Arab J Geosci DOI 10.1007/s12517-012-0749-5

ORIGINAL PAPER

Probabilities of earthquake occurrences in Mainland Southeast Asia Santi Pailoplee & Montri Choowong

Received: 7 February 2012 / Accepted: 31 October 2012 # Saudi Society for Geosciences 2012

Abstract The frequency–magnitude distributions of earthquakes are used in this study to estimate the earthquake hazard parameters for individual earthquake source zones within the Mainland Southeast Asia. For this purpose, 13 earthquake source zones are newly defined based on the most recent geological, tectonic, and seismicity data. A homogeneous and complete seismicity database covering the period from 1964 to 2010 is prepared for this region and then used for the estimation of the constants, a and b, of the frequency–magnitude distributions. These constants are then applied to evaluate the most probable largest magnitude, the mean return period, and the probability of earthquake of different magnitudes in different time spans. The results clearly show that zones A, B, and E have the high probability for the earthquake occurrence comparing with the other seismic zones. All seismic source zones have 100 % probability that the earthquake with magnitude ≤6.0 generates in the next 25 years. For the Sagaing Fault Zone (zones C), the next Mw 7.2–7.5 earthquake may generate in this zone within the next two decades and should be aware of the prospective Mw 8.0 earthquake. Meanwhile, in Sumatra-Andaman Interplate (zone A), an earthquake with a magnitude of Mw 9.0 can possibly occur in every 50 years. Since an earthquake of magnitude Mw 9.0 was recorded in this region in 2004, there is a possibility of another Mw 9.0 earthquake within the next 50 years. Keywords Earthquake probability . Earthquake catalogue . Frequency–magnitude distribution . Return periods . Mainland Southeast Asia S. Pailoplee (*) : M. Choowong Earthquake and Tectonic Geology Research Unit (EATGRU), Department of Geology, Faculty of Science, Chulalongkorn University, Bangkok 10330, Thailand e-mail: [email protected]

Introduction Earthquake activities in the Mainland Southeast Asia (MSEA) region are relatively high due to its location being close to the active tectonic plate boundary. A large number of earthquakes have been continuously generated in the past and have strongly affected the nearby countries. Beside the world’s largest instrumentally recorded earthquake with moment magnitude (Mw) 9.5 at Chile in A.D. 1960 (Kanamori 1977), the second largest earthquake with Mw 9.0 is the Sumatra-Andaman earthquake on December 26, 2004 (Fig. 1), providing the obvious evidence to remind the human communities about the earthquake potential and activity that should not be ignored. Practically, the mechanism of any earthquake is complex which limits the ability to effectively solve it by manual or simplified experiments. As a result, various mathematic or statistic concepts are currently being developed and employed in order to clarify the nature of earthquake occurrences in each individual earthquake source of interest, such as artificial neural networks (Bodri 2001; Alves 2006) and fractal dimensions (Maryanto and Mulyana 2008). In this study, the concept of frequency–magnitude distributions (FMDs) is applied in order to investigate the probability of occurrence of earthquakes in each defined area within the MSEA region. The results are mainly aimed at informing and arousing the public’s recognition of the earthquake potential in the MSEA’s countries and are useful for seismic hazard assessment of any area within or close to the MSEA region.

Seismotectonics and earthquake zonation The present-day seismotectonics in the MSEA region is a consequence of the interaction between the Indo-Australian

Arab J Geosci

and Eurasian plates (Polachan et al. 1991; Fenton et al. 2003). Persistent deformation within the Eurasian plate is illustrated by a large number of seismogenic faults and earthquake activities in this region (Molnar and Deng 1984) (Fig. 1). Based on the literature review, there are a few studies that propose the earthquake source zones in the MSEA region. The pioneering work of Nutalaya et al. (1985) defined 12 seismic source zones in Thailand and neighboring countries. However, they did not recognize some important seismic source areas, particularly those along the southern peninsular Thailand and the Sumatra Island. Thereafter, Charusiri et al. (2005) revised the seismic source zones of Nutalaya et al. (1985) extending them southwards into the areas of seismic sources in the southern peninsular Thailand, including northern Sumatra. Their revision of the seismic zoning was based on the epicentral distribution of earthquakes over the past two decades, present-day tectonic environments, active faults, and regional geomorphology. As a direct result, the number of seismic source zones increased to 21. However,

a

most seismic source zones proposed by the previous works are bounded by the boundary of the study area which is not representing correctly the geometries of the zones (Pailoplee et al. 2010). According to Kramer (1996), the exact locations and shapes of the earthquake sources are significant for seismic hazard analysis in terms of estimating the source-tosite distances and considering the earthquake events for earthquake potential investigation. This study, therefore, newly revised the earthquake source zones in a manner which is more reasonable and reliable for analyzing the probability of earthquake occurrences. The re-grouping and re-locating of the seismic source zones that had been proposed previously are accordingly clarified in this paper. Accordingly, from this detailed investigation, the MSEA region can be divided seismotectonically into 13 main seismic source zones: zone A: Sumatra-Andaman Interplate; zone B: Sumatra-Andaman Intraslab; zone C: Sagaing Fault Zone; zone D: Andaman Basin; zone E: Sumatra Fault Zone; zone F: Hsenwi-Nanting Fault Zone; zone G: Western Thailand; zone H: Southern Thailand; zone I: Jinghong-

b F7 L

China 25oN

F

F3

25oN

F6 7.4(1970)

I

8.0(1912) M

Myanmar

F1

Laos J

F4 6.5(1925)

K

F5 5.6(1975)

A G

C

Thailand

5.9(1983)

15oN

15oN

Cambodia B

5.0(2006)

Vietnam D

H

8.6(2005)

Malaysia

E

Su m at

ra

300 km

95oE

5o N

5o N 9.0(2004) 8.6(2012)

Is la nd

10 5 o E

Fig. 1 Map of the MSEA region showing a the 13 earthquake source zones, bounded by the black line, that are proposed newly in this study. The gray circles illustrate the earthquake epicentral distribution recorded during the period 1964 to 2010 for all magnitude ranges after

8.2(2012) 300 km 95oE

F2 105 o E

declustering with the algorithm according to Gardner and Knopoff (1974). b The significant fault zones (black lines) and earthquake events (black circles) mentioned in the text

Arab J Geosci

Mengxing Fault Zones; zone J: Northern Thailand-Dein Bein Fhu; zone K: Song Da-Song Ma Fault Zones; zone L: Xianshuihe Fault Zone; and Zone M: Red River Fault Zone (see Fig. 1 and Table 1 for details). According to Reading (1980), the MSEA is part of the Eurasian plate whose boundary is marked by the active eastdipping Sumatra-Andaman Subduction Zone (SASZ). The SASZ extends from northeastern India, passing southwards to western Myanmar, west of Nicobar Islands, and swinging southeastwards along the Sumatra Island, respectively. According to the estimated 43 mm/year rate of plate slip (Paul et al. 2001), the earthquake activities along this Sumatra-Andaman Interplate (zone A) are prominently strong, as demonstrated by the recent Mw 9.0 earthquake mentioned above. Beside the interplate tectonic activities, the east dipping of an intraslab also generated earthquakes beneath the area of western Myanmar, Nicobar, and Sumatra Islands. Based on three-dimensional investigation of the earthquake focuses, the earthquake events along the SASZ with the focal depths beyond 35 km are defined as the zone B: SumatraAndaman Intraslab. Moving eastwards of the SASZ, the Sagaing Fault Zone (zone C), striking in N–S direction along the central part of Myanmar (F1 in Fig. 1), is one of the most active seismic zones in the MSEA. This strike-slip fault has experienced several dozen large earthquakes, which have caused the loss of human lives and destroyed buildings. The highest slip rate occupied by this fault is around 23 mm/year (Bertrand and Rangin 2003), and the largest recorded earthquake, Mw8.0, (Fig. 1) occurred on May 23, 1912 (Brown and Leicester 1933). In addition to this great inland earthquake, at least ten events, Mw 7.0 to 7.6, have occurred in A.D. 1839, 1858, 1929, 1930, 1930, 1930, 1931, 1946, 1956, and 1991 (Brown and Leicester 1933; Tun 2008). Table 1 Calculated values of a, b, α, and β from the FMDs, including the most probable annual maxima (u), in the seismic source zones A–M

Mc is the magnitude of completeness

For the Andaman Basin (zone D), which is identified as the backarc region of the SASZ, the E–W rifting process is still active at present (Rajendran et al. 2003; Khana and Chakraborty 2005). As a result, present-day earthquake records from this area show that several small- to mediumscale earthquakes are commonly associated with this active zone. We therefore regard this region as an earthquake source in this study. Meanwhile above the southern part of Sumatra-Andaman Intraslab, the Sumatra Fault Zone (F2 in Fig. 1) generates a large number of shallow crustal earthquakes due to the compression force of the plate collision. This fault zone is, therefore, recognized separately to be zone E. Along the Thailand-Myanmar border (Fig. 1), some fault zones are proposed previously such as Pa Pun and Pan Luang (Nutalaya et al. 1985), Moei-Tongyi (Pailoplee et al. 2009), Mae Hong Sorn-Tak (Charusiri et al. 2004), Sri Sawat (Nuttee et al. 2005), and Three Pagoda (Rhodes et al. 2005) Fault Zones. Most of them are the strike-slip faults in the sense of movement which defined as zone G in this study. Some earthquakes occurred in this zone with a bodywave magnitude (mb) of 5.6 on February 17, 1975 (Le Dain et al. 1984) and mb of 5.9 on April 22, 1983 (Klaipongphan et al. 1991), respectively (Fig. 1). In northern Thailand, some active faults are also clarified such as the Lampang–Thoen (Charusiri et al. 2004), Phrae (Udchachon et al. 2005), and Mae Tha Fault Zones (Rhodes et al. 2004). These fault zones are associated with the intraplate basins in northern Thailand (zone J). In addition, some NE–SW-trending strike-slip faults, such as the Mae Chan and Dien Bien Fu Fault Zones, are cited by Fenton et al. (2003) and Zuchiewicz et al. (2004). The earthquakes in this zone are characterized by shallow focal depth as well as intermediate seismicity, and the seismic productivity is not so high when compared to the other zones. However, on December 22, 1925, this area experienced an earthquake up

Zone

Name

a

b

Mc

α

β

u

A B C D E F G H I J K L M

Sumatra-Andaman Interplate Sumatra-Andaman Intraslab Sagaing Fault Zone Andaman Basin Sumatra Fault Zone Hsenwi-Nanting Fault Zone Western Thailand Southern Thailand Jinghong-Mengxing Fault Zones Northern Thailand-Dein Bein Fhu Song Da-Song Ma Fault Zones Xianshuihe Fault Zone Red River Fault Zone

4.06 4.01 4.05 1.80 4.25 3.57 2.85 3.10 3.28 4.32 3.48 3.58 4.32

0.63 0.68 0.72 0.39 0.75 0.70 0.57 0.66 0.61 0.80 0.74 0.61 0.84

5.6 5.3 5.7 4.5 5.7 5.4 4.6 5 4.8 5.4 4.9 4.6 4.8

11,481.5 10,232.9 11,220.2 63.1 17,782.8 3,715.4 707.9 1,258.9 1,905.5 20,893.0 3,020.0 3,801.9 20,893.0

1.5 1.6 1.7 0.9 1.7 1.6 1.3 1.5 1.4 1.8 1.7 1.4 1.9

6.4 5.9 5.6 4.6 5.7 5.1 5.0 4.7 5.4 5.4 4.7 5.9 5.1

Arab J Geosci

to mb 6.5 in the vicinity of the Thailand–Laos border (Bott et al. 1997) (Fig. 1). For the northern part of the MSEA region, where northern Laos, northern Myanmar, northern Vietnam, and southern China are located, the Indo-Australian and Eurasian collision generates a wide area of NW–SE and NE–SW complex shear zones and strike-slip faults (Polachan et al. 1991). In this region, two earthquake source zones are proposed, which consist of zone F: Hsenwi-Nanting Fault Zone and zone I: Jinghong-Mengxing Fault Zone (Lacassin et al. 1998) (F3 and F4 in Fig. 1). Present-day earthquake records show that these complex zones have generated contemporaneous seismic activity continuously since the start of instrumental record keeping. Extending to the eastern part of MSEA (i.e., northern Vietnam), some fault zones have been proposed in this region, such as the Song Ca (Takemoto et al. 2005), Song Chay (Cuong and Zuchiewicz 2001), Song Da, and Song Ma (Phoung 1991) Fault Zones. These fault zones have a NW–SE orientation limited locally in the northern part of Vietnam. The instrumental records from this area show that earthquakes are commonly associated with these fault zones. Thus, these fault zones are classified to be the seismic source zone namely zone K: Song Da-Song Ma Fault Zones (F5 in Fig. 1). Moreover along the China–Vietnam border, the Red River Fault Zone is the longest and most prominent fault in this eastern part of MSEA (Duong and Feigl 1999). Due to this, 810-km-long fault zone has a much different of surface length, representing the much difference of the possible maximum earthquake magnitude, compared with those faults in zone K, the Red River Fault Zone (F6 in Fig. 1) is, therefore, proposed separately to be zone M. In southern Thailand (zone H), the seismicity is mainly related to two obvious NE–SW-striking faults, i.e., the Ranong and Klong Marui Fault Zones. The Ranong Fault Zone has a slip rate of 1 mm/year, which is tenfold higher than the slip rate of the Klong Marui Fault Zone at 0.1 mm/ year (Wong et al. 2005). An Mw 5.0 earthquake occurred on October 8, 2006 (RID 2006), in the vicinity of the northern end of the Ranong Fault Zone and extended into the Gulf of Thailand (Fig. 1). Within the Eurasian plate, and far away from the plate collision boundary, the N–S-trending Xianshuihe Fault Zone (Eleftheria et al. 2004) is situated in southern China (F7 in Fig. 1). The intra-plate shallow seismicity of this region is mainly due to the complex tectonic regime surrounding this zone. Based on the instrumental earthquake records, a lot of earthquakes are measured in this region. Therefore, this area is also regarded as a seismic source zone, herein called zone L. For the eastern and northeastern Thailand, the southernmost part of Laos, southern Vietnam and the whole of Cambodia, the earthquake activity is quiescent with no

Fig. 2 Graphs showing the FMDs for the earthquake source zones„ cited in Table 1. Each triangle indicates the total number of earthquakes for each magnitude; square represents the cumulative number of earthquakes equal to or larger than each magnitude. The solid lines are lines of the best fit. Mc is the magnitude of completeness

reported active faults or earthquake source zones. These areas are, therefore, excluded from this earthquake source investigation.

The applied method The FMDs, which are commonly used to assess the return periods of earthquake occurrences, known in the East as the Ishimoto and Iida (1939) relationship and in the West as the Gutenberg and Richter (1954) relationship, define the number of earthquakes with respect to their magnitude and represent one of the reliable empirical relations in seismology (Eq. 1). log10 ðNm Þ ¼ a  bm or lnðNm Þ ¼ ln a  bm

ð1Þ

For a certain region and time interval, the equation describes the number of events, Nm, with a magnitude equal to or larger (cumulative distribution) than m, where a and b are positive, real constants. The ratio of the occurrence of small to large earthquakes in a seismogenic volume is measured by the b value of the FMDs. The parameters α and β are constants, which are related to a and b by a ¼ expða lnð10ÞÞ and b ¼ b lnð10Þ If one assumes that the earthquake magnitude is unlimited, if the activity rate decreases in accordance with increases in the size of earthquakes, and that the pattern of occurrence of earthquakes is random (Yadav 2009), then the probability of non-exceedance of an earthquake of magnitude m in 1 year, G(m), will be expressed by Eq. 2. GðmÞ ¼ expðaðexpðbmÞÞÞ

ð2Þ

In addition to Eq. 1, several other useful formulas can be obtained from the model expressed by Eq. 2. For example, the most probable or most frequently observed annual maximum magnitude (u) can be estimated in terms of α and β as shown in Eq. 3. u¼

ln a b

ð3Þ

The most probable earthquake magnitude in t year periods (ut) can be presented by Eq. 4. ut ¼

lnðat Þ lnðtÞ ¼uþ b b

ð4Þ

In addition, the mean return period (i.e., mean interval in years) between earthquakes (Tm) is defined as the expected time interval for the occurrence of an

Arab J Geosci

A: Sumatra-Andaman Inter Plate

B: Sumatra-Andaman Intra Slab

C: Sagiang Fault Zone

D: Andaman Basin

E: Sumatra Fault Zone

F: Hsenwi-Nanting Fault Zone

G: Western Thailand

H: Southern Thailand

I: Jinghong-Mengxing Fault Zones

J: Northern Thailand-Dein Bein Fhu

K: Song Da-Song Ma Fault Zones

L: Xianshuihe Fault Zone

Arab J Geosci

M: Red River Fault Zone

Mw, mb, surface-wave magnitude (Ms), and local magnitude (ML), respectively. Empirically, any earthquake catalogue is the result of seismological signals recorded on complex, spatially, and temporally heterogeneous networks of seismometers, and processed using not only a variety of software and assumptions but also judgments (Habermann 1991; Habermann and Creamer 1994; Zuniga and Wiemer 1999). As a result, the critical issue to be addressed before seismicity analysis can be reliably performed is to assess the quantity, quality, and consistency of the obtained earthquake data. Consequently, the unified earthquake catalogue for this study is needed (Caceres and Kulhanek 2000). Earthquake magnitude conversion

Fig. 2 (continued)

earthquake with magnitude greater than or equal to m and is given by Eq. 5. Tm ¼

1 expðbmÞ ¼ Nm a

ð5Þ

Finally, the occurrence probability of earthquakes, Pt(m), the most important result of this theory, can be estimated for a particular magnitude, m, in t year period by Eq. 6. Pt ðmÞ ¼ 1  expðat: expðbmÞÞ

ð6Þ

The main objective in applying the FMDs concepts, as expressed by Eq. 1, in this study is to estimate the statistics of earthquakes for the MSEA region. The earthquake hazard is calculated in terms of the most probable largest magnitude of earthquakes in a given time period (ut), the expected time interval for the occurrence of an earthquake (Tm), and the occurrence probability of earthquakes in a specified time interval (Pt(m)), and so on. For this purpose, Eqs. 3–6 above are used for the estimation of the hazard parameters for the 13 different major seismic source zones (Fig. 1).

Homogeneous and complete seismicity database In this study, the earthquake database is obtained from the Incorporated Research Institutions for Seismology by downloading earthquake records in the study area within the latitude 1°S–30°N and the longitude 89°E–113°E. The time span of the earthquake records covers the period from 1964 to 2010 with all recorded magnitude ranges. This study area includes the SASZ, and the focal depth of the obtained earthquakes is down to 300 km in depth. The data are reported variously in the magnitude scales composed of

It is noted that each magnitude scale is derived from a specific assumption and analytical method that have a valid but different value and unique meaning. Therefore, the conversion of the different magnitude scales to only one standard, viz., the Mw scale, is inevitably important for the qualitative improvement of the composite earthquake catalogue. The magnitude conversions are performed according to the regression relationship between Mw to mb and Ms, as proposed by Pailoplee et al. (2009), and shown in Eqs. 7–8. For ML, we used the empirical relationship between ML and mb of Palasri (2006) (Eq. 9) and then converted mb to Mw using our Eq. 7. Mw ¼ 0:0167mb2 þ 0:8438mb þ 0:9071; mb  6:8

ð7Þ

Mw ¼ 0:028Ms2 þ 0:3364Ms þ 3:2574; Ms  7:6

ð8Þ

mb ¼ 1:64 þ 0:63ML; ML  6:8

ð9Þ

Earthquake de-clustering In nature, when any cluster of earthquakes occurs, the earthquake can be classified temporally into three types: foreshock, main shock, and aftershock. In quantitatively analyzing the earthquake occurrence, only the main shock, which represents the exact seismic stress released from the tectonic activities, is herein considered. To satisfy this requirement, the earthquake data obtained from the previous procedures need to be de-clustered by filtering main shocks from foreshocks and aftershocks. To the end, the assumptions of Gardner and Knopoff (1974) are applied in this study. Using the Gardner and Knopoff (1974) algorithm, 1,933 clusters from 102,693 earthquake events are distinguished.

Arab J Geosci

A: Sumatra-Andaman Inter Plate

B: Sumatra-Andaman Intra Slab

C: Sagiang Fault Zone

D: Andaman Basin

E: Sumatra Fault Zone

F: Hsenwi-Nanting Fault Zone

G: Western Thailand

H: Southern Thailand

Fig.3 Cumulative probability distribution curves of non-exceedance (G(m)) and exceedance (1−G(m)) for seismic source zones A–M

Arab J Geosci

I: Jinghong-Mengxing Fault Zones

J: Northern Thailand-Dein Bein Fhu

K: Song Da-Song Ma Fault Zones

L: Xianshuihe Fault Zone

M: Red River Fault Zone

Fig. 3 (continued)

Of these events, a total of 99,764 events (97 %) are classified as foreshocks or aftershocks and, therefore, are eliminated. Finally, the quantitative earthquake catalogue, derived from MSEA region contains 2,929 main shocks (Fig. 1). This meaningful catalogue which can be related directly to tectonic activities is used here to evaluate the earthquake source potential, as described in the next section.

Earthquake potential determination For each seismic source zone, we estimated the a and b values of the FMDs by observing a plot of the number of earthquake events, Nm, with a magnitude equal to or larger than m, as written in Eq. 1 (see also Fig. 2). The optimal values of a and b are calculated to yield the least square observed from the FMDs by using the ZMAP software

Arab J Geosci Table 2 Most probable largest earthquake magnitude (Mw) for different time periods in each of the 13 zones (A–M) Time (years)

Zone A

Zone B

Zone C

Zone D

Zone E

Zone F

Zone G

Zone H

Zone I

Zone J

Zone K

Zone L

Zone M

1 5 10 15 25 50

6.4 7.6 8.0 8.3 8.7 9.1

5.9 6.9 7.4 7.6 8.0 8.4

5.6 6.6 7.0 7.2 7.5 8.0

4.6 6.4 7.2 7.6 8.2 8.9

5.7 6.6 7.0 7.3 7.6 8.0

5.1 6.1 6.5 6.8 7.1 7.5

5.0 6.3 6.8 7.1 7.5 8.0

4.7 5.8 6.3 6.5 6.9 7.3

5.4 6.5 7.0 7.3 7.7 8.2

5.4 6.3 6.7 6.9 7.2 7.6

4.7 5.7 6.1 6.3 6.6 7.0

5.9 7.0 7.5 7.8 8.2 8.7

5.1 6.0 6.3 6.5 6.8 7.2

(Woessner and Wiemer 2005). The magnitude of completeness (Mc) is defined as the magnitude above which all earthquakes are considered to be fully reported (Fig. 2). The results of constant values of a and b are listed in Table 1 for the 13 seismic source zones defined here. Generally, the statistical methods need a large number of observations to estimate recurrence periods. In this analysis, 50-year data (i.e., A.D. 1964–2010) have been taken to estimate the seismic hazard parameters. Since the values of the model parameters a and b, which are used to estimate the probability of occurrences and return periods, do not significantly affect the short or long duration of the seismicity data (Shanker and Singh 1997), this study may be considered to be consistent and estimated return periods and probabilities may be used as a quantitative measure of seismicity, as shown below.

Results and discussion Based on the FMDs investigation (Fig. 2), the obtained values of a and b listed in Table 1 are then converted to the values of α and β, as described above. These parameters

(i.e., α and β) are useful to estimate the most probable largest magnitudes, the mean return periods, and probabilities of occurrences of different magnitudes in different time periods in this region. At first, from the obtained α and β values, the cumulative probability distribution curves for non-exceedance, G(m), and exceedance, 1−G(m), are estimated according to Eq. 2 and are presented in Fig. 3 for each of the 13 different seismic zones. For zone A: Sumatra–Andaman Interplate (Fig. 3a), the cumulative probability distribution for nonexceedance,G(m), of earthquakes with a magnitude of ≤6.4 is ≤50 %, whereas the cumulative probability distribution for exceedance, (1−G(m)), of the same magnitude is ≥50 % in the zone A. The same cumulative probability distributions are estimated for magnitudes 5.9, 5.6, 4.6, 5.7, 5.1, 5.0, 4.7, 5.4, 5.4, 4.7, 5.9, and 5.1 in zones B-M, respectively (Fig. 3). These magnitude values for each zone, where the cumulative probability distributions of non-exceedance and exceedance are 50 %, are closely related to the most probable largest annual magnitudes (u), and are estimated using Eq 3 (Table 1). The most probable largest earthquake magnitudes in different time periods (ut) have also been estimated and are

Table 3 Estimated annual number of earthquakes (Nm) in each of the 13 zones (A–M) Mw

Zone A

Zone B

Zone C

Zone D

Zone E

Zone F

Zone G

Zone H

Zone I

Zone J

Zone K

Zone L

Zone M

3 3.5 4

147.91 71.61 34.67

93.33 42.66 19.50

76.56 33.34 14.52

4.24 2.70 1.72

102.09 43.20 18.28

29.51 13.18 5.89

14.09 7.34 3.82

13.65 6.42 3.02

28.38 14.08 6.98

85.51 34.20 13.68

18.84 8.08 3.47

57.41 28.54 14.19

62.66 23.80 9.04

4.5 5 5.5 6 6.5 7 7.5 8 8.5 9

16.79 8.13 3.94 1.91 0.92 0.45 0.2.2 0.10 0.05 0.02

8.91 4.07 1.86 0.85 0.39 0.18 0.08 0.04 0.02 0.01

6.32 2.75 1.20 0.52 0.23 0.10 0.04 0.02 0.01 0.00

1.10 0.70 0.45 0.28 0.18 0.12 0.07 0.05 0.03 0.02

7.74 3.27 1.39 0.59 0.25 0.10 0.04 0.02 0.01 0.00

2.63 1.17 0.52 0.23 0.10 0.05 0.02 0.01 0.00 0.00

1.99 1.04 0.54 0.28 0.15 0.08 0.04 0.02 0.01 0.01

1.42 0.67 0.31 0.15 0.07 0.03 0.02 0.01 0.00 0.00

3.46 1.72 0.85 0.42 0.21 0.10 0.05 0.03 0.01 0.01

5.47 2.19 0.87 0.35 0.14 0.06 0.02 0.01 0.00 0.00

1.49 0.64 0.27 0.12 0.05 0.02 0.01 0.00 0.00 0.00

7.06 3.51 1.74 0.87 0.43 0.21 0.11 0.05 0.03 0.01

3.43 1.30 0.49 0.19 0.07 0.03 0.01 0.00 0.00 0.00

5.3 yrs l4.Oyrs 36.9 yrs 97.2 yrs 255.9 yrs 673.5 yrs l774.2 yrs 1.2 yrs 2.3 yrs 4.7 yrs 9.4 yrs 18.9 yrs 38.Oyrs 76.4 yrs 5.5 yrs 19.8 yrs 46.2 yrs 107.8 yrs 251.2 yrs 585.5 yrs 1364.6 yrs 2.9 yrs 7.lyrs 17.9 yrs 44.7 yrs ll17yrs 279.3 yrs 698.2 yrs 2.4 yrs 4.8 yrs 9.6 yrs 19.4 yrs 39.lyrs 78.5 yrs 158.9 yrs 6.8 yrs 14.4 yrs 30.5 yrs 64.9 yrs 138.Oyrs 293.4 yrs 623.7 yrs 3.6 yrs 6.8 yrs 13.2 yrs 25.3 yrs 48.5 yrs 93.2 yrs 179.lyrs 43 yrs 9.5 yrs 21.4 yrc 47.9 yrs 107.2 yrs 239.9 yrs 537.Oyrs 1.7 yrs 4.Oyrs 9.5 yrs 22.5 yrs 53.2 yrs 125.7 yrs 297.2 yrs 35 yrs 5.5 yrs 5.6 yrs 13.6 yrs 21.3 yrs 33.4 yrs 52.4 yrs 1.9 yrs 4.4 yrs 10.1 yrs 23lyrs 53.2 yrs 122.2 yrs 280.5 yrs dys days, yrs years

191.6 dys l.lyrs 2.2 yrs 4.6 yrs 9.5 yrs 19.7 yrs 40.7 yrs 6 65 7 7.5 8 8.5 9

1.2 yrs 2.6 yrs 5.6 yrs 12.3 yrs 26.9 yrs 58.9 yrs 128.8 yrs

5.8 dys 15.3 dys 40.4 dys 106.4 dys 280.1 dys 2.Oyrs 6.4 dys 12.8 dys 25.7 dys 51.7 dys 104.1 dys 209.3 dys 19.4 dys 45.2 dys 105.3 dys 245.4 dys 1.6 yrs 3.7 yrs 4.3 dys 10.7 dys 26.7 dys 66.7 dys 166.8 dys 1.lyzs 12.9 dys 25.9 dys 52.3 dys 1O5.4 dys 212.5 dys 1.2 yrs 26.7 dys 56.9 dys 120.9 dys 256.9 dys 1.5 yrs 3.2 yrs 25.9 dys 49.7 dys 96 6 dys 183.6 dys 1.Oyrs 1.9 yrs 12.4 dys 27.7 dys 62.Odys 138.5 dys 310.7 dys 1.9 yrs 3.6 dys 8.4 dys 20.Odys 47.2 dys I11.5 dys 263.5 dys 86.2 dys 135.1 dys 212.Odys 332.5 dys 1.4 yrs 2.2 yrs 4.8 dys 10.9 dys 25.ldys 57.7 dys 132.5 dys 304.3 dys 2.5 dys 5.1 dys 10.5 dys 21.7 dys 449 dys 92.7 dys 3 3.5 4 45 5 5.5

3.9 dys 8.6 dys 18.7 dys 41.Odys 89.6 dys 196.Odys

Zone L Zone K Zone J Zone I Zone H Zone G Zone F Zone E Zone D Zone C Zone B Zone A Mw

Table 4 Estimated return period (Tm) in each of the 13 zones (A–M)

listed in Table 2. For instance, in zone A (the SumatraAndaman Interplate) an earthquake with a magnitude of Mw 9.1 is revealed to occur in every 50 years. Since an Mw 9.0 earthquake was recorded in this region in 2004, then there is a possibility of another Mw 9.0 earthquake within approximately the next 40–50 years. For zone C (Sagaing Fault Zone), some earthquakes, Mw 7.0–7.6, have been reported in the past, as described above. The average return period of these Mw 7.0–7.6 earthquakes is around 17 years. It agrees well with the results obtained in this study which show a mean return period of 15 to 25 years for Mw 7.2–7.5 earthquakes. Therefore, the ut calculated in this study seems to be reasonably reliable. According to the fact that the latest Mw 7.3 earthquake occurred in 2003, the next Mw 7.2–7.5 earthquake may devastate this area within the next two decades. The yearly expected number of earthquakes (Nm) equal to or greater than 3.0 in magnitude and their mean return period (Tm), which are estimated by using Eq. 5, are listed in Tables 3 and 4 for each seismic source zone. In addition, these mean return periods, which are one of the most important hazard parameters, are also plotted clearly in Fig. 4. It is made clear that zones A, B, and E (i.e., the SumatraAndaman Interplate, Intraslab, and the Sumatra Fault Zone) have a higher expected number of earthquakes and a smaller mean return period than the other seismic source zones for particular earthquake magnitude. Thus, zones A, B, and E are more seismically active and vulnerable to future moderate earthquakes in comparison to the other zones. Indeed, zones A is the most seismically active and highly productive zones in the study region, where seismicity is mainly related to the subduction of Indo-Australian plate under the Eurasian plate. The Sagaing Fault Zone and Northern Thailand-Dein Bein Fhu (i.e., zones C and J) are the second largest seismically active zones of the MSEA study region, with return periods of around 50–120 years for an earthquake of magnitude 8.0–8.5. These regions, and in particular the Sagaing Fault Zone, experienced a great earthquake of Mw 8.0 on May 23, 1912 (Brown and Leicester 1933) which caused a huge loss of lives and damaged properties. Thus, it is susceptible to the occurrence of a similarly large earthquake in the near future along the Sagaing Fault Zone. Although the Red River Fault Zone (zone M) has a smaller expected number of earthquakes and larger return periods than the other zones mentioned above, suggesting it is far less vulnerable to future moderate earthquakes. However, this zone experienced a devastating Mw 7.4 earthquake (Fig. 1) on January 4, 1970 which caused a huge loss of lives and properties. Zones L, I, and G (i.e., the Xianshuihe, JinghongMengxing Fault Zones, and Western Thailand) are the third largest seismically active zone in terms of the return periods. The mean return periods for a magnitude 6.0 earthquake are estimated at 1.2–3.6 years in these zones. For the rest, zones

Zone M

Arab J Geosci

Arab J Geosci Fig. 4 Plots of mean return period (Rt) curves for seismic source zones A–M. a Return period 0–100 years, b return period 0–500 years, c return period 0–1,000 years, and d return period 0– 5,000 years

A: Return period 0-1000 years

B: Return period 0-5000 years

C: Return period 0-1,0000 years

D: Return period 0-5,0000 years

F (the Hsenwi-Nanting Fault Zone), H (Southern Thailand), and K (the Song Da-Song Ma Fault Zones), we estimate the mean return periods for an earthquake of magnitude 6.0 to be around 4.3, 6.8, and 8.5 years, respectively. The probabilities, Pt(m), of an earthquake occurrences have been estimated for the 13 individual major earthquake source zones for certain magnitude (m) and time period (t).The earthquake hazard curves, expressed in terms of the expected probability for earthquakes with the maximum observed magnitudes, are plotted for the magnitude and for the time span of 25, 50, and 100 years (Fig. 5). From these hazard curves, it is clearly observed that probability of occurrence of an earthquake with magnitude ≥6.0 in 25-, 50-, and 100-year return periods decreases exponentially with the magnitude in all 13 seismogenic zones. From Fig. 5, it is observed that zones A, B, and D have the highest probability (100 %) for the occurrence of an earthquake of magnitude 7.8–8.4 in the next 100 years,

whereas zones H, K, and M show the lowest probabilities (18–43 %) among the 13 zones.

Conclusions In this study, the earthquake hazard parameters, i.e., probable largest earthquake magnitudes, return periods, and probabilities of earthquake occurrences, are estimated for 13 seismic source zones of the MSEA. The comparative study is established by considering the FMDs relationships based theory of Ishimoto and Iida (1939) and Gutenberg and Richter (1954). For this purpose, a homogeneous and complete earthquake catalogue covering the time period from 1964 to 2010 is prepared according to the procedure of Caceres and Kulhanek (2000). From the analysis, it can be concluded that zones A, B, and E have a high probability for the earthquake occurrence

Arab J Geosci

A: Sumatra-Andaman Inter Plate

B: Sumatra-Andaman Intra Slab

C: Sagiang Fault Zone

D: Andaman Basin

E: Sumatra Fault Zone

F: Hsenwi-Nanting Fault Zone

G: Western Thailand

H: Southern Thailand

I: Jinghong-Mengxing Fault Zones

J: Northern Thailand-Dein Bein Fhu

K: Song Da-Song Ma Fault Zones

L: Xianshuihe Fault Zone

Fig. 5 Probability–magnitude curves for the seismic source zones A–M

Arab J Geosci M: Red River Fault Zone

faculty of science, Chulalongkorn University for a critical review and improved the English. We acknowledged thoughtful comments and suggestions by the Editor-In-Chief and anonymous reviewers which enhanced the quality of this manuscript significantly.

References

Fig. 5 continued.

in comparison to the other zones (Table 3). The largest annual earthquake magnitudes are close to 6.4, 5.9 and 5.7, respectively. Meanwhile for the other zones, the largest annual magnitudes are in the range of 4.0–5.0. In addition, the earthquakes with magnitude ≤6.0 have 100 % of occurrences in the next 25 year for all 13 seismogenic zones (Fig. 5). In particular for zones A, B, and D, there is ≥95 % of probability that the earthquakes with magnitude 7.8–8.4 may generate in the next 100 years. For the return periods of earthquakes, the Sagaing Fault Zone (zones C) has average return period around 15 to 25 years for the earthquake with Mw 7.2–7.5. This conforms to the 17-year return period observed empirically from the past earthquake records. Due to the latest Mw 7.3 earthquake occurred in 2003, the next Mw 7.2–7.5 earthquake may generate in this zone within the next two decades. Furthermore for the earthquake with Mw 8.0–8.5, the calculating return periods for Sagaing Fault Zone are around 50–120 years. Based on the latest Mw 8.0 earthquake in 1912, the area within or nearby the Sagaing Fault Zone should be aware of this prospective giant earthquake. Although these results have potentially useful implications for the probabilistic seismic hazard assessment (Kramer 1996) in the MSEA, some uncertainty should be recognized. According to Youngs and Coppersmith (1985), the numbers of large earthquakes sometime illustrate suddenly decreases (i.e., characteristic earthquake) which are not conforming to the common FMDs applied in this study. Thus, the earthquake parameters obtained from this study should be used with careful recognition in particular for the major or great earthquakes. Acknowledgments This work was jointly sponsored by Integrated Innovation Academic Center: IIAC Chulalongkorn University Centenary Academic Development Project (CU56-CC04), the Higher Education Promotion and National Research University Project of Thailand, Office of the Higher Education Commission (CC508B). Thanks also extend to T. Pailoplee for the preparation of the draft manuscript. We thank the Publication Counseling Unit (PCU), the

Alves EI (2006) Earthquake forecasting using neural networks: results and future work. Nonlinear Dynamics 44:341–349 Bertrand G, Rangin C (2003) Tectonics of the western margin of the Shan Plateau (central Myanmar): implications for the IndiaIndochina oblique convergence since the Oligocene. Journal of Asian Earth Sciences 21:1139–1157 Bodri B (2001) A neural-network model for earthquake occurrence. J Geodyn 32:289–310 Bott J, Wong I, Prachaub S, Wechbunthung B, Hinthong C, Sarapirome S (1997) Contemporary seismicity in northern Thailand and its tectonic implication. Proceedings of the International Conference on Stratigraphy and Tectonic Evolution of Southeast Asia and the South Pacific, Department of Mineral Resources, Bangkok, Thailand, pp 453–464 Brown JC, Leicester P (1933) The Pyu earthquake of 3rd and 4th December, 1930 and subsequent Burma earthquakes up to January 1932. Memoirs of the Geological Survey of India 42(1):1–140 Caceres D, Kulhanek O (2000) Seismic Hazard of Honduras. Nat Hazard 22:49–69 Charusiri P, Daorerk V, Choowong M, Muangnoicharoen N, Won-in K, Lumjuan A, Kosuwan S, Saithong P, Thonnarat P (2004) The study on the investigations of active faults in Changwat Kanchanaburi area, western Thailand. Technical Report, Department of Geology, Faculty of Science, Chulalongkorn University, Bangkok, Thailand (in Thai with English abstract) Charusiri P, Choowong M, Charoentitirat T, Jankaew K, Chutakositkanon V, Kanjanapayont P (2005) Geological and physical effect evaluation in the tsunami damage area for restoration and warning system. Technical Report, Department of Geology, Faculty of Science, Chulalongkorn University (in Thai with English abstract) Cuong NQ, Zuchiewicz WA (2001) Morphotectonic properties of the Lo River Fault near Tam Dao in North Vietnam. Nat Hazard Earth Syst Sci 1:15–22 Duong CC, Feigl KL (1999) Geodetic measurement of horizontal strain across the Red River fault near Thac Ba, Vietnam. J Geod 73:298–310 Eleftheria P, Xueze W, Vassilios K, Xueshen J (2004) Earthquake Triggering along the Xianshuihe fault zone of Western Sichuan, China. Pure Appl Geophys 161(8):1683–1707 Fenton CH, Charusiri P, Wood SH (2003) Recent paleoseismic investigations in Northern and Western Thailand. Ann Geophys 46 (5):957–981 Gardner JK, Knopoff L (1974) Is the sequence of earthquakes in Southern California, with aftershocks removed, Poissonian? Bull Seismol Soc Am 64(1):363–367 Gutenberg B, Richter CF (1954) Seismicity of the Earth and Associated Phenomena. Princeton University Press, Princeton, New Jersey Habermann RE (1991) Seismicity rate variations and systematic changes in magnitudes in teleseismic catalogs. Tectonophysics 193:277–289 Habermann RE, Creamer F (1994) Catalog errors and the M8 earthquake prediction algorithm. Bulletin of the Seismological Society of America 84:1551–1559 Ishimoto M, Iida K (1939) Observations sur les seismes enregistres parle microsismographe construit dernierement. Bulletin of the

Arab J Geosci Earthquake Research Institute, University of Tokyo 17:443–478 (in Japanese with French abstract) Kanamori H (1977) The energy release in great earthquakes. J Geophys Res 82:2981–2987 Khana PK, Chakraborty PP (2005) Two-phase opening of Andaman Sea: a new seismotectonic insight. Earth Planet Sci Lett 229:259–271 Klaipongphan S, Chakramanont V, Pinrode J, Chittrakarn P (1991) Geological and seismicity evaluation. Proceedings of the Second International Conference on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics, St. Louis, Missouri, pp 1357–1363 Kramer SL (1996) Geotechnical Earthquake Engineering. Princeton University Press, Princeton, New Jersey Lacassin R, Replumaz A, Leloup PH (1998) Hairpin river loops and strike-slip sense inversion of Southeast Asian strike-slip faults. Geology 26:703–706 Le Dain AY, Tapponnier P, Molnar P (1984) Active faulting and tectonics of Burma and surrounding regions. J Geophys Res 89:453–472 Maryanto S, Mulyana I (2008) Temporal change of fractal dimension of explosion earthquakes and harmonic tremors at Semeru Volcano, East Java, Indonesia, using critical exponent method. World Academy of Science, Engineering and Technology 42:537–541 Molnar P, Deng Q (1984) Faulting associated with large earthquakes and the average rate of deformation in Central and Eastern Asia. J Geophys Res 89:6203–6227 Nutalaya P, Sodsri S, Arnold EP (1985) Series on seismology. Volume II. Thailand. In: Arnold, EP (Ed.), Southeast Asia Association of Seismology and Earthquake Engineering, Bangkok Nuttee R, Charusiri P, Takashima I, Kosuwan S (2005) Paleoearthquakes along the southern segment of the Sri Sawat Fault. Morphotectonic and TL-dating evidences. Proceedings of the International Conference on Geology, Geotechnology and Mineral Resources of Indochina, Khon Kaen, Thailand, Kanchanaburi, Western Thailand, pp 542–554 Pailoplee S, Sugiyama Y, Charusiri P (2009) Deterministic and probabilistic seismic hazard analyses in Thailand and adjacent areas using active fault data. Earth Planets Space 61:1313–1325 Pailoplee S, Sugiyama Y, Charusiri P (2010) Probabilistic seismic hazard analysis in Thailand and adjacent areas by using regional seismic source zones. Terrestrial, Atmospheric and Oceanic Sciences 21(5):757–766 Palasri C (2006) Probabilistic seismic hazard map of Thailand. M.S. thesis, Department of Civil Engineering, Faculty of Engineering, Chulalongkorn University, Thailand Paul J, Burgmann R, Gaur VK, Bilham R, Larson KM, Ananda MB, Jade S, Mukal M, Anupama TS, Satyal G, Kumar D (2001) The motion and active deformation of India. Geophys Res Lett 28:647–650 Phoung NH (1991) Probabilistic assessment of earthquake hazard in Vietnam based on seismotectonic regionalization. Tectonophysics 198:81–93 Polachan S, Pradidtan S, Tongtaow C, Janmaha S, Intrawijitr K, Sangsuwan C (1991) Development of Cenozoic basins in Thailand. Mar Pet Geol 8:84–97 Rajendran CP, Earnest A, Rajendran K, Das DR, Kesavan S (2003) The 13th September 2002 north Andaman (Diglipur) earthquake: an analysis in the context of regional seismicity. Curr Sci 84(7):919–924 Reading HG (1980) Characteristics and recognition of strike-slip fault system. Spec Publ Int Ass Sediment 4:7–26

Rhodes BP, Charusiri P, Kosuwan S, Lamjuan A (2005) Tertiary evolution of the Three Pagodas Fault, Western Thailand. Proceedings of the International Conference on Geology, Geotechnology and Mineral Resources of Indochina, Khon Kaen, Thailand, pp 498–505 Rhodes BP, Perez R, Lamjuan A, Kosuwan S (2004) Kinematics and tectonic implications of the Mae Kuang Fault, northern Thailand. Journal of Asian Earth Sciences 24(1):79–89 Royal Irrigation Department (RID) (2006) Active fault investigation in Mae Yom fault zone, Kaeng Seur Ten Dam Site, Song district, Phrae Province, Thailand. Technical Report, Royal Irrigation Department, Bangkok, Thailand (in Thai with English abstract) Shanker D, Singh VP (1997) Seismic risk analysis for the occurrence of medium size earthquakes in Kangraregion of Himachal Pradesh, India. Proceedings of the Indian National Science Academy 63(A2):197–202 Takemoto K, Halim N, Otofuji Y, Tran VT, Le VD, Hada S (2005) New paleomagnetic constraints on the extrusion of Indochina: Late Cretaceous results from the Song Da terrane, northern Vietnam. Earth Planet Sci Lett 229:273–285 Tun ST (2008) Active faults in the Central Myanmar basin. Proceedings of the International Symposia on Geosciences Resources and Environments of Asian Terranes Bangkok, Thailand, p 63 (Abstract) Udchachon M, Charusiri P, Daorerk V, Won-in K, Takashima I (2005) Paleo-seismic studies along the southeastern portion of the Phrae Basin, Northern Thailand. Proceedings of the International Conference on Geology, Geotechnology and Mineral Resources of Indochina, Khon Kaen, Thailand, pp 511–519 Woessner J, Wiemer S (2005) Assessing the quality of earthquake catalogues: estimating the magnitude of completeness and its uncertainty. Bull Seismol Soc Am 95:684–698 Wong I, Fenton C, Dober M, Zachariasen J, Terra F (2005) Seismic hazard evaluation of the Tha Sae project, Thailand. Technical Report, Punya Consultants Co. Ltd Yadav RBS (2009) Seismotectonic modeling of NW Himalaya: a perspective on future seismic hazard. Ph.D. Thesis, Department of Earthquake Engineering, IIT Roorkee, India Youngs RR, Coppersmith KJ (1985) Implications of fault slip rates and earthquake recurrence models to probabilistic seismic hazard estimates. Bull Seismol Soc Am 75:939–964 Zuchiewicz W, Cuong NQ, Bluszcz A, Michalik M (2004) Quaternary sediments in the Dien Bien Phu fault zone, NW Vietnam: a record of young tectonic processes in the light of OSL-SAR dating results. Geomorphology 60:269–302 Zuniga FR, Wiemer S (1999) Seismicity patterns: are they always related to natural causes? Pure Appl Geophys 155:713–726

After this manuscript had been submitted for publication, the other great events of Mw 8.6 and 8.2 generated nearby the SumatraAndaman Interplate (zone A) on April 11, 2012 (Fig. 1). Comparing with the Mw 8.6 earthquake generated on March 28, 2005, the calculating 7-year return period of these magnitude levels are conform with the results obtained in this study which revealed the return period of earthquake 8.0–8.5 along the zone A are 9–20 years.