Earthquake hazard zonation of Sikkim Himalaya using a ... - CiteSeerX

7 downloads 0 Views 3MB Size Report
Nov 1, 2007 - eight thematic layers namely Geology (GE), Soil Site Class (SO), Slope (SL), Landslide ... International Seismological Center (ISC) and Global Centroid ... about 12% of the country, (ii) Zone IV: PGA 0.25 g and MMI VIII, covering 18% of the ...... Korte GB (1997) The GIS Book: understanding the value and ...
Nat Hazards (2008) 45:333–377 DOI 10.1007/s11069-007-9173-7 ORIGINAL PAPER

Earthquake hazard zonation of Sikkim Himalaya using a GIS platform Indrajit Pal Æ Sankar Kumar Nath Æ Khemraj Shukla Æ Dilip Kumar Pal Æ Abhishek Raj Æ K. K. S. Thingbaijam Æ B. K. Bansal

Received: 24 October 2006 / Accepted: 29 July 2007 / Published online: 1 November 2007 Ó Springer Science+Business Media B.V. 2007

Abstract An earthquake hazard zonation map of Sikkim Himalaya is prepared using eight thematic layers namely Geology (GE), Soil Site Class (SO), Slope (SL), Landslide (LS), Rock Outcrop (RO), Frequency–Wavenumber (F–K) simulated Peak Ground Acceleration (PGA), Predominant Frequency (PF), and Site Response (SR) at predominant frequencies using Geographic Information System (GIS). This necessitates a large scale seismicity analysis for seismic source zone classification and estimation of maximum earthquake magnitude or maximum credible earthquake to be used as a scenario earthquake for a deterministic or quasi-probabilistic seismic scenario generation. The International Seismological Center (ISC) and Global Centroid Moment Tensor (GCMT) catalogues have been used in the present analysis. Combining b-value, fractal correlation dimension (Dc) of the epicenters and the underlying tectonic framework, four seismic source zones are classified in the northeast Indian region. Maximum Earthquake of MW 8.3 is estimated for the Eastern Himalayan Zone (EHZ) and is used to generate the seismic scenario of the region. The Geohazard map is obtained through the integration of the geological and geomorphological themes namely GE, SO, SL, LS, and RO following a pair-wise comparison in an Analytical Hierarchy Process (AHP). Detail analysis of SR at all the recording stations by receiver function technique is performed using 80 significant events recorded by the Sikkim Strong Motion Array (SSMA). The ground motion synthesis is performed using F–K integration and the corresponding PGA has been estimated using random vibration theory (RVT). Testing for earthquakes of magnitude greater than MW 5, a few cases presented here, establishes the efficacy and robustness of the F–K simulation I. Pal  S. K. Nath (&)  K. Shukla  A. Raj  K. K. S. Thingbaijam Department of Geology and Geophysics, Indian Institute of Technology, Kharagpur, Kharagpur, West Bengal 721 302, India e-mail: [email protected] D. K. Pal Department of Geography and Environment Management, Vidyasagar University, Medinipur, West Bengal 721 102, India B. K. Bansal Seismology Division, Department of Science & Technology, Government of India, New Delhi 110 016, India

123

334

Nat Hazards (2008) 45:333–377

algorithm. The geohazard coverage is overlaid and sequentially integrated with PGA, PF, and SR vector layers, in order to evolve the ultimate earthquake hazard microzonation coverage of the territory. Earthquake Hazard Index (EHI) quantitatively classifies the terrain into six hazard levels, while five classes could be identified following the Bureau of Indian Standards (BIS) PGA nomenclature for the seismic zonation of India. EHI is found to vary between 0.15 to 0.83 quantitatively classifying the terrain into six hazard levels as ‘‘Low’’ corresponding to BIS Zone II, ‘‘Moderate’’ corresponding to BIS Zone III, ‘‘Moderately High’’ belonging to BIS Zone IV, ‘‘High’’ corresponding to BIS Zone V(A), ‘‘Very High’’ and ‘‘Severe’’ with new BIS zones to Zone V(B) and V(C) respectively. Keywords

AHP  Maximum earthquake  Earthquake hazard  GIS  Seismicity

1 Introduction The entire Himalayan region is a 2,500 km-long belt from Kashmir in the west to Arunachal Pradesh in the east. It can be divided into several seismotectonic blocks, including Darjeeling-Sikkim, where numerous moderate magnitude earthquakes (M C 5.0) had been recorded. By 2002, the Bureau of Indian Standards (BIS) mapped four seismic zones in India, namely (i) Zone-V: Peak Ground Acceleration (PGA) of 0.4 g (1 g = 980 gal) with 10% probability of exceedance in 50 years and MMI of IX, covering about 12% of the country, (ii) Zone IV: PGA 0.25 g and MMI VIII, covering 18% of the country, (iii) Zone III: PGA 0.2 g and MMI VII, spreading out at 26% of the country, and (iv) Zone II: PGA 0.1 g and MMI VI (BIS 2002). However, the design values for these zones cannot certainty reflect with what ground acceleration will act on the structures situated in those zones. It is the level of ground acceleration, coupled with site-specific effects, which actually buffets buildings due to the impact of an earthquake. Collapsing buildings during an earthquake have direct consequences on the casualties and economic losses. Earthquake hazard mapping essentially requires (a) Bedrock topography, (b) Subsoil profile, (c) Soil site classification, (d) PGA and Peak Ground Velocity (PGV) mapping, (e) Liquefaction potential mapping, (f) Geomorphological characterization, and (g) Probabilistic/Deterministic seismic hazard scenario using Geographic Information System (GIS), a primary working tool for the purpose. A typical earthquake hazard zonation framework is depicted in the roadmap given in Fig. 1. The GIS framework allows one to account for added levels of details and complexity. For ease of study the data attributes in the present analysis are subdivided into two major groups—Geomorphological and Seismological. It is important that the relevant data layers be consistent in their level of detailing, in order to successfully combine and cross analyze those at a later stage of integration. The GIS technology enables the complex spatial analysis usually associated with the earthquake hazard microzonation through data dissemination and its management through linking of databases on one-to-one relationship to end-user defined by common identification index or code (Marble and Pequet 1983; Korte 1997; Hohl 1998). In the present work, the geomorphological and seismological themes are initially reclassified as a first level geohazard and subsequently to final hazard zonation wherein the geohazard and seismological themes for the scenario of a magnitude MW

123

Nat Hazards (2008) 45:333–377

335

Fig. 1 Microzonation framework for earthquake hazard mapping

8.3 are combined through a fuzzy logic methodology aided by Analytic Hierarchy Process (AHP) to achieve operational simplicity in the evaluation of the hazard model. The scenario earthquake is assumed to be nucleating from the hypocenter of the seismic event of ML 5.6 largest amongst 80 significant events (Event 50 in Table 1) with the focal mechanism indicating 310°N strike and 35°NNE dip occurring just below the Main Boundary Thrust (MBT), as recorded by IIT Kharagpur Sikkim Strong Motion Array (SSMA) at 9 stations with hypocentral distance varying between 27 and 59 km at an elevation variation of 400–3,800 m. McHarg (1969) introduced multicriteria evaluation technique for systematic landuse planning. The idea of multi-criteria decision making was based on a simple matrix system for determining the degree of compatibility. AHP uses a hierarchal structure via pair-wise comparison based on forming judgments between two particular elements rather than attempting to prioritize an entire list of elements (Saaty 1980). A matrix of pair-wise comparison between the factors is thus built on a scale in a process of allocating weights in the participatory mode in which a group of decision makers may be encouraged to reach a consensus opinion about the relative importance of the factors.

123

Event (YYMMDDHHMM)

9903070614

9906140903

9906180936

9906190757

9907020601

9907100911

9907140601

0001010011

0004071026

0004181325

0005230352

0005310453

0005310621

0006020851

0006030555

0006031628

0006070910

0006080832

0006101207

0006130709

0006160612

0006181645

0006290426

0006300927

Sl. no.

1.

2.

123

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

27.330

27.400

27.390

27.680

27.500

27.180

27.240

27.000

27.270

27.210

27.200

27.550

27.270

27.230

27.520

27.380

27.300

27.250

27.360

27.250

27.350

27.350

27.470

27.250

Latitude (°N)

88.430

88.830

88.380

88.290

88.360

88.310

88.340

88.000

88.290

88.440

88.480

88.400

88.570

88.110

88.630

88.520

88.580

88.480

88.360

88.480

88.360

88.580

88.430

88.390

Longitude (°E)

3.6

4.6

4.3

5.2

5.3

5.1

4.2

4.9

4.2

3.0

5.1

5.0

3.8

3.0

4.1

3.5

4.0

3.7

4.0

4.2

4.2

3.9

3.9

4.6

Magnitude (ML)

6.1

10.0

14.9

10.0

10.0

23.4

10.0

18.9

10.0

10.0

22.3

7.4

7.5

10.0

3.0

10.0

10.0

17.4

14.7

17.9

14.8

10.0

23.1

23.6

Depth (km)

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

Singtam

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

Gezing

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

Mangan

*

*

*

*

*

*

*

Gangtok

*

Lachen

*

*

*

Chungthang

*

*

*

*

*

*

*

*

*

*

*

Jorethang

Table 1 IIT Kharagpur Sikkim Strong Motion Array recording history for 80 earthquakes with signal-to-background noise ratio C3 (after Nath et al. 2005) Aritar

Melli

336 Nat Hazards (2008) 45:333–377

0008070321

0008071359

0008201726

0008230700

0008280816

0009020715

0009041248

0009061907

0009080215

0009210751

0009250446

0010030502

0010181422

0011172135

0011230650

0012010355

0102090959

0101040236

0101051808

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

0111151432

0007270320

27.

48.

0007160757

26.

0111160424

0007041026

25.

47.

Event (YYMMDDHHMM)

Sl. no.

Table 1 continued

27.150

27.360

27.230

27.220

27.300

27.220

27.250

27.240

27.350

27.230

27.390

27.380

27.430

27.500

27.280

27.370

27.150

27.260

27.360

27.280

27.320

27.240

27.200

27.170

Latitude (°N)

88.300

88.160

88.380

88.360

88.280

88.310

88.300

88.540

88.480

88.480

88.370

88.520

88.440

88.520

88.350

88.270

88.300

88.310

88.280

88.330

88.420

88.250

88.480

88.450

Longitude (°E)

4.1

4.8

3.4

4.3

4.0

4.3

4.5

4.8

4.9

5.3

4.2

5.1

4.3

4.5

3.7

4.1

3.9

3.7

3.0

4.6

4.4

3.8

4.2

5.1

Magnitude (ML)

21.2

19.0

10.0

21.6

5.4

13.0

23.8

10.0

24.5

34.3

10.0

17.9

17.6

20.6

10.0

10.0

16.8

6.8

7.8

10.0

10.0

5.3

10.0

24.5

Depth (km)

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

Singtam

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

Gezing

*

*

*

*

*

*

*

*

*

*

*

*

*

*

Mangan

*

*

*

*

*

*

*

Gangtok

Lachen

*

Chungthang

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

Jorethang

Aritar

Melli

Nat Hazards (2008) 45:333–377 337

123

Event (YYMMDDHHMM)

0111231031

0112022241

0112030100

0203161126

0204081150

0204220936

0204241418

0204250458

0204250116

0204250821

0204251130

0204260304

0204260957

0204261551

0204271203

0204280548

0204290138

0204290628

0204291243

0204300646

0204301349

0205010244

0205011145

0205021028

0206150914

Sl. no.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

Table 1 continued

123

27.757

27.970

27.570

27.350

27.350

27.910

27.200

27.230

27.410

27.180

27.570

27.300

27.350

27.480

27.320

27.240

27.150

27.280

27.090

27.090

27.470

27.350

27.360

27.250

27.370

Latitude (°N)

88.714

88.870

88.540

88.580

88.800

88.540

88.700

88.580

88.390

88.710

88.660

88.640

88.580

88.400

88.300

88.780

88.830

88.630

88.860

88.860

88.340

88.580

88.230

88.460

88.430

Longitude (°E)

4.2

5.3

4.4

4.7

4.5

5.2

5.0

4.2

3.7

5.0

3.9

4.1

4.3

4.4

5.1

5.3

5.1

5.1

3.5

5.0

4.3

4.5

3.4

5.6

4.8

Magnitude (ML)

20.3

10.0

10.0

17.1

6.0

10.0

27.8

10.0

6.0

24.1

10.0

10.0

10.0

10.0

26.4

32.4

10.0

22.9

5.3

19.4

10.0

20.0

3.5

26.3

10.0

Depth (km)

*

*

*

*

*

*

*

*

*

*

*

*

*

*

Singtam

*

*

*

*

Gezing

*

*

*

*

*

*

*

*

*

*

*

*

Mangan

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

Gangtok

*

Lachen

*

*

*

*

*

*

Chungthang

*

*

*

*

*

Jorethang

*

*

*

*

*

*

*

*

*

Aritar

*

*

*

*

*

*

Melli

338 Nat Hazards (2008) 45:333–377

0206181021

0206260505

0206301522

0207081125

0208061916

0208210123

0208221612

74.

75.

76.

77.

78.

79.

80.

27.135

27.265

27.462

27.157

27.429

27.183

27.216

Latitude (°N)

88.388

88.611

88.701

88.478

88.664

88.359

88.774

Longitude (°E)

Event recorded at a station represented by asterisk (*)

Event (YYMMDDHHMM)

Sl. no.

Table 1 continued

4.8

3.6

3.0

3.1

4.0

3.8

3.5

Magnitude (ML)

16.5

10.0

23.1

10.0

16.3

10.0

12.2

Depth (km)

*

*

*

Singtam

*

Gezing

*

*

Mangan

*

*

*

Gangtok

Lachen

*

*

Chungthang

*

*

*

Jorethang

*

*

*

Aritar

*

*

*

*

Melli

Nat Hazards (2008) 45:333–377 339

123

340

Nat Hazards (2008) 45:333–377

2 Seismicity of the region The Sikkim Himalayan territory is characterized by intense micro-seismic activity. The earthquakes of magnitude ML 3.0–5.6 recorded by SSMA during 1999–2002 for a span of 3 years have been presented on the IRS-1C LISS III map of the Sikkim region as shown in Fig. 2a with the recording history given in Table 1. It is evident that the earthquakes have distributed occurrences, but the hypocentral depths are reasonably well constrained in the eastern and southern Sikkim Himalaya. They are generally shallower than 35 km but exhibit a clustering within the depth range of 10–25 km as depicted in the N–S depth section of Fig. 2b. A data catalogue is derived from the International Seismological Center (ISC 2007) covering the entire northeast Indian region bounded within 19°N 85°E and 32°N 101°E for the period 1906–2006. Several magnitude scales are found in the ISC catalogue of which the body wave magnitude mb,ISC, the surface wave magnitude MS,ISC and local magnitude ML,ISC are predominant. Acronym specifying the data source is added as an additional subscript to the usual label of the magnitude scales for the sake of identification. Only three events are found to have the moment magnitude MW,ISC. The correlation between mb,ISC and MS,ISC with the entries of both the magnitudes of the same event across the catalogue yields the following relationship, mb;ISC ¼0:595ð0:0471ÞMS;ISC þ 2:035ð0:19Þ;

for mb;ISC  6:2; MS;ISC  6:8:

ð1Þ

The Global Centroid Moment Tensor (GCMT) database (http://www.globalcmt.org) is consulted to establish correlations between different magnitudes mb,ISC, MS,ISC, and ML,ISC with MW,GCMT reported by GCMT. The correlation of MW,GCMT and MS,ISC yields the following relations, MW;GCMT ¼0:629ð0:06ÞMS;ISC þ 2:209ð0:30Þ;

for 4:1  MS;ISC  6:1:

ð2Þ

MW;GCMT ¼ 0:738ð0:22ÞMS;ISC þ 1:604ð1:48Þ;

for 6:1  MS;ISC  7:4:

ð3Þ

and

The correlation between MW,GCMT and mb,ISC is found to be MW;GCMT ¼0:3513ð0:2183Þm2b;ISC  2:837ð2:397Þmb;ISC þ 10:697ð6:551Þ; for 4:5  mb;ISC  6:6:

ð4Þ

Equation (4) is seen to have significant dispersion and, hence, is not preferred in the present analysis. With the consideration of the applicable range and using Eq. (1), mb entries in the catalogue have been scaled to MS and subsequently to MW. The correlation between MW,GCMT and ML,ISC entries is found to have higher scattering, thereby ruling out a suitable model fit. In an earlier case (Heaton et al. 1986), ML and MW were shown to be roughly equivalent up to magnitude 6.0. Deichmann (2006) also pointed out that theoretically the magnitude scales, ML and MW, are equivalent over the entire range for which ML can be determined. Furthermore, Ristau et al. (2005) implied that in the western Canadian region, the continental crust events (as is the case with the present study) indicated MW = ML for earthquakes with ML C 3.6. From the foregoing concepts and observations, ML is considered almost equivalent to MW in the present analysis. However, it may be noted that the procedure for the determination of ML that depends on the

123

Nat Hazards (2008) 45:333–377

341

Fig. 2 (a) Sikkim Strong Motion Array network on IRS-1C LISS III FCC, and (b) Depth profile along N–S direction

recording seismographs may imply observational differences between the magnitude scales in different cases and the appropriate region specific relationships be derived by analyzing relevant waveform data (e.g., Braunmiller et al. 2005) in a future study. An examination of the catalogue reveals that during the period 1964–2006, the maximum magnitude was 6.6 in mb,ISC, 5.0 in ML,ISC and 6.9 in MS,ISC. Four entries in the catalogue have values of mb,ISC [ 6.2 as listed in Table 2, so that scaling by Eq. (3) is not applicable. For the events of the years between 1988 and 1995, the corresponding MW values are obtained from the GCMT database. The 1970 event is assigned MW 6.8 correlating with that of the event on August 20, 1988. The largest earthquake record in the catalogue is the Great Assam earthquake of 1950, MS 8.6. Kanamori (1977) placed it as MW 8.6. Brune and Engen (1969) assigned a seismic moment of 4.31 9 1021 Nm to this earthquake. However, later workers associated higher values of 21.0 9 1021 Nm (Ben-Menahem et al. 1974) and 9.5 9 1021 Nm (Chen and

123

342

Nat Hazards (2008) 45:333–377

Table 2 The earthquakes with mb [ 6.2 in the ISC catalogue during 1964–2006 with the corresponding MW entry in GCMT database Date

Time

Lat (°N)

Lon (°E)

1970/07/29

10:16:20.40

26.0200

95.3700

1988/08/06

00:36:25.54

25.1297

95.1493

1988/08/20

23:09:10.14

26.7198

86.6261

1995/05/06

01:59:07.04

24.9605

95.2949

Depth (km)

mb,ISC

MW

68.0

6.4

6.8

100.1

6.6

7.2

64.6

6.4

6.8

117.6

6.3

6.4

Molnar 1977). The average magnitude computed with the latter values from the seismic moment magnitude relation (Hanks and Kanamori 1979) would be MW 8.715 implying a value of MW 8.7 for the earthquake in this study. The Great Nepal Bihar earthquake of 1934 in the central Himalaya with MS 8.3 is associated with MW 8.1–8.4 (Bilham and Ambraseys 2005; Nath et al. 2005; Lave´ et al. 2005).This earthquake is beyond the spatial scope of the present study, as we focus on the eastern and northeastern Himalaya. The earthquake of 1951 with MS 8.0 is connected with MW 7.8 correlating with the one in the Chinese catalogue (Rong 2002). The earthquakes with MS 7.5–8.0, whose corresponding MW equivalents are not correlated in any catalogue, are assigned MW values using a global relation from Scordilis (2006); MW ¼ 0:99ð0:02ÞMS þ 0:08ð0:13Þ;

for 6:2  MS  8:2:

ð5Þ

The regional seismic scaling relations are employed in applicable cases. However, in case of non-availability of the local observations, a fundamental correlation is drawn between the parameters from a global scaling relation. The above relation is adopted to overcome the paucity of data, as well as non-availability of relevant local relationships. The recorded events improved significantly from 1964 onwards (Fig. 3a). We employ a subcatalogue from 1964 to 2006 with coordinate bounds of 86°N–100°N and 19°E–31°E to estimate b-value and spatial fractal correlation Dimension Dc distributions. The b-value is the slope of the GR Frequency Magnitude Distribution (FMD) (Gutenberg and Richter 1944) given as, log10 ðNÞ ¼ a  bm;

ð6Þ

where N is the cumulative frequency of occurrence of magnitude, m in a given earthquake database. A square window of 2° 9 2° is rolled to cover the entire region with a slide of 0.5° degree every time. The magnitude of completeness, mt is estimated with the methodology described by Wiemer and Wyss (2000) which assumes the power law fit as the best fit for FMD at 90% confidence level. The estimation of mt is done in such a way that at least 100 events could be employed. The b-value and Dc are estimated only if a minimum of 50 events are found within the mt limit, and thereby is assigned to the centroid of the window. The 50 events criterion is essential for a meaningful statistical analysis (Utsu 1965).The estimation of b-value is done by maximum likelihood method (Aki 1965; Bender 1983; Utsu 1999) as, b¼

log10 ðeÞ ; mMean  ðmt  Dm 2 Þ

where mMean is the average magnitude and Dm is the magnitude bin size.

123

ð7Þ

Nat Hazards (2008) 45:333–377

343

Fig. 3 (a) A time series of events in the homogenized catalogue that indicates significant and stable recording during 1964–2006. However the recording of events with MW [ 5.6 is seen to be stable from 1924 onwards, and (b) the spatial distribution of b-value as contours overlaid on the Dc value zonation map

123

344

Nat Hazards (2008) 45:333–377

The spatial correlation dimension of the earthquakes is based on a power law exponent relating distance and the number of pair of points of either epicenters or focal depths within the distance (Kagan and Knopoff 1980; Kagan 2007). The correlated pair of points may be quantified by the correlation integral given by Grassberger and Procaccia (1983) as, CðN; rÞ ¼

N X N X 2 Hðr  jyi  yj jÞ; NðN  1Þ i¼1 j¼iþ1

ð8Þ

where H(x) is the heaviside step function whose value is 0 if x \ 0, otherwise 1. N is the total number of points in the query. Coordinates of the location of points are given by yi and yj. C(N, r) counts the number of points which are at a distance less than or equal to r with respect to total number of pairs of all the points. Great circle distances between the epicenter positions are used in the present study. The fractal correlation dimension is given as, Dc ¼ lim r!0

log10 CðN; rÞ : log10 ðrÞ

ð9Þ

A plot of log10(C(N, r)) against log10(r) is used to estimate Dc as the slope of the curve in its linear bound for a specific range of r. In case of an infinite two-dimensional distribution of the epicenters, the plot of log10C(N,r) against log10(r) is a straight line. But, practically for larger values of r a state of saturation is attained, thereby decreasing the gradient. For smaller values of r, an increase in the gradient is induced indicating a state of depopulation (Narenberg and Essex 1990). The specific range of r is projected from bounds of the saturation and depopulation limits. The spatial distribution of Dc value overlapped with b-value contours is depicted in Fig. 3b. The tectonic regime, spatial seismicity patterns: b-value and Dc distributions are considered to define four potential seismic source zones covering the region as presented in Fig. 4a. The four zones are Eastern Himalayan Zone (EHZ), Mishmi Block Zone (MBZ), Eastern boundary Zone (EBZ), and Shillong Zone (SHZ). The focal mechanism borrowed from GCMT database for moderate to large magnitude earthquakes in EHZ and its immediate neighborhood has been presented in Fig. 4b in order to highlight the type of tectonism associated with earthquakes occurring in this seismic territory of the Himalaya. Furthermore, it is seen that the strong events (M C7.0) in the study region mostly are of shallow depths. The Great Shillong earthquake of 1897 occurred in the neighboring seismic zone—the SHZ but at a considerably larger distance away representing this earthquake mechanism as a far-field source effect so far as Sikkim is concerned. As such, Himalayan earthquakes are likely to be more devastating as evident by the destructions caused to the Sikkim terrain by the Nepal Bihar Earthquake of 1934, MW 8.4 and the Bihar Nepal earthquake of 1988 MW 6.8 on the Himalayan terrain (Nath et al. 2005).

3 Methodology and results Earthquake hazard zonation employs the following variables: (i) Geomorphological characterization (a) Geology (GE), (b) Soil (SO), (c) Slope (SL),

123

Nat Hazards (2008) 45:333–377

345

Fig. 4 (a) The classified seismic source zones based on the seismicity analysis with ISC and GCMT catalogues for the period 1964–2006 is overlaid on the seismotectonic map prepared with the catalogue covering the period 1906–2006, and (b) The EHZ zoomed with the MW [ 4.5 earthquakes focal mechanisms from GCMT database depicted by beach balls

123

346

Nat Hazards (2008) 45:333–377

(d) Rock Outcrop (RO) and (e) Landslides (LS). (ii)

Seismological characterization (a) Prognosis of Maximum Earthquake (b) Site-specific ground motion studies, namely estimation of Site Effects (SE), Predominant Frequency (PF), PGA, Peak Ground Velocity (PGV) and Response Spectra.

3.1 Geomorphological thematic mapping and attribute generation The geomorphological inputs include IRS-1C LISS III digital data of March 2000, topo-sheets from Survey of India, geographical boundary for the State of Sikkim, surface geological maps, soil taxonomy map in 1:50,000 scale from National Bureau of Soil Survey (1994) and seismic refraction profiles. All the maps and topo-sheets are scanned at 200 dpi with a resolution of 6 m for a scale of 1:50,000 and rectified with a common base using Everest Polyconic projection system. A second-degree polynomial surface fitting during the rectification process removes any distortion in the scanned image. The respective features on each rectified raster image are digitized for conversion to a vector layer/coverage using ARC/INFO GIS software. The IRS-1C LISS III data are converted to False Color Composite (FCC) for the generation of multispectral images, which are also rectified to the same base. The themes thus generated are geographical boundaries of Sikkim and its districts, along with the geomorphological attributes, namely, surface geological units, soil taxonomy, site classification, ROs, and landslides. 3.1.1 Regional geology of the Sikkim Himalaya The Sikkim region is located in the earthquake-prone territory of the eastern Himalaya along Darjeeling-Sikkim tract, where fast and unplanned urbanization is still active with the record of a good number of moderate earthquakes in this terrain. Most workers have divided the Himalayas into a series of longitudinal tectono-stratigraphic domains such as (1) Sub Himalayas, (2) Lesser Himalayas, (3) Higher Himalayas, and (4) Tethys Himalayas, as shown in Fig. 5a, which are separated by major dislocation zones. In the Sikkim region, different lithological units are disposed in an arcuate regional-fold pattern (Gansser 1964). The ‘core’ of the region is occupied by the Lesser Himalayan low-grade metapelites, interbanded psammite belonging to Daling Group (Proterozoic to Mesozoic). The distal parts of the region are characterized by medium to high-grade crystalline rocks of the Higher Himalayan Belt (Higher Himalayan Crystalline Complex, HHC). A prominent ductile shear zone, the Main Central Thrust (MCT) separates the two belts. At places several subsidiary thrusts are present between MCT and MBT. The structural configuration of this foredeep region is architectured by a set of almost N–S faults, resulting in the development of alternate horst and graben structures. Along with the transverse faults, several lineaments cutting across the Himalayan belt are also present. Other prominent lineaments include NW–SE trending Tista and Gangtok lineaments (Dasgupta et al. 2000). In the surface geological unit layer of the geomorphological themes, the significant attribute comprises of Sub Himalayas, Lesser Himalayas, Higher Himalayas, and Tethys Himalayas, as shown in Fig. 5b.

123

Nat Hazards (2008) 45:333–377

347

3.1.2 Slope Since the study region belongs to the rugged terrain of the Himalayan territory, topography also comes into consideration. Triangulated Irregular Network (TIN) have been prepared for the percent slope mapping as shown in Fig. 6a with the help of other geomorphological data, namely IRS-1C LISS III digital data, Survey of India Topo-sheets, and administrative

Fig. 5 (a) Generalized Geological map of the Himalayas, showing different geotectonic domains and lithological units. Inset shows the location of the Sikkim Himalaya. MBT, Main Boundary Thrust; NP, Nanga Parbat; ND, Nanda Devi (after Gansser 1964), and (b) Schematic Geological map of the Sikkim Himalaya displaying detail petrographic provinces, morphotectonic features and drainage patterns. The location of strong motion stations are shown as solid triangles

123

348

Nat Hazards (2008) 45:333–377

boundary of the State of Sikkim. All the maps and topo-sheets are converted into vector coverage data through ERDAS Imagine and Arc/ INFO software.

3.1.3 Rock outcrop and landslide Rock outcrop and landslide scarp region had been identified and vectorized into two separate polygon coverage. Ground truthing was also done for the logistically accessible region. The ROs mapped into bi-color and the landslide zones encircled based on their aerial extent on the IRS imagery are shown in Fig. 6b and c respectively. 3.1.4 Soil Since soils are basically weathered products of rocks, there is a wide variation in the soil characteristics of Sikkim. According to the National Bureau of Soil Survey (1994), soils of Sikkim are broadly grouped under five physiographic zones as listed below. A. Soils on summit and ridge tops: (i) Steeply sloping ([30%) surfaces register mostly coarse-grained soils with rock fragments, and can be grouped under typical Hapludolls to typical Udorthents. (ii) Moderately steep sloping (30%) surfaces exhibit thick soil cover. They are well to somewhat excessively drain coarse-loamy to fine-loamy soils with little or no rock fragments. They vary from typical Haplumbrepts to Umbric Dystrochrepts. (iii) Moderately sloping (15%) surfaces are composed of well-drained fine-loamy soils with local strewn pebbles on the surface. They belong to typical Hapludolls. They are also associated with coarse loamy soils at places. B.

Soils on side slopes of hills: On steeply sloping surfaces (50%), they are excessively drained coarse-loamy to fineloamy soils with slight surface stoniness and can be classified under entic Hapludolls and typic Haplumbrepts. C. Soil on valleys: These are excessively drained loamy-skeletal soils with slight stoniness and moderate erosion. They are typical Haplumbrepts. D. Soils on cliff and precipitous slope: These are excessively drained loamy skeletal soils with strong stoniness and very severe erosion. These are permanent fallow lands, and can be treated as Lithic Udorthents. E. Soils on glacial drift moraines and boulders: These are associated with shallow excessively drained loamy-skeletal soils with strong surface stoniness and strong erosion. There is a great deal of variation in the physical properties of the soil here. On the basis of the origin of their constituents, soils can be divided into two major groups, those which are the results of chemical and physical rock weathering and those, which are of organic origin. If the products of rock weathering are still located at the place where they originated, they constitute a residual soil. As far as the genesis of soil of Sikkim is concerned, rocks of the Daling group contribute the most. The physiographic zones of soil of Sikkim are vectorized and stored as the

123

Nat Hazards (2008) 45:333–377

349

Fig. 6 (a) Slope coverage map, (b) RO coverage of the Sikkim Himalaya, and (c) Landslide coverage of the Sikkim Himalaya on IRS-1C LISS III image. Larger circle represents greater aerial extent

soil taxonomy coverage as shown in Fig. 7a. Using the soil classification based on composition, grain size, and lithology, the site classification is done as sites IB (S-wave velocity, b [ 1,500 m/s), IC (b = 700–1,500 m/s), II (b = 350–700 m/s) and III (b = 180– 350 m/s) by combining polygons of same broader taxonomy as depicted in Fig. 7b. 3.2 Seismological inputs The Seismological attributes considered for seismological thematic mapping are: (a) Site response (SR) at predominant frequencies (b) Peak Ground Acceleration for hazard scenario with a scenario earthquake magnitude (SEM) of MW 8.3 nucleating from the hypocenter of event 50 (Table 1).

123

350

Nat Hazards (2008) 45:333–377

3.2.1 Scenario earthquake magnitude Scenario earthquake magnitude (SEM) may be defined as the one that is employed to project a seismic hazard scenario. Either a large historical earthquake or possible Maximum Earthquake or Maximum Credible Earthquake (MCE) is adopted as SEM as situation demands. The initial analysis is done with the corresponding subcatalogue derived for the source zone covering from 1964 to 2006. The magnitude of completeness, mt is estimated to be 4.7 (at 90% confidence level) for a linear fit of GR law on the FMD plot as depicted in Fig. 8a which gives a and b-value to be 8.948 : 0.581 and 1.401 : 0.086, respectively. MCE is generally estimated from the linear extrapolation of the GR law and adopting Poisson’s distribution (Nath et al. 2005), f ðxÞ ¼ 1  ekt

ð10Þ

where f(x) is the probability of exceedance, k is the frequency of occurrence and t is the period in years. A recurrence of 10% probability of exceedance in 50 years which corresponds to a return period of 475 years in Eqs. (10) and (6) gives a MCE of magnitude 8.298 : 0.87. The lack of very large earthquakes in the subcatalogue allowed the linear extrapolation on the GR law to be constrained. However, a linear fit on GR law does not bound the magnitude of earthquakes and generally tend to predict extremely high magnitudes. In context of earthquakes as self organized criticality (SOC) phenomena (Bak and Tang 1989), three behaviors may be exhibited: critical, subcritical, or supercritical (Main 1995). The linear fit of GR law is seen in the critical behavior. In the subcritical situation, the largest events occur lesser than implicated by the GR law. Supercriticality is seen with the largest events exhibiting characteristics behavior and more occurrence than implicated by the GR law. The gamma distribution with power law scaling at the lower and exponential at the higher magnitudes has been seen to fit the best in each behavioral case (Main 1995; Kagan 1993). Further examination is carried out involving past records of large earthquakes in the estimation process. If we extend the subcatalogue backwards for the EHZ seismic zone; the records commence from 1924 onward. The moderate and large earthquakes magnitude [5.6) do have stable recording from 1924 onwards, as observed in Fig. 3a. We, therefore, consider a subcatalogue derived from the main catalogue for this zone that covers a period from 1924 to 2006. Adopting an approach similar to that of Koravos et al. (2003), incremental frequency distribution is preferred instead of cumulative one to avoid biasing to the lower magnitudes and overestimation of b-values. The incremental frequency accounts for the number of events in each magnitude bin, Dm = 0.1. The frequency function is applied to each unique data point in the data catalogue to perceive the gaps in the range of values. Following Main et al. (1999), a three point running mean is applied to the raw frequency data to minimize the effects of empty bins as given by, 1 Ni ¼ ðNi1 þ 2Ni þ Niþ1 Þ 4

ð11Þ

where Ni is each data point. The fit of a gamma model on the FMD plot as depicted in Fig. 8b shows a slight affinity toward super critical behavior. The gamma model can be given as,

123

Nat Hazards (2008) 45:333–377

351

Fig. 7 (a) Physiographic zones of soil in the Sikkim Himalaya (after National Bureau of Soil Survey 1994), and (b) Soil site classification coverage of the Sikkim Himalaya

ln NðmÞ ¼ a  bm  h expðCm þ DÞ

ð12Þ

where N(m) is the incremental frequency of earthquake occurrences with magnitude, m. a, b and h are model parameters. The values of C and D are computed as C = c ln(10) and

123

352

Nat Hazards (2008) 45:333–377

Fig. 8 (a) Frequency-magnitude distribution for EHZ from the corresponding sub catalog covering the period from 1964 to 2006. (b) The incremental frequency–magnitude distribution plots for the sub catalog covering the period from 1924 to 2006 for the seismic source zone EHZ. The thick line in the plot indicates the best fit curve. The dash lines give 95% confidence bounds on this fit. The error bars given for each magnitude are estimated as F/H N where F is the frequency and N is the number of events used to compute F (Koravos et al. 2003). (c) Magnitude against the total seismic moment contribution by each magnitude

D = d ln(10) respectively, where c and d are the constants used in the relation employed to estimate the moment from the magnitude; log10 ðMÞ ¼ cM W þ d:

ð13Þ

The b-value is given as b = b/ln(10). In order to constrain the estimation of the possible maximum earthquake by seismic moment release, we apply characteristic moment distribution for maximum moment given by Kagan (1993) as, 1B B Mmin X ¼ NT Mmax

B ð1  BÞ

ð14Þ

where X is the annual moment release, which we compute from the corresponding catalogue and B is given by b/c. NT is the annual number of events above a minimum

123

Nat Hazards (2008) 45:333–377

353

threshold moment, Mmin. Mmax is the maximum moment. Equation (13) is valid for B \ 1 (b \ 1.5) and Mmin  Mmax. Main (1995) observed that the MCE estimated with physical constraints from seismic moment rate are either dependent solely on the exponential tail of FMD or the events near the maximum magnitudes. The moment releases due to the lower magnitude earthquakes are negligible compared to those of higher magnitudes. The threshold or minimum cutoff magnitude of MW 6.4 is decided from Fig. 8c. The b-value is found to be 0.7435 : 0.182. The individual seismic moment for each earthquake is estimated using the empirical relation from Hanks and Kanamori (1979); log10 M0 ¼ 1:5MW þ 9:1;

ð15Þ

where M0 is in Nm. NT and X are computed and found to be 0.1429 and 2.0949 9 1019, respectively. The maximum earthquake as estimated from Eqs. (13) and (14) is found to be 8.34 : 0.238. Finally, SEM for EHZ, in order to simulate seismic scenario is considered to be MW 8.3 by rounding off the magnitudes estimated in the foregoing analysis to the one place of decimal. This earthquake has been designated to be nucleating from the hypocenter of the earthquake occurred on December 2, 2001, ML 5.6 (Event 50, Table 1), the largest amongst 80 significant events recorded by SSMA in a span of 3 years. This event occurred just below MBT at a depth of 26.3 km with a fault plane solution having the 310° strike and 35° NNE dip is similar to the focal mechanism of events of magnitude MW 4.8 at the latitude 86.68°E, longitude 26.93°N and MW 6.8 at the latitude 86.61°E, longitude 26.51°N.

3.2.2 Strong motion data analysis and SR estimation Several techniques to evaluate SR have been utilized and compared in recent studies (Field and Jacob 1995; Field et al. 1992). Two of the proposed methods, the standard spectral ratio (SSR) (Borcherdt 1970) and the receiver function technique (HVSR) (Langston 1979; Lermo and Cha´vez-Garcı´a 1993; Nath et al. 2000, 2002a, b, 2005) are based on a spectral ratio scheme. In both these techniques, the source and path contributions are removed from the seismic recordings by means of a deconvolution operation using a function free from site effects. The accelerograms of the selected events were first corrected for the system response. Then, the S-wave packets recorded by the accelerograph were windowed with a window width containing the maximum amplitude. The window length was selected following the results of Seekins et al. (1996). A Hanning taper is applied to the time windowed data and Butterworth bandpass filtered before the amplitude spectra were computed. The signal amplitude spectrum at the frequency fk can be expressed as, Aðrij ; fk Þ ¼ Oðrij ; fk Þ  Bðrij ; fk Þ;

ð16Þ

where O(rij, fk) and B(rij, fk) represent the S-wave spectral amplitude and that of the background noise respectively at the hypocentral distance rij. The corrected spectra are smoothened in order to reduce the data variance using a 5-point smoothing window and a spline interpolator at 0.1 Hz interval.

123

354

Nat Hazards (2008) 45:333–377

For a network with i events recorded by j stations, the amplitude spectrum of the ith event recorded at the jth station for the kth frequency, A(rij, fk) can be written in the frequency domain as a product of a source term SOi(fk), a propagation path term P(rij, fk), and a site effect term SIj(fk), (Lermo and Cha´vez-Garcı´a 1993, Nath et al. 2002a, b). Aðrij ; fk Þ ¼ SIj ðfk Þ  Pðrij ; fk Þ  SOi ðfk Þ;

ð17Þ

Lermo and Chavez-Garcia (1993) presented evidence that SR can be estimated by taking horizontal-to-vertical component ratios of the shear-wave spectra. This technique was originally introduced by Nakamura (1989) to analyze ambient seismic noise recordings. Langston (1979) applied a method of determining the velocity structure of the crust and upper mantle from teleseismically recorded P-waves. The vertical component is assumed to be relatively uninfluenced by the local structure, whereas the radial component contains P- and S-wave conversions from structural discontinuities below the site. Therefore, an estimate of the impulse response function or receiver function below the site can be obtained by deconvolving the vertical from the radial component. Site response can also be calculated by reference station spectral ratio technique. Site Amplification Factor, SIj(fk), in Eq. (17) estimated by Horizontal-to-VerticalSpectral Ratio, HVSRij(fk), can be computed at each jth site for the ith event at the central frequency fk from the root mean square (rms) average of the amplitude spectra (Nath et al. 2005), qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi p1ffiffi absHij ðfk Þj2NS þ absHij ðfk Þj2EW ð18Þ HVSRij ðfk Þ ¼ 2 absVij ðfk Þ where Hij(fk)|NS is the Fourier spectra of the NS component, Hij(fk)|EW is the Fourier spectra of the EW component and Vij(fk) is Fourier spectra of the vertical component. Since in the Sikkim region we lacked reference site data coverage due to logistic problem, we have considered the HVSR as representative of the SR (Nath et al. 2005) in the study region. Site amplification for each of the nine stations has been computed from 80 strong motion events given with signal-to-noise ratio greater than 3.0, as shown in Fig. 2a with the recording history given in the Table 1 for a span of 3 years during 1999–2002. It may be noted that the SSMA was initiated in May 1998. In order to analyze the azimuthal dependency of site amplification at different stations, we performed experiments using HVSR technique to study the effect on both the horizontal components of each event and dependency of the rms site amplification on minor and large variation in source azimuth. The results at Singtam and Gezing are presented here. In Fig. 9a–d, we have presented the HVSR plots at four source azimuths, 227°N, 36°N, 301°N, and 284°N respectively recorded at Singtam station for the NS and EW components. The similar analysis is performed at Gezing for the source azimuths 104°N, 122°N, 290°N, and 330°N as shown in Fig. 9e–h, respectively. Here also the NS and EW components follow each other closely thereby ruling out the dependency of site amplification on station azimuth, i.e., 90° difference between two components. The RMS site response components for the events coming from 104°N and 122°N azimuths are compared in Fig. 10a, those

123

Nat Hazards (2008) 45:333–377

355

coming from 290°N and 330°N in Fig. 10b and finally those coming from strikingly different average azimuthal directions of 115° (between 104°N and 122°N) and 310°N (between 290°N and 330°N) in Fig. 10c. It is evident from Fig. 10a and b that site amplification does not reflect the azimuthal variation for the events coming from the azimuthal deviation of 18°@40°. However it is not the same case when the events are coming from strikingly different azimuthal directions as exhibited in Fig. 10c for a deviation of 165° in source azimuth. The SR curves in the figure do not replicate each other. This significant difference in the spectral amplification may be attributed to different source radiation patterns, scattering, diffraction and topographic effects that influence the site effects in a hilly terrain. Considering the variation of SR, our consideration for simulation of seismic scenario with SR at all the stations for the source azimuth has been used with respect to the scenario earthquake source and fault plane solution (Fig. 11). Site response at different station azimuth for predominant frequencies is given in Fig. 12a while PF is depicted in Fig. 12b.

3.2.3 Synthesis of strong ground motion by wavenumber integration method An attempt has been made to simulate the scenario earthquake of magnitude MW 8.3 in the study region. Generally, a great earthquake of magnitude MW [ 8.0 is expected to generate a finite length rupture when it nucleates. The composite fault plane solution of an earthquake of magnitude ML 5.6 (Event 50, Table 1) and its proximity to MBT suggests thrust faulting with little strike slip component with a dip of 35°NNE striking at 310° as shown in Fig. 2b. Frequency–wavenumber integration method is useful if the range of possible fault rupture history is narrow enough to functionally constrain the predicted strong ground motion as is the case here. For computation of the synthetic accelerogram, an impulsive source has been used as a first approximation for the near-field effect. The wavenumber integration method of Herrmann and Mandal (1986) is then followed. The generation of synthetic seismograms for point sources in simply layered structures has made rapid advances in the past decade. Two approaches involving Laplace transform and Fourier transform are actively being pursued. The Laplace transform or Cagniard–de Hoop technique, usually referred to as the generalized ray method (Helmberger 1968), constructs the solution by tracking the individual seismic arrivals ray by ray from the source to receivers. This method is valid at high frequencies and works well at predicting particular phases, but is poorly suited to models with many layers and larger distances when a complete seismogram is desired. The other approach involves expressing the solutions in terms of a double integral transformation over wavenumber and frequency (Hudson 1969).The complete solution, rather than individual rays, is considered in such a full wave theory approach. This method can handle a larger number of plane layers, but requires considerable computational effort especially at high frequencies. Suppose that an earthquake can be represented by a double-couple without moment source model with the symbols ‘n’ for the vector normal to the fault and ‘f’ for the

123

356

Nat Hazards (2008) 45:333–377

Site Amplification

Site Amplification

(a) 10

10

1

10

0

(b)

1 SINGTAM- Source Azimuth: 277°N

N-S E-W

SINGTAM- Source Azimuth: 35.86°N

1

10

100

10

0

10

Frequency (Hz)

Site Amplification

Site Amplification

1

SINGTAM- Source Azimuth: 301°N

1

N-S E-W

10

1

100

1

Site Amplification

Site Amplification

10

0

10

0

10

1

0

10

N-S E-W

-1

GEZING - Source Azimuth:122°N

10

1

10

2

10

0

(g) 10

0

10

-1

GEZING - Source Azimuth:290°

10

N-S E-W 2

1

10

Frequency (Hz)

Site Amplification

Site Amplification

1

100

(f)

Frequency (Hz)

10

10

Frequency (Hz)

(e)

GEZING - Source Azimuth:104°N

N-S E-W

SINGTAM- Source Azimuth: 284°N

0.1

10

-1

2

10

Frequency (Hz)

10

10

(d)

10

1

1

Frequency (Hz)

(c)

10

N-S E-W

-1

10

1

(h)

0

10

N-S E-W

GEZING - Source Azimuth:330°

10

N-S E-W

-1

10

0

10

1

Frequency (Hz)

10

2

10

10

0

10

1

10

2

Frequency (Hz)

Fig. 9 Comparison of NS and EW component of SR at Singtam for the source azimuth (a) 227°N, (b) 36°N, (c) 301°N respectively, and (d) 284°N. Comparison of NS and EW component of SR at Gezing for the source azimuth (e) 104°N, (f) 122°N, (g) 290°N, and (h) 330°N respectively

123

Nat Hazards (2008) 45:333–377

Site Amplification

(a)

1

10

Site Amplification

1

10

357

0

10

-1

10

°

° °

0

1

10

-1

2

10

10

Frequency (Hz)

1

Site Amplification

10

0

10

°

GEZING: Source Azimuth 100 -130 N 104 N (RMS) 122 N (RMS)

(b)

10

°

°

GEZING: Source Azimuth 290 -330 N 290 N (RMS) 330 N (RMS)

° °

0

10

1

10

2

10

Frequency (Hz)

(c)

0

10

-1

10

GEZING: Average RMS 115 N Source Azimuth 310 N Source Azimuth

° °

0

10

1

2

10

10

Frequency (Hz)

Fig. 10 Variation of site amplification with different source azimuthal coverage at Gezing by HVSR. Comparative plot between the RMS site response for the events at (a) 104°N and 122°N, (b) 290°N and 330°N, and (c) average rms site response for the events at 115°N and 310°N azimuths (after Nath et al. 2005)

direction of force as used by Haskell (1963, 1964). The Fourier transform of vertical, radial and tangential components of the displacement can be given as, uz ðr; /; 0; xÞ ¼ZSS ½ðf1 n1  f2 n2 Þ cos 2/ þ ðf1 n2 þ f2 n1 Þ sin 2/ þ ZDS ½ðf1 n3 þ f3 n1 Þ cos / þ ðf2 n3 þ f3 n2 Þ sin / þ ZDD ½f3 n3 ;

ð19Þ

ur ðr; /; 0; xÞ ¼RSS ½ðf1 n1  f2 n2 Þ cos 2/ þ ðf1 n2 þ f2 n1 Þ sin 2/ þ RDS ½ðf1 n3 þ f3 n1 Þ cos / þ ðf2 n3 þ f3 n2 Þ sin / þ RDD ½f3 n3 ; u/ ðr; /; 0; xÞ ¼TSS ½ðf1 n2 þ f2 n1 Þ cos 2/  ðf1 n1  f2 n2 Þ sin 2/ þ TDS ½ðf2 n3 þ f3 n2 Þ cos /  ðf1 n3 þ f3 n1 Þ sin /

ð20Þ

ð21Þ

where ZDD, ZDS, ZSS, RSS, RDS, RDD, TSS and TDS are referred as Green’s functions. RSS and RDS in Eq. (20) also include near-field terms. These terms decrease faster than the others, and therefore, are important only at short distances.

123

358

Nat Hazards (2008) 45:333–377

Fig. 11 Site amplification at all the nine stations of corresponding source azimuth with respect to the source at event 50 (Table 1) of magnitude MW 5.6

The inverse Fourier transform (19), (20), and (21) on multiplication of @x 2, needs a convolution of the source spectra for the generation of acceleration time history of the vertical, radial and tangential components of ground motion as given below ur;z;/ ðr; /; 0; tÞ ¼

Z1

SðxÞur;z;/ ðr; /; 0; xÞ expðixtÞdx=2p

ð22Þ

1

where S(x) is the Fourier spectra of the impulse source function as described by Herrmann (1979). All the ten components of synthetic Green’s functions have been estimated at all the seismic recording stations and consequently the simulation of synthetic accelerograms that incorporates the fault parameters as well. The S-wave part of radial and transverse components of accelerogram thus obtained from the above simulation is convolved with SR of the station in order to compute the response on engineering bed rock. Finally the convolved accelerogram has been used to generate a seismic scenario for the SEM of MW 8.3. In order to establish the robustness of the simulation procedure, ground acceleration for two earthquakes of magnitude MW 5.1 (Event 57, Table 1) and MW 5.6 (Event 50, Table 1) have been simulated and compared with the recorded ones at the strong motion stations Chungthang, Mangan, and Jorthang for MW 5.1, while that for MW 5.6 at Peling, Singtam, and Gangtok in both time and spectral domains as presented in Figs. 13–18, respectively for Chungthang, Mangan, Jorthang, Peling, Singtam, and Gangtok. The simulated and

123

Nat Hazards (2008) 45:333–377

359

Fig. 12 (a) Site response distribution map of the Sikkim Himalaya at predominant frequencies, and (b) PF distribution map of the Sikkim Himalaya

123

360

Nat Hazards (2008) 45:333–377

observed accelerograms in both the radial and transverse directions match in both the time and frequency domains except for insignificant artifacts either toward very low or very high frequency ends. Unlike stochastic simulation, no high-cut filter is used to control and stabilize the corner frequency of the simulated ground motion spectra in the Frequency– wavenumber integration method. Wang and Herrmann (1980) suggested significant contribution to the waveform coming even from leaky modes related to the waves from (+ , @) sheet, i.e., leaking S-waves but trapped P-waves. The behavior at higher frequencies resembles that of the locked normal modes resulting in minor overestimation in cases wherever this condition is encountered. In the results given here, the acceleration spectra at Chungthang, Mangan, Jorthang, Peling and Singtam present a sharp spectral fall-off at higher frequencies, while those at Gangtok seem to suffer from leaky modes. It is to be noted that the accelerograms at distances less than 100 km include both near- and far-field terms; beyond this distance only far-field terms are predominant at the expense of enhanced computation time.

3.2.4 Estimation of PGA (Peak Horizontal Acceleration) using random vibration theory (RVT) Peak ground acceleration can be predicted using RVT (Vanmarcke and Lai 1980) without calculating time series. Although we have generated time series, this theory is preferred because we can avoid reliance on the additional comparison that may be required for calculating time series. Here E(arms) is determined by arms using following relations (Boore 1983). The kth moments mk of energy density spectrum of acceleration are defined as, 1 mk ¼ p

Z1

xk jAðxÞj2 dx

ð23Þ

0

The rms acceleration is given by arms ¼ ðm0 =T Þ1=2

ð24Þ

where T is the duration of the accelerogram. For an accelerogram in which the number of extrema N is less than 2, N is arbitrarily set equal to 2. If the value of N exceeds 20, an asymptotic approximation given below should be used for E(amax) (Cartwright and Longuet-Higgins 1956). Eðamax Þ ¼ ½2 lnðN Þ1=2 þc=½2 lnðN Þ1=2 arms

ð25Þ

where N is determined by N ¼ 2 feT

ð26Þ

1 fe ¼ ðm2 =m0 Þ1=2 2p

ð27Þ

and fe is PF and is calculated by

123

Observed

Simulated

-4

-3

-2

-4

10

-3

10

10

(d)

10

10

-2

0.1

0.1

Simulated Observed

Simulated Observed

1

Frequency (Hz)

1

Frequency (Hz)

10

10

100

100

Fig. 13 Comparison between the observed and simulated accelerograms of magnitude MW 5.1 at Chungthang. The radial component is depicted in (a) time domain, and (b) frequency domain. The transverse component is depicted in (c) time domain, and (d) frequency domain

(c)

Observed

Simulated

10

(b) Spectral Acceleration (g/Hz)

Spectral Acceleration (g/Hz)

(a)

Nat Hazards (2008) 45:333–377 361

123

123 Observed

Simulated

(d)

10

10

10

-4

-3

-2

-4

10

-3

10

-2

10

Simulated Observed

0.1

0.1

Simulated Observed

1

Frequency (Hz)

1

Frequency (Hz)

10

10

100

100

Fig. 14 Comparison between the observed and simulated accelerograms of magnitude MW 5.1 at Mangan The radial component is depicted in (a) time domain, and (b) frequency domain. The transverse component is depicted in (c) time domain, and (d) frequency domain

(c)

Observed

Simulated

(b) Spectral acceleration (g/Hz) Spectral acceleration (g/Hz)

(a)

362 Nat Hazards (2008) 45:333–377

Observed

Simulated

(d)

10

-5

-4

10

-3

10

-4

10

-3

10

-2

10

0.1

0.1

Simulated Observed

Simulated Observed

1

Frequency (Hz)

1

Frequency (Hz)

10

10

100

100

Fig. 15 Comparison of observed and simulated accelerograms of magnitude MW 5.1 at Jorthang. The radial component is depicted in (a) time domain, and (b) frequency domain. The transverse component is depicted in (c) time domain, and (d) frequency domain

(c)

Observed

Simulated

(b)

Spectral acceleration (g/Hz)

Spectral Acceleration (g/Hz)

(a)

Nat Hazards (2008) 45:333–377 363

123

123 Observed

Simulated

(d)

10

-4

-3

10

-2

10

-4

10

-3

10

-2

10

0.1

0.1

SImulated Observed

SImulated Observed

1

Frequency (Hz)

1

Frequency (Hz)

10

10

100

100

Fig. 16 Comparison of observed and simulated accelerograms of magnitude MW 5.6 at Gezing. The radial component is depicted in (a) time domain, and (b) frequency domain. The transverse component is depicted in (c) time domain, and (d) frequency domain

(c)

Observed

Simulated

(b) Spectral Acceleration (g/Hz)

Spectral Acceleration (g/Hz)

(a)

364 Nat Hazards (2008) 45:333–377

Observed

Simulated

(d)

1E-5

1E-4

1E-3

-4

10

-3

10

-2

10

0.1

0.1

Simulated Observed

1

Frequency (Hz)

1

Frequency (Hz)

SImulated Observed

10

10

100

100

Fig. 17 Comparison of observed and simulated accelerograms of magnitude MW 5.6 at Singtam. The radial component is depicted in (a) time domain, and (b) frequency domain. The transverse component is depicted in (c) time domain, and (d) frequency domain

(c)

Observed

Simulated Spectral Acceleration (g/Hz)

(b)

Spectral Acceleration (g/Hz)

(a)

Nat Hazards (2008) 45:333–377 365

123

123 Observed

Simulated

(d)

10

-4

-3

10

-2

10

-1

10

1E-4

1E-3

0.01

0.1

1

0.1

0.1

Simulated Observed

Simulated Observed

1

Frequency (Hz)

1

Frequency(Hz)

10

10

100

Fig. 18 Comparison of observed and simulated accelerograms of magnitude MW 5.6 at Gangtok. The radial component is depicted in (a) time domain, and (b) frequency domain. The transverse component is depicted in (c) time domain, and (d) frequency domain

(c)

Observed

Simulated

(b) Spectral Acceleration (g/Hz)

Spectral Acceleration (g/Hz)

(a)

366 Nat Hazards (2008) 45:333–377

Nat Hazards (2008) 45:333–377

367

In the PGA calculation, maximum cutoff frequency was taken as 15 Hz, since it was the maximum PF of the area as observed from the acclerometric data. The simulated PGA map is shown in Fig. 19 with the spectral acceleration of each station associated with it. 4 Integrated simulation of earthquake hazard Multi-criteria evaluation approach adopted in the present analysis employs a matrix of pair-wise comparisons between the factors built using Saaty’s Analytical Hierarchy Process (Saaty 1980, 1990). This matrix is constructed by eliciting values of relative importance on a scale of 1 to n wherein ‘1’ indicates that the two factors are equally important and ‘n’ indicates relative importance than the other. Similarly, a factor less important than others is indicated by reciprocals of the values ranging from 1 to n (i.e., 1/1–1/n). The process of allocating weights is a subjective one and can be done in participatory mode in which a group of decision makers may be encouraged to reach a consensus opinion about the relative importance of factors. The matrix developed by pair-wise comparisons between the factors can be used to derive the individual normalized weights of each factor. It is performed by calculating the principal eigenvector of the matrix. The weights for each attribute can be calculated by averaging the values in each row of the matrix. Addition of these weights will be ‘1’ and can be used in deriving the weighted sum of ratings or scores for each region of cells or polygon of the mapped layers. Since the values within each thematic map/layer vary significantly, they are classified into various ranges or types known as the features of a layer. These features are then assigned ratings or scores within each layer, normalized to ensure that no layer exerts an influence beyond its determined weight. Thus, a raw rating for each feature of every layer is allocated initially on a standard scale such as 1 to n (say 8 in this case) and then normalized using the relation, xj ¼

Rj  Rmin Rmax  Rmin

ð28Þ

where Rj is the raw score, Rmin and Rmax are the minimum and maximum scores of a particular layer. For earthquake hazard delineation, both geomorphological and seismological themes are broadly reclassified into a 1st phase geohazard map and 2nd phase earthquake microzonation map. The geomorphological themes include surface geology, soil taxonomy or site class, percent slope, RO and landslide, whereas the seismological themes include SR at PF, PF and simulated PGA for a SEM of MW 8.3. For the initial geohazard integrated model, the geology (Fig. 5b), site class (Fig. 7b), percent slope (Fig. 6a), landslide (Fig. 6c) and RO (Fig. 6b) are scaled in their contributing weights in 5:1, a maximum of rank 5 equivalent to a normalized weight 0.3333 has been attached to geology, rank 4 with an equivalent normalized weight 0.2666 assigned to site class, rank 3 with an equivalent normalized weight of 0.2 being attached to percent slope, rank 2 with an equivalent normalized weight of 0.1333 attached to landslide and rank 1 with an equivalent weight of 0.0666 attached to RO. The pair-wise comparison matrix for geohazard is shown in Table 3a wherein the Saaty’s Analytical Hierarchy Process has been used. The ranking value for geomorphological attributes is shown in the Table 4. All the geomorphological coverage has been prepared for the thematic mapping with a built up polygon topology having attributes classified in a range of assigned values. The vector overlay operation (Fig. 1) was performed with the Arc/INFO software application.

123

Fig. 19 PGA zonation map of the Sikkim Himalaya as estimated through the wavenumber integration simulation of the SEM of MW 8.3 assumed to be nucleating from the hypocenter of event 50 (Table 1). The spectral acceleration at each station is depicted in the subplots marked by arrows with the stations on PGA zonation map

368

123

Nat Hazards (2008) 45:333–377

Nat Hazards (2008) 45:333–377

369

Thus, the geohazard map is obtained through the integration technique considering all the geomorphological coverage in account using the following equation. GHI ¼ ðGEw  GEr þ SOw  SOr þ SLw  SLr þ LSw  LSr þ ROw  ROr Þ=Rw

ð29Þ

The notations have their usual meanings. The summarized geohazard index (GHI) ranging between 0 and 0.87, and the values thus obtained are suitably categorized into three classes with GHI ranged between 0.0 and 0.35 quantitatively termed as Low Hazard from 0.35 to 0.55 termed as Moderate Hazard, and greater than 0.55 assigned a qualitative definition as High Hazard. The geohazard map of Sikkim Himalaya is presented in Fig. 20, wherein the high hazard region is seen to dominate the lesser Himalaya, the moderate and intermediate hazard region in the terrain between the lesser and higher Himalaya Crystalline (Fig. 5a) just above MCT, while the higher Himalaya exhibits low geohazard index. The vector union operation is subsequently performed on the seismological themes one after another with the special distributions of F–K integration simulated (Green’s function) horizontal PGA (Fig. 19), PF (Fig. 12b), and SR at PF (Fig. 12a) respectively with the geohazard base layer presented in Fig. 20 in a supervised mode judging each thematic contribution toward earthquake hazard. The united site-condition vector coverage helped in the polygon interpolation of seismological attributes through least square error energy minimization criterion. Integration of PGA with geohazard following the relation X w ð30Þ EHIPGA ¼ ðGHIw  GHIr þ PGAw  PGAr Þ= gives the first level seismic guided earthquake hazard zonation as depicted in Fig. 21 with highest index being 0.3 and the lowest being less than 0.05. Integration of this layer with predominant frequency vector coverage enhances the hazard level to a maximum of 0.56 and a minimum less than 0.1. Even the high hazard index demarcated zone gets more concentrated toward the MBT encompassing Singtam. Further integration of this theme with PF using the relation   X w ð31Þ  EHIPGA þ PFw  PFr = EHIPF ¼ EHIPGA w r gives the second level seismic hazard index map which is depicted in Fig. 22. Finally, the seismic vector coverage site response has been overlaid and integrated to generate the earthquake hazard zonation of Sikkim Himalaya using the relation   X PF w ð32Þ EHI ¼ EHIPF w  EHIr þ SRw  SRr = The pair-wise comparison of seismic and geohazard matrix is presented in Table 3b, while the corresponding rank of attributes of seismological and geohazard themes are given in Table 5. A holistic earthquake hazard model evolved in this analysis delivering the output theme, presented in Fig. 23, which follows the relationship, EHI ¼

SRw SRr þ PFw  PFr þ PGAw PGAr þ GEw GEr þ SOw SOr þ SLw SLr þ LSw LSr þ ROw ROr

! =

X

w

ð33Þ The ultimate earthquake hazard microzonation map as depicted in Fig. 23 classifies the Sikkim Himalayan territory into six broader hazard zones with the maximum being 0.83

123

370

Nat Hazards (2008) 45:333–377

Table 3 Pair-wise comparison and weight distribution of (a) all the geomorphologic themes, and (b) seismological themes and the geohazard vector layer Theme

Geology

Site class (Soil)

Percent slope

Landslides

Rock

Weights

Geology

1

5/4

5/3

5/2

5

0.3333

Site class (Soil)

4/5

1

4/3

4/2

4

0.2666

Percent slope

3/5

3/4

1

3/2

3

0.2000

Landslides

2/5

2/4

2/3

1

2

0.1333

Rock

1/5

1/4

1/3

1/2

1

0.0666

(a)

(b) Theme

SR

PF

PGA

Geology

Weights

SR

1

4/3

4/2

4

0.4

PF

3/4

1

3/2

3

0.3

PGA

2/4

2/3

1

2

0.2

Geohazard

1/4

1/3

1/2

1

0.1

Table 4 Ranking table for geomorphological attributes

Theme

Weight

Rating

Normalized rating

Geology (GE)

0.3333

1

0.0000

2

0.2500

3

0.5000

4

0.7500

5

1.0000

1

0.0000

2

0.3333

3

0.6667

4

1.0000

1

0.0000

2

0.2000

3

0.4000

4

0.6000

5

0.8000

Soil/Site class (SO)

Percent Slope (SL)

0.2666

0.2000

6

1.0000 0.0000

Landslide (LS)

0.1333

1 2

1.0000

RO

0.0666

1

0.0000

2

1.0000

and the minimum 0.15. Quantitatively, we termed this classification ‘Low’ with hazard index less than 0.2, ‘Moderate’ between 0.2 and 0.25, ‘Moderately High’ between 0.25 and 0.35, ‘High’ between 0.35 and 0.60, ‘Very High’ between 0.65 and 0.75 and ‘Severe’ greater than 0.75 encompassing part of Gangtok, Singtam and touching

123

Nat Hazards (2008) 45:333–377

371

Fig. 20 Geohazard map of the Sikkim Himalaya

Jorthang, as well. Following BIS nomenclature for the seismic zonation map of India (BIS 2002) with PGA being the major indicator, we could classify the final earthquake hazard map into BIS Zone II with PGA less than 0.1 g having predominantly ‘Low’ hazard level, Zone III with PGA 0.2 g having predominantly ‘Moderate’ hazard level, Zone IV with PGA between 0.2 to 0.25 g encompassing both the ‘Moderate’ and ‘Moderately High’ hazard levels, BIS Zone V(A) with PGA between 0.4 to 1.0 g predominantly ‘High’ hazard level and BIS Zone V(B) and V(C) (New classification) covering both ‘Very High’ and ‘Severe’ earthquake hazard level in the hazard zonation map of Sikkim, while BIS (2002) puts the state in the Seismic Zone IV. Figures 20–23 essentially prove the progressive increase in hazard level due to incorporation of seismic themes in the integration process at successive levels.

5 Discussion and conclusions Sikkim Himalaya is seismically very active. All the 80 events recorded by the strong motion network also clustered in the same source zone (Fig. 2b). The well constrained hypocentral depths in the eastern and southern Himalaya exhibit a clustering within a range of 10–25 km. An analysis based on ISC catalogue and GCMT database provided region specific correlation relationships which are employed to derive a homogenous catalogue in MW. Based on the spatial b- and Dc-value distributions as depicted in Fig. 3b in conjunction with the underlying tectonic domain, the region is classified into

123

372

Nat Hazards (2008) 45:333–377

Fig. 21 Hazard zonation map as obtained on integration of geohazard and PGA vector layers

four seismic source zones. The maximum earthquake estimated with linear extrapolation of GR law using the subcatalogue for the period spanning from 1964 to 2006 is found to be 8.298(:0.87), while that estimated using a larger duration catalogue spanning from 1924 to 2006, and is found to be 8.34(:0.238). These results, therefore, put the Maximum Earthquake for the seismic scenario generation to a rounded off figure of MW 8.3. The hazard delineation and determination of the related risk cannot be fruitfully undertaken for macro-regions. This calls for a multi-disciplinary effort on the part of scientists and engineers to create a seismic hazard map through microzonation by incorporating a variety of factors including geology, topography, sub-soil condition, building morphology, earthquake ground motion amplification, etc. The rank and weight of the attributes were determined through the matrix operation of multi-criteria evaluation and AHP methods. Spatial operations like vector overlaying and integration for the various geomorphological and seismological coverage was performed in a GIS environment. The integration of litho-unit, soil site class, slope coverage, landslide, and RO provided the site-condition of the Sikkim region represented as a geohazard distribution (Fig. 20) on which the seismological attributes are overlaid sequentially. The SR in the region exhibited azimuthal dependency for a variation more than 40°, as shown in Fig. 10. The site response zonation map presented in Fig. 12a

123

Nat Hazards (2008) 45:333–377

373

Fig. 22 Hazard zonation map as obtained through the integration of geohazard, PGA and PF vector layers

exhibits high amplification at Mangan and its surrounding region, as well as at Jorthang. The state capital Gangtok and the commercial city Singtam experience intermediate ground motion amplification at PF. At those locations, on average, the south Sikkim exhibit higher SR compared to the higher Himalayan crystalline to the north. The PF, however, reverses the trend with a high to the tune of 17 Hz in the northern Sikkim covering HHC, intermediate in between the higher Himalaya and lesser Himalaya. The lesser Himalaya that actually provides human habitation with significant urbanization is seen to possess lower PF even less than 4 Hz.The strong ground motion is synthesized using frequency-wavenumber integration (Green’s function). On comparison of the simulated accelerogram for magnitude greater than 5 in both the time and frequency domains with those recorded by SSMA, the methodology adopted have been found to be robust. Thereafter, PGA has been estimated using RVT. PGA distribution in the Sikkim region presents a seismic scenario for an SEM of MW 8.3. The PGA map presented in Fig. 19 exhibits strong correlation with PF distribution in the region with a high of 2.15 g at Jorthang. Comparing the spatial PGA distribution with seismic zonation map of India, we could classify the entire region into five seismic hazard zones, the lowest one resembling BIS zone II, moderate one as BIS zone III and the moderately high-hazard region resembling BIS zone IV in PGA nomenclature. Further reclassification had been possible drawing

123

374

Nat Hazards (2008) 45:333–377

Table 5 Ranking table for seismological and geohazard attributes Theme

Weight

SR

0.4

PF in Hz

123

0.3

Value

Rating

Normalized rating

\1.4

1

0.0000

1.4–1.6

2

0.0435

1.6–1.8

3

0.0870

1.8–2.0

4

0.1304

2.0–2.2

5

0.1739

2.2–2.4

6

0.2174

2.4–2.6

7

0.2609

2.6–2.8

8

0.3043

2.8–3.0

9

0.3478

3.0–3.2

10

0.3913

3.2–3.4

11

0.4348

3.4–3.6

12

0.4783

3.6–3.8

13

0.5217

3.8–4.0

14

0.5652

4.0–4.2

15

0.6087

4.2–4.4

16

0.6522

4.4–4.6

17

0.6957

4.6–4.8

18

0.7391

4.8–5.0

19

0.7826

5.0–5.2

20

0.8261

5.2–5.4

21

0.8696

5.4–5.6

22

0.9130

5.6–5.8

23

0.9565

[5.8

24

1.0000

\4

1

0.0000

4.0–5.0

2

0.0769

5.6–6.0

3

0.1538

6.0–7.0

4

0.2307

7.0–8.0

5

0.3076

8.0–9.0

6

0.3846

10.0–11.0

7

0.4615

11.0–12.0

8

0.5384

12.0–13.0

9

0.6153

13.0–14.0

10

0.6923

14.0–15.0

11

0.7692

15.0–16.0

12

0.8461

16.0–17.0

13

0.9230

[17

14

1.0000

Nat Hazards (2008) 45:333–377

375

Table 5 continued Theme

Weight

Simulated PGA in g

0.2

Geohazard

0.1

Value

Rating

Normalized rating

\0.2

1

0.0000

0.20–.25

2

0.2000

0.25–0.40

3

0.4000

0.40–1.0

4

0.6000

1.0–2.0

5

0.8000

[2

6

1.0000

\0.3

1

0.0000

0.35–0.55

2

0.5000

[0.55

3

1.0000

Fig. 23 The earthquake hazard microzonation map of the Sikkim Himalaya

analogy of high, very high, and severe hazard indices with BIS PGA codal provision of zone V with a broader PGA range of 0.4–1.0 g, as zone V(A), while zones V(B) and V(C) put together exceeds even 1.0 g. Successive integration of seismic attributes with geohazard layer (Fig. 20) provided earthquake hazard maps as presented in Figs. 21–23 exhibiting enhancement of hazard from one level to the other after introduction of successive seismic layers viz. PGA, PF and SR. Acknowledgements The Department of Science and Technology, Seismology Division, Government of India supported this investigation vide sanction number DST/23(97)/ESS/95, DST/23(218)/ESS/98 and DST/23(574)/SU/2005. The continuous support received from the State Council of Science and Technology, Government of Sikkim in maintaining the Sikkim Strong Motion Array of IIT Kharagpur is gratefully

123

376

Nat Hazards (2008) 45:333–377

acknowledged. The critical and in-depth review of the manuscript and the constructive suggestions by Christos A. Papaioannou, the anonymous referee and the advisor greatly helped in improving the manuscript with enhanced scientific clarity and exposition.

References Aki K (1965) Maximum likelihood estimate of b in the formula log N = a @ bM and its confidence limits. Bull Earthq Res Inst 43:237–239 Bak P, Tang C (1989) Earthquakes as a self-organized critical phenomenon. J Geophys Res 94:15635– 15637 Bender B (1983) Maximum likelihood estimation of b-values for magnitude grouped data. Bull Seismol Soc Am 73:831–851 Ben-Menahem A, Aboodi E, Schild R (1974) The source of the great Assam earthquake, an interplate wedge motion. Phys Earth Planet Int 9:265–289 Bilham R, Ambraseys N (2005) Apparent Himalayan slip deficit from the summation of seismic moments for Himalayan earthquakes. Curr Sci 88:1658–1663 BIS (2002) IS 1893–2002 (Part 1): Indian Standard Criteria for earthquake resistant design of structures, Part 1—General Provisions and Buildings, Bureau of Indian Standards, New Delhi Boore DM (1983) Stochastic simulation of high frequency ground motions based on seismological models of the radiated spectra. Bull Seismol Soc Am 73:1865–1894 Borcherdt RD (1970) Effects of local geology on ground motion near San Francisco Bay. Bull Seismol Soc Am 60:29–61 Braunmiller J, Deichmann N, Giardini D, Wiemer S (2005) Homogeneous moment-magnitude calibration in Switzerland. Bull Seismol Soc Am 95:58–74 Brune J, Engen G (1969) Excitation of mantle love waves and definition of mantle wave magnitude. Bull Seismol Soc Am 57:1355–1365 Cartwright DE, Longuet-Higgins MS (1956) The statistical distribution of the maxima of a random function. Proc R Soc Lond Ser 237:212–223 Chen W, Molnar P (1977) Seismic moments of major earthquakes and average rate of slip in Central Asia. J Geophys Res 82:2945–2969 Dasgupta S, Pande P, Ganguly D, Iqbal Z et al (2000) Seismotectonic atlas of India and its environs. Geol Survey, India Deichmann N (2006) Local magnitude, a moment revisited. Bull Seismol Soc Am 96:1267–1277 Field EH, Jacob KH (1995) A comparison and test of various site response estimation techniques, including three that are not reference site dependent. Bull Seismol Soc Am 85:1127–1143 Field EH, Jacob KH, Hough SE (1992) Earthquake site response estimation: a weak motion case study. Bull Seismol Soc Am 82:2283–2307 Gansser A (1964) Geology of the Himalayas. Interscience Publ, New York, p 289 Grassberger P, Procaccia I (1983) Measuring the strangeness of strange attractors. Physica D 9:189–208 Gutenberg GB, Richter CF (1944) Frequency of earthquakes in California. Bull Seismol Soc Am 34: 185–188 Hanks TC, Kanamori H (1979) A moment magnitude scale. J Geophys Res 84:2348–2350 Haskell NA (1963) Radiation pattern of Rayleigh waves from a fault of arbitrary dip and direction of motion in a homogeneous medium. Bull Seismol Soc Am 53:619–642 Haskell NA (1964) Radiation pattern of surface waves from point sources in a multilayer medium. Bull Seismol Soc Am 54:377–393 Heaton TH, Tajima F, Mori AW (1986) Estimating ground motion using recorded accelerograms. Surv Geophys 8:25–83 Helmberger DV (1968) The crust-mantle transition in the Bering sea. Bull Seismol Soc Am 58:179–214 Herrmann RB (1979) SH wave generation by disclocation sources-a numerical studies. Bull Seismol Soc Am 69:1–15 Herrmann RB, Mandal B (1986) A study of wavenumber integration techniques. Earthq Notes 57:33–40 Hohl P (1998) GIS Data Conversion: strategies, techniques, and management. OnWord Press, Albany, New York, p 423 Hudson JA (1969) A quantitative evaluation of seismic signals at teleseismic distances—II Body waves and surface waves from and extended source. Geophys J 18:353–370 ISC (2007) On-line Bulletin, http://www.isc.ac.uk/search/index.html, International Seismological Centre, Thatcham, United Kingdom

123

Nat Hazards (2008) 45:333–377

377

Kagan YY (1993) Statistics of characteristics earthquakes. Bull Seismol Soc Am 83:7–24 Kagan YY (2007) Earthquake spatial distribution: the correlation dimension. Geophys J Int 168(3): 1175–1194 Kagan YY, Knopoff L (1980) Spatial distribution of earthquakes: the two point correlation function. Geophys J R Astron Soc 62:303–320 Kanamori H (1977) The energy release in Great Earthquakes. J Geophy Res 82:2981–2987 Koravos GCh, Main IG, Tsapanos TM, Musson RMW (2003) Maximum earthquake magnitudes in the Aegean area constrained by tectonic moment release rates. Geophys J Int 152:94–112 Korte GB (1997) The GIS Book: understanding the value and implementation of Geographic Information Systems. OnWord Press, Albany, New York, p 440 Langston CA (1979) Structure under Mount Rainier, Washington, inferred from teleseismic body waves. J Geophys Res 84:4749–4762 Lave´ J, Yule D, Sapkota S, Basant K, Madden C, Attal M, Pandey R (2005) Evidence for a Great Medieval Earthquake (*1100 AD) in the Central Himalayas, Nepal. Science 307:1302–1305 Lermo J, Cha´vez-Garcı´a FJ (1993) Site effect evaluation using spectral ratios with only one station. Bull Seismol Soc Am 83(5):1574–1594 Main IG (1995) Earthquakes as critical phenomena: implications for probabilistic Seismic Hazard Analysis. Bull Seismol Soc Am 85:1299–1308 Main IG, Irving D, Musson R, Reading A (1999) Constraints on the frequency-magnitude relation and maximum magnitudes in the UK from observed seismicity and glacio-isostatic recovery rates. Geophys J Int 137:535–550 Marble DF, Pequet DJ (1983) Geographic information system and remote sensing, in Manual of Remote Sensing. American Society of Photogrammetry, Falls Church, VA McHarg I (1969) Design with nature. National History Press, New York Nakamura Y (1989) A method for dynamic characteristics estimations of subsurface using microtremors on the ground surface. Q Rep Rail Tech Res Inst 30:25–33 Narenberg MAA, Essex C (1990) Correlation dimension and systematic geometric effects. Phys Rev A 42:7065–7074 Nath SK, Sengupta P, Sengupta S, Chakrabarti A (2000) Site response estimation using Strong Motion network: a step towards microzonation of Sikkim Himalayas, Seismology 2000. Curr Sci 79:1316–1326 Nath SK, Biswas NN, Dravinski M, Papageorgiou A (2002a) Determination of S-wave site response in Anchorage, Alaska in the 1–9 Hz frequency band. Pure Appl Geophys 159:2673–2698 Nath SK, Sengupta P, Kayal JR (2002b) Determination of site response at Garhwal Himalaya from the aftershock sequence of 1999 Chamoli Earthquake. Bull Seismol Soc Am 92:1071–1081 Nath SK, Vyas M, Pal I, Sengupta P (2005) A Hazard Scenario in the Sikkim Himalaya from seismotectonics, spectral amplification, source parameterization and spectral attenuation laws using Strong Motion Seismometry. J Geophys Res 110:B01301, doi:10.1029/2004JB003199 National Bureau of Soil Survey (1994) Master Plan for Irrigation Development, Sikkim, Govt. of Sikkim, Agricultural Finance Corporation Ltd, Report 2:35–47 Ristau J, Rogers GC, Cassidy JF (2005) Moment magnitude–local magnitude calibration for earthquakes in Western Canada. Bull Seismol Soc Am 95:1994–2000 Rong Y (2002) Evaluation of earthquake potential in China. PhD Dissertation, Geophysics and Space Physics, University of California, Los Angeles Saaty TL (1980) The analytic hierarchy process. McGraw-Hill International, New York, NY, USA Saaty TL (1990) An exposition of the AHP in reply to the paper remarks on the analytic hierarchy process. Manage Sci 36:259–268 Scordilis EM (2006) Empirical global relations converting MS and mb to moment magnitude. J Seismol 10:225–236 Seekins LC, Wennerberg L, Margheriti L, Liu HP (1996) Site amplification at five locations in San Francisco, California; a comparison of S-waves, codas, and microtremors. Bull Seismol Soc Am 86:627–635 Utsu T (1965) A method for determining the value of b in the formula log N = a – bM showing the magnitude–frequency relation for the earthquakes. Geophys Bull Hokkaido Univ 13:99–103 Utsu T (1999) Representation and analysis of the earthquake size distribution: a historical review and some new approaches. Pure Appl Geophys 155:509–535 Vanmarcke EH, Lai SP (1980) Strong motion duration of earthquakes. Bull Seismol Soc Am 70:1293–1307 Wang CY, Herrmann RB (1980) A numerical study of P-, SV-, and SH-wave generation in a plane layered medium. Bull Seismol Soc Am 70:1015–1036 Wiemer S, Wyss M (2000) Minimum magnitude of completeness in earthquake catalogs: examples from Alaska, the western US and Japan. Bull Seismol Soc Am 90:859–869

123