., C. Tyagi, and S. S. Rai, The attenuation mechanism of

467

seismic waves in northwestern Himalayas, Geophysical Journal International ,

468

167 , 354–360, 2006.

469

Mukhopadhyay, S., J. Sharma, E. Del-Pezzo, and N. Kumar, Study of atten-

470

uation mechanism for Garwhal-Kumaun Himalayas from analysis of coda of

471

local earthquakes, Physics of the Earth and Planetary Interiors, 180 , 7 – 15,

472

2010.

473

Mukhopadhyay, S., A. Kumar, A. Garg, E. Del-Pezzo, and J. Kayal, The atten-

474

uation mechanism of S-waves in the source zone of the 1999 Chamoli earth-

475

quake, Journal of Asian Earth Sciences, 79, Part A, 446 – 454, 2014.

476

477

478

479

Nelson, K. D., et al., Partially Molten Middle Crust Beneath Southern Tibet: Synthesis of Project INDEPTH Results, Science, 274 , 1684–1688, 1996. Paasschens, J. C. J., Solution of the time-dependent Boltzmann equation, Phys. Rev. E , 56 , 1135–1141, 1997.

480

Padhy, S., and N. Subhadra, Separation of intrinsic and scattering seismic wave

481

attenuation in northeast india, Geophysical Journal International , 195 , 1892–

482

1903, 2013.

483

Pujades, L. G., A. Ugalde, J. A. Canas, M. Navarro, F. J. Badal, and

484

V. Corchete, Intrinsic and scattering attenuation from observed seismic co-

485

das in the Almeria Basin southeastern Iberian Peninsula, Geophysical Journal

486

International , 129 , 281–291, 1997.

487

Rautian, T. G., and V. I. Khalturin, The use of the coda for determination

488

of the earthquake source spectrum, Bulletin of the Seismological Society of

489

America, 68 , 923–948, 1978.

490

Rodgers, A. J., and S. Y. Schwartz, Lithospheric structure of the Qiangtang

491

Terrane, northern Tibetan Plateau, from complete regional waveform model-

492

ing: Evidence for partial melt, Journal of Geophysical Research: Solid Earth,

493

103 , 7137–7152, 1998. 20

494

Royden, L., Coupling and decoupling of crust and mantle in convergent oro-

495

gens: Implications for strain partitioning in the crust, Journal of Geophysical

496

Research: Solid Earth, 101 , 17,679–17,705, 1996.

497

Sahin, S., M. Erduran, O. Alptekin, and O. akir, Intrinsic and scattering seis-

498

mic attenuation in southwestern anatolia, Pure and Applied Geophysics, 164 ,

499

2255–2270, 2007.

500

Sato, H., Energy Propagation Including Scattering Effects Single Isotropic

501

Scattttering Approximation, Journal of Physics of the Earth, 25 , 27–41, 1977.

502

Sato, H., and M. Fehler, S-wave attenuation in the crust in northern Greece,

503

504

505

506

507

BSSA, 85 , 1381–1387, 1995. Sato, H., and M. Fehler, Seismic Wave Propagation and Scattering in the Heterogeneous Earth, AIP Press/Springer, Verlag, New York 2000, 1998. Searle, M., Colliding Continents: A geological exploration of the Himalaya, Karakoram, and Tibet, Oxford University Press, 2013.

508

Sheehan, A. F., T. L. de la Torre, G. Monsalve, G. A. Abers, and B. R. Hacker,

509

Physical state of Himalayan crust and uppermost mantle: Constraints from

510

seismic attenuation and velocity tomography, Journal of Geophysical Re-

511

search: Solid Earth, 119 , 567–580, 2014.

512

Singh, C., A. Singh, and R. K. Chadha, Fractal and b-value Mapping in East-

513

ern Himalaya and Southern Tibet, Bulletin of the Seismological Society of

514

America, 99 , 3529–3533, 2009.

515

Singh, C., M. Shekar, A. Singh, and R. K. Chadha, Seismic Attenuation Char-

516

acteristics along the HiCLIMB Profile in Tibet from Lg Q Inversion, Bulletin

517

of the Seismological Society of America, 102 , 783–789, 2012.

518

Singh, C., P. Mondal, S. Singh, D. D. Mohanty, N. Jaiswal, and M. R. Ku-

519

mar, Lg attenuation tomographic models of Himalaya and southern Tibet,

520

Tectonophysics, 664 , 176–181, 2015. 21

521

Singh, S., and R. B. Herrmann, Regionalization of crustal coda Q in the con-

522

tinental United States, Journal of Geophysical Research: Solid Earth, 88 ,

523

527–538, 1983.

524

Singh, S., C. Singh, R. Biswas, and A. Singh, Frequency and lapse time depen-

525

dent seismic attenuation in eastern Himalaya and southern Tibet, Natural

526

Hazard review submitted , 2016.

527

Tripathi, J. N., and A. Ugalde, Regional estimation of Q from seismic coda

528

observations by the Gauribidanur seismic array (southern India), Physics of

529

the Earth and Planetary Interiors, 145 , 115 – 126, 2004.

530

Tuve, T., F. Bianco, J. Ibez, D. Patan, E. D. Pezzo, and A. Bottari, Attenuation

531

study in the Straits of Messina area (southern italy), Tectonophysics, 421 ,

532

173 – 185, 2006.

533

Ugalde, A., J. Tripathi, M. Hoshiba, and B. Rastogi, Intrinsic and scattering

534

attenuation in western India from aftershocks of the 26 january, 2001 Kachchh

535

earthquake, Tectonophysics, 429 , 111 – 123, 2007.

536

Unsworth, M. J., A. G. Jones, W. Wei, G. Marquis, S. G. Gokarn, and J. E.

537

Spratt, Crustal rheology of the Himalaya and Southern Tibet inferred from

538

magnetotelluric data, Nature, 438 , 2005.

539

Wu, R.-S., Multiple scattering and energy transfer of seismic waves separa-

540

tion of scattering effect from intrinsic attenuation i. Theoretical modelling,

541

Geophysical Journal of the Royal Astronomical Society, 82 , 57–80, 1985.

542

543

Xie, J., Lg Q in the Eastern Tibetan Plateau, Bulletin of the Seismological Society of America, 92 , 871–876, 2002.

544

Xie, J., R. Gok, J. Ni, and Y. Aoki, Lateral variations of crustal seismic attenu-

545

ation along the INDEPTH profiles in Tibet from Lg Q inversion, J. Geophys.

546

Res., 109 , 2004, b10308.

22

547

Yin, A., and T. M. Harrison, Geologic Evolution of the Himalayan-Tibetan

548

Orogen, Annual Review of Earth and Planetary Sciences, 28 , 211–280, 2000.

549

Zhao, L.-F., X.-B. Xie, J.-K. He, X. Tian, and Z.-X. Yao, Crustal flow pattern

550

beneath the Tibetan Plateau constrained by regional Lg-wave Q tomography,

551

Earth and Planetary Science Letters, 383 , 113 – 122, 2013.

23

552

Figure captions:

553

Figure 1: Tectonic map of our study area that includes part of Tibet and the

554

Himalayan region modified after Monsalve et al. (2008) along with the locations

555

of the earthquakes (circles) and stations (triangles) used in the present study.

556

The location of the study area with respect to India is also shown (upper panel).

557

Figure 2: Example of a seismogram (three components) of an earthquake

558

occurred on 24th March 2002 at 22 hr, 11 min and 21.25 s. The t ime of P , S

559

and Coda waves are marked. All the windows including Coda are measured for

560

a time window of 15 s as indicated by the boxes. Coda window is started at

561

tc = 60 s from the earthquake origin time.

562

563

Figure 3: Qc estimation with the least-square fits of chosen coda–window at different central frequencies. Standard deviations are shown by the error bars.

564

Figure 4: Coda normalized peak amplitude decay of S wave with hypocentral

565

distance at four central frequencies. The regression line from the least-squares

566

estimate is expressed by solid line.

567

Figure 5: Plots represent the normalized energy after correcting for the

568

geometrical spreading factor for the whole area at four frequencies. Different

569

colors stand for the different energy measurements such as red for the 0-15s,

570

blue for 15-30s and green for 30-45s respectively. The predictions for the best

571

fit to obtain B0 and L−1 e are represented by solid lines. Est.-Estimated; Predi.-

572

Predicted.

573

Figure 6: Misfit variation curve for 12 Hz.

574

Figure 7: Plots of mean values of Qc and Qs with frequency for the whole

575

region. A power law on frequency in the form Q = Q0 f n has fitted. Standard

576

deviations are shown by the error bars.

577

578

−1 −1 −1 −1 Figure 8: Figure represents the estimated Q−1 i , Qsc , Qt , Qs , QCobs and

Q−1 Cexp values as a function of frequency for the region.

579

Figure 9: Comparison of (a) Q−1 and (b) Q−1 sc with the other parts of i

580

Himalaya as well as reported worldwide. The reported values are as follows:

581

Southwestern Anatolia (Sahin et al. (2007)); Southern Italy (Tuve et al. (2006));

582

Southern Apennine zone, Italy (Bianco et al. (2002)); Central Egypt (Morsy and 24

583

Abed (2013)); Mt, Etna (DelPezzo et al. (1995)); Longvalley and central Cali-

584

fornia and Hawaii (Mayeda et al. (1992)); Israel (Meirova and Pinsky (2014));

585

Western India (Ugalde et al. (2007)); Garwhal-Kumaun Himalaya (Mukhopad-

586

hyay et al. (2010)); source zone of 1999 Chamoli earthquake (Mukhopadhyay

587

et al. (2014))

25

−1 −1 −1 Table 1: Values of Q−1 at different frequencies for the whole region. t , Qsc , Qi , B0 and Le

Freq(Hz)

Q−1 t

Q−1 sc

Q−1 i

B0

L−1 e

2

0.00747±0.00185

0.0024±0.00134

0.0051±0.00051

0.32±0.08

0.027±0.007

4

0.00207±0.00061

0.00032±0.00028

0.00176±0.00033

0.15±0.07

0.015±0.004

8

0.0017±0.00036

0.00053±0.00026

0.00113±0.00011

0.32±0.07

0.024±0.005

12

0.0013±0.00025

0.00054±0.00021

0.00075±0.00003

0.42±0.07

0.028±0.005

26

India

31˚

30˚ Indus Tsangpo Suture Zone 29˚

Tethyan Himalaya MC T

Nepal 28˚

Greater Himalaya

MB

T

Sikkim

Lesser Himalaya

MFT

Sub Hi

27˚

malaya

MFT-MAIN FRONTAL THRUST MBT-MAIN BOUNDARY THRUST MCT-MAIN CENTRAL THRUST

26˚ 84˚

85˚

86˚

87˚

88˚

89˚

90˚

Figure 1: Tectonic map of our study area that includes part of Tibet and the Himalayan region modified after Monsalve et al. (2008) along with the locations of the earthquakes (circles) and stations (triangles) used in the present study. The location of the study area with respect to India is also shown (upper panel).

27

2000

Counts

E−W component

O.T

1000 0

0−15sec

−1000 P −2000

0

15−30sec

30−45sec

Coda window

S

20

40

60

80

100

120

140

Time (sec)

2000

Counts

N−S component

O.T

1000 0

0−15sec

−1000 P −2000

0

15−30sec

Coda window

30−45sec

S

20

40

60

80

100

120

140

Time (sec)

2000

Counts

Vertical component

O.T

1000 0

0−15sec

−1000

30−45sec

15−30sec

Coda window

P −2000

S 0

20

40

60

80

100

120

140

Time (sec)

Figure 2: Example of a seismogram (three components) of an earthquake occurred on 24th March 2002 at 22 hr, 11 min and 21.25 s. The time of P , S and Coda waves are marked. All the windows including Coda are measured for a time window of 15 s as indicated by the boxes. Coda window is started at tc = 60 s from the earthquake origin time.

28

15

15

Frequency = 4Hz Q = 558 ± 67 c

ln [Ac/K(a,r)]

ln [Ac/K(a,r)]

Frequency = 2Hz Qc = 230 ± 21

10

5 30

40

50

60

70

80

10

5 30

90

40

50

60

15

90

Frequency = 12Hz Qc = 1661 ± 164 ln [Ac/K(a,r)]

ln [Ac/K(a,r)]

80

15

Frequency = 8Hz Qc = 1097 ± 103

10

5 30

70

Lapse time (sec)

Lapse time (sec)

40

50

60

70

80

90

10

5 30

40

50

60

70

80

Lapse time (sec)

Lapse time (sec)

Figure 3: Qc estimation with the least-square fits of chosen coda–window at different central frequencies. Standard deviations are shown by the error bars.

29

90

10

8

8

ln (As*r/Ac)

ln (As*r/Ac)

10

6 4 20

40 60 80 100 Hypocentral Distance (km)

120

f=4 Hz Qs =617±200 0

10

10

8

8

ln (As*r/Ac)

ln (As*r/Ac)

4

f=2 Hz Qs =409±120 0

6

6 4

f=8 Hz Qs =1255±330 0

20

40 60 80 100 Hypocentral Distance (km)

20

40 60 80 100 Hypocentral Distance (km)

20

40 60 80 100 Hypocentral Distance (km)

120

6 4

120

f=12 Hz Qs =1647±431 0

120

Figure 4: Coda normalized peak amplitude decay of S wave with hypocentral distance at four central frequencies. The regression line from the least-squares estimate is expressed by solid line.

30

Frequency = 4 Hz 10

8

8

Normalised Energy

Normalised Energy

Frequency = 2 Hz 10

6

4

2

0

0

20

40

60

80

100

6

4

2

0

120

0

20

40

Distance(km)

80

100

120

100

120

Frequency = 12 Hz 10

8

8

Normalised Energy

Normalised Energy

Frequency = 8 Hz 10

6

4

6

4

2

2

0

60

Distance(km)

0 0

20

40

60

80

100

120

0

Esti.1

20

40

60

80

Distance(km)

Distance(km) Esti.2

Esti.3

Predi.1

Predi.2

Predi.3

Figure 5: Plots represent the normalized energy after correcting for the geometrical spreading factor for the whole area at four frequencies. Different colors stand for the different energy measurements such as red for the 0-15s, blue for 15-30s and green for 30-45s respectively. The predictions for the best fit to obtain B0 and L−1 e are represented by solid lines. Est.-Estimated; Predi.-Predicted.

31

1000 900 800

Misfit value

700 600 500 400 300 200 100 0

5

10

15 Generation

20

Figure 6: Misfit variation curve for 12 Hz.

32

25

10000 Qc=(113±31)f(1.09±0.14) Qc fit Qs=(223±45)f(0.81±0.03)

Q

Qs fit

1000

100 1

10 Frequency (Hz)

Figure 7: Plots of mean values of Qc and Qs with frequency for the whole region. A power law on frequency in the form Q = Q0 f n has fitted. Standard deviations are shown by the error bars.

33

100

Qt-1 Qsc-1 Qi-1 QCobs-1 QCexp-1

0.01

Q-1

Qs-1

0.001

0.0001 0

4

8 Frequency (Hz)

12

−1 −1 −1 −1 −1 Figure 8: Figure represents the estimated Q−1 i , Qsc , Qt , Qs , QCobs and QCexp values as

a function of frequency for the region.

34

Southwestern Anatolia Southern Italy Southern Apennine, Italy This study Central Egypt Mt. Etna Longvalley California Central California Hawaii Israel Western India Garwhal-Kumaun Himalaya Source zone of 1999 Chamoli earthquake

0.1

Qi-1

0.01

0.001

0.0001 0

10

20

30

Frequency (Hz) Southwestern Anatolia Southern Italy Southern Apennine, Italy Central Egypt

(b)

0.1

This study Mt. Etna Longvalley, California Central California Hawaii Israel Western India Garwhal-Kumaun Himalaya source zone of 1999 Chamoli earthquake

Qsc-1

0.01

0.001

0.0001

1E-005 0

10

20

30

Frequency (Hz)

Figure 9: Comparison of (a) Q−1 and (b) Q−1 sc with the other parts of Himalaya as well as i reported worldwide. The reported values are as follows: Southwestern Anatolia (Sahin et al. (2007)); Southern Italy (Tuve et al. (2006)); Southern Apennine zone, Italy (Bianco et al. (2002)); Central Egypt (Morsy and Abed (2013)); Mt, Etna (DelPezzo et al. (1995)); Longvalley and central California and Hawaii (Mayeda et al. (1992)); Israel (Meirova and Pinsky (2014)); Western India (Ugalde et al. (2007)); Garwhal-Kumaun Himalaya (Mukhopadhyay et al. (2010)); source zone of 1999 Chamoli earthquake (Mukhopadhyay et al. (2014))

35

An estimate of seismic wave attenuation is made by using S and Coda wave for eastern Himalaya and southern Tibet.

MLTWA method is used further to estimate the relative contribution of intrinsic and scattering attenuation for the region.

At all the frequencies intrinsic absorption is predominant compared to scattering attenuation.

Our results may suggest the possible existence of trapped fluids in the crust or its thermal nature.